KT7E Core Voltage Mod?

[skligmund]

[Sorry about the double post as I have the same topic in Overclocking, but not enough people pay attention to those boards it seems]

Well, I have found out why I can't get my AXP 1800 T-Bred 'B' any higher than 1766 MHz without FUBARness going on...

When my bios is set to 1.85 volts (highest available) it only produces 1.8 volts (according to CPUID).

I have seen and read how to voltage mod a KT7A series, I was wondering if the same was true for the KT7E. I have looked at the location and solder points of the resister, and they are identicle, except there is what appears to be a micro resistor in one of the joints that isn't there on the KT7A series.

Does anybody know for sure if the KT7E is modded the same way? If not, I'll probably solder the resister in the next couple of days anyway, if nobody responds with a valid answer.


Thanks.

Here is a link to the KT7A mod: <A HREF="http://www.ocinside.de/html/worksho..._kt7_vcore.html" target="_new">http://www.ocinside.de/html/worksho..._kt7_vcore.html</A>

Water cooling is for the weak. Get liquid nitrogen.

 

[phsstpok]

AFAIK, the KT7, KT7E, KT7A all share the same basic board design differing only in chipsets plus capacitor revisions.

If the KT7E uses the same voltage regulator chip, HIP 6301, the mod should be the same.

Here's a good link.

<A HREF="http://www.vr-zone.com/guides/kt7/" target="_new">http://www.vr-zone.com/guides/kt7/</A>

I think it's the two stage mod.

The first mods just played with increasing Vcore but the max was limited to about 2 volts by the over-voltage control. The two stage mod also alters the over-voltage allowing up to 2.2+ volts.

Never tried it myself. Not very handy with a soldering iron.

<b>99% is great, unless you are talking about system stability</b>

 

[phsstpok]

By the way, what bus speed and multiplier did you manage to obtain on your KT7E?

<b>99% is great, unless you are talking about system stability</b>

 

[skligmund]

Right now I''m sitting on 16.5x and 106 MHz (1750 MHz). I could get about 108 FSB, but my CPU would get unstable over 1766 MHz.

And as you know, I can run my RAM at PCI + Host Clock (CPU FSB SPEED), so my RAM it running at 140 MHz.

Last night I again tried a few thins to get a 133 FSB on the CPU, but I got no joy. I couldn't get it to post.

Water cooling is for the weak. Get liquid nitrogen.

 

[phsstpok]

There are workarounds for getting the 13x-14x multipliers plus getting KT7A to work at 133 Mhz FSB. It's one of the wire mods. Works on other KT133A mobos too I here. The mods have been confirmed by a few people in newsgroups. Don't know if it will work on KT7E but it might be worth a try.

I used a variation of the mod on my EPOX 8KTA3PRO but I already had 133 Mhz available. I used the mod to get 14x working so that I could run my XP1700+ at 14 x 150 = 2100 Mhz instead of 15 x 140.

Check out this link. Look for the section, "KF traced Abit KT7's red circuit".

<A HREF="http://www.beachlink.com/candjac/Workarounds.htm" target="_new">http://www.beachlink.com/candjac/Workarounds.htm</A>

The simplest version involves inserting a loop of fine wire into the holes for two pins in your CPU socket. (The wires have to be fine enough to allow the insertion of the CPU).

You can also wire two pins on the chip, use conditive paint between the two pins, or wire the back of the motherboard.



<b>99% is great, unless you are talking about system stability</b><P ID="edit"><FONT SIZE=-1><EM>Edited by phsstpok on 04/26/03 04:16 PM.</EM></FONT></P>

 

[skligmund]

Well, when I get a little more time, I'll research that, thanks for the advise.

I have my VCore at 2 volts now, and the highest I can get it 1800 MHz. It will post ay 1900, or a little higher, but I get a WINXP error during start-up that says something like "Could not confirm the lisence" or something like that. So it looks like .2 volts increase only gets me 34 MHz. Did I just find the max of my CPU do you think? Or is something else limiting me? Right now at 1800 MHz, I am not overclocking my FSB, so RAM isn't the issue here, cause I normally run it at 140, not 133.

I think I'll drop my volatge a bit (getting a little warm) and go down into the 1700 MHz range again.

Water cooling is for the weak. Get liquid nitrogen.

 

[skligmund]

THE most SCREWED up thing just happened......

I went into the bios to lower my voltage, and bring my CPU back down to what it was.... It was all good.

I decided to get aother MHz out of the FSB, so I bumped it up to 8, making it 1766.... except it posted as 2166. So I went and dropped the multiplier, 2000 something. I set a few othe Multipliers, and I finally got a no post. I reset the CMOS, and started up, OKAY, same 2000 MHz it always posts at after CMOS reset. Set it back to my settings, 2166. Now I'm confused, but witty. I set the multiplier to 7X (I think, I already forgot) and it posts at 1300 MHz. I then say "ah what the hell" and set my FSB to 133. It posted at 1733 MHz, and that is what I'm on right now. WTF just happened? Suddenly my 133 FSB works!

Here is what I think happened......

Yesterday, in an attempt to fool with the CPU and get 133 FSB to work, I wanted to cut and paint the bridges to make it look like a different CPU, but I got no change. I'm thinking that the cut and paint didn't succede, but after heating the CPU up to 5X C, it made my paint connect or something. Now I get 133 FSB! A good screw up? I don't know.

Water cooling is for the weak. Get liquid nitrogen.

 

[skligmund]

It was actually 9X it is set at right now. so for some reason, 9X = 13X with 133 FSB capabilities. I bumped up my FSB to 135, and I'm scared to do much more with it, plus I have to go to work, so I'll talk at you later.

Water cooling is for the weak. Get liquid nitrogen.

 

[phsstpok]

If you get it working (at 133 Mhz) what you should see (in theory) is 7x=15x, 8x=16x, 8.5x=16.5x, 9x=17x (hence 2166 Mhz). The other mappings likely won't work because 1) timing issues 2)overclock too high for your XP1700+ (or is it XP1800+?)

On legacy mobos 13x-14x usually don't work because of timing issues. Has something to do with the mobo presumming when the chip is using 6x-7x (supposed to remap to 13x-14x) that it has an old, slow, 100 Mhz CPU installed and therefore set the incorrect timing values at POST resulting in an abort.

Those two issues are something the workaround adresses.

<b>99% is great, unless you are talking about system stability</b>

 

[skligmund]

I have an AXP 1800. I still don't understand why 9X is making a 13X. It should make something higher. If I go lower in the multipliers, the CPU speeds up, if I go higher, the CPU still speeds up. For some unknown reason, 9X is remapping as 13X. I'm not complaining though, I like having 133 FSB.

Water cooling is for the weak. Get liquid nitrogen.

 

[phsstpok]

When you were using 100 Mhz FSB did you get all the remappings? 7x-->15x? 7.5x-->22x? 8x-->16x? 8.5x-->16.5x? 9x-->17x? 9.5x-->18x? 10x-->23x? 11.5x-->19x? 12.5x/13x-->20x?

<b>99% is great, unless you are talking about system stability</b><P ID="edit"><FONT SIZE=-1><EM>Edited by phsstpok on 04/27/03 11:15 AM.</EM></FONT></P>

 

[skligmund]

Yes, everything was as it should have been, now it is all FUBAR. It won't post some multipliers, and others are super high, and 9X = 13X. Right now, that is the only one that makes any sense for me to run... Maybe I'll try to get the remap of all the multipliers.

Water cooling is for the weak. Get liquid nitrogen.

 

[skligmund]

O.K. I did all these with an increase of voltage and 100FSB.

5X==== NO POST
5.5X== NO POST
6X==== NO POST
6.5X== NO POST
7X==== NO POST
7.5X== 19X
8X==== NO POST
8.5X== 20X
9X==== 13X
9.5X== 13.5X
10X=== 14X
10.5X= 21X
11X=== NO POST
11.5X= 19X
12X=== NO POST
12X ABOVE = 20X


A little odd don't you think?

Water cooling is for the weak. Get liquid nitrogen.

 

[phsstpok]

Yup, it's odd.

I'm not all that good at it but let's see if we can analyze this.

<pre>5x-6.5x = NP (normal)
7X==== NO POST (should be 15x - no help)
7.5X== 19X (should be 22.5x, 4x too low on the table)
8X==== NO POST (should be 16x, again -4x)
8.5X== 20X (should be 16.5x, -4x)
9X==== 13X (should be 17x, -4x)
9.5X== 13.5X (should be 18x, -4x)
10X=== 14X (should be 23x, -4x)
10.5X= 21X (normally NP)
11X=== NO POST (OK)
11.5X= 19X (OK, but why?)
12X=== NO POST (OK)
12X ABOVE = 20X (OK, but why?)</pre><p>At first this looks inconsistant.

Suspect 4x bit is being pulled low.

Let's look at the multiplier bit break down

7.5x (BIOS) - 3x (offset) =
4.5x = 0.5x HI + 1x LO + 2x LO + <font color=red>4x HI</font color=red> + [8x HI (chip mod)]

8.5x (BIOS) - 3x (offset) =
5.5x = 0.5x HI + 1x HI + 2x LO + <font color=red>4x HI</font color=red> + [8x HI (chip mod)]

9x (BIOS) - 3x (offset) =
6x = 0.5x LO + 1x LO + 2x HI + <font color=red>4x HI</font color=red> + [8x HI (chip mod)]

9.5x (BIOS) - 3x (offset) =
6.5x = 0.5x HI + 1x LO + 2x HI + <font color=red>4x HI</font color=red> + [8x HI (chip mod)]

10x (BIOS) - 3x (offset) =
7x = 0.5x LO + 1x HI + 2x HI + <font color=red>4x HI</font color=red> + [8x HI (chip mod)]

All of these show 4x bit HI but you are getting it LO.

So far it seems I was right. 4x bit is being pulled LO.


Let's see if the other values confirm this.

11.5x (BIOS) - 3x (offset) =
8.5x = 0.5x HI + 1x LO + 2x LO + <font color=red>4x LO</font color=red> + [8x HI (chip mod)] (4x is already LO)

12.5x (NIOS) - 3x (offset) =
9.5x = 0.5x HI + 1x HI + 2x LO + <font color=red>4x LO</font color=red> + [8x HI (chip mod)] (again 4x is already LO)

This confirms my suspicion. The 4x bit is artificially pulled LO.

<b>[EDIT]</b>
I wasn't clear on this. You were getting the correct results for 11.5x and 12.5x. Because the multiplier codes for these values don't use the 4x bit this explains why your results were correct.

The likely cause is that your cut 4th L3 bridge (4x bit) is grounded to the substrate.


I would suggest brushing clean your mod'd bridges but this might cancel all of your results.

Don't know what to tell you.

<b>[EDIT]</b>
Oops. Just re-read your posts. I see you re-mod'd your chip. It is now obvious that you got paint in your original cut of the 4th L3 bridge.

Your mistake is like the workaround I mentioned earlier. LOL!


<b>99% is great, unless you are talking about system stability</b><P ID="edit"><FONT SIZE=-1><EM>Edited by phsstpok on 04/28/03 01:31 PM.</EM></FONT></P>

 

[skligmund]

LMAO!

OK, well, I'm glad I made the 'mistake' because I like having the 133 FSB. It is running stable and cool. And yes it was the 4th bridge.

CPU: 1785
Multiplier: 13X
FSB: 137
CVoltage: 1.86
Temps @ load: 46C (in a warm room)

Water cooling is for the weak. Get liquid nitrogen.