Sign in with
Sign up | Sign in
Your question

1.4ghz t-bird won't go 1.4gHz

Tags:
Last response: in CPUs
Share
June 22, 2002 8:05:57 PM

Greetings... newbie to this board here.

I just bought a new cpu for my system, the T-Bird 1.4GHz(133/266) with the "glorious" AYHJA designation. ;)  It replaces a 1.0GHz t-bird (100/200 fsb)

My problem: The chip won't run at 1.4GHz, let alone anything over 1.4 (ummm... actually I haven't tried anything over 1.4 yet, to be honest.) However it is stable at 1.33 GHz (133 fsb, 10x multiplier, 1.75V core).

My system:
*Abit KT7A-RAID w/ BIOS version 7N, which supports up to 1.4GHz T-birds, and 133MHz fsb. (this mobo is NOT the new version 1.3 KT7A, but the original with the latest BIOS).
*512MB PC133 Crucial RAM, CL=2 (2x 256MB dimms)
*Asus v7100 geforce2mx video, 32MB
*Windows XP pro.

The symptoms: When I set the speed to 1400 (133fsb/10.5x multiplier, 1.75V core) the system (usually) boots to windows, but locks up as soon as I start an app (like I.E.) It won't even recognize the case power button. I have to either hit the reset button or cycle the powersupply's switch.

The RAM is new also. It replaces a 256MB PC133 DIMM that was working fine, but was CL=3. I don't think the RAM is a factor, as I initially ONLY replaced the cpu, and had the same symptoms. Then I replaced the RAM and the symptoms did not change. Still locks up at 1400, stable at 1330.

I'm not experienced at all in overclocking or cpu's in general really. So, am I missing something? Did I not do something I should have? Or is the cpu just bad, and I need to send it back?

If this has been covered before, I'm sorry... I tried searching, but couldn't seem to narrow the hits down to anything relavent to this.

Thanks for any advice!

-Kieran

More about : 4ghz bird 4ghz

June 22, 2002 8:10:28 PM

Are you using the same heatsink which you used with the 1ghz tbird?

Amd 1ghz heatsinks wont cope with 1.4ghz cpus.

Also your psu may not be powerful enough to handle the powerdrain 1.4ghz causes.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 22, 2002 8:18:42 PM

Forgot to say... I did upgrade the heatsink to the Zalman CNPS3100-Plus. Not the best heatsink out there, but it's the quietest, and was rated for a 1.4GHz Athlon, so I figured it would work. My cpu temps are stable at 40C at 1.33gHz, but I can't even check the temps while at 1400 (except from the BIOS, not sure if that counts.)

How do I know if my power supply is the culprit? It's a 300W, from Sparkle Power Int'l, model # FSP300-60BT. In addition to what I already listed, it is powering a Western Digital 80GB eide drive, Pioneer dvd-105s, Lite-on 40x12x48x cd-rw drive, plain-jane floppy, soundblaster live, modem, NIC, and... that's about it.

-Kieran
Related resources
June 22, 2002 8:30:18 PM

OK, its not temp, I want you to try this.


Lower your vcore to 1.7v, then go to 1.4ghz, let me know if it runs, if it does, its your psu, if not try 1.725, if neither of those work the only way to test the psu is to get a new one and try that.


Sparkle psus suck, but it is 300w, it is possible that it is your issue, try my suggestion and get back to us.


:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 22, 2002 8:49:21 PM

Hey, thanks for the help. It's running now at 1.4gHz, with a core voltage at 1.7V. So, what does this mean? Is there a problem running at 1.7V instead of 1.75V? Do I need a new PSU? I was considering getting a new PSU anyway, because this one is *SO* loud. But decided to hold off for now. But if I need a new one to be stable, I guess I could get one.

Thanks,

-Kieran
June 22, 2002 8:52:07 PM

What it means is that your psu is on the edge of its ability, if the tbird runs at 1.7 volts and you have no lockups then you are fine(but dont add any more hardware.

Optimally you will want to buy a new 350watt psu, like a good enermax or something, but this solution will work for as long as you need it too assuming its stable under load, which you will need to test, and that you dont add anything new to your pc.


Your welcome for the help.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 22, 2002 8:59:30 PM

What's a good easy cheap way to test stability under load? The only thing I can think of is just to open a bunch of apps, play some 3d fps games, etc. Are there any freeware testing utilities for download that would do a better job?

edit:
BTW I just realized that even though in SoftMenu3 I set the voltage to 1.7, the VIA Hardware Manager shows it at 1.72v... not sure if this means anything.

Thanks,

-Kieran<P ID="edit"><FONT SIZE=-1><EM>Edited by Kieran on 06/22/02 05:00 PM.</EM></FONT></P>
June 22, 2002 9:01:15 PM

Download an application called prime 95, then run the torture test, if your computer dosent lock, or give errors in the torture test, then you can be safe in its stability.


If the test passes and your core is at 1.7volts, then you can keep your current psu untill you want to add new hardware or overclock or anything like that.

Good luck.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 22, 2002 9:40:18 PM

Well... 2 things...

1) Is Prime95 this prime number search tool that I found? If so, it's taking an HOUR to do the initial test/setup. :p 

2) I don't really need to test things, as my system locked up again while I was browsing to find Prime95. I shutdown and put it back to 1.33gHz/1.75V for now. So, I guess I really do need a new PSU?

-Kieran
June 22, 2002 10:33:45 PM

Quote:
2) I don't really need to test things, as my system locked up again while I was browsing to find Prime95. I shutdown and put it back to 1.33gHz/1.75V for now. So, I guess I really do need a new PSU?


Yep, you need a new psu, before you do though, set it to 1.4 and 1.725, see if thats stable, if it isnt then a new psu is in your future kemosabe.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 22, 2002 11:25:10 PM

how do you figure sparkle sucks? i have used them a lot, we use sparkle at work in all our machines, have for quite a long time, only had 2 or 3 have the fan go out and start rattling and one up and died out of 600 or so.

how do you shoot the devil in the back? what happens if you miss? -verbal
June 22, 2002 11:42:42 PM

Quote:
how do you figure sparkle sucks? i have used them a lot, we use sparkle at work in all our machines, have for quite a long time, only had 2 or 3 have the fan go out and start rattling and one up and died out of 600 or so.


I have only had one 300w psu that couldnt handle 1.4ghz+, it was a sparkle, I got an enermax and never looked back.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 12:18:34 AM

thank for the info the a new

Never stop learning

cheap, cheap. Think cheap, and you'll always be cheap.AMD version of semi conducteur industrie
June 23, 2002 12:20:18 AM

Well, it booted to windows, got IE started, but locked up as soon as I started the prime95 software, at 1.4gHz & 1.725V (which the VIA hardware manager showed as 1.74, by the way.) And as I mentioned before the VIA HM showed the 1.7V setting in the bios as 1.72V.

Right now, I'm running IE, outlook, the hardware manager, and the torture test on prime95, with the cpu at 1400 & 1.675V (which the HM shows as 1.68 -- close enough).

If its stable at 1.68V is there any danger in leaving it here for now? Or should I just run it at 1.33gHz/1.75V until I get a new PSU?

-Kieran
June 23, 2002 2:27:52 AM

Quote:
If its stable at 1.68V is there any danger in leaving it here for now? Or should I just run it at 1.33gHz/1.75V until I get a new PSU?


if it runs stable there is no danger running it at a lower voltage, also your core temp will be lower(slightly).


After the prime95 tedst play some games and compress some files, if you dont get a lockup then you are 100% good to go at this voltage.

Remember, if you want to overclock any or add new hardware you will need a new psu, because this one is at its limit.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 5:57:08 AM

Well, it locked up after 5 minutes or so on the torture test. So, I'm giving up for now, and leaving it at 1.33gHz and 1.75V until I can get a new PSU.

Thanks again for your help. I'll post back here with results after I get a new PSU.

-Kieran
June 23, 2002 6:03:19 AM

Im sorry to hear that, you can try to set the voltage even lower, give it one more college try, your stepping overclocked very well and thusly can run on lower voltage.

All in all 1.33ghz isnt that much slower than 1.4, enermax is the best psu you can buy btw.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 6:17:39 AM

Quote:
your stepping overclocked very well and thusly can run on lower voltage

What do you mean by the above? "stepping overclocked very well" ? What's stepping? I didn't think I had done any overclocking. Again... total newbie :)  (or "stranger" even...) ;) 

Not that I know what I'm talking about, so correct me if I'm wrong, but from what I've read of others trying to overclock various chips... doesn't one usually need to INCREASE the voltage when overclocking to get stability? (thus the need for better cooling...) So, I assume that as you decrease the voltage for a given clock/fsb speed, you are approaching instability. You've had me dropping the voltage, which makes sense, since you think my psu is at its limit. But am I right to assume that as I drop the core voltage lower and lower, I'm also approaching the limit of what the cpu needs in order to maintain a 1.4gHz speed?

Also, what's the I/O voltage for on the SmartMenuIII setup? (I know I/O means input/output... but what's the voltage setting for this, used for?) I've left it alone through out all this.

-Kieran
June 23, 2002 6:55:48 AM

Quote:
Lower your vcore to 1.7v, then go to 1.4ghz, let me know if it runs, if it does, its your psu, if not try 1.725, if neither of those work the only way to test the psu is to get a new one and try that.


Sparkle psus suck, but it is 300w, it is possible that it is your issue, try my suggestion and get back to us.

lower the vCore for 0.25v against a 300w psu plus a slight configuration?!?
i guess this willn't work.
June 23, 2002 6:59:03 AM

FMI, what is exactly your configuration?
June 23, 2002 7:09:20 AM

to test your psu & to know if it is the issue,

disconnect the alim from the Pioneer dvd-105s, the Lite-on 40x12x48x cd-rw drive, the plain-jane floppy. unplug your soundblaster live, your modem card (if it is) and your NIC.

set your cpu to 1.4Ghz and retry.
June 23, 2002 7:17:44 AM

have you modified your bios settings or update it?
another test is to reset the bios and load bios default.
a b à CPUs
June 23, 2002 7:48:05 AM

Do you like my signature?

<font color=blue>At least half of all problems are caused by an insufficient power supply!</font color=blue>
June 23, 2002 7:49:22 AM

Stepping is the AG*** coding on the CPU. It represents, to AMD at least, revisions made to the core. The T-Birds (pre-Athlon XP) cores had a A*** codes, and the so-called best OC'er is the AXIA, which I own, but with too much overclocking, I am no longer able to reach to 1.4-1.47GHz speeds I used too... I'm stuck at 1.27 on a 1.2GHz CPU.

But I digress. Your PSU woes, if that is indeed the case, should be solved if you get an Antec TruePower or Enermax, preferably around 350W... but the most important part of the PSU is not the wattage, but the amps the PSU can push along the 5V rail. According to AMD's latest system builder's guide (the current Tech Tour that AMD is putting on), the PSU must be able to supply 45 amps along the 5V line. Strangely enough, I have not been able to find a single PSU that is able to fulfill this requirement, unless you're talking 550W or more. So, it is either a typo at [H], or AMD is not talking about the 5V rails in the same way that PSU makers talk about them. Regardless, a 330W or 380W TruePower from Antec or similar wattage PSU from Enermax should be enough... it'll run you around $75. Or, you can do what my brother did and take advantage of <A HREF="http://www.newegg.com" target="_new">Newegg's</A> good PSU prices and get a 430W TruePower, which is way more than anyone would need right now, unless you are a die-hard overclocker and/or case modder (which my brother <i>says</i> he will do, though he has none of the tools needed). I think that ran him $85 after shipping.

As to the Vcore, generally, the highest overclocks you see are due to extreme cooling (liquid nitrogen) and voltage mods that allow +2V to be pushed through the core. This forces the electrons to follow their intended path more closely, but has the undesired effect of greatly increasing core temps and shortning, sometimes drasitically, the life of the CPU. One way to test the overclockability of any given CPU is to underclock it. Usually this is done by dropping the Vcore as low as it will go, then seeing the fastest that the CPU can run stable. If you have something to compare this too, you can get a good idea of how well the system will overclock, as the theory is that if a CPU rated at 1.4GHz@1.75Vcore can run 1.4GHz@1.65Vcore, then it should be able to run the default Vcore past 1.4GHz no problem, which would be considered by some to be a safer overclock than upping the Vcore past default just to get an extra 66MHz. Though, I will attest that just because the Vcore isn't upped, the heat alone can cripple the CPU after a while... *mutters something about overclocking with an AMD retail HSF and 60C temps*

Anywho, if you have a supposedly good OC'ing core, you should be able to drop the Vcore without problem, just make sure to run Prime95. And finally, just to confuse you more, the VIO setting, at least from what I remember, is used to adjust the voltage for the RAM. In theory, just like with the Vcore, bumping that up .1V might be just enough to make a flaky overclock into a solid one. That's a tweak, though, I wouldn't recommend unless you get yourself some heat spreaders for the RAM, and a lot of case cooling.

-SammyBoy
June 23, 2002 8:29:33 AM

Wow... lots of great info. Thanks.

Actually Newegg is where I bought the cpu and RAM. If only they were out of state for me. ;) 

Enermax's "whisper" models of PSU's were recommended to me for their low noise. Not sure which models shown on Newegg's site are the quiet ones though. I don't mind spending an extra $20 on a PSU to go up to 400+ Watts if it means problems like this will be less likely to occur.

-Kieran
June 23, 2002 8:33:11 AM

Good idea about disconnecting all peripherals and re-trying the cpu at 1400MHz. I think I'll try that in the morning.

To answer your questions re. bios... I have not fiddled with the BIOS much at all. It is the latest version from the manufacturer, and hardly any performance tweeks were done. In fact I don't think I did any at all (yet). Just changing cpu speed and voltage throughout this ordeal.

-Kieran
June 23, 2002 9:20:32 AM

umm...

to be slightly more fair...play, i would say something like 5 percents & at least 10...maybe.
June 23, 2002 9:39:13 AM

Quote:
I just bought a new cpu for my system, the T-Bird 1.4GHz(133/266) with the "glorious" AYHJA designation. ;)  It replaces a 1.0GHz t-bird (100/200 fsb)


Quote:
at do you mean by the above? "stepping overclocked very well" ? What's stepping? I didn't think I had done any overclocking. Again... total newbie :)  (or "stranger" even...) ;) 



Ayjha, thats the stepping. Some tbird steppings are legendary for their oc ability, axia and ayjha are the two best.



Quote:
Not that I know what I'm talking about, so correct me if I'm wrong, but from what I've read of others trying to overclock various chips... doesn't one usually need to INCREASE the voltage when overclocking to get stability?


Yes once one gets past a certain point you usually have to increase the voltage,however on the axia and ayjha steppings that point is higher than normal, and it follows a curve, so since one can go higher on the stock voltage, one can reasonably expect to be able to function with lower than stock voltage at stock speeds.

Quote:
So, I assume that as you decrease the voltage for a given clock/fsb speed, you are approaching instability. You've had me dropping the voltage, which makes sense, since you think my psu is at its limit. But am I right to assume that as I drop the core voltage lower and lower, I'm also approaching the limit of what the cpu needs in order to maintain a 1.4gHz speed?


Exactly, the trick is to find out if the minimum the core needs to run at 1.4ghz is less than or more than the maximum your psu can put out, since the lockup for underpowered psus occurs in exactly the same manner as the lockup from an underpowered cpu the only way to find out is to test it. NO harm can come to your cpu or components due to lowering voltage, nor will it harm them to run at said lower voltage for extended periods.

Quote:
Also, what's the I/O voltage for on the SmartMenuIII setup? (I know I/O means input/output... but what's the voltage setting for this, used for?) I've left it alone through out all this.


Thats the power to the pci and agp cards iirc, it draws from a different line than the cpu so lowering that probably wont help, but its worth a try as it cannot hurt.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 9:40:18 AM

Quote:
lower the vCore for 0.25v against a 300w psu plus a slight configuration?!?
i guess this willn't work.


It did work as you can see, just not enough. The fact that when he lowered his vcore .25 it ran longer than @1.75 proves his psu is at fault.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
a b à CPUs
June 23, 2002 9:40:53 AM

Actually it was 15% max in 3D apps with 256MB, and 5% average with 512MB. But I only have 2 256MB DIMMS and wanted to use the other one in another system. Funny thing, 98SE worx GR8 with 128MB!

<font color=blue>At least half of all problems are caused by an insufficient power supply!</font color=blue>
June 23, 2002 9:41:12 AM

Quote:
disconnect the alim from the Pioneer dvd-105s, the Lite-on 40x12x48x cd-rw drive, the plain-jane floppy. unplug your soundblaster live, your modem card (if it is) and your NIC.

set your cpu to 1.4Ghz and retry.


The hdd's etc are on a different rail than the cpu, this may help but it is by no means a sure test.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 9:42:36 AM

Quote:
disconnect the alim from the Pioneer dvd-105s, the Lite-on 40x12x48x cd-rw drive, the plain-jane floppy. unplug your soundblaster live, your modem card (if it is) and your NIC.

set your cpu to 1.4Ghz and retry.


The hdd's etc are on a different rail than the cpu, this may help but it is by no means a sure test.


Furthermore, if he had a hardware issue with one of those components removing them would remove the hardware issue and give him the false impression he had a psu issue.

No offence, but as you can see in the above posts him and I were well on our way to completion and have obviously identified the psu as the issue, thanks for the help but its under control.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 9:44:42 AM

Quote:
Enermax's "whisper" models of PSU's were recommended to me for their low noise. Not sure which models shown on Newegg's site are the quiet ones though. I don't mind spending an extra $20 on a PSU to go up to 400+ Watts if it means problems like this will be less likely to occur.


Even a 300w enermax psu would not have your issue, the problem is quality not quantity, enermax psus have great quality, the moral is. You dont need a 400w psu. 350 is more than enough if the psu you buy is quality.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 10:57:04 AM

ok. that starts again. no matter!
June 23, 2002 10:59:12 AM

</font color=red>
<pre>there is no proof for the moment. wait the end of your helpful argumentation! dude!</pre><p><b>nothing is under control. the issue isn't solved.</b>
June 23, 2002 11:24:31 AM

is there a compatibility problem between your motherboard and this 1.4Ghz cpu or maybe from the motherboard itself?
this could be also your motherboard which doesn't support this processor due to an issue from your motherboard itself? (you said before you never tried this frequency)

there is also another point to check. it is the release of the motherboard. sometimes you can't upgrade your motherboard with a newer cpu because a next motherboard release is needed with the next bios release.

from abit faqs:
Quote:
2002/01/30
Why the system can't boot when I set my Duron 600 at 500MHz after loading the fail-save or optimized default option? (KT7-RAID)

If you experience a 600MHz Duron failing to boot when set to 500MHz (either manually or by fail-safe/optimized default loading), clear the CMOS via jumper and manually set the CPU to the correct frequency (600MHz).

did you try this with your cpu?

another faq:
Quote:

2002/01/10
Does ABIT KG7 / KT7A / KT7E mainboard support AMD Athlon XP processors?

Statement: ABIT KG7 / KG7-RAID mainboard is hardware compatible (ready) and can support AMD Athlon XP processor via BIOS update.
ABIT KT7A / KT7A-RAID V1.3 mainboard is hardware compatible (ready) and can support AMD Athlon XP processor via BIOS update.
ABIT KT7 / KT7-RAID / KT7E all version and KT7A/KT7A-RAID V1.0/1.1/1.2 cannot support AMD Athlon XP processor.
1.ABIT has release BIOS for KG7 and KT7A V1.3 to patch the CPU model name. If user did not upgrade to the latest bios, users can also manually set the suitable working frequency for AMD Athlon XP processors in SoftMenu BIOS.
AMD Athlon XP Processors SoftMenu III Setup -> CPU Operating Speed ->User Define
With above setting users can use AMD Athlon XP processors flawlessly. The incorrect CPU information during POST can be fix via BIOS update.
2.As for KT7 and KT7A old version mainboards, which launched earlier than AMD Athlon XP processors, they are not hardware compatible with new CPU.


and this other faq:
Quote:
2002/01/07
What kind of power supply unit is suitable for the KT7, KT7-RAID, KA7, and KA7-100?
You can find listings of AMD certified power supply units at the following links. If you have high power consumption devices (TV out or capture card) as well as many IDE drives, a stronger PSU is always preferable.
<A HREF="http://www1.amd.com/athlon/power" target="_new">http://www1.amd.com/athlon/power&lt;/A>

June 23, 2002 12:09:31 PM

Quote:
there is no proof for the moment. wait the end of your helpful argumentation! dude!


he lowered his voltage, the cpu ran longer @ 1.75 and locked under heavy load, before lowering the voltage it locked during boot, that is PROOF his psu is the cause.

I know your trying to help but telling him other possible issues (which arent) is only going to confuse the issue and make him spend time looking at things which are not related.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 12:18:09 PM

Quote:
is there a compatibility problem between your motherboard and this 1.4Ghz cpu or maybe from the motherboard itself?
this could be also your motherboard which doesn't support this processor due to an issue from your motherboard itself? (you said before you never tried this frequency)

It ran longer at a lower voltage than at stock, there is not a motherboard issue.


Quote:
there is also another point to check. it is the release of the motherboard. sometimes you can't upgrade your motherboard with a newer cpu because a next motherboard release is needed with the next bios release.

any motherboard which can run a 1.33ghz tbird can run a 1.4ghz tbird, there are no physical differences between those two chips save for .5 of a multiplier.


Quote:
2002/01/30
Why the system can't boot when I set my Duron 600 at 500MHz after loading the fail-save or optimized default option? (KT7-RAID)

If you experience a 600MHz Duron failing to boot when set to 500MHz (either manually or by fail-safe/optimized default loading), clear the CMOS via jumper and manually set the CPU to the correct frequency (600MHz).

That is because of the 5x multi being reused by amd for 14x, it has nothing to do with his issue.


Quote:
2002/01/10
Does ABIT KG7 / KT7A / KT7E mainboard support AMD Athlon XP processors?

Statement: ABIT KG7 / KG7-RAID mainboard is hardware compatible (ready) and can support AMD Athlon XP processor via BIOS update.
ABIT KT7A / KT7A-RAID V1.3 mainboard is hardware compatible (ready) and can support AMD Athlon XP processor via BIOS update.
ABIT KT7 / KT7-RAID / KT7E all version and KT7A/KT7A-RAID V1.0/1.1/1.2 cannot support AMD Athlon XP processor.
1.ABIT has release BIOS for KG7 and KT7A V1.3 to patch the CPU model name. If user did not upgrade to the latest bios, users can also manually set the suitable working frequency for AMD Athlon XP processors in SoftMenu BIOS.
AMD Athlon XP Processors SoftMenu III Setup -> CPU Operating Speed ->User Define
With above setting users can use AMD Athlon XP processors flawlessly. The incorrect CPU information during POST can be fix via BIOS update.
2.As for KT7 and KT7A old version mainboards, which launched earlier than AMD Athlon XP processors, they are not hardware compatible with new CPU.


your second faq link is about athlon xps, he has a tbird not an xp, it does not apply and is only confusing the poor guy.

And your final link, about approved psu's. The sparkle corp is on the approved list, but that dosent mean his 300w is able to cut it. (I personally as I said, had a sparkle which could not handle a 1.5ghz tbird)

I am trying not to be rude, but the issue is clearly visible, he needs to upgrade his psu, there is no other causal factor.


This guy came for help and he got it, having him take things out of his system or upgrade bios's or worse may make the situation worse or cause confusion which hides the true problem.

Trust me on the psu call, I have been doing this for years and I have helped many many people with issues similar to the ones he is having.

Note how my initial reccomendation immediatly made his situation better(and tested the psu hypothesis at the same time).

Giving him a bunch of non related faq links and telling him he may have a motherboard incompatability only serves to obfuscate the true issue, which is his psu cannot cope@ 1.4ghz.

I appreciate your attempts to help though, too many people just troll/flame and dont spend time helping, on this particular case however your advice is not needed as we have resolved the issue friend.




:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 12:22:56 PM

Quote:
he lowered his voltage, the cpu ran longer @ 1.75 and locked under heavy load, before lowering the voltage it locked during boot, that is PROOF his psu is the cause.

we can play with words, if you want to.

"he lowered his voltage, the cpu ran longer"
this is not to say the cpu ran but ran "longer"; a cpu which run longer is a non-sense. a cpu run or not run & that is all.

we must not mix a weird phenomenon and a real issue.

btw, this is a confusion at all.

proove really the cause is your psu with a tangible manipulation & you would have find its issue. but for the moment there is nothing but wachiwacha.
June 23, 2002 12:26:51 PM

Quote:
we can play with words, if you want to.

No we can use logic to deduce a problem, I will spell it out for you.

Quote:
"he lowered his voltage, the cpu ran longer"
this is not to say the cpu ran but ran "longer"; a cpu which run longer is a non-sense. a cpu run or not run & that is all.

@1.75 volts the cpu locked during boot, getting to windows was rare.

@1.7 volts windows runs fine and it only locks during heavy load when the components are strained sucking more voltage than the psu can provide.

@1.75volts@1.33ghz the extra 66mhz worth of cpu activity is enough to bring the needed power below the issue level.

Running the cpu@ 1.4ghz with a lower voltage achieves more stability than with the higher voltage, thus excluding motherboard issues as the cause(because if the mobo had an issue with a 1.4ghz chip the result would be the same no matter what voltage was used).

The fact that changing the voltage at all resulted in more stability is the big sign its a psu issue.




Quote:
btw, this is a confusion at all.

proove really the cause is your psu with a tangible manipulation & you would have find its issue. but for the moment there is nothing but wachiwacha.

Your not making sense, and your faq links are all off topic non related, with exception of the approved psu list which really has no bearing on the analysis.

Please stop trying to help the poor guy because all of your suggestions are not relavent to the problem, I have tried to be nice about it but you are accusing me of playing with words?



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 12:42:29 PM

Quote:
<b>Trust me on the psu call</b>, I have been doing this for years and I have helped many many people with issues similar to the ones he is having.

Trust & Logic are completely different, friend.
June 23, 2002 12:58:08 PM

Quote:
Trust & Logic are completely different, friend.


Indeed they are, which is why I overly laid out the reason the psu is the issue, there is no other logical conclusion.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 23, 2002 1:09:34 PM

ok you win. i trust you. (haha)
June 23, 2002 6:01:06 PM

Well, that was fun! :wink:

Seriously though, thankyou very much to all of you for your help and advice. Much appreciated. Even if those "OT" links that "8235k8hta" posted weren't applicable to this, they were still informative. Although, I already knew about the mobo&cpu compatibility, which I sort of noted early on in the thread when I said that my mobo had the latest BIOS available, but was NOT the newer v.1.3 kt7a, which is AthlonXP compatible. Upgrading my BIOS to the latest version last month was actually what prompted me to upgrade the cpu... because the new BIOS added support for T-Birds up to 1.4GHz, which it turns out is the fastest cpu that I'll ever get for this mobo, since AMD is discontinuing the t-bird line (or so I've heard.) So, I figured I may as well max out my performance with this mobo while I can. :smile: Of course, what was to be a simple cpu upgrade grew: new HSF, new RAM (hey why not?!), and now a new PSU. :tongue: (sigh) It's never simple is it?

Anyway I am fairly convinced that the PSU is the culprit, from the tests Matisaro suggested, and the comments of others in the thread, and my own (limited) knowledge of electronics (I'm *only* a mechanical engineer... but did study a few electronics & power courses!)

Besides, this gives me a good excuse to get a quieter PSU anyway! This Sparkle PSU gives my girlfriend's hair dryer a run for "loudest fan" award. :wink:

So anyway... thanks again, and I will post back here when the new PSU is installed. Then we will know for SURE! (place your bets, people! :lol:  )

-Kieran
June 24, 2002 10:42:28 PM

I just ordered the Enermax EG365P-VE FCA 350W PSU, for $69 +tax +shipping from newegg.com... $80 out the door. Should have it installed by Wednesday night. :) 

BTW, spoke to one of the EE's here at work who does this stuff for a living, and he explained to me that while what Matisaro said is 100% correct, that it may not be the cpu that is failing as I drop the voltage down. The cpu may infact be quite stable at that voltage. He said that it's VERY likely that some of the 5V devices (many components on the mobo) are failing as their voltages drop to less than 5V, and it is this "brownout" condition that causes a system lockup. He said that the PSU will try to keep the current supply to the cpu up, and to do that, since it is limited by P=I*V ~= 300W, it will drop voltages throughout the system, (not just the cpu), as the cpu demands more current as processing demands rise. So a lot of 5V components (chips) on the mobo will start to see 4.8, 4.5, volts, and will fail, causing a system lock. He said it COULD also be the cpu, but less likely.

So, this is second hand, so if you disagree, don't shoot the messenger! ;)  Just thought it was interesting, and that some of you might find it interesting/informative also.


-Kieran
June 25, 2002 12:15:59 AM

thats the problem with many cheapass PSU's. ive been a victim myself, and ive seen it happen to a friend.

they struggle on, pumping out the heat, right on the edge. an extra half volt here, new device there can push em over the edge.

low quality PSU's cause a large proportion of problems encountered with amd systems.

strange really. people get high quality parts for everything else, but neglect the PSU. Companies who assemble PC's are also guilty of this practice too. they do it to save a few $$$ mostly.

<font color=blue>All religions are true, for a given value of 'true' - Terry Pratchett.</font color=blue>
June 27, 2002 6:59:02 AM

Okay, I installed the new PSU: An Enermax Whisper 350W EG365P-VE(FCA). The PSU seems to be working fine, and it's much quieter than my previous Sparkle PSU. However, the CPU is NOT running fine.

Same symptoms as before. Locks up at 1.4GHz w/ 1.75V core voltage, with very little action in the OS. It's more stable at 1.4GHz w/ 1.70V core, but still locked up after about 20 minutes of prime95 torture testing.

Temperature is not an issue. CPU temp is stable in around 42C, and hardly budges even with the torture test.

Any ideas? My full configuration is listed earlier in this thread.

Thanks! :) 

-Kieran
June 27, 2002 9:18:15 AM

Quote:
My system:
*Abit KT7A-RAID w/ BIOS version 7N, which supports up to 1.4GHz T-birds, and 133MHz fsb. (this mobo is NOT the new version 1.3 KT7A, but the original with the latest BIOS).
*512MB PC133 Crucial RAM, CL=2 (2x 256MB dimms)
*Asus v7100 geforce2mx video, 32MB
*Windows XP pro.


Try this, lower the vcore as low as it will go in your bios, and see if it is more stable or less stable, if its less stable increase in .25 incriments untill you reach the MOST stable point.


Then return and tell me what that point was.


Now that we have ruled out the psu, you have a VERY weird issue, lowering the voltage giving more stability hints at either a psu or heat issue, since you replaced the psu, we are left with heat, which seems unlikely given the fact your@42C which is nice.

You may in fact have a bad tbird, which should be covered by warranty, however if you can get it running at 1.4ghz stabilly, even if its below stock voltage, then that is a working solution.


Also try this experement, try running it with .5 more of a multiplier(1.466) with stock and 1.7 volts, tell me what happens.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
June 27, 2002 4:22:46 PM

Thanks. I'm at work now, but I'll try to implement your suggestions tonight when I get home.

-Kieran
June 27, 2002 5:17:12 PM

I just found this at <A HREF="http://www.viahardware.com/faq/kt7/kt7faq.htm" target="_new"> Paul's Unofficial KT7 FAQ </A>:
Quote:
* Why does Windows lock up when I use a 1.4GHz Thunderbird?
It has been reported that with processors running at above 1.33MHz it is necessary to set the K7 CLK_CTL to Optimal in the BIOS in order for the system to be stable. An alternative to this is to use BIOS WW or earlier, which does not provide this option. If K7 CLK CTL is set to Default then the machine becomes unstable. Another user wrote to me that the reason for this is that at high speeds the ACPI HLT function does not work correctly and as soon as the processor begins to cool it will lock up. The best solution (since changing K7_CTL to optimal is not completely stable) is to disable ACPI completely under these circumstances. Note that ABIT's beta BIOS 3Rb02 has ACPI disabled to resolve this stability issue

I don't even know what ACPI is, or what disabling it will do to my system. Any ideas? Does this sound like something that might be applicable to me?

-Kieran
!