Purely anecdotal...

LawnMMLawnMM Colorado
edited June 2004 in Hardware
The cpu I was having problems with...it seems to run fine underclocked by 66mhz. Drop the multiplier from 13.5 to 13 and it runs fine. Had it sitting here running the prime95 torture test for more than a day. Tried it at 13.5 as a control test to see if that was it...froze in about 10-15 mins.

Clearly the chip is bad, it doesn't work at stock speeds, period. Though it does seem to be okay one multi lower, and she'll never see the difference. I just thought it was interesting what a finicky piece of hardware it turned out to be. Thought I would share it with you all.

Comments

  • muddocktormuddocktor
    edited May 2004
    What kind of chip is it and did you buy it from a store or from someone else, Lawn? If you bought it from a store, then return it for a good proc. ;)
  • edcentricedcentric near Milwaukee, Wisconsin Icrontian
    edited May 2004
    as you are well aware it might actually be a issue with the mobo.
    I have one CPU/mobo combination that will not work. Swap parts around and all is fine.
    Oh well
  • LawnMMLawnMM Colorado
    edited May 2004
    Its a 2200+ chip that came stock on the gf's HP :banghead:

    Warranty's expired so its either find a way to make it work or replace it outright. If I had bought it I'd have already sent the thing back, but alas. Its not a mobo issue, which was honestly my first thought...I had it narrowed down to the crappy HP mobo or the sweet AMD cpu. I figured the mobo was the likely suspect so I bought some replacement parts off newegg.

    Threw it altogether and it was still acting funky till I started messing with the multiplier for the hell of it. 13.5 = lockup. 13 or lower...runs fine. Ain't that some shiz? I'll probably just run at it 1730mhz rather than 1800. Real world its not even worth mentioning, there's no tangible difference realized with 70mhz.

    I just thought it was kinda funny in that flukey never would have believed it would happen on an AMD chip sort of way...
  • FlintstoneFlintstone SE Florida
    edited May 2004
    Maybe the power supply is crapping out and not providing enough for the proc at 13.5? It is a crappy HP after all.


    Flint
  • LawnMMLawnMM Colorado
    edited May 2004
    I tried upping the voltage when I stuck it in the new mobo and case, no effect. =(
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited May 2004
    I think what Flint means is that the power supply may not be putting out enough power for the 2200.
  • FlintstoneFlintstone SE Florida
    edited May 2004
    Geeky's right as in HP mostly uses the cheapest they can get away with and it just may not be up to the task.

    Flint
  • LawnMMLawnMM Colorado
    edited June 2004
    Any suggestions on how to test that theory? What is uATX?

    Here are the specs as listed by HP for her motherboard...

    http://h10025.www1.hp.com/ewfrf/wc/document?product=90387&lang=en&lc=en&cc=us&dlc=en&docname=bph07899

    Same for her sister's motherboard, who's pc is now also acting up...

    http://h10025.www1.hp.com/ewfrf/wc/document?product=296562&lang=en&lc=en&cc=us&dlc=en&docname=bph07941

    Can anybody review that and tell me if I can plug a regular ATX power supply into these mobo's without frying it? Can I test to make sure with a multimeter? Never done it so I'll need some details. It looks like a cheapo KT266 knockoff, probably isn't even Kt266a, regardless it seems like a normal socket a mobo...

    Trying to get these things running in the cheapest manner possible. As to the theories on the GF's 2200+ not getting enough power, I believe that to be unlikely as it still locks up at default settings. I have to underclock it by stepping the multiplier down one notch at a cost of about 66mhz clock speed for it to work. Interestingly enough, it runs 100% stable with no problems once thats done.
  • TheBaronTheBaron Austin, TX
    edited June 2004
    im fairly sure uATX is proprietary, thats what you get with companies like HP and Dell.

    if underclocking by a mere 66 mhz works, i'd count my blessings and leave it alone. sounds like more pain than its worth
  • muddocktormuddocktor
    edited June 2004
    Lawn, if you have a voltmeter handy, you can easily check the pinout voltages on the ATX psu while the machine is running, to make sure that HP didn't pull a Dell and change the pinouts on the psu. Just get a manual from Abit, Asus, etc that shows the correct voltages on each line for comparison, then boot up the machine and check the voltages on each wire on the ATX connector by sticking the meter probes into the back of the ATX plug where the wires go into the plug. If the measured voltages are the same as shown in the manual for the ATX headers on a standard mobo, then you should be able to sub just about any ATX psu in that fits the case.

    BTW, I just subbed a regular ATX psu into a Compaq case the other day, in place of the matx psu. It was a fairly tight fit, but it worked. As long as the pinouts are the same, you are good to go.
  • LawnMMLawnMM Colorado
    edited June 2004
    TheBaron wrote:
    im fairly sure uATX is proprietary, thats what you get with companies like HP and Dell.

    if underclocking by a mere 66 mhz works, i'd count my blessings and leave it alone. sounds like more pain than its worth

    Well, the problem with that is...its 200+ bucks worth of parts to be capable of doing that. The HP bios won't let you change the multiplier. So I had to take her cpu, memory, and the other non crap parts and put them in a new mobo with a new case and new video card.

    It needs to be done, she needs to cut loose of HP anyway and join the ranks of the do it yourselfers who can fix problems for 50 bucks and be done with it rather than paying ridiculous replacement costs to HP. However, her sister's machine is acting up now too...and I don't have another 200 bones laying around to fix her sister's machine TOO.

    I'll have to borrow my father's multimeter and check the pin voltages with each machine running to see if they're the same.
  • qparadoxqparadox Vancouver, BC
    edited June 2004
    uATX = micro ATX. The pinout voltages should be the same. I haven't tested that theory on a modern HP though so it may not be.
  • qparadoxqparadox Vancouver, BC
    edited June 2004
    uATX = micro ATX. The pinout voltages should be the same. I haven't tested that theory on a modern HP though so it may not be.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited June 2004
    If it runs the NF7 without frying it, the pinouts are the same. The problem may be that the PSU can't provide enough power for the system.
  • LawnMMLawnMM Colorado
    edited June 2004
    Geeky1 wrote:
    If it runs the NF7 without frying it, the pinouts are the same. The problem may be that the PSU can't provide enough power for the system.

    The NF7 is running on a new PSU that came with the new case. I haven't tried crossing them from one to another. Perhaps a simpler answer is available though, I'm gonna take a look at the HP PSU itself, if its a micro atx mobo the PSU should say ATX on it somewhere right? I know the antec does and every other PSU...maybe I overlooked the easy answer to this question.
  • Geeky1Geeky1 University of the Pacific (Stockton, CA, USA)
    edited June 2004
    Yeah, Croc_'s Dell PSU said ATX on it too... ask him how plugging that into a P4C800 worked out... ;D
  • LawnMMLawnMM Colorado
    edited June 2004
    Well, I popped her case open. Won't need the multimeter, her PSU had the wire colors broken down with the voltages for each wire right on the PSU. It matches up perfectly with the pin diagram in the NF7-S manual except the NF7-S calls for a -5vDC line 3rd pin down on the right hand column (you can double check me fellow NF7 people). Her PSU has no -5vDC line. I compared it to the ATX PSU in the new case the NF7 is in right now, it does have a -5vDC line. Anybody explain the significance of that? My uncommon sense tells me no go :(

    **Update** - This looks somewhat promising...found the manual and a compatibility report at the FIC site...

    http://www.fic.com.tw/support/download.aspx?dl=motherboard%2fmisc%2fcompatibility%2fsocketa%2fam35%2fcomp_am35_pcb12.zip&title=comp_am35_pcb12.zip&info=compatibility

    Lists compatibility tests with a number of different things, one of them being power supplies...its on page 21 and they're all listed as ATX. Trying to find the pin diagram...

    **Update** - Bingo! Found the manual...
    http://www.fic.com.tw/support/download.aspx?dl=motherboard%2fmanual%2fsocketa%2fam35%2fam35_a0_eng.zip&title=am35_a0_eng.zip&info=manual

    Their pin diagram is identical to the NF7 and the manual states that it uses ATX power. I wonder if the absence of this wire has anything to do with the system's instability
  • croc_croc_ New
    edited June 2004
    Geeky1 wrote:
    Yeah, Croc_'s Dell PSU said ATX on it too... ask him how plugging that into a P4C800 worked out... ;D


    Shhhhhhhh you.
  • LawnMMLawnMM Colorado
    edited June 2004
    Upon closer inspection I count 20 wires from the HP PSU and 20 from the working ATX PSU. Interestingly, there's that missing -5v wire on her connector, I found the extra wire attached to one of the +3.3v pins...it has the 3.3v wire leading into the connector as per the color breakdown on the PSU. It also has another wire, brown, which isn't on the breakdown and isn't used elsewhere on the connector also in that pin of the connector. We looking at a possible manufacturing error? Anybody know what the significance of the -5vDC wire is?
  • muddocktormuddocktor
    edited June 2004
    I noticed the same missing wire on a Compaq/HP Celeron computer a friend is working on. We changed that weak-kneed mATX psu out for a regular ATX psu and it ran just fine with the Compaq mobo, which is oem'ed by Asus in that box. I don't think the missing wire makes a difference, at least on the mobo my friend was working on.
  • LawnMMLawnMM Colorado
    edited June 2004
    I tend to agree. Everything I dig up points to it being a regular ATX power supply with identical pin configurations. I'm not gonna rip the new machine apart for spite, but what I will do is test the theory on her sister's machine. I'll swap the processor over to the stripped down HP I have in the other room, and I'll plug a good ATX psu into it and see what happens.

    If its the same problem, it'll lock and freeze up anyway when I start folding or prime95 and let it run. If that happens, I'll try moving her cpu over to the NF7-S rig I'm building the gf. I'll try droppin the multiplier one notch and see if that stabilizes her cpu as well. Either way, I do not believe the motherboard was the original cause of the fiasco at this time.

    I believe we're either dealing with a bad cpu that needs to be slightly underclocked to regain stability (definitely the cause of the problem with my gf's pc) or a power supply that isn't up to the job thats causing problems for my gf's sister. They both bought them around the same time so it wouldn't suprise me if I'm running into a batch of faulty cpu's. Maybe it was a bad run or problems with shipping or on the line when they were built.

    Either way I'll know soon whats wrong with her sister's comp. I'm hoping its just the cpu because thats the easiest and cheapest to fix. I wouldn't mind if its the PSU either provided I don't run into trouble there. I'm gonna plug the atx psu into my gf's mobo first since its already been replaced by something better and if I burn it out...well, no harm no foul. I don't think that'll happen though.

    I'll keep you guys posted. Thanks to everybody for tossing out ideas and muddling through this with me.
Sign In or Register to comment.