im burning pots on my vtt mod any ideas??

[flapperhead]

as posted previously ive just finished up hipro's 5 in mod for the ic7.. everthing went well. however, i have burned up 3 potentiometers on the vtt mod. soon as i up the the vtt voltage so it is 1/2 of the memory voltage, i burn up the pot.. i cant get a good reading on the miliamps cause the damn thing burns open soon as i hit about 1.5v. any suggestions??

[Hell-Fire -XS-]

Interesting.

I heard the same thing from DAOS after modding his board.

[flapperhead]

i have discovered that i have to run the pot at 1 ohm or less to get the vtt up to 1.5-1.6 that causes a tremondous amount of current flo thusly heating up the pot and burning it.. hey hellfire can u give me the ohm reading on ur pot when the vtt is around 1.5 1.6? i suspect i have a high resistance comming from the vtt wire feeding the pot. therefore the only way to get the vtt to respond is to run the pot at almost a dead short/ thus causing it to burn...

[sandman001]

What about trying like 3 of them in series with each being 1/3 the resistance?

I'm not sure it would work, but it could alleviate the heat problems.

Or heatsink one of them.

[flapperhead]

thnx for the input, but unfotunately it would either burn all 3 or just 1 then it would open up the circuit. i think i may have a bad connection back toward the vtt voltage(too much resistance) so the only way i can get the vtt sencor to respond is to run the pot at almost a dead short which of course makes the pot a lightbulb. thats why i wanted hellfire togive me his ohm reading toward the vtt mod b4 it gets to the pot.. flapper...

[Hell-Fire -XS-]

You know, one thing worth thinking about maybe be to consider finding a way to add a pot to lower the voltage on the 3.3v rail so that its around the Vtt that you want and feed that directly to the Vtt pin on the dimm slot??

Another thing to consider is to make it so that Vref = Vtt or Vdd = Vtt. Since the Vdd ranges from 1.4-2v or so, then setting Vtt/Vbt equal to that amy be a good answer....or Vref = Vtt/Vbt would work as well. My understanding is that Vref automatically tracks to 1/2 of Vdimm on many boards and so that may be an avenue to consider.

What ya think?

[flapperhead]

how many ohm does ur pot read to get 1.6v? i have to have damn near a dead short.. gotta think a little on ur suggestions.....
thnx...

[Hell-Fire -XS-]

I do not have a board with a Vtt mod on it at the moment. The one I had was daos's IC7 that also was burning off pots.

I am working on a NF7 and will try the other moethods and keep you updated bud.

Good Luck.

[flapperhead]

ok, i thought u had an ic7 w/vtt to check out... maybei can use a ceramic vr
capable of hi wattage. only thing is i dont want to burn the circuit
somewhere else on the mb...

[Hell-Fire -XS-]

Naw, sent it back to the owner.

I have thought about that as well...burning something else up due to the high current.

If you are talking about Vtt at 1.6v, then at 1/2ohm the current is a body melting 3.2A. Thats enough to kill a person if I recall.

No wonder we are melting pots left and right.

This is why I am thinking there is another way around this.

[flapperhead]

just got me a hi temp 1/2 ohm 1/2 watt resistor, but im afraid of burning the motherboard..

[Hell-Fire -XS-]

Well, at 1/2w and 1/2ohm and still talking 1.6v, it could withstand 0.3A or so I think.

If I recall that:

P = I*V
I = V / R

But then again, if you are saying that 1.6v @ 1/2ohm..that is 3.2A which is quite a big current.

[flapperhead]

how about a 5 watt 1/2 ohm resistor?

[crotale]

Hello Flapp!
I'm going to look a little more into Hipros method and see if I can find a good way to overcome the pot-burning problem.

Hellfire: Do you have a board to try something? Flapp and I have been talking a little about this and one suggestion I had was to try using a 20ohm resistor instead of a 10ohm. That way it shouldn't be neccessary to turn the pot down so much.

I get back here tomorrow with some drawings and schematics to explain further what I mean.

[flapperhead]

Originally posted by crotale
Hello Flapp!
I'm going to look a little more into Hipros method and see if I can find a good way to overcome the pot-burning problem.

Hellfire: Do you have a board to try something? Flapp and I have been talking a little about this and one suggestion I had was to try using a 20ohm resistor instead of a 10ohm. That way it shouldn't be neccessary to turn the pot down so much.

I get back here tomorrow with some drawings and schematics to explain further what I mean.


really, really appreciate your help, you have forgotten more than i know... lol flapper...

[crotale]

Originally posted by flapperhead
you have forgotten more than i know...

I didn't understand that one...

Anyway... This schematic is showing what we're trying to accomplish:
http://www.nordiccrew.com/crotale/vttmod.gif
Orientation:
R1 is our fixed resistor, which is 10ohm initially.
R2 is our variable resistor/pot, 100ohm, 25 turns or similar.
Monitoring is where the IC read the voltage from.
Vtt is the real vtt voltage.

We're trying to fool the IC to think it put out to little vtt with these two resistors, and it will try to compensate for it.
What makes this diffrent from other vmods is that this one has much smaller voltages to play with. To get the appropriate effect, our curcuit needs to handle high current.

With the two resistors we split the voltage between them, and the monitoring will sense the voltage over the pot/R2.

One thing I found strange is when you say that you need around 1ohm the get any good increase of the vtt.
According to my calculations, the monitoring should read ~0.14V with a fixed 10 ohm, and 1ohm over the pot. And this is way under the reference value of 1.4v. It should try to boost the vtt alot.
Can anyone of you meassure the voltage over the pot. From ground to the other side of it that is...

I think it would be possible to change the 10ohm fixed resistor to one 20ohm instead. This will bring down the current, and it shouldn't be needed to turn the pot all the way down to zero.

(This makes me wanna mod my own board :cool: )

[flapperhead]

hey man thnx alot, im at work now soon as i get home tonite ill, check this out...

[joemac1]

Why not try a higher power rated pot and flame proof resistors?

[Hell-Fire -XS-]

There is also a Vtt fix out for the IC7. Thinking maybe that is bypassing this porblem we are having. Most likely this custom circuit will make it so that we dont have to worry about the current across our variable/fixed resistor.

crotale, I agree with you that adding a larger fixed will eleviate the need to dial the VR down so far to achieve the needed Vtt, but that still means that a base resistance is needed to reach say 1.6Vtt....which means that the same current will be flowing thru the fixed and VR...so we are still at the same cross roads with that problem.

I have another Max3 headed my way and I will be using the custom circuit this time.

http://www.ocforums.com/showthread.php?s=&threadid=287510

[crotale]

Originally posted by Hell-Fire -XS-
crotale, I agree with you that adding a larger fixed will eleviate the need to dial the VR down so far to achieve the needed Vtt, but that still means that a base resistance is needed to reach say 1.6Vtt....which means that the same current will be flowing thru the fixed and VR...so we are still at the same cross roads with that problem.

I'm not sure about that. The only thing we're after is lowering the voltage enough. We can either decrease resistance (higher current) or decrease current.
There is only one source so if we got a fixed 20ohm resistor and turn the resistance down to zero on the pot, it won't flow more current than vtt/20ohm.

[flapperhead]

how about if i add another 10 ohm resistor in series b4 the pot? that way it would be 20 ohms.only thing is im not sure if the original surface mount 10 ohm resistor is after or b4 the voltage feedback sensor...

[Hell-Fire -XS-]

Originally posted by crotale


I'm not sure about that. The only thing we're after is lowering the voltage enough. We can either decrease resistance (higher current) or decrease current.
There is only one source so if we got a fixed 20ohm resistor and turn the resistance down to zero on the pot, it won't flow more current than vtt/20ohm.

If thats the case where we have the 20ohm in parallel with the VR, then if the VR is set to a lower resistance, wouldnt the bulk of the current go across the VR? I recall something about path of least resistance somewhere in all this.

Not being rude, just cant recall everything that has been shoved into my brain at school.

Also, the 20ohm fixed would still keep the Vbt/Vtt a little to low for our needs yes?

[flapperhead]

when i set the vdimm voltage to 2.5, i measured 1.25v at the input to the vr with the vr set to its lowest resistance (1.5 ohms) . this gives me about .05v more vtt when measured at the correct vtt mossfet. which aint much... if i use a vr of lower resistance and set it to .5 ohms the vtt will go much higher ie 1.6v, but it burns open quickly...

[Hell-Fire -XS-]

Originally posted by flapperhead
when i set the vdimm voltage to 2.5, i measured 1.25v at the input to the vr with the vr set to its lowest resistance (1.5 ohms) . this gives me about .05v more vtt when measured at the correct vtt mossfet. which aint much... if i use a vr of lower resistance and set it to .5 ohms the vtt will go much higher ie 1.6v, but it burns open quickly...

We will keep looking at this, and I will post my findings on the new Vtt Fix after performing on the incoming Max3.

[flapperhead]

cool, meanwhile when i get home tonite im gonna experiment with different valu resistors in parrallel and maybe a couple pots in parrallel...

[flapperhead]

good news, (of a sort..) i placed 2 pots(1.5 ohms-lowest they will go) in parrallel and i was able to bring the vtt voltage up. however they have to be almost perfectly matched in resistance, cause otherwise one or the other gets hot as hell. so, i stopped before i burned one of them up cause its almost impossible to perfectly match each in resistance. my first thought is to simply put a 10 ohm resistor behind these pots to inhibit some of the current flo. but it seems like the vtt sensor needs a very strong current flo away from it to work, and if i introduce more resistance into the circuit, wouldnt that keep the vtt from sensing if i even set the pot to 0 ohms? anyway im going to try it and see how it works....

[flapperhead]

well ran a 10 ohm resistor in line b4 the pots and all it did was act like a 10 ohm pot, thereby not allowing enuff current to flo for the vtt sensor.

i am starting to think that the trace used for this mod is actually the the main feed to the memory vtt and not the trace coming off of the vtt sensor circuit. that would explain why i have to almost set the pots to 0 to get the vtt sensor to add voltage. for example my nb mod (which starts out at 1.50v and similar to the vtt in voltage )starts to respond by upping the voltage at about 50 ohms! without a print of this board this is the only logical conclusion i can come up with..

[flapperhead]

ok, if madcats max3 voltmod wont work im gonna put a 1/2 ohm 5watt resistorin . that will definitely hold without burning

[crotale]

Originally posted by flapperhead
well ran a 10 ohm resistor in line b4 the pots and all it did was act like a 10 ohm pot, thereby not allowing enuff current to flo for the vtt sensor.

Do you mean inline with the pot or inline with the other 10ohm fixed resistor?

[flapperhead]

in series with the pot cause i cant get to the resistor cause i epoxied it(what a dummy i thought the mod was workin, so i epoxied it)

[flapperhead]

well im running a 1/3 ohm hi watt resistor in place of the pot and it seems to be holding fine, and it brings my vtt up to 50% of whatever voltage i choose.