VOGONS


First post, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie

I just got myself a good deal on a Abit BX133-RAID, a 866MHz P3 (SL4CB) and a 256MB PC133 SDRAM stick - seller claims they were working fine when he tested them.
Along with an otherwise good config (Seasonic S12II 520W - good 3.3V and 5V rails on it, along with a Connect3D Radeon 9250), I get a working boot screen but can't get into the BIOS (XK revision currently). The only thing showing is a blue bar at the bottom of the screen when I try.
Tried resetting CMOS (seeing the battery was flat when I bought the package, I replaced the battery anyway) and running all DIP-switches set to 0 (for Soft Menu III setup). It seems to set it to a 433MHz by default when I boot, so I guess it'd be an easy fix in the BIOS settings... provided I can get into them.

The manual and archived product page state support for 133MHz bus CPUs, but seeing it's a 440BX board, maybe that's the issue? Got no 100MHz bus P3s to test with, sadly - only a few Tualatins. (1.2GHz and 1.4GHz)
Would I be better off just getting a 100MHz bus P3 first and try that?

I've seen one other thread here with the exact same issue, but it's 10 years old and I was wondering if there's any new input on this? (Didn't want to necro the old thread either.)
My board has the same crappy Jackcon caps all across the board (see below), but I see no visible bulging or leaking whatsoever.

11x 6.3V 1500µF 8x17mm* - Just above the CPU socket
4x 6.3V 1000µF 8x15mm* - Between the CPU and the AGP port
7x 10V 1000µF 8x15mm* - Surrounding the CPU socket and chipset, between it and AGP/DIMM/USB ports
7x 16V 100µF 6x10mm* - Spread out, between PCI slots, between ATX power and DIMM sockets, close to CMOS batt.
1x 25V 10µF miniscule :p - Next to Super I/O chip

Seeing how I haven't got the equipment or skill to recap a multi-layer PCB, it's gonna cost a bit to get it done, so I'd love to hear if anyone else has any suggestions as to what it could be?
Not keen on flashing the newest (modded or unmodded, picked up both) BIOS if there's a cap issue for risk of crash/misflash - but if that's a likely culprit, I could try.

Would love any input!

EDIT/ADD:
There are Panasonic FR replacement caps for pretty much all of them (bar the small ones, but I don't think I need to replace the 10µF at least, and I'm hesitating whether it'd be any idea replacing the 100µF ones too), and I might go for it anyway - but it'd really suck to get it back with new caps and the issue still there, so I'm open to other thoughts. 😀

Reply 1 of 17, by shamino

User metadata
Rank Oldbie
Rank
Oldbie

Tried resetting CMOS (seeing the battery was flat when I bought the package, I replaced the battery anyway) and running all DIP-switches set to 0 (for Soft Menu III setup). It seems to set it to a 433MHz by default when I boot, so I guess it'd be an easy fix in the BIOS settings... provided I can get into them.

Your 866/133 CPU has a 6.5X multiplier. When you reset the CMOS it sounds like it defaulted to 66MHz FSB, giving you 433MHz CPU speed. If it's not working properly at 433/66 then buying a different CPU isn't going to help.
As you've mentioned, this is a 440BX board so officially the highest bus speed it should be used at is 100MHz. 66MHz will minimize the stress and give it the best chance of booting.
This is a late 440BX board that was marketed to overclockers who wanted to run the 440BX at 133MHz, but it's still an overclock, so don't worry about 133FSB until you get it working at lower speeds first.

This is the 2nd motherboard I ever recapped, and the first was just a junk board I only swapped about 3 caps on.
With the BX133-RAID I did a terrible job, but somehow it works anyway. It's an awesome board.
On my board the symptoms were very simple - it powered on to a black screen, no POST at all. A few of the caps near the ATX connector (probably connected to the 5V input) were bulged. Once recapped it worked perfectly. Caps are the typical problem of ABit boards in this time period, and I've had good luck with them coming back to life with new caps.
I doubt there's a single example of these BX133-RAID boards that hasn't had the caps fail. I can't guarantee that recapping it will fix it, but I'd estimate the odds at 80-90%.

I made the mistake of not replacing all the 1000uF caps, so I had to go back and finish the job later when the rest of them went bad. Just replace them all. Even if they haven't bulged they're probably faulty - bad caps don't always bulge and when they do, it happens *after* failure has already occurred.

The small 10uF and 100uF caps are probably not under the same stresses so they might not require replacement. If you're doing the work yourself and feeling nervous about it, then you might skip those. For someone who feels comfortable doing the work though, or if you're paying for the job, then just to be thorough I'd prefer to see them all replaced. It probably doubles the time the job takes though.

If you have any junk boards around, you could practice removing and replacing caps on those, and put away the BX133-RAID until you feel ready to do it yourself.
The BX133-RAID seems to be a rare board so I regret that I butchered mine when I recapped it, but on the other hand it was a learning experience and I'm glad I did eventually get better at it.

Reply 2 of 17, by The Serpent Rider

User metadata
Rank l33t
Rank
l33t

I have this board and it does not behave correctly with some PSUs and can just lock up in BIOS.

Get up, come on get down with the sickness
Open up your hate, and let it flow into me

Reply 3 of 17, by shamino

User metadata
Rank Oldbie
Rank
Oldbie

True enough. I skipped a step. I'm a believer that the first step of troubleshooting (or simply validating) any system is to check voltages with a multimeter.
The Jackcon caps on these boards are garbage, but even so, I'd start by checking the supply rail voltages first with a multimeter just to be sure there's not an obvious problem there.

If you don't have a multimeter, well, I also believe that anyone who works with any kind of electronics should have one. 😀

If the DC voltages are correct, then you might still observe some PSU sensitivity as a result of bad caps on the motherboard. The board might be relying too much on the PSU to do all the filtering.

Reply 4 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie

I unfortunately lack a decent multimeter (I know, I really should get one ^^), my PSU tester shows good output on all rails though. 😀
Sure, that's without any load, but it's practically a brand new Seasonic unit (I'd ballpark it at around a hundred or so hours of usage in a P4 build, and it worked flawlessly in that) that I've come to like for its stability and good rated output even on the minor rails - running three of them in retro rigs without any issue now. 😀

And while I see no visible bulging or leaking of the caps on the board, I think I'll take the advice presented both here and in the other thread and have the board recapped - there's a local business doing it for ~50 bucks plus cost of caps, and I've got way too little soldering experience and equipment to do it myself. ^^ I think I'll skip the 10µF cap, seeing how low the capacitance is, but I'll probably replace the others.

Shamelessly copy-pasting my question from badcaps here too (no input there yet) - Panasonic FR's are readily available and I've used them in previous recaps, but I've found some alternatives.
Any series mentioned below that are way better than the FR's?

United Chemi-Con KZH
United Chemi-Con KZM
United Chemi-Con KZN
United Chemi-Con LXZ
Nichicon UHV (this is the one I'm most curious about, seeing that if these'd be ok for replacing the 6.3V 1500µF ones, I could order all caps from the same place)
Nichicon UHE
Nichicon HZ
Rubycon ZLH
Würth WCAP-ATLL
Würth WCAP-ATG8 (these seems like general purpose, when I looked a little closer)

Those are the options I've found for the 1000 and 1500µF caps - and I haven't dived into the specs and electrical properties enough to know what series are most suitable for recapping boards, so any input here would be greatly appreciated too! 😀
Thanks a lot for the feedback, electrical stuff isn't exactly my area of expertise, and I'm really grateful for the replies!

Reply 5 of 17, by jakethompson1

User metadata
Rank Member
Rank
Member

Looking at the other thread, it seemed everything worked perfectly fine on their system except getting into the BIOS, until they tried another cpu and it miraculously worked.
Is that true for yours too?
As others in that thread stated, wouldn't it be odd for a capacitor issue to block getting into BIOS setup but have no effect whatsoever on booting into Windows, which is much more intensive on the cpu?
I wonder if it's something in the bios's code to identify the cpu or whatever that goes into an infinite loop.

Reply 6 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie
jakethompson1 wrote on 2020-09-05, 17:13:
Looking at the other thread, it seemed everything worked perfectly fine on their system except getting into the BIOS, until they […]
Show full quote

Looking at the other thread, it seemed everything worked perfectly fine on their system except getting into the BIOS, until they tried another cpu and it miraculously worked.
Is that true for yours too?
As others in that thread stated, wouldn't it be odd for a capacitor issue to block getting into BIOS setup but have no effect whatsoever on booting into Windows, which is much more intensive on the cpu?
I wonder if it's something in the bios's code to identify the cpu or whatever that goes into an infinite loop.

Yeah, I've thought a bit (no pun intended) about that as well... unfortunately, I have no other S370 P3s (except the faster Tualatins, and I don't wanna risk them or the board) I can test with. 😒
I can try to find one used, I guess... hmm. Seeing it's an 866, I could try to get my hands on an 800MHz:er, seeing how that's the fastest 100MHz bus CPU I can configure the board to use, using the DIP switches on the board.
If that works, it's not much lost in terms of performance. Then again, the caps aren't exactly known to be good either. 😒

Reply 7 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie

Update:
Tried removing everything, checking for any visible damage whatsoever (there's none on the motherboard, the CPU, the PSU, the RAM or any of the two graphics cards I've been using), and I thought I'd try to set the settings manually via the board's DIP switches. (6.5x133, AGP at 2/3 of bus and Soft Menu III off.) Lo and behold - since that, the board wont POST at all. Even when turning all DIP switches off to enable software config via Soft Menu III again. Checked and double checked everything from cabling to connected devices (got no PCI or ISA boards installed, tried with a Connect3D Radeon 9250 and an ATi Rage XL), and it hard crashes after a few seconds after power-on. (Got Num Lock control on PS2 keyboard for a few seconds.)

Not looking too good right now, so it seems recapping is pretty much my only viable option at this point considering the crap quality of the caps and the tough to diagnose issue. (Got loads of SDRAM sticks lying around... all 512MB, so can't check with other RAM. And two Tualatins which are no good for testing the board either.)
If anyone got any other ideas, it'd be great - otherwise it seems like I'm shelling out for a recap. 🙁 Thanks for the input so far!

Reply 8 of 17, by PC Hoarder Patrol

User metadata
Rank Oldbie
Rank
Oldbie

What happens if you manually set it to 6.5 x 100MHz / Soft Menu III OFF using the DIPs?

Strangely, the Abit FAQ for the board says 133 FSB processors aren't supported, even though they clearly are!

BX133-RAID FAQ.jpg
Filename
BX133-RAID FAQ.jpg
File size
100.83 KiB
Views
357 views
File license
Fair use/fair dealing exception

Still think it might be worth a BIOS refresh from XK (to ZW or 72) if the system will run stable, except fot the BIOS issue (it's a boot floppy upgrade only though). Think 72 means PIII Coppermine-T support in the notes.

BX133-RAID BIOS.jpg
Filename
BX133-RAID BIOS.jpg
File size
58.07 KiB
Views
357 views
File license
Fair use/fair dealing exception

Reply 9 of 17, by swaaye

User metadata
Rank Moderator
Rank
Moderator

The AGP slot is going to be troublesome with 133 MHz FSB because of 440BX's 2/3 ratio. It's questionable product labeling.

I'm fairly sure I saw a BX133-RAID with bulging caps already in 2004 or so, and my BF6 was leaking at that point too. It is amazing to see some of these boards for sale with no bulging caps but I doubt they are still the correct capacitance.

Reply 10 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie

Tried a 100MHz bus speed (6.5x multiplier, and 2/3 AGP ofc), with no luck there either. 🙁 It's frustrating to not get any video signal at all anymore, considering I got one when booting it up the first time.
The AGP overclock is of some concern, I had thought that at least one of my cards would manage though. 😒 (Tried reading tests on what cards and/or GPUs can handle it, seems like it's hit or miss.)
And seeing it wont POST anymore with a 66MHz AGP either... I just don't know. If it'd POST, I might've given it a shot at flashing the 72 BIOS, but as it stands that's not happening.

If I inspect the caps as closely as possible, I can maaaaybe see the slightest hint of bulging (we're talking barely perceptible) on maybe two of the 30 caps on the board, which indeed seems too good to be true. (Seeing how pretty much every board from early 2000's are highly susceptible to crap caps.)
That said though, both capacitance and ESR values are probably not exactly gonna be great anymore. I wish I had some better measuring equipment, but I feel I can't throw an unlimited amount of time and money at the issue right now. 😒

Might turn in an MSI 694T Pro board that I've never been able to get to POST (old caps, but no visible damage) for recapping at the same time, if one of them works, I at least get something out of it.
Bit more resources than I would've liked to invest in getting one rig running though, but I was definitely aware of the risk with old caps. 😒

Again, thanks all for the input - been a frustrating couple of days of troubleshooting, and I really hope recapping will solve the issues.

EDIT: I got recommended United KZM and Rubycon ZLH for another motherboard recap, so I might grab those in lieu of Panasonic FRs. 😀 Fingers crossed that something working comes out of all this...

Reply 12 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie
shamino wrote on 2020-09-06, 23:51:

Panasonic FR, Nichicon HV, and Rubycon ZLH are all pretty much the same specs, at least in that 1500uF 6.3V 8x20mm size.

What's the difference between the Nichicon HV and UHV series? Tried looking through the spec sheets, maybe I'm missing something?

Reply 13 of 17, by shamino

User metadata
Rank Oldbie
Rank
Oldbie
PeTTs0n wrote on 2020-09-07, 10:06:
shamino wrote on 2020-09-06, 23:51:

Panasonic FR, Nichicon HV, and Rubycon ZLH are all pretty much the same specs, at least in that 1500uF 6.3V 8x20mm size.

What's the difference between the Nichicon HV and UHV series? Tried looking through the spec sheets, maybe I'm missing something?

The "U" is in front of the part number but it's not part of the series code. I think "U" represents "liquid electrolytic capacitors" or something like that. It's a more general category of the product.
So in other words I think "UHV" and "HV" should be the same thing.

Reply 14 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie
shamino wrote on 2020-09-07, 18:04:
PeTTs0n wrote on 2020-09-07, 10:06:
shamino wrote on 2020-09-06, 23:51:

Panasonic FR, Nichicon HV, and Rubycon ZLH are all pretty much the same specs, at least in that 1500uF 6.3V 8x20mm size.

What's the difference between the Nichicon HV and UHV series? Tried looking through the spec sheets, maybe I'm missing something?

The "U" is in front of the part number but it's not part of the series code. I think "U" represents "liquid electrolytic capacitors" or something like that. It's a more general category of the product.
So in other words "UHV" and "HV" should be the same thing.

Ahh, thanks. 😀 Makes navigating the jungle a bit more easier. ^^ (I see it now that I look at the breakdown of the product designation as well - my bad on missing it first time around!)
I ended up ordering United KZM's (if I would've been more certain when I placed the order, I would've gone for UHVs/HVs), I hope they're gonna do the job.
So FRs wherever I can, a bunch of KZMs and a few ZLHs. 😀 Thanks again for the info!

Reply 15 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie

Quick update: caps and recap job seem fine - been fighting with the other board (MSI 694T Pro) for more than the past day (far more hours than I'd like - I'll blame being an aspie) to try and get it working properly.
It posts now, which is more than it did when I last tried it, but when using a 1200 P3 Tualatin I get memory errors and halts in Memtest86+ (tried 4 different modules). With a 1100A Celeron (Tualatin), it works completely fine *if* the CPU is properly seated. Had to reseat it a couple of times, not sure if there's a contact issue with the socket, perhaps. Tried some different BIOS settings ("normal" and "turbo" modes, timings and RAM frequency, CPU Vcore, disabling pretty much every integrated device and removing all other stuff, USB compatibility on/off, etc.), but no difference at all. Of course tried reseating the 1200 as well.

Seeing the PSU is a good one (24A on both 3.3V and 5V, 130W combined - and I've got some faith in Seasonic), and the CPU power consumption should be about the same, I'm thinking it's a defective 1200. 😒

TL;DR: Recap seems ok, will probably post some update when I get around to testing the BX133 a bit more - need a couple of days of sleep first. 😀

Reply 16 of 17, by SpectriaForce

User metadata
Rank Oldbie
Rank
Oldbie

Besides making sure you choose the right capacity, voltage (can be higher than original) and dimensions, ideally you recap the board with capacitors that have the same or lower ESR as the original capacitors. You can look up the ESR of capacitors in the datasheet of each series. Electrolytic capacitors on motherboards are (ultra) low ESR. Replacing them with generic ones can cause stability issues.

The Seasonic S12 II is a load balanced power supply and is of a newer ATX12V revision. It can reset under circumstances, causing your classic computer to freeze or reboot. It's not ideal for this sort of old hardware. I recommend a power supply with DC-DC conversion instead or one with a really old ATX12V specification but which is new.

I sell 'ridiculously expensive' retro hardware to people like you. I feel sorry for that.

Reply 17 of 17, by PeTTs0n

User metadata
Rank Newbie
Rank
Newbie
SpectriaForce wrote on 2020-09-17, 09:32:

Besides making sure you choose the right capacity, voltage (can be higher than original) and dimensions, ideally you recap the board with capacitors that have the same or lower ESR as the original capacitors. You can look up the ESR of capacitors in the datasheet of each series. Electrolytic capacitors on motherboards are (ultra) low ESR. Replacing them with generic ones can cause stability issues.

The Seasonic S12 II is a load balanced power supply and is of a newer ATX12V revision. It can reset under circumstances, causing your classic computer to freeze or reboot. It's not ideal for this sort of old hardware. I recommend a power supply with DC-DC conversion instead or one with a really old ATX12V specification but which is new.

Yeah, I've been diving into it a bit since I had an MSI 848P board recapped a few years back - and capacitors are a damn jungle at best. Finding ESR values in spec sheets isn't exactly easy (and far from all manufacturers and series spec it), but at least found a few guides regarding suitable series here and on badcaps. 😀 Panasonic FRs do the trick (they're marketed as low ESR too), and the other series are known good ones for motherboard applications. 😀 I appreciate the input though - it's definitely something that can be said twice.

As for the PSU, I've checked with a multimeter (only got a basic one though, but still) and 3.3 and 5V rails seem stable enough under load - and I prefer running a modern PSU with modern protections and components. (Sure, it's possible to recap old PSUs as well, but I have neither unlimited energy, time or money to put into it.)

This can be a bit of a headache too though, especially if trying to find a new PSU that doesn't break the bank for an old rig. Tried a Corsair CV550, specced for 24 and 20A on the 3.3V and 5V rails respectively (120W combined) and that creates hiccups long before those numbers are reached. (P4 Prescott w/ ATX12V running a Matrox Parhelia 256MB AGP8x crashes at boot with artifacts, whereas an S12 runs that system completely fine.) The S12s I'm running have managed to run both Socket A, P4 and P3 machines fine so far though - that said, I'm not running the most power-hungry components in those systems either. If you've got any suggestions (preferably not Corsair 1000W+ units, I know they've got good rated output on minor rails though 😉 ) for suitable modern PSUs, I'd be really grateful! 😀

Managed to get a hold of a relatively cheap CX650M which is specced for 25A on both the minor rails, 130W combined, will see how that runs things. No DC-DC topology, but let's hope the VRMs in it can deliver. We'll see - they should at least be better than in the CV550. 😀

Again, thanks for the input!