Reply 20 of 79, by Mamba
- Rank
- Oldbie
Now I really want to know how it goes.
Now I really want to know how it goes.
Thanks for all the replies, lots of useful info.
For bad capacitors - I did look them over and they all appeared pretty normal, but I'm not really that experienced at spotting that sort of thing so I'll take some pictures of the board later to post. Is there any way to electronically test them in case they may have failed invisibly? I have a multi-meter and just enough knowledge to be dangerous.
I would quite like to get a TNT2 card anyway, and it sounds like that might be a good idea to test the motherboard so I think I'll go ahead and do that. I'll also go ahead and get the S3 properly installed in Windows and stress it a bit with some benchmarks, and see if it crashes when drawing more power.
One thing about the POST beeps I am getting with the Voodoo3s - I am not 100% they are the 1 long plus 2/3 short "VGA failure" one. The reason I say that is I accidentally got the "real" VGA failure beeps with the S3 when I forgot to plug in the monitor cable, and it sounded very different from the one I get with the Voodoo3s - much more well defined individual beeps with pauses between them, whereas with the Voodoo3s I just get one long beep with some "stuttering" towards the end that I assumed was meant to be the "short beeps".
Wait.... why the hell does your PC not POST without a monitor connected? Never seen that kind of behaviour.
I was wondering the same alexanrs. Perhaps during testing he is not screwing the cards down in the slots, and on that run the card was not seated fully?
The more I learn, the more I realize how much I don't know.
OPL3 FM vs. Roland MT-32 vs. General MIDI DOS Game Comparison
Let's benchmark our systems with cache disabled
DOS PCI Graphics Card Benchmarks
I think it does POST without the monitor, it just puts out the VGA error beeps cause no VGA makes it angry.
Sometimes beep-codes gives you the answer on what to look for.
Not that they give you one solid answer. They are more like a tool.
The BIOS is an award one? Right? Have googled it, and it looks like
there is both 1-long + 2-short and 1-long + 3-short.
https://www.google.dk/search?q=award+post+bee … eNqe3ygOt-LygCg
Don't eat stuff off a 15 year old never cleaned cpu cooler.
Those cakes make you sick....
My blog: http://to9xct.blogspot.dk
My YouTube: https://www.youtube.com/user/brostenen
001100 010010 011110 100001 101101 110011
wrote:I think it does POST without the monitor, it just puts out the VGA error beeps cause no VGA makes it angry.
All motherboards I've ever handled boot just fine without a monitor attached to the video card. In fact, there is no way for a VGA card to inform the motherboard whether the monitor is attached or not before specific drivers are loaded. If you are testing without screwing it to a case then perhaps it is relying on the VGA cable for grounding... but that should not happen (the slot has ground rails)
wrote:Is there any way to electronically test them in case they may have failed invisibly? I have a multi-meter and just enough knowledge to be dangerous.
With just a multi-meter the answer is sorta, but not really. There is an old school check but it does not fully test the cap.
IMHO it is best done with an analog multi-meter because the movement of the needle is informative.
With a digital multi-meter the numbers jump around so you can't tell if the cap is charging or discharging.
Despite what some people think you can't check caps in-circuit on a motherboard. You must remove them to test them.
This is because on motherboards the caps are often in parallel with other devices including other caps.
To really test caps you need:
An ESR meter that uses a 100kHz test signal and is readable down to 0.01 ohms. (Digital)
A basic capacitance meter.
A basic multi-meter. (With an ESR meter also the type doesn't matter.)
[An ESR meter is basically a resistance meter that only checks AC resistance. Ripple is AC and is generally around 100kHz.]
They can also be tested with an O'Scope but that's a PITA. I have 3 O'Scopes and I never do it that way.
To tell the truth, it's been so long I don't even remember how to do it that way.
.
.
RANT
Aluminum Electrolytic Capacitors were the electronics part with the highest failure rate long before the famous "Capacitor Plague", and they still are.
They are a roll of aluminum foil stuffed into a metal can and immersed in a electro-chemically active liquid. Not exactly solid state.
And, the older they are the higher the failure rate is. - Retro electronics are not getting any younger.
So if you are going to be into any kind of old electronics as a hobby you ARE going to run into bad caps now and then so it behooves you to know about them and how to replace them.
.
GRUMPY OLD FART - On Hiatus, sort'a
Mann-Made Global Warming. - We should be more concerned about the Intellectual Climate.
You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
Yes. A computer will boot without a monitor but you may be missing helpful error information.
GRUMPY OLD FART - On Hiatus, sort'a
Mann-Made Global Warming. - We should be more concerned about the Intellectual Climate.
You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
What was the username of the seller of the Voodoo cards? I ran into a bad batch of them from the seller of the Compaq version of the V3 3500.
Didn't get a change to do any work on it last night (the retro machine currently shares a monitor with my wife's PC, and she was busy gaming on it); tonight I should have time to double check the caps and get some pictures of them. It sounds like the effort of actually testing the caps is such that it's a better idea just to replace them with nice new ones, should they be at all suspect.
What you guys are saying about the monitor disconnection POST beeps is interesting - it seemed odd to me that it would complain about it too, but thought it perhaps just an idiosyncrasy of this board. I'll give this aspect a more deliberate test this evening - perhaps the card just wasn't seated right and plugging the cable in pushed it down into good contact.
One of my friends is an electronics engineer, and he has suggested I (carefully) test the voltages running through the appropriate pins in the AGP port on the motherboard and make sure they match what they should be. Does that sound reasonable?
The seller I bought the Voodoo3s off is combinationinc on eBay; the cards are listed as being tested and with a 30 day warranty, so I dropped the seller a message letting them know I was having trouble just in case the cards do actually all prove to be faulty. I presume all those Compaq Voodoo3 3500s (working ones, anyway) are long gone by now?
wrote:What was the username of the seller of the Voodoo cards? I ran into a bad batch of them from the seller of the Compaq version of the V3 3500.
Same here, US seller, dead cheap price. 2 cards that both booted and even played games for about an hour before the PC crashed. After that it started to beep on power up (as if there was no VGA) and they don't function anywhere any more. I couldn't believe it, I tried them again and again on 3-4 different PCs before tossing them.
Did you try any more? I mean did we get by bad luck the few defective ones, or are they all like this?
Wild idea: Could there be something that fries these cards on some motherboards? For me it was an Intel SE440BX-2. Then I got a few 3000 & 2000 that still function on the same M/B after many hours/days of benchmarking/stressing/playing a lot on purpose.
I hope you guys won't consider this discussion completely off topic, the pattern is so much alike that in fact is the reason I'm following this thread. Maybe something will come out of it for all of us to learn...
EDIT: yeap, I just read the seller's name. It's the same for me. Those are the Compaq OEM 3500 cards that boxpressed mentioned. So that's why the pattern looked soooo familiar, I guess that makes three of us now. Sorry man, but I don't think your card will ever function again. Now the question is if every card is like that or something fries it. I told you we're gonna learn something from this thread 😉
If 3 Vogoners experienced the same issues from the same cards from the same seller, then what could be the problem here?
Were all those cards old RMAs that were never destroyed? Were they really the Compaq 3500s or were they reflashed 3000s? Do all of these use the same PCB or are they different?
I was even considering (very briefly!) getting a couple myself (even though I don't actually need more V3s) because of the good price and good feedback, but this seller was in the USA, right?
Yes, it's a US seller. As far as I know and can judge, they were all brand new, unused, proper Compaq OEM 3500s. Never inserted into a slot (no marks on the contacts, so certainly not tested), extremely clean (100% sure they are new basically) and proper Compaq stickers with serial #.
For me the problem is the cards. What I don't know and would like to learn is
-whether there is a possibility of them getting fried under certain conditions
-whether this was indeed a defective stock that is getting re-sold for cheap (intentionally or not)
-whether these cards have some kind of time-triggered weak point
-if the latter is the case, if it's a replaceable component.
btw, I'm sure I've read more about these cards on another thread, just the seller's name and the cards weren't identified that precisely. So I guess we're much more than 3.
Count me in as someone who bought one of those $10.00 Compaq OEM Voodoo3 3500 cards from combinationinc. I also bought it to use in a SS7 build (Asus P5A 1.04 motherboard). Initially, mine would not POST (got the 'no-video' beep code) so I set it aside and swapped in a Voodoo3 3000 for a while. Then maybe about a month ago I tried plugging the Compaq 3500 back in and it worked! I haven't done any serious stress testing, but it is still working after numerous reboots and running dxdiag a few times. Keeping my fingers crossed.
AMD386/IIT387DX40, 32MB, ATi Mach64, AWE64
Compaq Prolinea 4/33, 32MB, Tseng ET4000, PAS
AMD X5, 64MB, S3 Virge/Voodoo1, AWE64
AMD K62+550, 256MB, Voodoo3, AWE64 Gold
P3 1.2Ghz, 512MB, Radeon 7500/Voodoo2 SLI, SB Live!
Someone with one of these bad cards from combinationinc please try the following experiment:
Install the card.
Start the system and just let it sit doing nothing for about 2 hours.
(Ideally at the BIOS screen but that isn't mandatory. Just needs power to the card.)
Power it down and let it sit until it cools completely.
Then try the card again.
.
ultimate386's experience suggests this will work.
That scenario is not that uncommon when parts with caps have been improperly stored long term.
If it does work I will explain what I think happened with these cards in detail - and most of these cards could be brought back.
.
GRUMPY OLD FART - On Hiatus, sort'a
Mann-Made Global Warming. - We should be more concerned about the Intellectual Climate.
You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
Seems like a lot of people had bad experiences with that seller!
I'll give PCBONES suggestion a try now; would obviously be great if these could be revived.
wrote:Seems like a lot of people had bad experiences with that seller!
Yes.
And you'd never know it because eBay's feedback system has been bogus and fraudulent since John Donahoe took over eBay in 2008.
.
GRUMPY OLD FART - On Hiatus, sort'a
Mann-Made Global Warming. - We should be more concerned about the Intellectual Climate.
You can teach a man to fish and feed him for life, but if he can't handle sushi you must also teach him to cook.
So I've tried PCBONEZ suggestion of running the system for a few hours then letting it cool a few times, and with two different Voodoo3s, and in neither case did it seem to fix or have any effect on the problem.
The seller has replied they are happy to refund me if I return the cards and they test them and find them faulty - but given I only have the one motherboard to test with I'm still not convinced enough that it is the Voodoo3s at fault and not the mobo; certainly not enough to be international shipping costs on it.
I have a TNT2 Pro based AGP card arriving next week - I figure if I try that in the motherboard and it works it makes it a lot more likely that the Voodoo3s are at fault, given the TNT2 should in theory be drawing just as much if not more juice.
I'm actually tempted to just buy a Pentium4 + socket 478 mobo with AGP just to have something else to test them in - can be had for about £15, and you never can have enough retro hardware...