Major KX7-333 w/vmods probs...

[Marci]

OK, many of u know my rig...

ABit kX7-333R vmodded
Athlon XP2000+ Unlocked
256Mb Corsir XMS3200
Radeon 8500 vmodded
Enermax 431w PSU
Maxtor D740XL HDDs

Now, previously on air without vmods I was struggling to get anywhere. Once the board was vmodded I was away quite happily clocking my little ass off, just limited by air-cooling.

Now, got watercooling in and without vmods the system will run wonderfully at 200x9.5, 1.85v, Normal Decode. Ram at the moent after tonights stupid escapades at Cas 2-4bank-3-5-3-3T-2T-840ps at 2.85v.

CPU Idle temps at 34 deg C. Now, here comes the problem...

As soon as i hook up the vmods it all goes to hell. The system doesn't reset properly from windows, doesn't turn on properly at the best of time... with vmods cpu is at 1.97 and ram at 2.98v,and idle temps are 35 deg C

On air the vmods helped, on water however I find I have problems maintaining the speeds I'm at now, let alone getting higher. These p[robs have only arisen over the past 2 days, but I have checked everything... no shorts, variable resistors are 100% perfect working order. Had board out, no signs of any damage at all, examined CPU and looks like its no more than a day old... ram seated correctly... I literally cannot find a single thing wrong with anything, except as soon as I hook up the vmods performance drops. I've checked PSU rails and they're all fine, actually doing better now than when I was on air cooling so I'm guessing thats due to the reduction in fans on the system...

I've flashed to 9k tonight to try and resolve things, and the best I could get with vmods is 180x10. As SOON as I disconnect all the vmods, its straight back up at 200x9.5 no probs... anyone got any clues whatsoever I'd be most appreciative!!! I managed to get 2Ghz once without vmods but not stable, figure with 1.95v it'll be fine, but looks like it just ain't possible thanks to this arse-pain of a problem...

[shadco]

Originally posted by Marci
OK, many of u know my rig...

ABit kX7-333R vmodded
Athlon XP2000+ Unlocked
256Mb Corsir XMS3200
Radeon 8500 vmodded
Enermax 431w PSU
Maxtor D740XL HDDs

Snipage

I've flashed to 9k tonight to try and resolve things, and the best I could get with vmods is 180x10. As SOON as I disconnect all the vmods, its straight back up at 200x9.5 no probs... anyone got any clues whatsoever I'd be most appreciative!!! I managed to get 2Ghz once without vmods but not stable, figure with 1.95v it'll be fine, but looks like it just ain't possible thanks to this arse-pain of a problem...

This matches my experience almost exactly.

I'd been cruising along no probs with an unlocked 1600 and had the system Rock solid and stable upto 210 fsb fast command decode Cas 2.5 2 6 3 2T and all the enhance options turned on, with a clip on vmod bumping the vcore by .2 volts.

Late last week I swapped out the 1600 for an unlocked tbred and all the sudden No fast command decode not even at 180. No 200fsb unless I backed all the way off on the timings and went normal on the command decode, and don't even think about anything over 200. I also had restart problems and would have to power down and it would usually come up at that stage.

Once early on when I tried to go over 200 I got a CMOS checksum error, then things were really screwed up. I reflashed to 9k without the new HPT bios same same, back to 7m ditto. I even popped a 1600 back in early during the process and had the same results.

I haven't pulled the vmod but will tomorrow.

I hope the board isn't trash. It's the silvercap / metal lever version.

Shad

btw the memory still does 237fsb in my IT7

[mdzcpa]

I'd try applying one voltage mod at a time. I suggest doing the CPU Vcore mod first (as your temps look fine). See if you can get the chip clocked up there to 2Ghz by using a higher multiplier and lower (known stable) fsb speed.

If that works okay, then do the Vmem mod. I have a suspicion the Vmem mod may be borking the system.

[shadco]

Originally posted by mdzcpa
I'd try applying one voltage mod at a time. I suggest doing the CPU Vcore mod first (as your temps look fine). See if you can get the chip clocked up there to 2Ghz by using a higher multiplier and lower (known stable) fsb speed.

If that works okay, then do the Vmem mod. I have a suspicion the Vmem mod may be borking the system.

I don't have a vmem mod and don't think I need one.

» http://service.madonion.com/compare?pcm=562286

Shad

[mdzcpa]

Sorry....should have been more clear. I was referring to Marci's problem:)

[shadco]

no prob

Shad

[Marci]

Er, not quite exactly the same prob shad...

1. This is an XP2000+ I've been using for months...
2. The problem only occurs when I attach the vmods. Either of them.

The entire system works fine as soon as I remove them.
Each individual vmod resistor multimeters as fine. The wiring multimeters as ok, no pins are shorting anywhere. This hasn't been caused by anything being added or removed or changed recently as all that has been changed is my Alpha removed and the swiftech screwed straight back in it's place. The Mobo wasn't removed or anything, and like I say, the vmods are physically and electrically sound.

[shadco]

Same here when I remove the vcore mod.

I put a couple of washers under the head of each screw that mounts the hs to the mobo. I wonder if there is an effect where increasing pressure around the cpu socket is affecting this.

Shad

[Marci]

Why would that make a difference to me? Removing voltmods has nothing to do with pressure the heatsink has on the block... so if I remove a vmod which is located nowhere near to my cpu socket the pressure suddenly vanishes between my block and cpu and it magically starts working again??

[mdzcpa]

Marci,

It is possible that your voltage regulation on that particular mobo just isn't up to the task of managing the higher voltage. This could easily lead to large momentary fluctuations in core voltage and cause stability issues.

Have you tried only one mod at a time to discover which mod is causing the trouble?

[Marci]

System is in same state no matter which vmod is attached... have tried running independently.... it could manage the voltages quite easily about a month ago, and nothing has changed since then other than the removal of my AlphaPAL and fitting of my swiftech... neither of which can cause these probs...

[mdzcpa]

Not sure if I've read the thread correctly, but did you say that you removed the waterblock and swapped back to the Alphato try with the voltage mods again? Is it possible that it is just coincidence that the voltage regulation went south at the same time you added water cooling? The only way to know for sure would be to put the Alpha back on and try the vmods again.

If the vmods start working again with air cooling, I might be inclined to beleive the additional airflow over the mosfets and throughout the case is helping keep the overvolted system stable. A stretch....but a possibility I suppose.

[Marci]

Still have active cooling on the mosfets as I had previously... one of the Blue Ramsinks that comes with the blue orb, with a 40mm ys-tech attached. I have a window mounted Zalman 92mm which also blows over these, and the usual rear mounted 80mm ystech more orless beside them so they should be getting plenty of cooling from there I would've thought... rather than swap back ot the alpha it was easier just to dangle an 80mm delta over the mosfets temporarily, but no change :( however, i think now we're moving in the right direction of where we need to be looking...

I think it is all coincidence with timing of it all, which is what I'm trying to point out to shadco.... the change from air to water isn't really involved its just that now I'm on water I can get the most out of the vmods, and suddenly they've started doing this... as u can imagine, fairly pi**ed off!!!

[shadco]

Originally posted by Marci
Why would that make a difference to me? Removing voltmods has nothing to do with pressure the heatsink has on the block... so if I remove a vmod which is located nowhere near to my cpu socket the pressure suddenly vanishes between my block and cpu and it magically starts working again??

Marci

Obviously there is no connection.

Pardon me for having the temerity of cluttering up your thread with useless comments.

It's clear that such a noob as me couldn't possibly have anything to add.

You won't have to worry about any further annoying particpation by me in any further threads you participate in on this forum.

As far as I'm concerned you can get lost.

Shad

[mdzcpa]

I hear ya brother....it plain just sucks:(

If you have good airflow like you say I'm inclined to beleive the mobo is just starting to give it up under the pressure of increased voltages. Not a good time for that to happen no doubt.

I'll keep thinking about it, but at this point I'm kinda stumped.

[Marci]

shad, just struggling to see the relevance of your suggestions to the problem I'm having when my posts make it clear that it's nothing to do with hs fitting or anything...

neways have come to conclusion that it's redo the vmods from scratch with new resistors and if no change then the voltage controllers on the board have gone fruit-loopy, so then it's new mobo time... have pm'd a few other guys and they all come to same response, stumped / never seen before hence the decision based on their recommendations... if it ain't on ething its another with this sodding system!

[shadco]

Sorry for flying off the handle. That's what I get for owning stocks.

The only common thing is that mine is flaking out the same way and straightens out when I remove the vcore mod in that I can go higher on the fsb turn command decode to fast at and above 200 and can get more aggressive on the ram timings... but I can't take the processor as high without the vcore mod.

My old board which I sold once I got the new one up to 215 semi aggressive with fast decode, never gave me any hassle for 3 months but I wanted a rev 1.1 and now I'm regretting it. The old board was also had the black rubycons and the new one has silver nichicons.

Is yours rev 1.1 or 1.0?

Shad

[Marci]

This has silver... can't spot the revision at the mo without stripping it all out so will check that tonight... It was bought about 2 weeks after they came out in the UK, so knowing Abit I'd say it's probably a 1.1 but possibly Icee may be able to help on that front... ICee, it's got the 2x Silver twin towers by the mosfets what rev does that drop into being as the most liekly contender!?

[Player0]

If you got the big silver caps and the metal lever handle, than you should have the newer revision. (sorry if i misread anything, i was kind of skimming).

I didn't have these problems with the vcore mods, but I did have with the Vmem mods. I use radioshack variable resistors, the surface mount tweak pots. They are hideous.

They can be pretty heat sensetive, and their resistance can fluctuate because of this. This causes minor voltage swings, or even small spikes that you would never see without the proper equipment. Anyway, the pot i soldered on for my Vmem just started going insane after me playing with it. Unfun. I had to remove it, and I haven't yet replaced it.

I dont know if your using pots or static resistors, but either way, get high tolerance, high quality parts. Im not sure if this is really you're problem, but might be worth swapping the pot.

Keep your leads as short as possible, and twist them if you can, this will reduce any interference. Small wire leads can act like mini radio antennas, so if you have anything causing RFI near the system (like a compressor, air conditioner, vacuum cleaner), these leads might be introducing interference in to the system which causes instability.

Maybe your PSU is taxed or something, and it cant handle the extra wattage well? I dunno...just throwing out ideas :)

[Marci]

Hooked the lot up to a Fluke Hydra today, and discovered NOISE SPIKES!!! The voltage regulator looks to be generating noise spikes as the vmods weren't even on at the time. Plugged them on a few mins ago and the noise spikes got worse, so combination of the 2... looks like something on the board has leaned itself to microphonic / radiophonic tendencies... anyways, gonna wait til it gets so bad that it does it all at stock voltages and clock speeds then RMA it, as it'll be easier to demonstrate the prob to the supplier...

FYI, these are the resistors I'm using:

http://www.over-clock.co.uk/acatalog/vMod_Gear.html

Ta fer the suggestions!

[Player0]

Huh...I wonder if thats normal for all these boards lol :)

Do they get any better at lower voltage settings?

[MrIcee]

marci:)

The Nichicon cap version with metal lever zif socket came out several weeks after the KX7's release as its final iteration of the board as far as components are concerned.

The newer version 1.1 which I currently have and which has the hardware fix for recognizing the T-Bred cpu, has only been out I'd guess 3-4 weeks..maybe a wee more. The only way to be 100% sure of your revision is to check the tag on the back of the board, as there are no physical or discernable differences between silver cap version 1 or version 1.1. BTW..my revision 1.1 was a first shipment board..as the version # was penned in from 1.0 to 1.1.:)

Randi:D

[Marci]

Will have a look for that when I strip it out in a few days.... cheers!