• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

running Tualatin on CuMine MB w/o Powerleap

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
martinus said:
well, I've done the mod today, twice actually. First I did my Asus slotkey. I opened the socket and took out AN3, AJ3, AK4. No connection between AK4 and AK26! Voltage selection is handled by jumpers on the slotkey (vid pins). Runs like a charm in my AOpen AX6BC Pro - whereas the powerlep adaptor seems to be incompatible. ;)

I got to 133 FSB at 1.5V, but my Matrox G400 starts crapping out at that speed, so 129 is the highest stable FSB at the moment. The cpu is obviosly capable of so much more!

Next I modded my trusty Asus CUBX system, which I know to do 150 FSB stable. Same method. Unfortunately I miscounted and took out a Vcc pin. Annoying, but there are enough left, so no worries. ;)

The celery 1000A did not want to run in the CUBX. I then proceeded to connect CMOS PWRGOOD with VTT PWRGOOD, using wrapping wire. I don't know, maybe it was too ghetto, or maybe it just doesn't want to play. No go.

Anyway, I'll have to get a new video card and try to get the AOpen board higher... :cool:
Click to expand...

It would seem that many can get the slockets to work if they are not generic but when you do not connect ak4 to ak26 your performence and OC'ability takes a hit. If i remeber correctly earlier in this thread someone could only hit 133fsb and queue depth of 1. This was before they connected ak4 to ak26. Once they did this it OC'd much better and they could take it to queue depth of 4 or 8 not sure which. Anyway their mem scores were just like they should be on a BX board.
 
ol' man said:


It would seem that many can get the slockets to work if they are not generic but when you do not connect ak4 to ak26 your performence and OC'ability takes a hit.
Click to expand...

well, so far I cannot verify that. also not obvious from the meaning of the signal. but I'll keep an eye on it.
 
belinassu said:
have a look , I'm so happy!!! :D

Ciao belinassu
I'm Italian as you are
I'm interested to mount a tualatin on a Cusl2-c, but I don't know if is better to remove the copper shim on the core, or leave it in place..
I've a Waterblock cooler and use artic silver.. I think that whitout the copper shim wuold be easier achieve lower temperatures
What's your experience?
And what the method you've used to remove the shim?
Thanks all
and sorry for my bad english

Romeo:cool:
Click to expand...
 
Coulnt get a generic to work...
going to pick up another slotket today hopefully...
see if i can get the new one to work
 
Romeo said:
belinassu said:
have a look , I'm so happy!!! :D

Ciao belinassu
I'm Italian as you are
I'm interested to mount a tualatin on a Cusl2-c, but I don't know if is better to remove the copper shim on the core, or leave it in place..
I've a Waterblock cooler and use artic silver.. I think that whitout the copper shim wuold be easier achieve lower temperatures
What's your experience?
And what the method you've used to remove the shim?
Thanks all
and sorry for my bad english

Romeo:cool:
Click to expand...

If you are gonna be water cooling, I would leave the IHS on the core. Since you really don't know how much presure to use to secure the waterblock, you might tighten it too much and crush the core. And since the core is small, it could be harder to get the waterblock perfectly flat on the core. I am watercooling also and these are the reasons why I left the IHS on.
Click to expand...
 
"Did you rip some pins out and not set the vcore to 1.5? WIth the vid pin 25 connected the vcore wil be 2.1v. Maybe you cooked yer cel-t's."

Thanks ol'man for the reply... All of the pins on the CPU are intact. However, I did noticed that all of the VID pins are opened (20K ohm) up internally in the CPU after I plugged it in the sloket. The VID25m (AK36) is physically connected to Vss( ground) on this sloket. Are you suggesting that I should open this pin to the slocket as well???

Sorry for the late reply as I spend the entire weekend setting up a 1.8a on the Ga-8irxp. By the way, this MB so far don't like to be overclock. I was told that Gigabyte is working on a BIOS fix right now... Other than that, It's seems to be solid.
 
Just tried it on a MSI MS6905 Master Slotket -
All the same setup as before - same cpu with three pins removed.
Before I shut down the last time before the change, I set the FSB to 100 to give better boot odds- set the slotket jumpers to 1.7v -
booted into the bios - set the voltage to 1.75 and rebooted.
Seemed to take too long to get into windows, so I rebooted and got the
safe mode menu, but chose normal and boot gave a windows protection, which I learned to expect occasionally as normal, went back to bios and found I hadn't changed the 1/4 PCI divider, so did that and set and set FSB to 112 - booted into safe mode - rebooted normal
went back to the bios and saw the "in order que depth had change to 1"
I set the FSB to 133, 1.75v, and booted normal -
went back and change the que depth to 8 - rebooted normal
Ran the sandra cpu bench burn in for several hours with no problems.
So this slotket works fine for this cpu mod.

Another reason I finally decided to just remove the three cpu pins was that even though the cpu cost a little more, they are plentiful, and it appears the good slotkets aren't anymore. Wouldn't be surprised if Abit, Asus, and MSI, don't even make them now.

To sum up my experience to date:
BE6-2
Celeron/Tualatin 1.0A with AN3, AJ3, AK4 pins removed on the cpu
Retail heatsink
Win98 - non SE
Abit Slotket!!!, Asus S370DL-133, and MSI 6905 Master
All of these ran stable at [email protected]

I believe they would run as well on a BX6r2 up to 133FSB.
I would try it on mine but my BE6-2 is sitting on top of it, and
behind my K6-3+450/600, so will probably be awhile. I already have
a Celeron 1.0G@10x133 in it, so not much to gain. Never could get the BX6-2 to run above 133FSB.

For those who might not know, Tom's, the original B21 Doctor, hardware has an article
on how to convert a pre-coppermine PPGA to FC-PGA, by just
connecting AH4 and X4, and disconnecting AM2
Here is the article:
FC-PGA Pentium III on Celeron PGA370/Slot1 Converter Cards
http://www6.tomshardware.com/cpu/00q1/000229/
I'm going to try that next on an pre FC-PGA Asus S370,
and then maybe on an MSI.

Then I'm going to try my celatin 1.2A and use a better than
retail heatsink.

Has anyone compared benchmarks with this super cool mod for BX,
with the same cpu's on a Tualatin board like the Tusl??
I'm really curious to see how they compare.
I wouldn't be surprised if they are better on a BX.

Hope you all are having as much fun as I am!

Note to Catcha22: I like your site, but takes forever to load and couldn't get on to register.
 
Last edited:
Hey guys - just want to pass this on.
I apologize if someone else already brought it up
.
I mentioned on a previous post, that I found my Tualatin 1.0A was high on the edges of the slug, and I used a file to get it flat around the edges.

I just opened my new Tualatin 1.2A. First thing I did was place the slug against the heatsink I'm going to use, and wiggled it while pressing - Wow, it's really high, and would definitely keep the center of the slug off the heatsink by some amount. Maybe Intel thinks people will use the thermal pad they put on their heatsinks - who does that?

Any way, although this is just my second one, I'll bet most all of them them have this problem.
Think it is worth checking yours.
 
Ruiner i have a soyo 7vca and wondering exactly what you did
i am guessing you dissabled the three pins,
did you connect ak4 and 26
vid mod?
 
belinassu- nice page
I am not familiar with details of the BX133, so I have a few questions.

Do you think it is possible that when you tried with just the three pins removed, that the core without the slug might not have been flat against the heatsink?

What was the highest voltage you could adjust on your board with just the three pin mod?

Was the reason you did the additional mods to get a higher voltage setting?

You did a nice job reinstalling the slug - do you think this improvement in the core to slug interface helped you to get the FSB that high?


Just looked at the Overclockers.com CPU database - who is going to be the first to post your BX data next to all the Tualatin boards?
 
Belinassu: awesome page! Be prepared for some huge bandwidth costs once word gets out!
I wish my chip would go past 133fsb. :(

sammy: I was the first on the database with a modded 1.0a (via, not bx) :)

dmcollin: I used wire insulation 'condoms' on the 3 pins, and used enamel coated 26g wire to jump ak4-26. I did the 'pin hole trick' to jump Vss to Vid4. It would not work without ALL those mods. Default Vcore is 1.45 done like that.
 
sammy said:
belinassu- nice page
I am not familiar with details of the BX133, so I have a few questions.

Do you think it is possible that when you tried with just the three pins removed, that the core without the slug might not have been flat against the heatsink?

What was the highest voltage you could adjust on your board with just the three pin mod?

Was the reason you did the additional mods to get a higher voltage setting?

You did a nice job reinstalling the slug - do you think this improvement in the core to slug interface helped you to get the FSB that high?


Just looked at the Overclockers.com CPU database - who is going to be the first to post your BX data next to all the Tualatin boards?
Click to expand...

excuse me , my english is so poor and i could not understand some questions...
1)the core and the HS are perfectly flat iI do have to connect
AK4 to AK26 to get stability it wasn't a hight T° problem
2)The answer is in the page i linked 1,45v default 1,7v max in bios
3)I did Vid4->Vss to improve stability (strange but true) and Vid2->Vid1(also Vss) to get 1,65vdef 1,9vmax
4)yes i think in the pics you can't see the intel's compound (i removed it before) under the slug
but it was worse than artic silverII ;) so it gained some °C
5) escuse me but i couldn't understand the question :(
 
wirehead said:
The VID25m (AK36) is physically connected to Vss( ground) on this sloket. Are you suggesting that I should open this pin to the slocket as well???
Click to expand...

AK36 on an FCPGA socket is ground. I don't know where the myth comes from that it has anything to do with a slot1 pin called vid4 and could potentially send the voltage above 2V. just leave AK36 as it is and you'll be fine. :cool:
 
in intel c 1.2 datasheet the ak4 specification
ak4
The VTT_PWRGD signal informs the system that the VID/BSEL signals are in their correct logic state. During Power-up, the VID signals will be in a indeterminate state for a small period of time. The voltage regulator or the VRM should not sample and/or latch the VID signals until the VTT_PWRGD signal is asserted. The assertion of the VTT_PWRGD signal indicates the VID signals are stable and are driven to the final state by the processor. Refer to Figure 6 for power-up timing sequence for the VTT_PWRGD and the VID signals
Click to expand...

the ak26(pwrgood signal) is similar to the ak4(Vttpwrgd),so connect them make more stable.

Vid25mv(vid4) connect to Vss is prevent from getting high core voltage for some mobo don't connect them to Vss by circuit.in some mobo just connect Vid25mv(vid4) to Vss get 1.3v core voltage,so you must connect Vid2 to Vss also to get 1.5v core voltage.
 
Last edited:
We've been discussing the data sheets in the ocau forum, and I'd like to share this question with you:

when looking at the FCPGA data sheet, intel does already mention the vtt_pwrgd signal. It is only relevant to the VRM and serves to sample and latch the vid and bsel signals from the cpu. this typically happens 1ms after power-up, and pwrgd is asserted after another 1ms by the VRM. The signal vtt_pwrgd is not routed to the FCPGA socket.

Interestingly, in the 0.13um FCPGA2 data sheet, vtt_pwrgd plays exactly the same role. However, for some reason it now appears in the socket (pin AK4). Can anyone shed some light, why the cpu should care about this signal at all?
 
ebc said:

Vid25mv(vid4) connect to Vss is prevent from getting high core voltage for some mobo don't connect them to Vss by circuit.in some mobo just connect Vid25mv(vid4) to Vss get 1.3v core voltage,so you must connect Vid2 to Vss also to get 1.5v core voltage.
Click to expand...

I think we can solve the confusion by using clear terminology. ;) Let's not mix up the terms vid25mv and vid4.

vid4 is a slot 1 pin used for voltages above 2.05V. It has never existed in any socket370, since there have never been socket cpus that were specified for more than 2.05V. the first PPGA cpus were mendocino core, which is the celeron version of the deschutes core (2.0V).

vid25mv has been introduced with the 0.13um FCPGA2 socket, it occupies pin AK36 which is vss in PPGA and FCPGA sockets.

From the point of view of a standard FCPGA slot key, vid4 only exists on the slot1 side (usually pulled to ground or jumpered), whereas AK36 only exists on the socket side (always vss).

edit: typo
 
martinus said:
We've been discussing the data sheets in the ocau forum, and I'd like to share this question with you:

when looking at the FCPGA data sheet, intel does already mention the vtt_pwrgd signal. It is only relevant to the VRM and serves to sample and latch the vid and bsel signals from the cpu. this typically happens 1ms after power-up, and pwrgd is asserted after another 1ms by the VRM. The signal vtt_pwrgd is not routed to the FCPGA socket.

Interestingly, in the 0.13um FCPGA2 data sheet, vtt_pwrgd plays exactly the same role. However, for some reason it now appears in the socket (pin AK4). Can anyone shed some light, why the cpu should care about this signal at all?
Click to expand...

as you said in the 0.13um FCPGA2 data sheet, vtt_pwrgd plays exactly the same role.I guess,now the ak4 had been open,if not connect to pwrgood(ak26),because open meaning logic 1,it action before pwrgood(ak26),so would made some unexpected unstable. sorry for my poor english again.
 
ebc said:


as you said in the 0.13um FCPGA2 data sheet, vtt_pwrgd plays exactly the same role.I guess,now the ak4 had been open,if not connect to pwrgood(ak26),because open meaning logic 1,it action before pwrgood(ak26),so would made some unexpected unstable. sorry for my poor english again.
Click to expand...

yes, it is the "unexpected" I am trying to narrow down. ;) Because from the description, the signal is only relevant to the VRM, and the coppermine doesn't need it, so what does the tualatin need it for?
 
You must log in or register to reply here.

Similar threads

M
running Tualatin on CuMine MB w/o
Replies
0
Views
813
Min
M
Back