View Full Version : Why is my 5V so low? Read this
Page 1
I see a lot of people discussing the fact that their +5V is low and suspecting that their power supply is the culprit. Some times this is the case, but many times it is not. I will try to explain what I have observed regarding this problem. I have tried to keep it simple and in a context that non-technical people can follow. This is at the sacrifice of some technical accuracy. Those of you into Electronic Academia, look away.
First a small bit of electronics theory. Ohms law tells us that for a given amount of current passing through a resistive path, a given amount of voltage drop will occur. We usually think of copper as a good conductor. Any conductor, regardless of the material will exhibit some resistance to current flow. Obviously, the lower the resistance, the better.
The +5V coming from your PSU goes through a path comprised of several resistances that occur in the different parts of the path. Some are material resistance, like that mentioned above and some are resistance incurred as a result of insufficient contact area.
First, lets look at how the +5V is distributed from the PSU. Some of the conductors go off to the Molex connectors associated with disk drives and some go off to the ATX connector that plugs into your motherboard. Regardless of where they end up, they all originate from a common point inside your PSU. The voltage at that point is sensed by the regulator circuitry in the PSU and is used to control it at a (hopefully) fixed level.
http://hill195.home.mchsi.com/ocpix/5vdist.jpg
Page 2
We will only concern ourselves with the ATX conductors as they carry the +5V over to your motherboard, though the conductors going to the Molex connectors can be put to use later on.
The following illustration represents the different path resistances that occur. I have placed the schematic symbol of resistors in them to drive home the point.
http://hill195.home.mchsi.com/ocpix/res1.jpg
The first resistance is one many people never ponder. It is the source impedance of the +5V output inside the PSU. Generally speaking, it is a small value and if the PSU wattage is sufficient for your system setup, it does not enter into the equation. I mention it only in the interest of technical accuracy.
The second resistance is the copper conductors that go from the +5V output to the ATX connector. PSUs use several of them, in parallel, to reduce the path resistance. Resistors in parallel divide by the number of them in the circuit. Even with four conductors in parallel, their resistance is still a factor and they contribute to part of the voltage drop. The longer those conductors are, the greater the resistance.
Now the main contributor to the voltage drop. The ATX connector itself. Though the conductors coming to the connector may be large, their size is somewhat negated by being "funneled" down to the actual female pins in the connector and their associated male pins on the motherboard. Much like the surface of a rough finished heatsink, mating to your CPU core, the do not necessarily make good contact. The connector has to be loose enough to be able to plug and unplug it without tearing the pins out of the connector and motherboard, so they are somewhat loosely fitting.
The more you plug and unplug that connector, the more of the mating surfaces erode and the looser they become. The looser the fit, the less contact area, the greater the resistance, the greater the voltage drop. As these pins deteriorate, they embark on a downward spiral in performance. The less the surface are in contact and the resultant resistance the more they heat up from the current passing through them. The more they heat up, the greater the resistance and the spiral continues. So pins will get so hot as to melt the ATX connector.
Lastly, there is the resistance on the motherboard of the copper voltage trace from the ATX connector to the MosFets used for regulating the Core Voltage of the CPU. Though the +5V goes to places all over the motherboard, for the various circuits, the bulk of the current is delivered to that regulator. This resistance, though typically much less than that seen in the ATX connector, does contribute to some of the voltage drop. Like the conductors coming from the PSU to the ATX connector, the longer the path to the core regulator, the greater the resistance. That's why you see some motherboard manufacturers moving their ATX connector as close as physically possible to the core voltage regulation circuit lately. Even they learn from experience.
Page 3
So, what can you do about the ATX connector, which contributes the greatest to your +5V drop? Not a lot. You obviously could solder all the ATX conductors to the motherboard, but that is a lot of work and you would have to undo it and redo it every time you took the motherboard out and put it back in. Short of finding a different kind of connector, more suited for the current, about the only thing you can do is "bandaid" the problem by supplementing the ATX connector with one or more +5V bridging paths around it. By this, I mean adding an additional conductor or two directly from the PSU to the core voltage regulation circuit. I have done this on every motherboard I have used for overclocking and just one additional conductor really helps.
.http://hill195.home.mchsi.com/ocpix/res2.jpg
Page 4
If you're lucky and have an unused 4-pin Molex connector cable available, simply rob the +5V conductor from it. You can still use it to power your case fans. In the case of some PSUs, there is a cable that goes to some odd connector whose function I am not familiar with. It looks like this. I robbed its +5V conductor for my modification.
http://hill195.home.mchsi.com/ocpix/oddplug.jpg
Page 5
Once you locate the core voltage regulating Mosfets, typically six of them, you will see them typically grouped in threes. On three of them the center pin is attached directly to the +5V path coming over from the ATX connector and three of them are not. You can determine this by taking a multimeter and checking each of them with the power on. If you can not determine, beyond a shadow of a doubt, which is which, do not attempt this modification! If you put the +5V booster wire on the wrong Mosfet, you motherboard will appear in the Classifieds section as a "fixer upper" :D
Once you have determined the correct MosFets, you can take one of two approaches. You can solder the 5V booster conductor to the center pin of the central MosFet like this:
http://hill195.home.mchsi.com/ocpix/mosfet1.jpg
Page 6
-or-
You can do it like this, which is how I do it.
http://hill195.home.mchsi.com/ocpix/mosfet3.jpg
Page 7
Admittedly, the spade lug connector in the second example adds a little resistance again, but it is designed for high current and has a tight fit and large contact area. Also, it is only 1 of five conductors, the other four being those in the ATX circuit. It sure makes changing out the motherboard a lot faster.
Earlier, I mentioned that you could put one of your 4-pin Molex connectors to use. To determine whether the +5V drop is the result of a PSU that has insufficient wattage for your system setup, or an ATX connector problem, all you need to do is take an unused Molex connector and measure the +5V on it while running your CPU at high load. Since the Molex connector is connected to nothing, no current is flowing through it and hence no voltage drop is occurring. Under that circumstance, the voltage you read onit will be exactly what the PSU is putting out at the +5V common distribution point. If that voltage is nearly as low as that read on your motherboard, then your PSU is not powerful enough. If it reads near to 5V and distinctly higher than what you read on the motheboard, then the ATX connector is your likely suspect and you can benefit from this modification.
One last note of caution. It is a dangerous practice to simply turn up the +5V output from your PSU, assuming it is adjustable, to keep it at spec when your CPU is working hard. Once it goes back to idle, the +5V output will go up, sometimes to a level that will damage your other components.
If I can give you any better advice, please let me know, either by a reply to this post or a PM.
I apologize to kloster64a for removing his post, which appeared while I was still writing this. I did that so the thread would be uninterrupted in its entirety. Kloster64a, please repost your message now that I am done.
73, Hoot
*Rasies Hand*
Um.. Yes I have a question. about the volatge regulateing Mosfets, I have the same board as you,KT7a, though it is dead from a dumb accident, I was wondering, Since there are 6 Mosfets, Do you just connect the added conductor coming from the PS to only 3 Mosfets, or can you do all 6, or are only 3 of them connected to the +5V line?
Very well written by my english is horrible as I'm only in grade 9 and english hasn't kicked in this semester but will start soon so I don't get a majority of the post.
Very well written though.
Just noticed your detication to this sticky where you spent near an hour setting this up.
Congrats Hoot to your contribution of the forums.
Originally posted by Angry
*Rasies Hand*
Um.. Yes I have a question. about the volatge regulateing Mosfets, I have the same board as you,KT7a, though it is dead from a dumb accident, I was wondering, Since there are 6 Mosfets, Do you just connect the added conductor coming from the PS to only 3 Mosfets, or can you do all 6, or are only 3 of them connected to the +5V line?
From Page 5:
"Once you locate the core voltage regulating Mosfets, typically six of them, you will see them typically grouped in threes. On three of them the center pin is attached directly to the +5V path coming over from the ATX connector and three of them are not. You can determine this by taking a multimeter and checking each of them with the power on. If you can not determine, beyond a shadow of a doubt, which is which, do not attempt this modification! If you put the +5V booster wire on the wrong Mosfet, your motherboard will appear in the Classifieds section as a "fixer upper" :D
73, Hoot
Great info from the man:beer: Looks like them " En******* is a POS 'cause my 5V is low" wont show up too often I hope. Do you still think that people should still test the PSU before they buy? I still do just to have peace of mind that the unit is good:beer:
Wookie8662
01-14-02, 03:21 AM
Hey,
Cool job Hoot!
Very informative posting.
I am an electrician and understand everything in your posts.
Makes alot of sense on the ATX plug being the "weakest link" of the possible problems.
I'm gonna look over my 3 rigs, and check out the 5v supplies to see how they compare. Who knows I may get a couple more MHZ on the FSB after I'm done. :)
Thanks for a very good post,
Wookie8662
Originally posted by Hoot
From Page 5:
"Once you locate the core voltage regulating Mosfets, typically six of them, you will see them typically grouped in threes. On three of them the center pin is attached directly to the +5V path coming over from the ATX connector and three of them are not. You can determine this by taking a multimeter and checking each of them with the power on. If you can not determine, beyond a shadow of a doubt, which is which, do not attempt this modification! If you put the +5V booster wire on the wrong Mosfet, your motherboard will appear in the Classifieds section as a "fixer upper" :D
73, Hoot
Oh ok, Thanks. Very good info.
Originally posted by Hoot
Page 4
If you're lucky and have an unused 4-pin Molex connector cable available, simply rob the +5V conductor from it. You can still use it to power your case fans. In the case of some PSUs, there is a cable that goes to some odd connector whose function I am not familiar with. It looks like this. I robbed its +5V conductor for my modification.
I have one of those.
WHat can i do with it?
DodgeViper
01-14-02, 10:07 PM
I see a lot of people discussing the fact that their +5V is low and suspecting that their power supply is the culprit. Some times this is the case, but many times it is not.
Hi Hoot, I think in my case it was the supply. With everything being the same and only changing the PS. I increased my rail from 4.78 to 5.03. I went from a 300w PS that AMD had on their site to a Antec 400w.
WarriorII
01-15-02, 08:27 AM
YOU DA MAN HOOT !!!!!!!!!!
:cool:
Breadfan
01-15-02, 09:41 AM
Wow...that pretty much explains it...geez, thats awesome. Hmm, if motherboard manufacturers and psu manufacturuers start putting out secondary 5v connectors striaght to the board then we know who to thank!
My 5v's are ok on everything I have (I think I'm lucky), but if I ever have a problem I'll definitely do that...
Mike
Originally posted by Breadfan
Wow...that pretty much explains it...geez, thats awesome. Hmm, if motherboard manufacturers and psu manufacturuers start putting out secondary 5v connectors striaght to the board then we know who to thank!
My 5v's are ok on everything I have (I think I'm lucky), but if I ever have a problem I'll definitely do that...
Mike
they already have them
that odd connector that hoot had a picture of (the thing that looks like half of an AT power connector) is the auxilairy power connector, its on p4 boards (dont know if its on all of them or not).
its used to keep the voltage levels stable when the computer is under load :)
Dissolved
01-16-02, 02:04 PM
how could i tell what mosfets to use?
my board has so many and a few are in odd places :confused: :confused:
Originally posted by Dissolved
how could i tell what mosfets to use?
my board has so many and a few are in odd places :confused: :confused:
My guess would be one of these two sets of three.
Verify which set with a DMM per Page 5 near the
beginning of this thread.
Hoot
Great post Hoot! One question, and please forgive my electronics ignorance; what a multimeter? I assume it's some kind of device that measures the power (volts?) running through something.
Would I use one of these tools to measure the power in the un-used 5v connector, per your directions on page 7, or would this require a different measuring device?
Thanks!
-Ace-
That was just beautiful. When I hooked up my core volt mod months ago my 5v line dropped, I assumed it measured so low (4.79v) simply because of the mod but when I tested with a multimeter I got something like 0.08v higher at the drive connector compared to the monitoring software. I had assumed the meter was not calibrated accurately, you may have saved me from buying a new powersupply unnecessicarily. Thank you.
Btw: When are you and the Blowfish going to put out another album???
Dissolved
01-17-02, 04:17 AM
Originally posted by Hoot
My guess would be one of these two sets of three.
Verify which set with a DMM per Page 5 near the
beginning of this thread.
Hoot
well i have an antec 300w, and the 5v line is 5.1 idle.. i havent tested load tho.. but i dont think ill need to mod..
my 3v line is a tad low tho.. 3.29 idle..
Originally posted by AceTKK
Great post Hoot! One question, and please forgive my electronics ignorance; what a multimeter? I assume it's some kind of device that measures the power (volts?) running through something.
Would I use one of these tools to measure the power in the un-used 5v connector, per your directions on page 7, or would this require a different measuring device?
Thanks!
-Ace-
A Multimeter is one of many names for a device that every overclocker should have. You can find them everywhere for varying prices and consequently features and accuracy. The technology has evolved to the point where you can get one for the cost of a new Panaflo 120mm fan, which is both accurate and feature rich. They all, typically have the capability to measure AC & DC Voltage, Current and Resistance. Additional features may include Continuity, Diode/Transistor junctions, Frequency, Temperature and assorted lesser features. I can not imagine performing experiments in electronics et al, without one. Kind of like exploring caves without a flashlight. You don't have to be particularly adept technically to use one and it is very handy to have around the house for many uses other than overclocking.
To answer your question, yes this is the device you would use to measure the voltage on your 4-pin molex connector or to determine which MosFets are the ones you would attach the +5V booster wire to.
Hoot
I just popped my door off and checked the system out. The line from one of the drive connectors is 0.09v higher than what the board has. I shut down and checked out the board layout. My board is an ASUS A7V-E, it has 4 mosfets, not 6, and using the circuit continuity testor it shows that all 4 are for the 5v line. I tested from the ATX socket to the middle pins on all 4 mosfets, is that an accurate method of testing or is there something else I should do to verify?
Usually that is a sound approach but I prefer to actually read the voltage on those pins while the unit is running. The resistance between the +5V and ground is so low, your continuity tester may think they are all connected, but they might not be directly connected to the +5V. Instead, you are seeing the path back around through the ground. That probably sounds a tad abstract. Trust me, read the voltage, not the resistance.
Hoot
Originally posted by Hoot
Usually that is a sound approach but I prefer to actually read the voltage on those pins while the unit is running. The resistance between the +5V and ground is so low, your continuity tester may think they are all connected, but they might not be directly connected to the +5V. Instead, you are seeing the path back around through the ground. That probably sounds a tad abstract. Trust me, read the voltage, not the resistance.
Hoot
Hoot you just saved my bacon. I checked all four, two were 4.86v (which is within 0.01v of what my 5v line is using monitoring software) and the other two were 1.94v. If I'd have hooked 5v up to those two I'd have cooked my system. Any idea what the other two are for? They're close to core voltage but I don't think close enough. My core voltage measured by internal sensors is 1.86v
It's 4.5 hours after my last post and I've only now got the system back together. I put 2 extra wires coming right off the cluster of 5v wires on the power supply circuit board. Unfortunately my skills as a solderer are insufficient to get wires attatched to the stumps of the middle pins on those mosfets. Any suggestions??
freshy98
01-18-02, 08:24 AM
Stupid question maybe, but when you measure the Volt lines of the mosfets with the multimeter do you do that with power on or power off?
Also, in what range do you set the multimeter? Typiclly spoken (I know, first word is wrong!)
Do it with the power on so you have something to measure. :D
Since you are measuring 5V, set the scale to anything greater than 5V, IE 10V, 20V, etc.
Hoot
freshy98
01-18-02, 09:24 AM
:D Have printed out the pages and saw on a page to measure with the power on. Already thought that, but wasn't sure.
What I still don't understand is why in the first mod you solder the +5V wire to the center mosfet, and in the second mod you take all three of them. Why is that?
Btw: I know that there are a lot of people using the MSI K7T 266Pro2 here which have the problem. I will make some pictures tonight on where to find the mosfets and upload them here.
Think that will be appreciated. :)
No suggestions on how to get the wires attatched to the stump of the pins?? Anyone know of any clamping pamona grabbers that will hold onto the entire transistor package or such?
Hi,
I have taken some voltage measurments from my system, and i was wondering if i would benefit from this 5v mosfet mod. My system specs are: 300W psu, Abit KG7, Tbird 1.2@1.23 (9x137)
Here's the volts info:
With numbers like that I would think so. Man that's low!
freshy98
01-18-02, 02:00 PM
Thought this might be nice for MSI K7T 266Pro2 users. The picture is not too good, but I only have a webcam to take pictures with.
Anyways... The green arrows point to the mosfets Q25 through Q28. The red arrows point to the mosfets needed for the mod (Q26 and Q28).
This info is not by me, but another guy in this forum (sorry, forgot your name).
freshy98
01-18-02, 02:11 PM
What I still do want to know is where you connect the other side of the measuring wire of the multimeter too?
Seems obvious that you connect the red side to the middle pin of the mosfet, but where does the other side go? The black wire of the multimeter that is.
It is said so many times to measure the 5V line, but without an black I can't measure anything. It's called electrotechnics and had been here for ages.....get my point?
So how does everyone measure? In the air? With their finger? Tell me!!!! :D
WarriorII
01-18-02, 02:41 PM
Seems obvious that you connect the red side to the middle pin of the mosfet, but where does the other side go? The black wire of the multimeter that is.
- The "Black" wire goes to case ground. :eek:
If you are measuring the 5v on the mobo, still in the case, that is.
:cool:
freshy98
01-18-02, 03:06 PM
aaah okay, now I'm getting somewheres.
I heard that the ground is the top left screw is also for the ground? Bit weird though since every screw is attahced to the case?
Or can I just hold the measuring stick right on the case?
ElectricMan
01-19-02, 06:15 PM
I checked out my system , I hooked up the multitester like this. one lead going to a blk wire from one of my drives and the other lead going to the center lug in the mosfet. I had 5.09volt's. I then played a music cd and put a dvd movie in and them ran 2001 mark and it went down to 5.02 .I am a electrician (Building's) but I think from what you guys are saying that is good and needs nothing, no mod like hoot did is needed. I have a Aopen Case and power supply 300 watt. What do you all think????
freshy98
01-19-02, 06:44 PM
Well, found out what NOT to do... It scared the living daylights out of me, and it was night! ;)
I tried to use the multimeter, but somehow due to my not to steady hand I accidently touched another pin of the mosfet.
Thus creating a link between the gate and an anode or kathode.
My system was of in a flash! (not a real one).
restarting wouldn't work, so I waited a while. Still nothing.
Then I cleared the CMOS and luckily everything worked again :D
So I didn't fry my mobo :D :D :D
It's working happily now again, but learned my leasson.
be carefull!
I have had 5V problems ever since I built this system, ( sig below) and am now on my 3rd PSU. Macron 300W ( 30 days), Fortron 300W (60 days) and lastly the Enermax ( about 4 months). The Enermax 5V has slowly slid downwards and am now having stability and startup problems. Have been in touch with MSI who blame Enermax and have been in touch with Enermax who said that I was not having any problems at that stage but to contact them agian if I did. Have contacted them and now awaiting a response.
I have tested the MOSFETS on my MSI K7T Turbo Raid V3.0 and for those who are interested the pic below shows the relevant ones (4 in all). If you use this info check your mobo first. I don't take responsibility for mistakes. The other MOSFETS were all giving readings of +-2V. I haven't done the mod yet. Waiting to see if Enermax have suggestions before I resort to that.
Hoot is it okay to take a single connection for all 4 from a molex connector? I have molex connectors for fans and thought of just joining all 4 wires and connecting via one of these molex connectors.
The more dedicated conductors going to the MosFets, the less the drop through them. That, having been said, use as few as possible to get the job done. I'd start with one and see how it goes. Only add more if, for some reason, one did not cut it. Keep it simple.
Lets say that your Enermax puts out 5.1V under load, as measured at one of the unused Molex connectors, but you're only seeing 4.75 at the MosFet. If by adding one 5V booster wire, your voltage at the MosFets goes up to 5.0V. It would be a waste to add three more 5V booster wires to just to get it up to 5.05V. More wires equals more opportunity for an error.
Hoot
Originally posted by eobard
No suggestions on how to get the wires attatched to the stump of the pins?? Anyone know of any clamping pamona grabbers that will hold onto the entire transistor package or such?
I've been stretched pretty thin lately and only now saw all the posts that need addressing.
Some core regulators use 3 MosFets per leg and some use only 2. By leg, I mean that of all the MosFets in the core voltage regulator, half of them switch the plus side of the signal and half switch the ground side. I don't want to go into a big explanation of that design, hence the simple explanation. You only want to supplement the plus half of the signal path. There are plenty of paths on the ground side already.
If you don't feel comfortable soldering to the center "stubs" of the plus side MosFets, the Tab of them, that is soldered to the motherboard is the same connection. Inside the MosFets, the center pin is connected to the Tab. IE They are the same connection.
Hoot
Originally posted by ElectricMan
I checked out my system , I hooked up the multitester like this. one lead going to a blk wire from one of my drives and the other lead going to the center lug in the mosfet. I had 5.09volt's. I then played a music cd and put a dvd movie in and them ran 2001 mark and it went down to 5.02 .I am a electrician (Building's) but I think from what you guys are saying that is good and needs nothing, no mod like hoot did is needed. I have a Aopen Case and power supply 300 watt. What do you all think????
For the most accurate measurement of the +5V at the MosFet, the meter minus lead would best be placed on a point on the motherboard that corresponds to ground, preferably as close to the MosFets as convenient. There are plenty of access spots. Yes that should be the same as the case or even a black lead on one of the Molex connectors, but with such high current flowing, the further away you are from the area of the motherboard where the MosFets are mounted, the greater the chance of inaccuracy.
Hoot
Originally posted by J37X
Hi,
I have taken some voltage measurments from my system, and i was wondering if i would benefit from this 5v mosfet mod. My system specs are: 300W psu, Abit KG7, Tbird 1.2@1.23 (9x137)
Here's the volts info:
Looking at your graphic, if your PSU has an internal trimpot for turning up the +5V output, you should adjust it to ~5.1V at idle.
Then recheck your readings under the different scenarios. If the voltage measured at an unused 5V line on a Molex still drops to below 5V under load, chances are your particular PSU is running at or very near its capacity. At that point, you have to accept that fact and live with it, or buy a higher capactity PSU.
Hoot
Just shoot
01-21-02, 10:39 AM
Hoot you did a very good job of writing this but I guess I'm alittle confused as why wouldn't you recrimp the connections on the atx connector to ensure a good connections before voiding your warranty on a good O/C Board. This should only be used if after testing with a multimeter that you cannot get a good connection at the atx connector and the recrimping doesn't work.
I had found this when I replaced my 300 watt PSU to a 400 Watt Psu that the atx connector was a little sloppy (and 5 volt rail still hung around 4.8 volts at load) so I just took a eye glass screw driver into the female side of the atx connector and closed the spade back up on all and wala voltage at mosfets 5.07 under full load and 5.1 at idle.
Granted this won't fix all the connectors as being the weak link but it should rule out most of them.
That's an excellent idea, but very technique sensitive. You, obviously have the technique down, as do I and I'm sure some other technically inclined people, but for the majority of the users, its hard to say. You can take that technique too far and disfigure your pins. Then when you try to undo it, they break off or become more disfigured. You know what I mean. I've had 30 years, in electronics, to do it wrong and learn how to do it right the second time. :D
Hoot
Just shoot
01-21-02, 11:19 AM
Yes, you are right patience is a virtue, and ignorance just takes longer. LOL:D :p ;)
Originally posted by Hoot
I've been stretched pretty thin lately and only now saw all the posts that need addressing.
That's what happens when you do a good job on something that a lot of people can use. I can think of worse problems than being compotent.
If you don't feel comfortable soldering to the center "stubs" of the plus side MosFets, the Tab of them, that is soldered to the motherboard is the same connection. Inside the MosFets, the center pin is connected to the Tab. IE They are the same connection.
Hoot
Comfort isn't the issue, capable is. I'm just too ham fisted to make it stick but I do have an idea on what to do. As for using the tab/backing of the transistor they are not really an option. Most "TO-220" transistors I've seen have a length of metal with a hole in it to allow the transistor to be screwed to a sink for a tab but the MOSFETS in my board barely have a ridge of 2mm coming off the top, no screw hole or anything else to get a good solder conection.
Originally posted by Hoot
I've had 30 years, in electronics, to do it wrong and learn how to do it right the second time. :D
Hoot
I don't suppose you could help me try to fix my monitor could you? The red is gone on it so I'm thinking it's probably the red gun but I'd hate to give up on it if it's only a $2 chip that's giving up. I'm pretty sure I've got the physical skills (in spite of my ham fistedness) and base electronic knowledge to make repairs if I knew what repairs to make but I don't. If you actually do have any ideas I am willing to hear them.
btw: I've done the "shake the cable" thing and the "dust it out with compressed air" thing already.
Most monitor manufacturers don't give out schematics and without a schematic, you may as well turn off the lights and just throw darts at the circuit card. Even I won't touch a monitor without a schematic. I could have had a 17" Sony Trinitron for free from work. It had lost its Red also. I putzed around with it for an hour, trying the obvious things, but no revelations. So, it went back on the recycle skid. Sounds like yours is headed that way also.
Hoot
I finally hooked mine up. It has apparently yeilded no benefit, although I'm runing at 1155mhz my 5v has dropped to 4.813v, a voltage level that has caused my system to stall before. :( Still, it hasn't stalled yet. Should I no longer trust my 5v readings as reported by PROBE? Does this mod bypass the sensor?
Well I've rebooted from my crash. :rolleyes: :rolleyes: :rolleyes: Well it was worth a try. Got to 1102mhz though, upped the resistance on my volt mod, core voltage went down a bit, 5v line went up a bit. An acceptable trade.
freshy98
01-29-02, 09:52 AM
I was thinking. (yes, I can...)
Since soldering would leave you with the troubles of removing the motherboard when needed, I have this idea.
Why not solder wires to the gates of the mosfets you need to mod and wrap them together at the end and solder it.
Then, take something which is able to connect and disconnect. Like molex connectors, only different.
I'm sorry, my English is OK but I don't know every word out of my head ;)
When I'm visiting my parents I will make the stuff and take some pictures of it.
Reason for this idea is that soldering the wires permenantly isn't a real option because of the troubles removing the motherboard which tweakers do.
And secondly is that getting a grabber like Hoot shows on page 6 is almost impossible to get here in The Netherlands.
Thirdly, soldering the wires to the gates of the mosfets is better in my opinion. Offcourse, the connection between the wires coming from the gates and the wire coming from the PSU are linked together in the middle which could cause problems, but I think this will be better in general.
Let me know what you guys/girls think about it.
freshy98
Originally posted by freshy98
I was thinking. (yes, I can...)
Since soldering would leave you with the troubles of removing the motherboard when needed, I have this idea.
Why not solder wires to the gates of the mosfets you need to mod and wrap them together at the end and solder it.
Then, take something which is able to connect and disconnect. Like molex connectors, only different.
I'm sorry, my English is OK but I don't know every word out of my head ;)
When I'm visiting my parents I will make the stuff and take some pictures of it.
Reason for this idea is that soldering the wires permenantly isn't a real option because of the troubles removing the motherboard which tweakers do.
And secondly is that getting a grabber like Hoot shows on page 6 is almost impossible to get here in The Netherlands.
Thirdly, soldering the wires to the gates of the mosfets is better in my opinion. Offcourse, the connection between the wires coming from the gates and the wire coming from the PSU are linked together in the middle which could cause problems, but I think this will be better in general.
Let me know what you guys/girls think about it.
freshy98
That's what I did. It is a trade off. It upped my 5v but not by much. But it was enough to let me get to 1102mhz stable. The slight boost allowed me to set my core volt mod to 1.94v and my 5v line was running no lower than 4.857v, which is just good enough. I got half of what I wanted. I think if I had soldered the connections on to the mosfet I may have gotten to 1150, or even 1200 stable. mmmmmmm 1200!!!! :D
Voodoo Rufus
02-01-02, 12:28 PM
I did something similar for my board. It didn't need extra voltage, but the Molex ATX connector melted, caused by the red ATX wires getting so hot that they burned apart. My father and I cleared some of the plastic on the back of the motherboard PCB to expose the copper plate that provided the power to the mosfets. We used two 13 gauge wires soldered to that plate and used a 2-pin Astroflight gold connector to bypass the ATX connector so that we were able to supply power to the proc.
All I think of that went wrong was a too small contact patch for the poewr connectors for the power that was going through them, or bimetallic corrosion accelerated by the current flow.
Instead of using another Molex POS connector for these mods, something even better would be to use any hobby/radio control connector from companies like Deans or Astroflight. They can easily handle 50 amps per pin.
Are the red wires on ATX connectors 5V or 12V? It would make sense for them to be 12V I would think.
The red is 5v, the yellow is 12v. I agree that the red would make more sense as 12v. Red is the danger color, yellow is the caution color. Neither 5 nor 12 volts will kill you but 12v is (microscopically) more likely to.
Voodoo Rufus
02-01-02, 01:14 PM
There would be less electrical resistance to use the 12V line to power the CPU, especially with the small wires they use for the connector.
Ichelo351
02-02-02, 11:47 PM
the color doesn't really have to do with danger to a person, like 5 or 12 volts would do very little to a person. it has to do with the sensentiveness to the computer, the 12 v line is the "rough" current, the only thing it is used for in the computer is really to power fans and motors. and thus if you watch the +12 v line, it isn't very stable or exactly at 12 v. less monitoring or filtering is on the 12 line because it is just for motors and fans which will not have major problems with "rough" current, so its "caution". while the +5 v line runs pretty much everythig else, all the expensive stuff that if it had "rough" current it could potientually die. so to the computer, the +5 v line is the dangerous one, because if somethign happens to that, like it being spiked with something it would have worse effects than with the +12 v line. at least thats how i know it, if that makes sense.
hoot, that is a great piece of work, i may do that at some point, because right now my +5 is down alittle, but also my +5 is down from the molex but not as much. i am suffering from a sucky power supply problem, a barely 300 watt total generic, but it is atleast alittle better than my old 250 watt one! i am hopfully getting a nice 410 watt one that will last my long time, then do nice mod to it hehe.
-Aaron
I have a epox 8KHA+ and there are 3 mosfets on top of the board and 3 mosfets at the back side of the motherboard. Is the mosfets on the top or on the back that I am suppose to connect the +5 cable?
Page 5
Once you locate the core voltage regulating Mosfets, typically six of them, you will see them typically grouped in threes. On three of them the center pin is attached directly to the +5V path coming over from the ATX connector and three of them are not. You can determine this by taking a multimeter and checking each of them with the power on. If you can not determine, beyond a shadow of a doubt, which is which, do not attempt this modification!"
If you dont have access to a voltmeter, find someone with an 8KHA+ who does and have them determine it for you.
Hoot
WarriorII
02-05-02, 10:42 AM
This really works great ! I just need a desent PS now ! lol.
:cool:
I can fix a voltmeter no problems but if somebody already have done this mod to an 8KHA+ then I figured he could let me know which mosfets it is so I don't need to play around because the ones at the back is hard to get to.
I have taken the necessary readings now.
I have 5,13V from my 2x300W power supplies on idle and 5,09V at full load.
I have determent that the 3 mosfets on top of the board are the ones whit +5V.
At idle I had 5,02V on the mosfets and when I was running prime 95 I hade 4,94- 4,95V.
According to mbm5 and bios I have 4,89-4,95V idle and at full load I have 4,76-4,84V.
Is it dangerous to supply the mosfets whit 5,13V or is it within the level of tolerance?
Ichelo351
02-07-02, 08:06 PM
just wondering,
has anyone done this with the soyo dragon +? i looked on my board and it has like 9 mosfets! i was just wondering....
MikeBD35
02-08-02, 03:14 PM
With Flounders recent requiem for a motherboard posting where he had some solder fly around and dislodge a resistor, I thought it might be a good idea to throw in a suggestion to use a piece of paper as an apron around the mosfets to avoid any more mishaps. Great article, appreciate the time you put into it.
Mike
I submitted it as an article for the front page, for people who don't make it to these forums.
Hoot
Originally posted by pa1983
I have taken the necessary readings now.
I have 5,13V from my 2x300W power supplies on idle and 5,09V at full load.
I have determent that the 3 mosfets on top of the board are the ones whit +5V.
At idle I had 5,02V on the mosfets and when I was running prime 95 I hade 4,94- 4,95V.
According to mbm5 and bios I have 4,89-4,95V idle and at full load I have 4,76-4,84V.
Is it dangerous to supply the mosfets whit 5,13V or is it within the level of tolerance?
Well within the tolerances. Go for it.
Hoot
freshy98
02-15-02, 02:11 AM
I did the mod last week too. Went very well thanks to my dad.
Got the wire Hoot used too on his mod (I have a Enermax 431W (newest type of PSU form Enermax I believe?))
Everything stays stable. No high or low during idle and load, though no +5V. I only get it between 4.89V and 4.95V. But it seems enough to get it stable at 1650MHz (original is 1465Mhz).
Even at 1705Mhz the voltages stay the same, just need to increase the VCore a bit more (also modded that last week) but that is for a different thread :-)
I found this link. http://www.hardware-unlimited.com/articles/5vrailmod/
On this mosfet mod they also connect the 0V cable to the mosfets to another pin that is ground, same as 0V. Is this necessary? While there be any differens
freshy98
02-15-02, 09:49 AM
Originally posted by pa1983
I found this link. http://www.hardware-unlimited.com/articles/5vrailmod/
On this mosfet mod they also connect the 0V cable to the mosfets to another pin that is ground, same as 0V. Is this necessary? While there be any differens
I don't think that will be nescassery as the 0V line is only for the way back. get what I mean?
You only want to supply more +5V to the mosfet's gate. The 0V line is not really interresting in this case.
freshy98
02-15-02, 01:07 PM
GRMBL...
stupid +5V line is still dropping to much.
running 1705MHz under load freezes my machine. VCore is not the problem. that's ok at 1.95V ~ 2.00V (gotta test that out thoroughly yet.
at 1650MHz the +5V line drops to 4.89V under load. when running 1705MHz under load the +5V line drops to 4.81V and eventually freezes my machine.
what else can I do to get a better +5V line? I just know my XP 1700+ can do so much better than 1705MHz!
I don't mean to "stir the pot", but how do you know that your 1705 overclock lack of stability is due to the +5V beyond a shadow of a doubt? You could only make that observation after improving the +5V regulation and it becomes stable. There are a lot of other variables in an overclock besides the contribution the +5V gives to the picture. At some point, component tolerances, add-on cards, memory, etc enter into the equation. There are no certainties when you are pushing the upper limit of your system. By system, I mean the overall behavior of the other devices combined as a whole. Once you are pushing them beyond their specified rating, the chain becomes only as strong as its weakest link.
Hoot
freshy98
02-17-02, 08:27 AM
My compenents are up to it. I'm sure of that. Been running Celeron's before this XP 1700+ and had high FSB's and bus speeds.
And I'm also very positive that the +5V is the problem.
before the mosfet mod the line would drop to low 4.80's and with the mod it would hang at about 4.92V.
when trying to run 1705 the +5V drops again to low 4.80's and freezes my system.
what else is there to consider???
Just shoot
02-17-02, 10:40 AM
I beleive there was a power supply mod for enermax's or antecs with five volt rail problems do a search and I'd bet you'll find it. One of the two I think also have a pot that allows you to tweak that rail any hoot ( no pun intended).:D Do a search you should find the thread.
freshy98
02-17-02, 01:12 PM
I did a search on Enermax Mod but couldn't find a real thread about modding the Enermax. And I only have 56K which sucks in speed :mad:
can you help me searching?
thnx!
Originally posted by Just shoot
I beleive there was a power supply mod for enermax's or antecs with five volt rail problems do a search and I'd bet you'll find it. One of the two I think also have a pot that allows you to tweak that rail any hoot ( no pun intended).:D Do a search you should find the thread.
Just shoot
02-17-02, 08:48 PM
Here (http://www.amdmb.com/vb/showthread.php?threadid=54084) and here (http://www.hardforum.com/showthread.php?threadid=201241)
Good luck he thread was one that I started and was called "kt7a and 5 volts"
freshy98
02-18-02, 02:45 AM
hehe saw that thread about "KT7A and 5 volts" but thought it didn't applied to me :)
thnx! will check it out tommorow evening. I'm having a course every monday till 22:00h
Interesting article. How much higher can you overclock with stability after the mod Hoot?
I've heard that for IC's you want to maintain a voltage of +/- 5%. Have you noticed unstable settings when the voltage dropped below 4.75 or was greater than 5.25?
The Highlander
02-22-02, 07:14 PM
I measured the Voltage on my MOSFETS and a ground and it says 6.06v
but my bios and my HWMonitor say I have 4.66at high speed.
I measured it at molex 4 pin and I get 6.16v and about 17v!!! no wonder my deltas screamed that high!!! jejejeje I use a rheostat and it lower to 11.5v!!!!
Anyways...
WHy is this happening?
the reading on the 5v regulators seems to be MORE THAN ADEQUATE...
Thanks
It's not so much how much higher in frequency you can clock, but how much higher you can take the core voltage, if needed to achieve a stable overclock. Of course, you have to also have good cooling to dissipate that additional wattage. Without the modification, when I ran my core above about 1.75V, I could see the +5V start dropping. By the time the core voltage was at 1.9, the +5 was down to 4.85 and at a core voltage of 2.0, it was down to 4.75. Both of those drops even though the PSU output only went from 5.10 to 5.05.
So, to answer your question. If you have adequate cooling for 100+ watts of heat, then increasing the core voltage will allow you to increase your CPU speed, up to the point where your other system components, even the motherboards characteristics start to be a limiting factor. Keep in mind that every CPU has a sweet zone where increasing the voltage allows you to increase the speed. Once you are outside that sweet zone, it takes a lot more voltage to get a small additional gain in performance. I strongly believe you should not operate your CPU in that region, but rather hang around the top of the sweet zone. I typically test a setup for every additional mhz increase I can squeeze out of it, to find the absolute maximum it can obtain. I look at how hard everything is working to achieve that. Then, with an eye on reliability and longevity, I usually back it down a little bit and leave it there. To some overclockers, that may sound like heresy, but I don't run my car at the Red Line continuously either.
Hoot
Voodoo Rufus
02-22-02, 07:41 PM
What would you say is the minimum safe voltage for the 5V line?
The Highlander
02-22-02, 09:57 PM
I measured the Voltage on my MOSFETS and a ground and it says 6.06v
but my bios and my HWMonitor say I have 4.66at high speed.
I measured it at molex 4 pin and I get 6.16v and about 17v!!! no wonder my deltas screamed that high!!! jejejeje I use a rheostat and it lower to 11.5v!!!!
Anyways...
WHy is this happening?
the reading on the 5v regulators seems to be MORE THAN ADEQUATE...
That doesn't sound right. Are you sure you're measuring the Mosfets associated with your core voltage regulator or the Vagp regulator?
Hoot
The Highlander
02-25-02, 08:15 AM
YESSS i did that but you were also right it sounded strange... I bought a new 9v battery to my voltimeter... and guess what.. my PS is throwing now 4.99V vs 4.73 on the mosfets now...
My other question is... I did the volt mod with just one cable...
it had no effect... Must I do it with all 3?
I checked with the volt meter on that mosfets and I was getting 4.99v but my bios and HW mon kept saying 4.73.
Let me know.. Thanks
Thanks
freshy98
02-25-02, 08:49 AM
Hoot: Now I understand why my +5V line drops again.
Will try to fix it at 1.85V and then see what the +5V line does. Right now I modded it to 1.95V.
I'm just so irritated that I can't get 1705Mhz stable under load!
Little thought. Would it be a nice idea to take 2 small PSU (125W from an old 486'er) and use those +5V on each of the mosfets gates of my mobo? that way there is no problem when using the main PSU for it.
What do you people think about it?
Originally posted by The Highlander
YESSS i did that but you were also right it sounded strange... I bought a new 9v battery to my voltimeter... and guess what.. my PS is throwing now 4.99V vs 4.73 on the mosfets now...
My other question is... I did the volt mod with just one cable...
it had no effect... Must I do it with all 3?
I checked with the volt meter on that mosfets and I was getting 4.99v but my bios and HW mon kept saying 4.73.
Let me know.. Thanks
Thanks
Assuming you have a good battery in it, :D always trust your voltmeter above the on-board monitoring circuit. I must confess that I've never seen that large of an error in an on-board monitoring circuit. It's usually on the order of .05V or less, not .26V. One additional 5V conductor to the Mosfets should be sufficient.
Hoot
The Highlander
03-03-02, 01:58 PM
Ok... So i did the voltage modification to another mosfet of those... Now I have 2 of them wired...
My system is again stable... Something happened that the 5V dropped to 4.60... I had removed the 5v mod... Now I did the mod again to 2 and it raised to 4.71 under seti...
Everything stable again...
But ... I think I'll have to do it to the 3rd one to get to the 5.00v mark..
freshy98
03-04-02, 07:59 AM
stupid question maybe, and maybe also already answered, but with multiple mosfets do you take mutiple wires too or just one single wire and split them to the various mosfets?
Originally posted by Hoot
If you dont have access to a voltmeter, find someone with an 8KHA+ who does and have them determine it for you.
Hoot
Which brings me to this, for all to whom it's relavent:
res0r9lm
03-04-02, 03:57 PM
I measured the mosfet on my shuttle ak31 with a analog multimeter it's showing a little over 5 volts under load the same as psu is putting out so there isn't much of a voltage drop. mbm is reporting +5 volt line at 4.81-4.87 the only thing I can figure is multimeter isn't reading too accurate but since there isn't any drop I'm thinking my problem is I need a bigger psu right now I'm using a enhance 300w.
DaVolfman
03-07-02, 01:08 PM
Originally posted by ElectricMan
I checked out my system , I hooked up the multitester like this. one lead going to a blk wire from one of my drives and the other lead going to the center lug in the mosfet. I had 5.09volt's. I then played a music cd and put a dvd movie in and them ran 2001 mark and it went down to 5.02 .I am a electrician (Building's) but I think from what you guys are saying that is good and needs nothing, no mod like hoot did is needed. I have a Aopen Case and power supply 300 watt. What do you all think????
I've been doing the figuring too, and as best I can figure out the resistance can't be in the connector (at least not the most of it). This mod may do enough to work, but I think the effective resistance of the PSU itself may actually be the primary culprit. If it really is that much of a problem for some people, maybe they should shell out the money for a PSU from PC Power and Cooling. One of the nifty things their powersupplies do is flatten that voltage/load curve so it is always 5v.
Any way I had some Ideas of my own on the subject though I started them on a thread on AthlonMB (http://www.athlonmb.com/vb/showthread.php?s=&threadid=108300)
Anyone want to check my math for me?
DaVolfman
03-07-02, 03:52 PM
I think I just figured out the flaw in my reasoning. When the motherboard or other devices are trying to draw more power in most cases it is probably an act of flipping on a switch to power something creating anothe path for current and thereby DECREASING the resistance across the five volt rails. Since the resistance of the PSU and the cables remains the same this means that the devices alonf the five volt rails have become a lesser portion of the total resistance and thereby drop a lesser portion of the total voltage. This means that my logic that most of the resistance can't be in the cabling is false (sorry, wrong side of the curve).
Still my idea might work for those who are borderline and merely suffer from rebooting and such when other devices briefly draw more power. The tapping of a capacitor onto the circuit would help stabilize the line voltage and might prevent problems from all but the lenthiest drains.
My suggestion would be to try a number of things before this mod:
Remove some of the possible oxide layer from the power connector pins by disconnecting and reconnecting it several times.
Try the screwdriver trick previously described where you re-narrow the recpeticles for the pins on the connector to make the friction conection less resistive.
The Highlander
03-10-02, 03:11 AM
HOOT!!!!
I did the voltage mod to all my mosfets...
Now I am using a KR7A instead of a KG7.
The 5V at stock VCore was 4.95v now is at 4.84v under load with the modification... why???
MY PSU is delivering... but why dont I see it near 5v on the MBM5?
Originally posted by The Highlander
HOOT!!!!
I did the voltage mod to all my mosfets...
Now I am using a KR7A instead of a KG7.
The 5V at stock VCore was 4.95v now is at 4.84v under load with the modification... why???
MY PSU is delivering... but why dont I see it near 5v on the MBM5?
I think I'd get another opinion on what that resultant voltage is, like from a multi-meter. That is not what should happen. I take it you were thorough in determining which Mosfets to tap in to, or else you system would not be running. You did not say what the voltage was you read at one of the unused molex connectors. That would give you the voltage present at the output of the PSU, before it makes the journey to the Mosfets.
Hoot
The Highlander
03-10-02, 10:47 AM
5.05v on the power supply
on the multimeter it says 5. but Its a direct connection with the system, so... I think its a bit obvlious that it will show 5v on the multimeter...
Let me know!!!
Pardon me if these questions have been asked already- I looked thru the rest of this and the other PSU thread but I cannot find what I seek! Maybe I missed 'em- sorry if I did.
First- how low is your 5V line allowed to drop? I see folks with it at 4.9, 4.8, 4.7.... what is an acceptable drop, and what would be considered 'dangerous'?
Second- what are the symptoms of a too-low 5V line? If your 5V is too low, it's gotta have some effect on your system. Does it reboot itself, does it lockup alot, would it get 50+mhz more OC if it had the right voltage, what? Basically- how do you know for SURE that you need to up your 5V?
I just want to get this straightened out before I even think about modding my mobo. I'm not worried about frying it, I just want to be sure it is worth doing.
res0r9lm
03-11-02, 03:27 PM
Originally posted by Monster of Rock
Pardon me if these questions have been asked already- I looked thru the rest of this and the other PSU thread but I cannot find what I seek! Maybe I missed 'em- sorry if I did.
First- how low is your 5V line allowed to drop? I see folks with it at 4.9, 4.8, 4.7.... what is an acceptable drop, and what would be considered 'dangerous'?
Second- what are the symptoms of a too-low 5V line? If your 5V is too low, it's gotta have some effect on your system. Does it reboot itself, does it lockup alot, would it get 50+mhz more OC if it had the right voltage, what? Basically- how do you know for SURE that you need to up your 5V?
I just want to get this straightened out before I even think about modding my mobo. I'm not worried about frying it, I just want to be sure it is worth doing.
reboots or lockups are common with low 5v line. Don't know what you would call acceptable mine is 4.78 and I can't overclock any higher 1618 and if I raise vcore past 1.85 it won't boot into windows. As long as your happy with your stablity at whatever speed you have your cpu running then your fine but if your running into problem a low 5v line might be the cause. I'm sure I could go higher if 5v line was a little higher.
ok that makes sense, thx res0r9lm
my situation is this- my celeron 566 is super happy at 961, very slightly flaky at 975, and very unhappy at 995, even when boosting vcore to 2.05. It flakes out by rebooting or jumping straight to the W2K blue screen at odd times- it will run folding fine for hours then crash hard as you open notepad, that kind of thing.
Does this sound like a power supply issue? The C566 stepping I have usually runs out of juice right around there, so I am unsure if it is my CPU or PSU holding me back.
res0r9lm
03-11-02, 08:05 PM
sounds like a power problem to me.
Vfrjim1
03-14-02, 07:17 PM
Hoot, great article, just wish it could have helped me. My 5v at the molex connector is 5.5v, bios says 4.85v, measured at the Mosfets is 5.2-5.3v, looks like either my board is reporting a wrong voltage or something is messed up in my board, BTW, my board is a K7T Turbo V3 with a 650watt(peak) Enermax, also tried a cheapo Enlight 300watt, with VERY little difference.
Jim
Tismedt
03-17-02, 06:45 PM
OK Hoot I read all post on this thread and it seems to me if the ATX connector is the main culprit due to loose fit couldnt you just add a dab of thermal paste to each female connector to increase the contact? You compared this situation to a HSF and a CPU connection. Is my thinking wrong? Could there be side effects? It seems it would be worth a try. Let me know what you think.
Originally posted by Tismedt
OK Hoot I read all post on this thread and it seems to me if the ATX connector is the main culprit due to loose fit couldnt you just add a dab of thermal paste to each female connector to increase the contact? You compared this situation to a HSF and a CPU connection. Is my thinking wrong? Could there be side effects? It seems it would be worth a try. Let me know what you think.
Thermal paste is a heat conductor, not necesscarily an electricty conductor. Arctic Alumina is designed to be electrically inert. It's not supposed to conduct at all. Even if a conductive paste were used it might not be completely conductive and it could lower the connection's effectiveness.
Tismedt
03-17-02, 09:37 PM
thanks for response. I wasent sure about the paste being conductive or not. I seem to have low 5v reading for a 400 sparkle ie 4.87 @ 100%. It is also 4.87 at a molex connector. seems like I got a so-so psu for my $75.00.:( Oh well live and learn Thanks for the reply
res0r9lm
03-17-02, 10:20 PM
Originally posted by Tismedt
thanks for response. I wasent sure about the paste being conductive or not. I seem to have low 5v reading for a 400 sparkle ie 4.87 @ 100%. It is also 4.87 at a molex connector. seems like I got a so-so psu for my $75.00.:( Oh well live and learn Thanks for the reply
some of the psu have adjustable pots inside. I don't see why it no one has tried to solder a wire on back side of atx connection and than use a spade to tap into 5v on psu.
Originally posted by Tismedt
OK Hoot I read all post on this thread and it seems to me if the ATX connector is the main culprit due to loose fit couldnt you just add a dab of thermal paste to each female connector to increase the contact? You compared this situation to a HSF and a CPU connection. Is my thinking wrong? Could there be side effects? It seems it would be worth a try. Let me know what you think.
Not thermal paste, contact restorer/enhancer like DeoxIT.
Hoot
Tismedt
03-18-02, 11:38 PM
Originally posted by res0r9lm
some of the psu have adjustable pots inside. I don't see why it no one has tried to solder a wire on back side of atx connection and than use a spade to tap into 5v on psu.
i don't know much about psu' so i wouldnt know where to look for a adjustment pot. Is it inside? I dont see anything on the outside.
Originally posted by Tismedt
i don't know much about psu' so i wouldnt know where to look for a adjustment pot. Is it inside? I dont see anything on the outside.
Yep, inside. And if you crack your psu open to look around tread lightly, you can kill yourself if you touch the wrong thing while poking around.
res0r9lm
03-19-02, 06:58 AM
make sure you stay away from the capsitors b/c they store electricity. not all psu have the pots.
Tismedt
03-23-02, 05:48 AM
I think I will stay away from the psu completely. btw I put together a new rig w/ a enhance 300w and the 5v reads @5.1, 12v @ 12.0. Go figure huh. I have another enhance 300w laying around I have never used. I'll check it now. If it has the same 5v @ 12v would you recommend i use it instead of the 400w sparkle?
My rig has
1) cd-rom and cdrw
2) digidoc 5
3) 2 hdd
4) fdd
5) 4case fans (power by psu molex)
6) swiftech mcx hsf
7) gpu fan, mobo chip fan
AudiMan
03-23-02, 09:03 PM
Actually, I just ordered the Antec PP412X 400 Watt Power Supply, and apparently it has two pots to adjust the 12V and the 5V lines to your liking. All you have to do is to open the casing.
I guess that'll save me some soldering! :D
T'wolves
04-07-02, 07:14 PM
I might do this mod, but I want to verify the mosfets. I have an Abit KT7E and the mosfets are in the same arrangement as the KT7A in the pictures. So can I assume the mosfets will be the same? If nobody can verify that or tell me which do what for sure I'll skip it. My mobo and Motherboard Moniter 5 are both reading around 4.87 or so. If that isn't enough to go by I'll skip, too.
Larry Quinn
04-08-02, 12:20 PM
Good Job HOOT....
I may just attempt this mod....
I like your connector..is it just a standard male/female spade???
That gauge.?
Has anyone tried this on an Epox board???
Are the correct Mosfets the ones on the front of the board (HOPE..HOPE)..?
If it's the ones on the back it's going to be a ROYAL PAIN to perform!
And one last Q
Hoot....how about conductive contact Gel put into each wire's connector in the ATX plug??
Could that improve the connection to the point that the resistance is low enough to not be a factor??
res0r9lm
04-08-02, 02:06 PM
how about soldering from the backside of motherboard on atx plug and using spade to connect to 5v line? anybody tried?
maxima88
04-10-02, 10:46 PM
I'm gonna open my Antec 403 and see if it also has the pots for increasing my +5v and 3.3v.
My +5v is currently at 4.96~4.99 and my 3.3v is at 3.22
What are safe values to raise them to?
res0r9lm
04-10-02, 11:17 PM
Originally posted by maxima88
I'm gonna open my Antec 403 and see if it also has the pots for increasing my +5v and 3.3v.
My +5v is currently at 4.96~4.99 and my 3.3v is at 3.22
What are safe values to raise them to?
I would keep it around 5.05 but what you already have is good
I have a really strange outcome. I did the mod to one of the mosfets. The 5V reading is unchanged at about 4.8V under load but the vcore reading has increased from 1.96V (set at 1.85 in the bios) to 1.99V. I know the MSI mobos generally add 0.1V.
Insthink
05-20-02, 10:25 PM
I got an Epox 8K7A, my 5v is at 4.71, very low I know
Right now I cant check cause I dont have a voltmeter whatever its called. Ill get one from my friend tomorow, but I checked my mobo and only found 3 mosfets, 2 at top right near ram, 1 bottom left amongst the PCI slots.
I cant check the back because i got a Antec 1030 case. Will any mosfets link with a voltage near 5v do? Does any other parts of mobos can look like a mosfet and I screw up? Some mosfets have the middle pin pulled inside and some other have its middle pin extended over the board like the other pins on the sides.
btw, you showed a pic with a strange connector which u said u didnt know what it was for. I got one of those also unused, which wire is the 5v? and the 4pin molex, even if they are not used, does it matter if they are in serial (they are a connector not used but only because a part using the wire has an exit connection to use in serial) Or I have to take one going directly to the box without anthing plugged on it?
Originally posted by Tiger
I have a really strange outcome. I did the mod to one of the mosfets. The 5V reading is unchanged at about 4.8V under load but the vcore reading has increased from 1.96V (set at 1.85 in the bios) to 1.99V. I know the MSI mobos generally add 0.1V.
Here is a good test. To measure the voltage drop between the entire 5v run, stick the negative probe of multimeter onto the 5v lug of FET (I found it to be the base aswell, easier to hit) and the posivie 5v pin of an UNUSED molex. That will tell you the total voltage drop. If its something small, like .1 volts, then the source of your problem is either poor grounding of your motherboard (perhaps it isnt in a case? Or you have used insulated motherboard standoffs everywhere?) or you simply need to jack the output voltage of your PSU.
Just adjust the output of your psu, by adding some pots or useing the ones that is already there. Just watch the 12 volt line while you are at it. My 12 volt line is pushing 13 volts, which is stopping me.
Pyrotechnic
05-27-02, 12:37 AM
watch out for extreme low voltages, my 350 enermax was running a 5v line of 4.5v, and i get 1/2 melted 5v wires in the ATX cable and the ATX connector gets fused to my mobo, ruining it.:mad:
Now the first thing out of everyones mouth is going to be "YOU OVERLOADED THE PSU !," however this is not the case. If i did infact overload it, and overload protection circuit in the PSU would have been tripped and no damage would have been done.
My specualtion is that the mobo started drawing more amps because of the lack of voltage. It coudn't even run folding without crash. Also, the system was at stock speed at the time, since i was having some trouble with voltages running at 1500.
Yes i'm a dumbass for letting something like this slip by, and it cost me, but it really tells you about the quality of enermax psu's.:mad:
watch out for extreme low voltages, my 350 enermax was running a 5v line of 4.5v, and i get 1/2 melted 5v wires in the ATX cable and the ATX connector gets fused to my mobo, ruining it.
Now the first thing out of everyones mouth is going to be "YOU OVERLOADED THE PSU !," however this is not the case. If i did infact overload it, and overload protection circuit in the PSU would have been tripped and no damage would have been done.
My specualtion is that the mobo started drawing more amps because of the lack of voltage. It coudn't even run folding without crash. Also, the system was at stock speed at the time, since i was having some trouble with voltages running at 1500.
YOU OVERLOADED THE PSU !
Just kidding. What I believe happened is that your ATX connector became oxidized or the connections were simply poor, resulting in a whole lot of power with nowhere to go. This explains both symptoms: Your low voltage and your melted ATX power connector. How many pins burned out/melted? I would imagine it was only one or two since your +5V voltage was not abysmal... Just pretty bad.
This is also my first post here, so... Hello everybody. I post over @ Icrontic and my attention was drawn to the article by a newer article, etc. etc. Great job on the article Hoot. I'm employing it somewhat, but going to solder the inductor to the capacitors that are before the MOSFETS in order to minimize line noise on the +5V.
!-=sky=-!
06-07-02, 08:32 AM
ummm.....
how would you do this on a epox 8kha board........
i m..running 4.5x on my 5v rail line with a enermax 451watt psu
i dont have access to a voltmeter at the moment but i think i hsould buy one
thx
btw my computer setup is
t-bird 1ghz running at 1472 (9.5x155)
with
quantum fireball lm 30gb 7200rpm hdd
elsa geforce2 pro 64mb ddr video card
winfast 6x sound card
8 case fans
creative infra cdrom and an old hp cd writer
packratbob
06-08-02, 02:40 AM
Originally posted by !-=sky=-!
ummm.....
how would you do this on a epox 8kha board........
i m..running 4.5x on my 5v rail line with a enermax 451watt psu
i dont have access to a voltmeter at the moment but i think i hsould buy one
thx
btw my computer setup is
t-bird 1ghz running at 1472 (9.5x155)
with
quantum fireball lm 30gb 7200rpm hdd
elsa geforce2 pro 64mb ddr video card
winfast 6x sound card
8 case fans
creative infra cdrom and an old hp cd writer
woah 4.5 is pretty low man!! and isn't it cosing u stability problems??
Sky - I think the general opinion is that Enermax is not the high quality PSU it once was. From the posts I read pretty regularly, this sort of sag is almost expected.
You might try a different (better) PSU. Personally, I love the Antec I'm running. There are several brands suggested in other topics within this forum.
!-=sky=-!
06-08-02, 11:33 AM
Originally posted by packratbob
woah 4.5 is pretty low man!! and isn't it cosing u stability problems??
no....for some reason my computer is STILL runnning fine
also, since i live in hong kong i dont have as much selection of hardware as people in other countries do....and...i dont have money...haha
thx for the replies..but....i m still wondering on how to do this mod
i dont get most of the stuff in the article
Can someone please poast a pic of the right Mosfets to use on the Abit KR7A raid.
IceMan3928
06-08-02, 07:32 PM
what do you guys consider low? according to ASUS PC probe my 5v line is 4.919, is that ok?
I didn't have problems with my old power supply until the 5v dropped below 4.83v. Depending on what you have in the system, 4.91v should be fine.
flyfisher
06-09-02, 05:40 PM
My ps is only putting out 4.85v. Is there someway to connect this wire from the mofsets to the 12v line with a rheostat or potentiometer or something and adjust the 12v down to a GOOD 5.0v?
Did the mod and it worked great, voltage went up from 4.74v to the 4.85v the ps is kickin out, am now able to run at 160fsb, up from 146 max before!
HOOTIE HOOT:D
check out my psu stats...they suck don't they....300w fsb group
stats in this post.. http://forum.oc-forums.com/vb/showthread.php?threadid=96451
JFettig
06-16-02, 01:55 PM
anyone done this on a 8k3a? because im interested in knowing what ones it is so i dont have to take the time to figure out what ones it is...
JFettig
06-18-02, 11:44 PM
i did it!
its the 3 mosfets on the front of the board, now at stock voltage i get 4.99v, and at 2.2v i get 4.92, and my voltages dont move a bit...
JFettig
06-24-02, 11:04 AM
well i had a huge scare, and it seems noone is reading this, but one of cords came off the center fet, and almost fried my board....
The Coolest
07-01-02, 09:00 AM
Hey hoot, I had a thought, if your PSU isn't powerfull enough on the 5V rail then why not use the 12V? Just connect the +12V to a "converter" (forgot how its on english) to 5V and u'll be getting a constant flow or very stable 5V? Could this work?
freshy98
07-03-02, 01:14 AM
Originally posted by The Coolest
Hey hoot, I had a thought, if your PSU isn't powerfull enough on the 5V rail then why not use the 12V? Just connect the +12V to a "converter" (forgot how its on english) to 5V and u'll be getting a constant flow or very stable 5V? Could this work?
Couple of weeks ago that thought was already presented on the main page of overclockers.com. Forgot his name but he used 2 7805's for it.
But it's still a good idea :)
The Coolest
07-03-02, 07:39 AM
Yea, that's what I meant so the converted 5V would be good enough to use in this mod?
freshy98
07-03-02, 08:25 AM
yeah, in fact it's even better.
the 7805's compensate when the +5V drops.
so basicelly (yes I know, misspelled) it will always be kept at +5V.
the 7805 draws 12V and converts it to 5V, but never lower than 5V. mind though that you need a big heatsink for the 7805's 'cause they get real hot!!!
The Coolest
07-03-02, 04:47 PM
What if I put a little thermal paste and glue them to the top part of the case to use the chassis as a really large heatsink?
freshy98
07-04-02, 01:24 AM
Originally posted by The Coolest
What if I put a little thermal paste and glue them to the top part of the case to use the chassis as a really large heatsink?
hehe, I wouldn't do that if I were you!!!
the side connected to the heatsink is also under current.
so putting the back of the 7805 to the case would mean that you put 12V to the case ;)
really, use a heatsink and don't put the heatsink on the case!!!
packratbob
07-04-02, 06:05 PM
Originally posted by freshy98
hehe, I wouldn't do that if I were you!!!
the side connected to the heatsink is also under current.
so putting the back of the 7805 to the case would mean that you put 12V to the case ;)
really, use a heatsink and don't put the heatsink on the case!!!
i thought the back of those were equivalent to the ground??
freshy98
07-05-02, 01:42 AM
According to my dad, who is an electro engineer with a high grade in electronics, it's not the ground. I'm not really sure what exactly it is, but with a ground there will always run current over it. When I am back in my hometown next week I will scan a schematics about the 7805.
The Coolest
07-05-02, 03:28 PM
OK, thanks a lot for bringing that info up to my attention. I will not use the case as a big heatsink then :p
I'm not going to do this mod just yet, I'll have a couple of questions when I will be planning it tho.
res0r9lm
07-05-02, 04:09 PM
I just got a Epox 8k7a and I'm using a enhance 350w psu and my 5v line is 5.05 with my vcore set to 1.95 :D
packratbob
07-07-02, 09:04 AM
Originally posted by freshy98
According to my dad, who is an electro engineer with a high grade in electronics, it's not the ground. I'm not really sure what exactly it is, but with a ground there will always run current over it. When I am back in my hometown next week I will scan a schematics about the 7805.
i just check with a multimeter the middle pin is the same as the back of thoses. on some heat sinks they supply a earth to the sink as well so the bottom of the 7805 r deffinetlly ground
freshy98
07-22-02, 06:25 AM
Originally posted by packratbob
i just check with a multimeter the middle pin is the same as the back of thoses. on some heat sinks they supply a earth to the sink as well so the bottom of the 7805 r deffinetlly ground
Er, you're right. I have been mistaken with some things.
Anyways, I did it a bit different now thanks to my dad.
Look in the post below.
freshy98
07-22-02, 06:36 AM
This is how I did it on my MSI KT3 Ultra with thanks to my dad.
Will update with more info on the used electronics.
Here are some pictures of the final result.
http://members.home.nl/t.weustink/overclockers/mb1.jpg
http://members.home.nl/t.weustink/overclockers/mb2.jpg
The component used brings the voltage down to 5V @ 3A max.
It will only deliver the needed Amps thus not supplying 3A constantly.
http://members.home.nl/t.weustink/overclockers/mb3.jpg
Here you see the two connectors used. One for a yellow wire, and one for a black. Both used from the P4 connector.
freshy98
07-23-02, 09:36 AM
Right, here is the schematics used for my 5V mod.
I tried it last night and it's great! I have exaclty 5V, as expected offcourse, and it's holding under stress :)
http://members.home.nl/t.weustink/overclockers/schema5v.jpg
freshy98
07-24-02, 03:09 AM
Originally posted by silvershad0w
Great mod freshy98. My linear IC mod posted on the front page is still going strong, so this is a surefire way to raise your 5V line simply and cheaply. Along the way, I have had a few noteworthy suggestions from a few people. The one I like the most so far is using an adjustable linear regulator, i.e. LM317 to provide for a tunable setup, allowing you to fine tune the linear regulators output. This is an interesting way to go, because after the month or so I've been running this mod, the voltage can drop as low as 4.92 volts.
Aside from the LM317 approach, Providing adequate cooling should be the highest concern for a mod like this, which I see you've taken care of! Just make sure that you keep these IC's within thier operational limits, as stated earlier, they do become quite hot.
According to my dad, who has a lot more knowledge about this than me, the LM317 should be better along the way. Can't confirm this yet as it only runs for 2 days now. Will try to keep you posted about this in a month or so.
As for the cooling. I have a 120mm Papst 4412 in the front of my Vapochill case sucking air in. Thus blowing it over the heatsink.
Will try to make a picture of it this week and post it here.
freshy98
07-26-02, 12:58 AM
I used a guide last night to tweak my BIOS. It's a lot more stable now! Running at 1600MHz right now. The 5V mod holds very nicely as expected :) But I am wondering if the -/+12V and -5V are allright.
Can someone give me an answere wether these are OK?
Btw: The idle temp was taken last night 10.30PM and the stress temp was taken 6.00AM this morning. The VCore is at 1.85V (BIOS) but even the BIOS says 1.9V in the Health Status.
thnx!
freshy98
John Jr
07-31-02, 09:40 PM
Wow you got the news down. I love it but I'd burn my 'puter to the ground if'n I got near it ith X-tra wire and stuff. Cool that the info is there dude, way outta my league for a mod. Hopefully one day I may be confident enough for this kinda mods. Good job...
freshy98
08-01-02, 02:29 AM
Originally posted by John Jr
Wow you got the news down. I love it but I'd burn my 'puter to the ground if'n I got near it ith X-tra wire and stuff. Cool that the info is there dude, way outta my league for a mod. Hopefully one day I may be confident enough for this kinda mods. Good job...
He he he well, the idea it self wasn't mine. I'm not that good in electronics, but my dad is.
Soldering is not a problem. really. just follow the schematic real close and it should be no problem.
After 2 weeks it still runs very nicely! :D
wildfrogman
08-03-02, 09:27 PM
Hey, uh was i the only one.....but on your MBM5 reading your voltage on your 3.3 was uh LOW 1.52 volts!?:eek: i think that could be a big problem right there. What is up with a 3.3 rail being THAT low and uh how can your computer stay on?
hey guys sorry interrupt but I was wonder if you could help me out a bit. I have xp1700@1716mhz, fsb156 and 3 80 mm case fans @50cfm each. I am running this on my standard 300w psu and I just checked my 5v line and it is at 4.7v. The funny thing is that it is completely stable at this speed but I am kinda a bit worried about that lower voltages
v.core: 1.89
aux: 0.06
+3.3: 3.52
+5: 4.7
+12: 12.28
-12: -12.25
-5: - 5.04
I was thinking of getting enermax true watt 450.
Any I deas?
(Cheaper is better)
Thanks Benno
Benno, Enermax is not what it used to be. Their quality seems to have declined a bit. There are numerous threads in here about Enermax PSUs having problems with sagging 5V rails.
Personally, I like Antec. There are a bunch of threads that mention other good names (Sparkle, RaidMax, etc.). Please search and read. Not being rude ... okay ... not MEANING to be rude. :D
res0r9lm
08-04-02, 08:17 AM
I have an 350w enhance psu and epox 8k7a 5v line is 5.05
freshy98
08-05-02, 05:46 AM
Benno, I have a Enermax 430W and I still had a low +5V.
After using Hoot's approach on the +5V mod I used My dad's approach. (after examening Silver Shadow's)
It works great.
I purchased a 2200+ last saturday and at 150MHz FSB it's still a very nice +5V :D
Skiing Squirrel
08-06-02, 05:48 PM
My +12 is at 12.30. Is that good? Or bad?
tseongsun
08-12-02, 03:04 PM
I just did this mod on my kk266R and it works great! Thanks Hoot. Anyways for those who don't have a multimeter but want to mod this same motherboard, the 5v mosfets are the 3 on the bottom row right above the AGP slot. (note: I have the kk266R and not the KK266R+ I'm not sure if the spots are the same on the + version)
x x x
x x x <--5v mosfets
==== <--- AGP slot
pcwhisper
09-07-02, 04:40 AM
Hy everybody,
i'm kinda new here. :)
I find this thread is very interesting but i dont see anybody who did this on a KX7r . I dont have measure equipment so i was wondering if the location of the 5Vmosfets is the same as of the KT7A. (forgive me if my english is bad)
I have a SamCheers/Topower 470w triple fan psu and with a Vcore of 1,95 running toast the 5 V goes down to 4,60 and system freezes.
Thanks in advance:cool:
John Jr
09-07-02, 08:53 AM
I would like to welcome you here. I can't offer any help on this topic bit I have learned alot here.
pcwhisper
09-09-02, 09:24 AM
Originally posted by John Jr
I would like to welcome you here. I can't offer any help on this topic bit I have learned alot here.
Thanks John Jr,
seems like you where the last persone alive in this thread. :confused: I'l take it yo another one:)
BigMofo
09-22-02, 05:22 PM
Woot Woot, Hoot..you're the man. after the mod, my 5v jump from 4.67 to 4.91:D
wolfman
10-04-02, 05:45 PM
I used award 6 bios voltage reading on 8k7a+ board and tweaked the control in my generic psu to 4.99v. But remember increasing the it to 5v will raise your 12v line. Most psu's feedback from the 5v line. It was 4.79v, computer seems more stable.
Doc_Skurlock
10-26-02, 03:44 PM
I'm kinda worried about my +12 volt readings. I'm using MBM5 and my +12 volt readings are 0. My -12v readings are -11.77volts. Everything else is good to go. My +5 is a steady 4.99 and my 3.3 is right at 3.30 -/+ .4. CPU volts are steady at 1.75. Any idea why my +12 is not showing anything?
I have a Intel MB, D845BG. Guess I should get a new mobo?
everything else is in my sig.
*EDIT* Using my Enterprise System Information, I apparently don't have a +12v sensor. Oh well. Guess I need a new mobo.
It was brought to my attention that the original pictures at the beginning of this thread were lost. I have restored them. It was not easy.
Hoot
BoonieRat
11-05-02, 11:26 AM
Hoot!, I've been told that soldering a wire to the middle leg (the stubby~cut off one?) actualy bypasses the mosfet!,not just the ATX/traces. can you elaberate?...
I got different readings when useing the case, and the black wire on a molex next to the red/5V wire (is it -5V or just earth??). should the mobo be earthed to the case better through the standoffs?. I'm asuming it's a bad earth causing the difference in readings?...
On the KX7333R it's the three bottom(nearest AGP slot) mosfets that read 5V at the middle/stub pin. the ones above read whatever the V~Core is set to?, is that right?...
a1cnolan
11-21-02, 03:20 AM
Okay Hoot, i don't know if this is a silly question(nor should i care, but i do,....anyway)
i was running the multimeter tests on my mosfets last night. i found that when i do 180X9.5 (my limit) my 5v drops to 4.89 on the bottom mosfet .4.90 on the middle and 4.91 on the top. so are these run in a chain like mosfet capacitor mosfet capacitor mosfet so on.like the like bottom feeds middle feeds top?is this an error checking device?anyway if this is true why not connect the 5v to the top one and have it go straight through ?do you do the middle so the voltage can be checked by the last one? and if that is true why not do the bottom one to ffed the capacitor that feeds the middle one? am i totally wrong on this one? i know i need to do this mod if want to go any further or a new power supply.... i could use some help on these questions. kinda like measure twice solder once...also on the first version of the mod i couldn't really see where that was soldered to. wouldn't i want that in the front like on the version you use? have you really seen any added benefit between the first and second version?if so what effects did you have.....the more you know i guess
here is an aprox of my layout
m m
cpu
m
m
m
m atx connecter
(m=mosfet)
also when i check my voltage with the multimeter i do my reading by touching the left and right leg of each mosfet
the power supply is a generic 320watt with no adjustable pot
in page 5 of the guide i see only one wire and you say to connect it to the one in the middle also by looking at the next post attachement can you tell me wich one you would consider the one in the middle
http://pages.infinit.net/ataxy/mosfetvoltage.jpg
ok i tried something i filed my atx connecter with a stuff called ox-guard it is a sort of grease that is made to "Guards againts oxidation, improves conductivity and produces a cooler connection." so since i applied it to my connecter my +5 line whent from 4.89idle to 4.92idle and from 4.78 under prime95 to 4.83 under prime95 torture test(sometime going as low as 4.81 but still better than the 4.76 i saw before) all of those reading where took using speedfan and and are the same that i get when i check the bios
i tried to connect my 5v to each of the 5v mosfet but it did not change anything is it possible that i should connect it to the three of them
RonnieG
12-05-02, 10:23 AM
It is not advisable to bypass the input inductor with 5V booster cables. The input inductor is the one just right after the ATX connector, NOT the ones in the output phases. So if you have a 3-phase system there will be 4 inductors: 3 tightly wound ones and 1 loosely-wound one -- the input inductor. And it's there for a reason...
This is my version of the 5V Booster Cable:
Get some thick wire and solder it BEFORE the input inductor:
http://www.theforumisdown.com/uploadfiles/1102/boost1.jpg
Make some sort of connector, preferably a beefy screw-in type. I don't trust those you just pull in and out
:
http://www.theforumisdown.com/uploadfiles/1102/boost2.jpg
Solder another wire to the PSU board's 5V terminal:
http://www.theforumisdown.com/uploadfiles/1102/boost3.jpg
so do you connect yours the the loose one and the inducter is solder at two point on the board to wich one do you connect your wire
RonnieG
12-06-02, 09:23 AM
I solder my wire on the "leg" of the inductor connected to the ATX connector. I recommend you solder it directly to the inductor's "leg", not to the PCB traces or ATX connector pins.
BoonieRat
12-13-02, 07:01 AM
RonnieG!, is an 'inductor' the same as a 'ferrite ring'?, they are responsible for 'cleaning' the current are'nt they?. what kind of increase on the 5v rail did you get from this mod?. I have a lousy ATX conection and am thinking of soldering wires directly to the ATX pins on the back of the mobo and using a 'block connector',(a 240V~AC 3AMP type), to bypass the ATX connector. the connector is a bigger loss than the mobo traces IMO, if I 'jiggle' the ATX connection I get fluctuations on the 5&12v lines, my 5&12v lines have both dropped over the last couple of months, 4.84~4.76v & 12.15~11.96v respectivley, this is an Enermax350watt PSU on an Abit KX7R...
shadowdr
12-20-02, 08:48 PM
a little something i learned a long time ago dring a truck.we use a pigtail to connect power from the truck to the trailer.over time the male ends from the trailer corrode and have poor quality connections.to remedy this we put a dollop of white lithium grease on the female end of the plug and force some down inside.it acts somewhat like arctic silver on a heatsink and makes for a much better connection also protecting the metal connectors from corroding.since i allready use it on the atx connector i can't say it increases the flow of electricity but for those of you that cannot do a 5 volt modd it might be helpful.
RonnieG
12-26-02, 07:22 AM
Originally posted by BoonieRat
RonnieG!, is an 'inductor' the same as a 'ferrite ring'?, they are responsible for 'cleaning' the current are'nt they?. what kind of increase on the 5v rail did you get from this mod?. I have a lousy ATX conection and am thinking of soldering wires directly to the ATX pins on the back of the mobo and using a 'block connector',(a 240V~AC 3AMP type), to bypass the ATX connector. the connector is a bigger loss than the mobo traces IMO, if I 'jiggle' the ATX connection I get fluctuations on the 5&12v lines, my 5&12v lines have both dropped over the last couple of months, 4.84~4.76v & 12.15~11.96v respectivley, this is an Enermax350watt PSU on an Abit KX7R...
Yes the inductor is that ferrite ring with a fat wire looped around it. It's also called a "choke".
Solder your wires to the inductor.
BoonieRat
12-28-02, 02:00 AM
Cheers peeple!, I was especialy glad to find out the name of a grease that conducts!, trying to find something like that (specific/brand/product) for improving connections is impossible where I live...
PsYcO CyBrArIaN
02-13-03, 02:30 AM
Direct gold connections soldered(smelted?) make the case come apart, not the board and its components. Hooot you know your shiznik!:)
DiaperGravy
02-24-03, 03:42 AM
In response To (pcwhisper)I find this thread is very interesting but i dont see anybody who did this on a KX7r . I dont have measure equipment so i was wondering if the location of the 5Vmosfets is the same as of the KT7A. (forgive me if my english is bad)
I have the KX7333R as well, and the 3 Mosfets your looking for are the one's located on the bottom row... these are your +5v one's to mod!
Good luck! It Fixed my problem~was @ 4.81+/-.02 Went up to 5.01v+/-.01
DpGravy says> Thanks Hoot!!:D
Hi i have found this post very interesting and helpful. On my board i am experiencing slightly out rails 3.26, 4.87, 12.16 at stock clocks and voltages. On my board the atx connector is loacted to the left of the board just above the agp slot which made good cabling and airflow hard. This combined with the dyson like sound the twin fan psu made me move the psu out of the case, now the atx cable comes out a hole just under the exhaust fan and the psu only requires 1 7v fan sucking out the top of the psu case, this gives me much better temps and cabling. Before i done this my +5v was 4.91 and 3.3 was 3.26 to 3.31 and 12 was 12.22. I obviosly have encounered resistance by increasing the cable size but i did use good quailty copper speaker cable. My 3.3, 5 and vcore have fluctuated up to 0.3v either way even before i relocated the psu. I have measured the 5v rail with a multimeter under full cpu load and it is 4.97. Would u guys recommend doing this mod? Where is a good place to connect the ground on the multimeter to when measuring the mofsets? a black psu wire? or a mobo fan connecter ground?
Thanks
DiaperGravy
02-24-03, 04:03 PM
I have measured the 5v rail with a multimeter under full cpu load and it is 4.97. Would u guys recommend doing this mod? Where is a good place to connect the ground on the multimeter to when measuring the mofsets? a black psu wire? or a mobo fan connecter ground?
Hey, 4.97v isn't bad at all, but what the hell? the mod can't hurt it. As a matter of fact, it can only help as far as I can tell :D.
You can either use the ground wire = Black wire next to 5v red on a free molex connector, Orrrr you can use a ground basically anywhere on the case, I use on the case upper left mounting screw of the power supply because this is the com ground for it :D
But it doesn't really matter just as long as you ground somewhere, and take your measurements while your system is ON :p
Hope that helps a bit!
DpGravy
Damn. to think, my '70s era analog multimeter just broke. Well time to get up with the times! :cool:
Very informative article hoot-- I always had the feeling that there was other resistance going on, just not the PSU being crappo dappo for undervolting the +5v rail. :)
I have an enlight 230W "ultraquiet" ( atleast it seems so.. very quiet cpu ;) ) ..
Core0/1 voltages: 0.3 / 1.06 ( this is off because of slotket )
+3.3 = 3.47V, +5 = 5.11V, +12V = 11.55V
-12 = -11,47, -5V = 4.56V
Im mildly worried about the +3.3 and +12/-12/-5 voltages being off by so much.. tho, the system is very stable..
Should I be worried in the long run?
(have already run the PSU for ~2 years now tho :) )
heavy666
08-20-03, 02:01 PM
I know that its a stupid question but in my 8rda+ 1.0 where should I put the +5v cable?
witch morfet should I use?
thanks
ver1.000001
09-16-03, 05:53 PM
Awesome mod!!!
I just did it on my ECS K7S5A. I also used this method for the
3.3V line, I used the ram and agp voltage regulators for this.
Before,
idle->load
4.89->4.80V
3.20->3.00V
After,
4.96->4.91V
3.22->3.12V
It's better then before but still not what I got with the loose cable. What I had was 5.02->4.99V, 3.28->3.26V. I don't really care for the 5V as it is pretty well in spec but my 3.3V I wish I could improve. I'm feeding 2 seperate extra 3.3V lines, one going to the agp voltage regulator and the other to the ram voltage regulator.
Cemal Gurel
09-23-03, 06:17 AM
Hey Hoot,
There is another way of increasing the real power of (+) 5Volt line. But you need to open your PSU. The capacitor connected to (-) pole of (+) 5 Volt line has a 10Volt 1000uf to 4700uf capacitor. If you increase the uf of the capacitor, expecially with a low ESR one, will do the trick. Low ESR capacitors will allow lesser resistance, and therefore allow higher Volt to pass on them. (As you know caps can not allow all the power on them because of the space between the aluminum foils.) If you can not find a Low ESR one, incresing the uf value can also allow a bit of lower resistance.
If you replace 1000uf 10Volt cap with a 3300uf 10Volt one (normal, non-Low ESR), can possibly gain 0.2Volt. If you replace with a 3900uf or 4700uf Low ESR one, can gain upto 0.5Volt. But do not hope more than 0.5Volt!
Actually, your mod is very very risky! Your motherboard may have diodes before the mos-fets, and thus can cause a bit of lover (+) 5Volt to pass in to the mos-fets. Some other motherboards may have bobbins before the mos-fets too. You are by-passing your board manufacturers specifications. This is so much risky! I DO NOT RECOMMEND YOUR WAY TO ANYONE.
If you are uncomfortable with (+) 5Volt value that you see on your hardware monitor, what about the success of the hardware monitoring IC (the chip) on your motherboard? Maybe you have more than (+) 5Volt from your PSU?
Do not regard&believe to those hardware monitoring tools too much, sometimes they can give lower fan RPM, such as 2000RPM instead of 3000RPM etc or vice versa. My Asus Probe Hardware monitor also senses my Vantec Aeroflow VA4-C7040 misleding, about 8000RPM, instead of 5600RPM stated on the specs. I was nearly believing to 8000RPM because of the noise the fan generates, but obviously it is not. (38.0dBa is too high for a fan.)
Your motherboard guarantee is also off as you know. Finding a PSU is easier than motherboard, in case of death, when playing moddings on them. If you kill your 2~3 year old motherboard you may not find a working one and therefore you may pay for a higher&different CPU, Ram and Graphics card such as not a 2x AGP one.
PSU guarantee is ****-off because no one really likes to repair them, or believe their fixing within the guarantee. Motherboards have better repairing successes than the PSU's, but if you can&only carry them to their original repairing services and it takes a while. So if you blow the PSU can easily look for a newer one, but the motherboard issue is not so easy. Therefore do not play with your motherboard because of a little nasty bug at (+) 5Volt reading...
Cemal Gurel
09-23-03, 06:28 AM
By the way, you can do that to 3.3Volt line using 6.3Volt caps upto 4700uf, and 16Volts for 12Volt line upto 3300uf. The administrator automatically disabled my f**k-off above but I am sure you know that the gurantee of PSU's are untrustworthy.
threeputt
10-14-03, 03:29 PM
Great read Hoot!
Question. If my voltage reading at the MOSFET at full load w/ 2.0V Vcore is 4.4V, but my molex reading is 4.9V, would this mod help?
threeputt: that's a big drop :o yeah, it'll help.
Cemal: all capacitors have infinite resistance (or at least those that work) and no current passes through them. The capacitors on the DC side are solely for keeping the voltage steady.
(edit) I don't like that smiley up there...
vulcan5d
11-11-03, 12:36 PM
Server motherboards do the same thing as described by this mod, that's actually what the aux connector is for. I wish desktop PC's had this connector aswell, would have saved the extra work :).
I was wondering, wouldn't connecting ground also be recommended to improve stability because I'm seeing a pattern here. 1 aux connector has 3 ground, the 2nd aux connector has 2 ground and each molex connector has 2 ground aswell. I mean why would you do that if it didn't have a purpose?
What do you guys think, wouldn't it also be a good idea to solder ground wires to those voltage regulators?
asw7576
12-24-03, 08:39 AM
I think this mod is lame.... and people who already done this mod are dumber too. Sorry to say that .... but I have to straighten up the truth: This mod is not the way to increase 5 volt.
When you get lame 5V on monitoring program or bios hardware monitor.... all you need to do is increase 5V line, therefore you need to open the PSU and find something to modify the 5V output. Maybe by adding resistor and play with resistance value.
There are some website discussing about PSU modification.
Here:
http://www.xtremesystems.org/forums/showthread.php?s=&threadid=16050
http://www.xtremesystems.org/forums/showthread.php?s=&threadid=14059
http://www005.upp.so-net.ne.jp/octo/pc/
Here the proof: My PSU
I don't usually respond to posts like this. A Wise man from your neck of the woods once said, "It is better to be silent and thought a fool than to open your mouth and prove it."
The voltage drop across the ATX connector is a well documented phenomena and my work-around mod has stood the test of time in a lot of overclockers PCs. Ohms law is not open to debate. We all live with it, sometimes we suffer under it.
Your advice of turning up the +5V source to offset voltage drop during times of high current consumption associated with high Vcore settings is a dangerous one. There are many devices in a PC that run off of the +5V rail. If you turn it up during times of high current draw to offset its drop, what do you think happens when the current draw drops, such as during idle? That +5V goes up, sometimes to a dangerous level, depending upon how high you adjusted it to under load. A spiking high +5V rail can kill devices that share it with the Vcore regulator circuit.
Sheesh!!!
Hoot
larrymoencurly
12-26-03, 12:37 AM
How does adding a resistor in series with the +3.3V sense line increase the +3.3V by any significant amount? Because inside the PSU there should be an approx. 50 ohm resistor between the +3.3V lines and the chip that senses the voltage, to keep this voltage within range in case the sense line breaks.
PaperKat
01-02-04, 10:14 PM
Hi. I'm a senior member of the K.I.S.S. club and also I have this problem w/ +5V dropping below 4.44V when in full load, 4.52 otherwise. Would not be much easy to attack this problem at the very source instead of sticking wires onto the poor motherboard? I've got at the end of my Molex conectors -- any one of them -- a beautiful and hard-rock stabile 5.02V so I know the problem resides in the ATX conection between PSU and mobo. Also I disagree with Hoot, the cabel itself is not responsible for THAT MUCH voltage loss. I say let's concentrate on that weak link, the male-female conector. I'm takeing a wild guess here, but... how about lube it? You know, filling the gaps between those pins with something that will maximise the contact, thus reduceing --- if not eliminating -- voltage loss? Maybe something based on silver??? Is my idea crazy or what?
Oklahoma Wolf
01-02-04, 11:18 PM
Originally posted by PaperKat
Maybe something based on silver??? Is my idea crazy or what?
In a word, yes. Not being sealed for that sort of thing, any conductive compound you put in there will eventually (depending on viscosity) find its way down around the motherboard connector's pins, get onto the board beneath, run down and find the other pins, and kapow.
Your 5v is some cause for concern though... I would check the ATX connector for indications of overheating, and if anything looks fishy the motherboard's voltage regulators may be failing. It's not too common for Molex connectors to fail - it's usually a symptom of a problem elsewhere in the circuitry causing too much current draw. However, if all seems ok and you can run the system perfectly fine and stable despite the low readings, you likely have nothing to worry about - my own board reports 4.50 at times under a heavy o/c.
sangram
01-30-04, 01:54 AM
While the mod itself makes a lot of sense, regardless of whatever anyone may say, there is a complete filtering circuit that would be in series with the MOSFET to ensure that the power going in is clean. Soldering directly to MOSFET will bypass those circuits, and unless the power supply is a guaranteed source of whistle-quiet power, this may be dangerous to some components. Where I live there is not much difference in the price of an Antec 430 Solution series and an A7N8X deluxe, I would rather void the board warranty than splurge on a new PS. But that's just me, so I'm gonna go all the way.
I would look at adding further connections much in the way we see hard-core audio amplifiers constructed, specially those in the 70s and 80s.
What I would (and will) do is three things:
1. Improve the ATX connector itself. This would involve contact cleaner and slightly re-crimping all the connections.
2. Create parallel path to the filtering circuit using Hoot's method, but only till the ATX connector/filtering components and not directly to the MOSFET. Each parallel path shall also be self-wound to minimise EMI generation in its path. This can be achieved using ferrite beads and twist the wire a few turns through them in the middle. The whole thing can be potted with epoxy or silicone sealant to keep it insulated.
3. Create another path from the input filtering till the regulators, if deemed necessary (the inductors may be very close to the MOSFETs). We have seen the use of solder on PCB tracks to improve current-carrying ability. A small solid-core wire can be also laid out onto the track before soldering to further enhance the capability of the path.
The anally retentive (and I am not one) can use wire wrapping instead of soldering, as soldering multi-strand high current wire results in extremely poor connections electrically, compared to direct physical contact as experienced with screw terminals such as used in cars (that carry hundreds of amperes) or wire-wrapping.
Let's see how far the mod goes. Under OC (2.2 GHz) my 5 V line drops to 4.5 volts. My AXP definitely uses the 5V line, not the 12V...
sangram
02-04-04, 02:07 AM
****Disclaimer****
I AM NOT RESPONSIBLE FOR ANY DAMAGE YOU DO TO YOUR PC OR PERSONAL SELF DUE TO THE ATTEMPT AT DOING THIS MOD. iT IS COMPLETELY AT YOUR RISK
*END OF DISCLAIMER*
OK here's the lowdown. Voltages @2.37 GHz running Prime95
Before: 4.46 volts (not a typo) leading to crash
After: 4.84 volts
Voltages@ stock under Prime95 load
Before: 4.67
After: 4.84 volts
The cons first: You are going to have a very high 12V after this mod, if your supply uses one regulator for the 5 and 12 (most PS designs have this limitation). Mine has been to 13.85V under heavy stress, as the regulator pumps up the voltage to compensate for the load. It is important to have a good PSU for the mod, as you can see. Polishing turd PSUs like mine will make it smell less, but it is still a turd.
On to the mod itself. Like Hoot I have robbed that little connector of its 5V line. It is not the P4 ATX connector, which has two 12V lines. This has one 5V and two 3.3 Volt connectors. I assume it is for server class boards or something. The robbed 5V line looks to be of significantly better guage than the ones going to the ATX connector.
The ATX connector gets hot during operation. Very hot. So recrimping those connections is a must. And there is little worry of overcriming. The males on the board connector are, umm, well, shaped to penetrate. a small sharp tool (I used a reamer) on the females jammed between the plastic and metal and forced in, does the job nicely enough.
On the mobo itself, one notices that there is a good drop of about 0.2 volts (!) between the ATX connector and the MOSFET. This drop is due to the board itself, and is the rationale for the mod. What I have done different from Hoot is not soldered to the MOSFET as it would bypass the on-board filtering.
Which, in the case of the A7N8X DLX, is wonderfully located. There is a inductor (coil, choke, filter) just above the fan connectors near the AGP slot. There are three capacitors next to it, followed by the two MOSFETS comprsing the first phase of the regulator. We need to wire our 5V line to this inductor.
I first attempted soldering on the PCB, but aborted that quickly. There is little room and solder balls floating about on a motherboard do not excite me. In my case the inductor is mounted vertically, presenting a nice smooth long leg for soldering. The inductor has two pins, one which canot be seen (and does not matter for this case as it is the 'output' leg of the inductor) and the other can be seen as a long stiff wire going to the board.
You will need to scrape the enamel off the inductor leg, or solder will not grip it. Then solder the bare wire to it. Use grippy tape to hold the whole thingjigamig together, so that loose wires or a bad solder job do not fry your board or other components. This is very close to the AGP slot and a Radeon 9800 Pro will not like a 5V line coming into contact with the PCB side of things.
I have now my monster power supply coming in - dual fan custom built PS with 330 watts on 3.3 + 5V, 30 + amps on 12 V, individual regulators per rail, separate low-guage wires for 5V lines for hard soldering - I have purposely kept away from total rating but it shoud be around 600 watts. I have a manufacturer building this for me specially.
Dragonprince
04-09-04, 08:36 AM
I am just wondering how many people this thread has killed or seriously injured :eek:
If you have worries about clean power to your system components I recommend two easy steps. Buy a high quality power supply and buy a high quality motherboard..... "You get what you pay for...."
mustrum
09-10-04, 04:07 PM
I am just wondering how many people this thread has killed or seriously injured :eek:
If you have worries about clean power to your system components I recommend two easy steps. Buy a high quality power supply and buy a high quality motherboard..... "You get what you pay for...."
I have an Asus a7n8x Deluxe 2.0 and an enermax 431Watt PSU with extremely decent Amps on the rails.
My 5v rail dips down as low as 4.45v!!!!!!
:bang head
Doulton
12-06-04, 08:17 PM
Hello, So I'm having a unique problem, currently my PSU is 350W, I recently added a Radeon 9800 Pro and my current -12V is at (-13.4)-(-13.5) I understand this is bad considering most people here are freaking out about some small varianences of Voltages. I figure a new PSU is necesary, but could you tell me what uses the -12V ? Thanks...
maybe we can make a sticky thread about "Too many of you are too worried about -v rails! " :) ;)
ok, if -v rails are really used by some device/pc components than i wrong to believe that -v rails are useless in the atx/modern system...
maybe some tech savy people here can explain more details about it...
Tw3ak3d
05-02-05, 03:32 AM
Mine used to drop also , I modded the line to stabalize it and a pot to increase it , no it drops no more than .05v ever ..
Banthur
07-08-05, 03:15 AM
Will have to try this...
Nice, informative guide. Im gunna have to try this. :)
ochungry
02-23-06, 04:55 AM
Aren’t those 5V lines from ATX power go through a series of capacitors(and possibly the chokes) before connecting to the mosfets? and if so, wouldn’t the mosfets encounter increased ripple and noise interferences, that might damage the core regulator, and the caps and resistors interacting with it?
Just wondering. I am seeing a few small caps along the 5V path before connecting to mosfets. I wonder why they are there and what would happen if by passed doing this mod.
Thanks
They are there to smoothing the input load for the mosfet (vrm). If they fail on job, mosfet with get an excessive ripple input and it will get hot as its inefficiency rise..
As for the inductor (coil/chokes), they store the energy, and functioning the same as capasitor.
Failing caps on PSU can also make the output ugly, and the mobo will get an excessive input. cmiiw..
Beware of the caps condition in your PSU & motherboard. See this forum : www.badcaps.net/forum
so let me get this all strait
first to test your psu you check an unused molex on the +/-5v lines
then you check on the mosfets using the middle tab and a ground point as close as possible to the mosfet? such as a mounting screw?
then once you find your mosfets that get the +5 you solder the +5 volt wires onto the center tab and that is it
or did i miss something?
Hi Hoot,
I have a suggestion regarding your original post, here.
Have you heard of this stuff called Stabilant-22?
SUPER CONDUCTOR
"If you've ever had to track down an elusive bad or intermittent connector-a computer card connector, an IC socket or even an RCA jack-Stabilant 22 is for you. The slightly 'goopy' liquid is sparingly brushed onto connectors or IC pins (or whatever) and becomes conductive when 'activated' by an electric current. The manufacturer claims S-22 'provides the reliability of a soldered joint without bonding the contact surfaces together.'
Illustration of bottles of Stabilant 22/22A If you're worried about shorting -out adjacent contacts with this 'liquid conductor,' don't be. S-22 is only conductive across a very narrow gap: The contacts must be in physical proximity for Stabilant 22 to work. (To prove the point, the manufacturer's trade-show-computer, housed in a see-through plastic case, has its mother board totally submerged in S-22!) Winner of Byte Magazine's Product of the Year Award, S-22 has for some time been the audio and broadcast engineer's secret weapon. It's useful from 'dc to daylight,' and one application lasts for several years. S-22 has thousands of uses in Amateur Radio, and is now available from the manufacturer in smaller, more affordable amounts. (Yes, it is expensive, but a small vial will probably last a long time. My miniature sampler vial has already 'tamed' an intermittent RTTY terminal plug-in connector and restored function to a 'potted' RF converter module that would otherwise have been unfixable.) Contact the manufacturer for information and application notes."
http://www.stabilant.com/llsting.htm#rrevv01.htm
I had found about this long ago when a distributor back in the late 80's/90's would dilute it, and repackage it and called it "Tweek". It was sold to the audiophile market. That distributor and product no longer exists, but the real manufacturer does, and apparently they are still selling the stuff mostly to professional industry like auto and emergency medical. You might check it out.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.