Changing Caps?

[JayDee]

I have changed the Caps on my VP6 board, that is, all the ones that were recommended to be changed.
My question is, why not change ALL the Caps on the board?.
If the suppliers of the Caps are known to be producing faulty components,
surely we should change all of them.

[Topcat]

The main reason I don't recommend it is because the caps over by the CMOS/BIOS are VERY SMALL, and so are the holes!! Soldering in that area can be a very delicate matter, and extremely easy to make a mistake. The ones by the VRM's, AGP, and RAM are the ones known to go, so to keep it simple and effective, those are the necessary ones. I've NEVER seen any others on the VP6 go bad, and I've recapped about 100 board so far...

[stretch0069]

When manufacturers build boards, not just mobos, they will get different components from different sources. They might get a 2200 µf cap from...say...Panasonic and a 1000 µf cap from...oh...Nichicon. It depends on cost.

It could also be that whoever made the bad caps didn't offer the caps in the value range for CMOS (as an example) or might not have made the best deal. There are a lot of variables.

I know when I was working putting together components for mass board assembly, I would purchase the cheapest (price) components that were known to be reliable. Nothing is ever 100% though.

Topcat makes a good point also. Unless you have the right equipment, messing around on a multilayer board can be dangerous.

[JayDee]

The

[Topcat]

The actual reason the smaller caps don't go bad is this:

The 1500uF abd 1000uF caps are known as 'low ESR' caps, which is SUPPOSED to be their high-quality cap... The others on the board that survive are not. They use different electrolyte as the others on the board. If you notice on your VP6, right by the VRM, there are 8 1500uF caps and 4 1000uF caps... If you notice, those 1000uF caps NEVER EVER go bad, but the same valued low ESR ones located by the AGP are usually always dead...

What is ESR?!?
ESR is 'Equivalent Series Resistance' within the capacitor. ESR acts like a resistor in series with a capacitor thus the name Equivalent Series Resistance. Low ESR capacitors are not supposed to create this problem, but in the case of our VP6's they simply blew up... The caps that don't fail on the other hand were considered their lower-end or general purpose quality... Those caps will probably live forever. Thier \"high quality\" low ESR caps are the ones that are dying.

I hope that helped clear that up a little...

[ncdave]

No, the fact that 1000uF and 1500uF caps are failing does NOT mean you should worry about smaller ones. According to what I've read, different sizes and types of aluminum electrolytic caps use different electrolyte formulas, and apparently only one particular electrolyte type was bad. The bad caps are all low voltage (probably just 6.3v & 10v, and maybe 16v), high capacitance (probably 1000 uF and above), low ESR (usually identified by gold writing on green or black plastic sleeve). Only those caps are likely to be bad, and only those made with electrolyte purchased by the capacitor maker from one particular vendor. So if your big Jackcon caps with gold writing are failing, your little Jackcon caps with white writing are probably nevertheless okay.

See http://www.burtonsys.com/bad_BP6/

-Dave

[JayDee]

Thank you all very much, I stand corrected,- - -and not as worried!!
JayDee.

[Socket2Me]

I recently won a nice BP6-based box on eBay for $81, plus $15 for S&H! It included a pair of Celeron 366 CPUs w/ Alpha 7HO HS/fans, a Mitsuko ATX-250 PS, & a Maxtor 33073H3 30.7GB HD, w/ W2K installed. It POSTs fine, but loses its CMOS settings when powered down completely, even though I replaced the 3V CR-2032 cell w/ a new one. Also, it returns a \"DISK BOOT FAILURE, INSERT SYSTEM DISK AND PRESS ENTER\" message after I reset the CMOS and proceed, after saving the settings. Since I haven't the sys disk, the unit having come w/ W2K installed, I can't do that, and I hesitate to run Maxtor's PowerMax utility to check the HD's integrity (which, I believe, will ruin the W2K installation in the process). The HD does spin, albeit somewhat noisily. Does this mess sound like bad ECs to you guys? The fellow I got it from insists that it booted up fine a number of times before he sent it out, & I'm rather inclined to believe him, since there is no visible cap leakage or swelling, although its ECs are the Taiwanese Tayeh gold-on-green low-ESR jobs, & white-on-black standard ESR ones. I also realise the PS is not too big, but the 'puter is not maxed out by any means. (Mitsuko is an unknown quantity to me, power-regulation-wise.) Any thoughts on the matter that anyone here might have would be very much appreciated, since I'm at a standstill with this Digital Dragster (I even got some Celeron 533s to take it to its non-PowerLeaped SMP limits!) Also, if I do decide to re-cap it, has anyone around used tantalums instead of 'lytics to re-do an Abit (or other motherboard)? What are the pros & cons of that? Thanks a heap for the help!

[ncdave]

Socket2me, there's one other clue you can look for. Go into the BP6's BIOS Setup, and look at the Chipset Features screen. Look at the CPU temperatures and the various voltages. Are they fluctuating?

If they are all fluctuating wildly, then you definitely have either bad caps or a bad power supply, but probably bad caps. (Actually, this symptom is probably caused by one particular bad cap, but I don't know which one.) If VCORE and/or VTT are fluctuating, then you probably have bad caps. If the supply voltages from the power supply are all stable (and correct), then the power supply is probably okay. If no voltages or temperatures are fluctuating, then you haven't learned much about the caps: at least some of your electrolytic caps are probably okay, but other caps might or might not be bad.

Note that a voltage that bounces back and forth between two approximately correct values is not \"fluctuating wildly.\" That is just an effect of the rather coarse granularity of the voltage sensors. E.g., a \"5v\" supply that is alternately 4.97 and 5.01 is fine.

Also, the negative voltages don't matter much. +5 and +12 are the main supply voltages on this computer (and +3.3 for the memory).

Power supplies are cheap, and they are easy to swap out, so swapping power supplies is a worthwhile experiment, too, if the voltages appear to be wrong or fluctuating. (Besides which, everyone should keep a spare power supply or two on hand, right?)

As for the caps, I would suspect the gold-on-green Tayeh 6.3v 1500uF caps rather than the white-on-black 10v 1000uF caps. My guess is that the white-on-black 10v 1000uF caps are probably okay.

I've never heard of anyone substituting tantalums for the electrolytics. I doubt that it would work. A tantalum in parallel combination with a standard ESR electrolytic would probably be an excellent substitute for a low ESR electrolytic, but there's barely room for the electrolytics alone on that crowded motherboard, so good luck cramming in all those tantalums, too.

One other simple thing to try with that machine is to remove the 2nd CPU (the one at the corner of the motherboard). I have a BP6 with bad caps that still runs almost reliably with only one CPU installed. (However, if this lets your machine run, that doesn't necessarily mean that the problem is bad caps -- I'm sure that there are other problems which might have this effect, too.)

-Dave

[Topcat]

The BP6 is a strange animal... Back prior to the VP6, I was an avid BP6 user/worshipper.... I was running coppermine CPU's in it before the VP6's release in November of 2001. I had great luck with it with a few mods, which included recapping.

Recapping the BP6 was a little tricky, but not any more difficult than any other board... There's one capacitor on the BP6 who's value needs to be changed as per a factory revision on the version 1.1 boards. The capacitor is EC10, close to the top edge of the board, between the CPU sockets. EC10 needs to be replaced with a 1500uF 10v cap. The 1500uF caps need to be changed to 2200uF caps also. Replace all the 1000uF caps with their original values. This made all my BP6's PERFECT! It fixed issues ranging from POST issues, stability, loss of CMOS settings, to helping overclocking successes....

I would not use anything but radial electrolytic capacitors. USE LOW ESR caps only, that is VERY important on the BP6!!

Good Luck!

[ncdave]

Thanks, Tomcat, for that excellent info!

BTW, will a VP6 run PPGA (pre-coppermine) Celeron CPUs? Or does it require Coppermines?

Also, do you know whether a hard drive with Win2K or Linux installed for a BP6 would run on a VP6 without a full reinstall?

-Dave

[Topcat]

A VP6 will run PPGA celerons with no problems, in single mode only. The VP6 will even run Tualatin CPU's, although, it is not 100% stable. It requires socket adapters, and SMP (dual processing) is very quirky. I've successfully ran 2x 1.13GHz 512k L2 Tualatins in one of my VP6's, however, it would NOT run my 1.4GHz 512k L2 Tualatins at full speed.

As for drive swapping, I know that Win2k wouldn't have a problem being swapped between the systems. Plug it into the standard IDE controller first, and install the RAID drivers (if used), and the VIA 4in1 Hyperion drivers, and it should take off like a rocket!!! I did exactly the same thing when I migrated from the BP6 to the VP6 in my servers. I have SCSI, so installing drivers was a minimal thing. The OS I swapped like that is still in operation now after 2 years, as a matter of fact, it's the system that hosts this site and forum.

As for linux, I don't really know. I am not very Linux literate yet. Perhaps Dustoff could help you there, as this boy lives and breaths Linux!

Dustoff?? A hand here??

[ncdave]

Again thanks for the wealth of info, Topcat!

Too bad it will only run one PPGA Celeron. Abit just reneged on their promised RMA of my BP6, so either I switch to newer boards or get out the soldering iron. I didn't really want to discard my old CPUs, though I realize they aren't worth much.

-dave

[ncdave]

Well, Abit just changed their mind again, and issued an RMA number for my BP6... and for free, too. I'm gonna send it in quickly, before they change their minds again. :-)

[Socket2Me]

Dave, thanks for suggesting that I check my BP6's Chipset Features' readings. Excepting the fractional periodic wavering of CPU Temps, VCORE, etc., all values are normal. As you said, this would seem to indicate that the caps aren't yet at a point of catastrophic failure, but could still be on enough of a decline to affect other functioning. (Since the previous/original owner said he didn't run it OCed, and that it had been out of use for about the past year, its caps probably haven't been fully cooked yet.)

Also, thank you, Topcat, for your specific instructions about repairing a board. I had previously read on one BP6 web page that the cap values you listed are the ones to use. Your statement that re-capping made your BP6s run perfectly, and solved the CMOS setting loss and booting problems is very encouraging.

This brings me to what I'd like to ask Dave - could you briefly describe how you got Abit to agree to RMA your board? Since I'd have to buy a low-voltage, ESD-resistant soldering station, along with caps, to attempt a successful, first-time mobo repair (even though I've been soldering since 1965), I'd rather get them to do the work, and, presumably, test the board before sending it back.

On that note, Dave Burton (<http://www.burtonsys.com>) stated that he'd read several usenet messages saying that Abit will: RMA and fix BP6 boards with bad caps free of charge; charge $25 to repair boards sent to them without original purchase receipts; replace only those caps that are visibly bad, i.e., bulging or leaking. What are your thoughts on those statements? Is the first one true, and the others not (I hope)? I'd say that's likely, since Abit's RMAed Dave's BP6.

Thanks again, T.C. & Dave, for your assistance - I'm thankful for this forum, and those who give of their time and information! - S2M

[ncdave]

Dave, thanks for suggesting that I check my BP6's Chipset Features' readings. Excepting the fractional periodic wavering of CPU Temps, VCORE, etc., all values are normal. As you said, this would seem to indicate that the caps aren't yet at a point of catastrophic failure

Well, it is a good sign if the readings are stable, but I would not break out the champagne just yet. My guess is that stable voltage and temperature readings simply means that certain capacitors are okay, not that all of them are okay.

I have another BP6 which is becoming flakey, but which has no visibly leaking or bulging caps. The voltage and temperature readings are quite stable, too. Yet I still strongly suspect that the problem is bad caps. On this machine, the 1000uF caps appear to be regular (white writing, so probably not low ESR), so they are probably not likely to fail, but the 1500uF 6.3v caps appear to be the same gold-on-green Tayeh brand caps that went bad in the other two BP6s.

This brings me to what I'd like to ask Dave - could you briefly describe how you got Abit to agree to RMA your board? Since I'd have to buy a low-voltage, ESD-resistant soldering station, along with caps, to attempt a successful, first-time mobo repair (even though I've been soldering since 1965), I'd rather get them to do the work, and, presumably, test the board before sending it back.

On that note, Dave Burton (<http://www.burtonsys.com>) stated that he'd read several usenet messages saying that Abit will: RMA and fix BP6 boards with bad caps free of charge; charge $25 to repair boards sent to them without original purchase receipts; replace only those caps that are visibly bad, i.e., bulging or leaking. What are your thoughts on those statements? Is the first one true, and the others not (I hope)? I'd say that's likely, since Abit's RMAed Dave's BP6.

To get Abit to RMA the BP6, I first emailed them and asked them for replacement caps (I said I would replace them myself). I got no reply in 24 hours, so I phoned them. After waiting through the hold queue, they told me that I needed to call their RMA department (phone number 1-510-492-0882). I told their RMA guy the same thing: that I'd replace the caps myself if they'd send me some caps. He said they couldn't do that, but that they'd RMA it for free, even though I bought it more than 3 years ago! He told me to use the \"eRMA\" feature on their web site to request an RMA.

Then they finally sent an email reply to my email, which also said they would not give me capacitors but would RMA the motherboard (but it didn't mention whether or not there would be a cost.)

I then went to their web site to do the eRMA process, and read the info there, which indicated that it would cost me $25. So, I think it was on 10/14/03, I filled out the eRMA form, and in my message which I entered on the web form I mentioned that they had told me on the phone that they would RMA it for free. I described the problem as \"bad electrolytic capacitors (bulging & leaking)\" or something like that. (I also included a link to some pictures of the bad caps on that motherboard, but that was probably overkill.)

The next day they sent me an email that said \"The requested item has long been phased out. We are sorry that no further support on this item will be available.\" (In other words, tough luck!)

I grumbed about that, but hadn't yet gotten around to calling them back when, the following day (10/16/2003), they sent another email, saying that they had issued the RMA, and here's the RMA number! I didn't do anything to change their minds, they did it on their own.

Yipee! So I sent in the motherboard. On 10/21/03, their eRMA system's \"status\" feature showed that they had received it, and that's where things stand right now. So far, it does not appear that they are going to charge me the $25, but who knows?

Since I don't have the repaired/replacement motherboard yet, I can't tell you anything about it.

-Dave

[Topcat]

The last time I RMA'd a VP6, it took about a month and a half to get the replacement. This board had problems beyond bad capacitors though. If it only had bad caps, I'd have done it myself.

Along with the $25.00 fee from Abit, it cost me about $15.00 to ship it, for a total of about $40.00 for the whole affair. I shipped it 2-day Priority, thinking it would expidite the RMA... HMMM, that was a bad idea, since the replacement came 1.5 months later...

To my disgust, the replacement they sent me was full of those damn Jackcon caps. I don't know if Jackcon had fixed their issues at that point. Since the board was for a server, I recapped it anyway so not to have to fool with it again.

If all you're facing is bad caps, and if you're able, I'd still recommend recapping it yourself. Your board will be back up and running in a day, rather a month or more...

I do have to give Abit credit for the fact that they will still RMA the BP6 and VP6. Although, the $25.00 fee is extremely absurd, correcting an issue that they in part created.

[ncdave]

Well, Abit has had the first of my bad BP6 motherboards for more than 2 weeks, and they still haven't shipped a repaired replacement, and they aren't answering my email. So, getting desperate, I decided to recap one of the motherboards myself, using the replacement capacitors which Jackcon sent me.

It went pretty smoothly except for three problems:

Problem #1 (solved): I had trouble getting the solder out of the 48 tiny holes on the motherboard, after removing the 24 bad capacitors. For your amusement, and with apologies to Tim \"The Toolman\" Taylor, here are a couple of photos of my solution:
http://www.burtonsys.com/BP6_desoldering1.jpg
http://www.burtonsys.com/BP6_desoldering2.jpg
Yes, they are hooked \"in series,\" with the bags removed for extra power. They made a most impressive roar, lemme tell you! Note that the power cords snake off in different directions, because I plugged them into different circuits, so that they wouldn't trip a breaker.

Problem #2 (not solved): After I'd already removed about half of the bad capacitors, I realized that the two small inductors, L17 and L18, are also toasted. I don't know whether or not they are still functional, but they sure look bad. Here are some pictures of them:
http://www.burtonsys.com/BP6_bad_coils1.jpg (before I removed the caps or noticed the inductor problem)
http://www.burtonsys.com/BP6_bad_coils2.jpg (after removing the bad capacitors)
http://www.burtonsys.com/BP6_bad_coils3.jpg (another view after removing the bad capacitors)
I don't have anything to replace them with, so I left them alone. If I'd noticed them before starting on this project, I'd have left the board alone and RMA'd it to Abit to worry about. I doubt that they'll take it back, now. :-(

What do ya'all think? Are L17 and L18 gonners? Any idea where I can get replacements?

Problem #3 (the big one): It doesn't work. After replacing the capacitors, I plugged in a couple of sticks of memory, a pair of CPUs, video card, monitor, power supply, and keyboard. Then I turned on the power supply, and momentarily jumpered the power switch connector. The CPU fans started up, but nothing else happened. No POST at all. So I hooked up a speaker to check for beep codes, and was rewarded with a European police siren: Beep, Boop, Beep, Boop, etc.. So I powered down, removed CPU #2, and powered up again: instant siren (no change). So I powered down, and removed the other CPU, and tried powering it up with no CPU: the siren again (no change).

Ideas, anyone?

-Dave

[Topcat]

Dave,

I don't be mad, but I was on the floor chuckling at the desoldering 'rig'.............. That's the best stroke of \"Macgyverism\" I've seen in a LONG time!!! I NEVER use a solder sucker on a board, but rather a needle pick to remove the solder, it's safer for the board. I'll take some pics and post them for ya.

As far as coils, I have a BP6 I hosed a long time ago I've been bilking parts off of, if u need the coils, let me know, I'll be happy to send them to you. It still has all the coils in it.

As far as it working, if it is beeping at you, thats a really good sign actually... That means it is trying, and things are for the most part getting power. find out what beep code it is giving, and that will tell you where the malfuntion it is. More than likely, the burt coils have a hand in the problems...

Let me know if you want the coils!!

[ncdave]

Oh, thank you, thank you, thank you! Yes, I do want the two small coils (L17 and L18).

As for the beeping... that's what I thought, too, at first: that it must be BIOS beep codes. But it isn't. It is some sort of hardware-generated alarm indicating that the CPUs aren't running.