Wash it, chuck it in the sun, and see what happens when it's dry

I have an i810 system that won't boot (no video) with PC133 ram sticks. It only works with PC100. It's a dell so I would've got different beeps.

So, while I'm waiting on the new RAM/GPU/VGA Header to come in to test, I'm trying to figure out a bit more about this board, specifically whether it's i810 or i810e, since that will determine whether I can use Coppermine CPUs.

The little bit I've been able to dig up leads me to believe that the base i810 can only support up to 100 MHz FSB and chips clocked up to 533MHz while the 810E can go up to 133MHz and faster chips. The manual for the most part just says i810, but in other sections it mentions the 133MHz FSB (with the crazy 'CPU booster') and that it supports '550, 600 MHz and higher speed Pentium II/III and Celeron processors'. So I'm really confused.

Maybe my eyes are going, but I can only find 2 Intel stamped chips on the board, neither of which show the Chipset number itself. One is stamped Intel FW82801AA / SL3Z2 (see image below). The other has Intel N82802A88, which I think is the firmware hub. I'm pretty sure the SL3Z2 chip is the actual chipset. Digging up that, I've found all sorts of references to the SL3Z2, even up to saying it's i815...

Any thoughts on what the actual chipset might be and what I'm looking at for max FSB / CPU compatibility?

MeatballB wrote on 2024-10-03, 12:00:

Thanks for all the good ideas, have tried a few other things:

Tried a new CPU, Celeron 366/Mendocino, and gotten the same result/beeping.

OK, then CPU compatibility isn't the issue.

Tried pressing all the keys to see if there was a stuck key and even pulled the keyboard completely to check if it was a stuck key, same result/beeping.

Possibly some traces leading from the ICH to the DIN connector are damaged or something in the connector itself is damaged/shorted.

Changed the toggle switch for the FSB from Auto and tried to force both 66MHz and 100MHz, same result/beeping.

This is interesting, as both the CPUs you have tested with are 66MHz parts and neither is likely to work with 100MHz overclock. I would expect different behaviour (no happy beep for starters) at 100MHz, particularly with the Celeron 700 (no way that it would run at 1050MHz; some C366 will run at 550MHz though).

MeatballB wrote on 2024-10-03, 20:58:

So, while I'm waiting on the new RAM/GPU/VGA Header to come in to test, I'm trying to figure out a bit more about this board, specifically whether it's i810 or i810e, since that will determine whether I can use Coppermine CPUs.

No it doesn't.

Coppermine CPU support is determined by voltage support (sub-1.8V) and in the case of So370 the correct FC-PGA pinout. Chipset is irrelevant. Difference between i810 and i810E is 133MHz FSB support. Note that RAM still runs at 100MHz regardless.

The little bit I've been able to dig up leads me to believe that the base i810 can only support up to 100 MHz FSB and chips clocked up to 533MHz while the 810E can go up to 133MHz and faster chips.

Wrong and right. The "533MHz" is a reference to PPGA socket supporting only Mendocino Celeron, which - as explained above - is a physical/electrical characteristic of the socket, not of the chipset. The only difference between the chipset versions is FSB support.

Also note that there are P3 Katmai CPUs with 133MHz FSB, just as there are 66MHz FSB Coppermine Celerons and 100MHz Coppermine P3 CPUs. CPU core and FSB are not connected.

The manual for the most part just says i810, but in other sections it mentions the 133MHz FSB (with the crazy 'CPU booster') and that it supports '550, 600 MHz and higher speed Pentium II/III and Celeron processors'. So I'm really confused.

I can't quite get my head around the "CPU booster" either, particularly as it claims to boost memory to 133MHz as well. That sounds like it's just overclocking an i810 set to 1:1 memory divider to 133MHz FSB...

Maybe my eyes are going, but I can only find 2 Intel stamped chips on the board, neither of which show the Chipset number itself. One is stamped Intel FW82801AA / SL3Z2 (see image below). The other has Intel N82802A88, which I think is the firmware hub. I'm pretty sure the SL3Z2 chip is the actual chipset. Digging up that, I've found all sorts of references to the SL3Z2, even up to saying it's i815...

The FW82801AA is the ICH - i.e. the southbridge I/O controller. It is shared between multiple Intel chipsets in the 800-series. The GMCH (Intel's new name for northbridge) is under the heatsink next to socket and slot.

Any thoughts on what the actual chipset might be and what I'm looking at for max FSB / CPU compatibility?

It's either i810 or i810E, and given the manual makes no reference to i810E or to having to run memory async at 4:3 divider with 133MHz FSB CPU installed, I'd say the 810DTC has an i810. What the DTA has is anyone's guess. Until such time as you get it to post, the only way to find out is to pop off that heatink - but it's glued on so you'd have to have more thermal glue to put it back if you do that.

Ah, got it. So I should probably be able to run Coppermine based on the information in the manual on the voltage capability and since I got the normal first 'beep' whether I'm using the 366 Celeron (Mendocino) or the 700 Celeron (Coppermine) I'll going to assume that I might be able to use either.

I'm definitely not going to pop off the heatsink to validate the chipset, but I see it now, the big copper/gold block. Guess I'll just have to wait and see if I get lucky with any of the new parts.

Also, I'm nearly positive it's not the keyboard. It is a PS2 keyboard with a PS2 to AT adapter, but I have before and after trying to build this, used that same keyboard and adapter setup on another machine and it worked fine.

Oh, and from an Addendum I found buried online, the 810DTA is basically the 810DTC with the only difference being a different onboard Audio codec. I can't vouch if the document is real, but from what it says the DTA uses an 'AD1881A'.

Does the beep change when the kb is unplugged?

Nope...same beeping sounds pretty much whatever I do. The only time I get something different is if I pull the RAM. Then I just get a long beep, long pause, repeating, or if pull both RAM/CPU, then I get no beeps at all. So, as long as the CPU/RAM I have are in, it's the same sound I posted the recording of, regardless of anything else on/not on the board.

I even just tried an ATX PSU just in case it was the AT PSU.

New RAM should be here Saturday, so if that's the problem, I should know this weekend. Otherwise, new GPU and the VGA header won't be here til Mon-Tues. Pretty much out of things to try until those come in.

MeatballB wrote on 2024-10-03, 20:58:

So, while I'm waiting on the new RAM/GPU/VGA Header to come in

Did you really find VGA header dedicated to this AZZA board? They are not universal, and manual doesnt mention the pinout 🙁 the only way to make it work is to either use oscilloscope or take off black plastic shroud and count on traces being obvious enough to guess correct pinout.

So the order of the error beeps slightly confuses me.
It could be keyboard controller/no keyboard detected error, but that would not prevent PCI devices init (second VGA would had its bios executed = at least blank screen).
It cant be corrupted bios - that check is very early and skips any ram detection.
I think it cant be ram, because that single beep is after everything was correctly detected.

rasz_pl wrote on 2024-10-04, 02:27:

Did you really find VGA header dedicated to this AZZA board?

No, I did not. It was a generic VGA plug that I had from another GPU that had a VGA port on a short cable to the PCB. Could very well be the wrong cable, but the seller did show me a picture of the one he used, and it looks like just a generic 16 Pin header to 15 pin plug/VGA outlet and I don't see any twists in the cable. Agreed that it could be in the shroud, and it definitely wasn't the one shown in the manual.

That being said, since I wiped the CMOS, the default setting in the BIOS should init the PCI display before the onboard AGP/header port. (See image below). I've tried the PCI GPU in every slot, in case it was a bad slot, but still the same.

After that first beep there's that first series of fast beeps that I think is the Mem Test, but the more I think about it, does that seem _really_ fast for it to be testing 128/256MB of RAM, or is that normal for something this era? I know my 4MB on my 486 counts up slowwwww. Granted, it could very well be anything else, including the board. Just hoping that for some reason this board is finicky and doesn't like PC133 like Spark mentioned the one system he had a similar issue, or it just doesn't like the TNT2 like renejr902 mentioned in his post.

I still say give it a wash.
I know everyone's scared of water and electronics, but they're washed in the factory.
I've washed several completely dead things and they've recovered. Zero have ever got worse.
One was a completely dead rx580. Dead to full function, furmark etc.

Full sun with light wind will dry it completely in about 5 hrs if you flick it dry with your arms at full length first. If there's no sun, an oven at around 90c will do it in about 15mins.
If you're ultra paranoid, rinse in isopropyl or metho to displace the water first.

Grab some distilled water from your local supermarket. You don't know what's in the water coming out of your tap. All my boards have had a bath in distilled water. Never had an issue.

If the new RAM and GPU don't fix things, I think a good wash might be in order. 😀

Why not do it now when you have days for it to dry?

+1 for giving it a good wash

However, some things to consider:

- If you haven't tested your TNT2 in conjunction witn ATX power supply, it might be the case that such card requires 3.3v present in pci bus (especially if it's an OEM card like Dell).

AT power supplies do not show 3.3v to the board, and even some AT boards do not connect 3.3v from ATX PSU to pci at all.

- Try an USB keyboard. i810 belongs to the era when USB human interface devices were starting to be popular so no bugs expected. I have an i810 from a system that came with USB keyboard by default. No problem on using that option.

- Try to get a proper bracket if possible for onboard video, or use an older card to test, such a S3 from Pentium era or similar.

With regards to cpu my plain i810 runs either a celeron 900 or a P3 900 with 100mhz fsb. Didn’t try anything faster as they start getting expensive.

Man...the plot thickens.

Got new RAM and a VGA adapter/header today. Still waiting on the other PCI GPU and it probably won't be here til Monday. I popped the new RAM in with the TNT2 and started up to the same thing, no video out and the same beep codes. So then, I pulled the TNT2 and put in the VGA adapter/header. It's just a generic VGA Port Adapter I got on eBay. Well, for the first time since I've had it, I got video out, but I was even more confused. I hear the first beep like normal, then I hear the second series of multiple beeps when the screen below pops up.

What in the world? Anyways, that sits there for about 5-10 seconds and then I hear the crazy machinegun beeps as the screen blanks out. It does seem like it's keeping the video signal alive though, I just see a blank screen with some artifacting up top, but the monitor doesn't go to sleep. Going to that website, it looks like some sort of background music/digital sign/image system for commercial places like Dentists, restaurants, etc. Is it possible this thing has some sort of custom firmware/bios that was part of a some system? I bet the crazy beep stream is it looking for whatever other components were part of the system. I don't see any way to get into any BIOS, and I'm definitely not one for flashing my own BIOS chips. I'm gonna email that company, but I'm wondering if I'm sorta stuck here...

Yeah looks like it likely has a custom BIOS chip. Your only resort is to find or make a BIOS chip with official firmware for this motherboard.
(edit)... and find someone to replace the chip on the motherboard, because it looks like it's soldered. 🙁

Or he could force a checksum error and flash the BIOS via floppy. IIRC he has to short data lines to force a BIOS checksum error, then have a floppy drive with an autoexec.bat file that flashes the BIOS automatically.

PcBytes wrote on 2024-10-05, 16:45:

Or he could force a checksum error and flash the BIOS via floppy. IIRC he has to short data lines to force a BIOS checksum error, then have a floppy drive with an autoexec.bat file that flashes the BIOS automatically.

Assuming that the custom code in the chip still retains that functionality.

The manual https://theretroweb.com/motherboard/manual/um … df119228817.pdf mentions there is a proprietary connector "AIR bus" into which some device they call "keybios" can be inserted and allows manipulation of the BIOS. I haven't found any information on the web about "keybios" except it being mentioned in a review of another Azza motherboard on Tomshardware https://www.tomshardware.com/reviews/heir-bx,241-9.html

If I were to take a guess, it's probably how that custom BIOS was made/installed. And it would probably need that "keybios" to have it replaced. Good luck finding one.

Otherwise, it's likely that the BIOS chip needs to be physically replaced.

(proprietary hardware sucks)