Nice catch!
TheRetroWeb doesnt even have a "Ergo Electronics Co., Ltd." manufacturer entry 😮
Contemporary review https://www.classic-computers.org.nz/bits-and … -5-4.pdf#page=8 from January 1987
Another Ergo Electronics XT mobo for different chassis https://www.finn.no/recommerce/forsale/item/474647536

What a weird layout with ram above ISA slots.
CGA onboard 😮 cic8645 CGA controller. Here example ISA card using same chip https://www.vgamuseum.info/images/vlask/misc/cic8645bef.jpg
SCX6218 quite big for 1988 Gate Array by National Semiconductor. What does it do there?

>10 cpu/fpu pin21

reset pin. Matches classic 82C84A https://theretroweb.com/chip/documentation/vl … 94141474056.pdf https://www.renesas.com/ja/document/dst/82c84 … asheet?r=496476

>14 ISA pin30

OSC 14.something MHz, matches 82C84A

>16 crystal which then goes to cpu pin20

cpu pin20 is ground, but also matches 82C84A 😀

>18 cpu/fpu pin40

cpu pin40 is 5V

Imo its safe to assume U2 is 82C84A.

Now looking at that other Ergo Electronics XT mobo I linked it looks like second crystal is 22MHz and both chips surrounding SN74LS74AN are NEC μPB8284AD https://www.alldatasheet.com/datasheet-pdf/pd … /UPB8284AD.html

Oh, and welcome to the forum 😀

Wow... Thanx for this.
I forgot to mention that I already guessed that U2 and U3 are 82c84a compatible. Based on the image of the other Ergo board, there are 2 18pin chips above U1 and U3 both PB8284AD you mentioned. They should be compatible, but maybe older one has something specific. It could be worth checking if I have that in some other broken board or find it online. I did order new 82c84a chips and tried it all configurations between U2 and U3 but did not changed anything. Maybe new ones are crappy for this. 😀

Also the crystal is definitely probable 22MHz. On the board from your link there are 22 and 14 MHz crystals. On mine one I can read to be 14 and unlabeled might be 22MHz.

Not sure if that would change anything but it is super easy to test. Just need to order it.

Attached are some pictures of the front panel and case. There is a map of switches. Board came with a ram expansion (the small connectors on left/right of the RAM chips. But nothing regarding those switched did not changed anything.

I will do more testing.
Thank you so much for a warm welcome and this info. 😀

>I did order new 82c84a chips and tried it all configurations between U2 and U3 but did not changed anything.

What brand? maybe they are too slow for fitted crystal.
Do you have a scope? Is that an Aneng AN8009? afaik tops out at 6MHz = should be fine for checking 4.7MHz clock
I would populate just one 82c84a, set TURBO jumper so its the one being used, then verify CPU gets clock and reset.

rasz_pl wrote on 2026-08-28, 13:59:

>I did order new 82c84a chips and tried it all configurations between U2 and U3 but did not changed anything. […]
Show full quote

>I did order new 82c84a chips and tried it all configurations between U2 and U3 but did not changed anything.

What brand? maybe they are too slow for fitted crystal.
Do you have a scope? Is that an Aneng AN8009? afaik tops out at 6MHz = should be fine for checking 4.7MHz clock
I would populate just one 82c84a, set TURBO jumper so its the one being used, then verify CPU gets clock and reset.

I ordered ones I could find localy. It is labeled UM82c84ae which I think should be UMC. But no idea how legit that is. It wasnt very expensive 😀

I only have a cheap small digital osciloscope which is too crappy to give me anything. I ordered a UNI-T UPO1102, 2x100MHz today. It should arrive in few weeks and hopefully help me a bit more.
And yes it is a AN8009. Tbh did not occured to me to try it. Never used it for frequency before. Thanx for the tip.
I did try no chips, both and only one but did not tried the TURBO jumpers.
Will check it all out tomorrow.

Thanx 😀

I tried messing up with jumpers. Nothing changed. Also AAN8009 did not give me absolutely anything on CPU, 82c84a and SN74LS74. Also did not get anything on all the crystals on board.
Before I tried with that cheap digital osciloscope I have and it did not gave me nothing. I assumed its because its crappy one it could not registered it but on another board it could read 10MHz.

So I guess something else is the issue. Hope to get more info once the UNI-T scope comes.

CPU, like combustion engine, needs few things to run.
ICE needs Spark, Compression, Fuel.
CPU needs Reset, Clock, Data.

82c84a generates both clock and reset so is most likely your problem, thankfully its easy to diagnose

clydeRad wrote on 2026-08-30, 08:45:

I tried messing up with jumpers.

jumper, only one TURBO jumper does anything for this circuit. Im guessing it controls SN74LS74

clydeRad wrote on 2026-08-30, 08:45:

AAN8009 did not give me absolutely anything on CPU, 82c84a and SN74LS74.

neither one of 82c84s ?

clydeRad wrote on 2026-08-30, 08:45:

Also did not get anything on all the crystals on board.

you wouldnt, thats not how crystals work, even with a scope its hard to measure crystal directly because its a tiny signal and any load can disrupt oscillation

clydeRad wrote on 2026-08-30, 08:45:

Before I tried with that cheap digital osciloscope I have

model?

Here an example modern XT clone with turbo handled by 74LS74 http://www.baltissen.org/images/5160clone.png
here explanation of what the flip flops and those gates in front are doing https://www.eetimes.com/techniques-to-make-cl … ng-glitch-free/
I dont see any nand/nor gates near U2 U3 on your board 😐 that would make pressing turbo risk glitching the clock and crashing?
Looking at another contemporary turbo circuit http://bitsavers.informatik.uni-stuttgart.de/ … ug86.pdf#page=9 makes me think that maybe syncing clocks is not needed after all as long as switching signal itself is glitch/bounce free and 82c84 handles the proper switch internally.

On your board U3 is the Main low speed 14MHz OSC clock source. Its not used for anything else other than generating that 14MHz clock and feeding it to:
- isa pin30 (OSC)
- U2 pin 14. In your notes you have "U2 pin 14 ISA pin30" so that matches 😀

Now U2 is responsible for
- listening to pin 13 FREQUENCY/CRYSTAL SELECT (F/C) and switching clock sources between connected 22MHz crystal and external 14MHz
- F/C High generates 22/3 = 7 MHz turbo
- F/C Low generates 14/3 = 4.7 MHz clock

This means that you need both U2 and U2 plugged and working. U2 for 14MHz OSC and U3 for CPU clock.

EDIT: messed up which chip I was thinking of. Again:

Lets start with U3. The problem here is we need something to measure 14MHz. Frequency red probe pin 12, black probe pin 9.
Afaik Aneng cant display above 9.999MHz 🙁 but maybe it will display something that lets us know that it sees frequency but its too high? if not then try that digital scope. It might also be possible to measure red probe pin 8, black probe pin 9. that would be only 4.7MHz and meter should show that.

Next U2.
>U2 pins:
>16 crystal which then goes to cpu pin20

I disregarded that earlier, but on a second look this actually looks sus. Maybe you DO have short on one of the resistors feeding 22MHz Crystal.
Im guessing those resistors are R11 R12, they do look green-brown-brown in your photo.
Is one of them really shorted to ground? That would be great news.

clydeRad wrote on 2026-08-28, 21:51:

It is labeled UM82c84ae which I think should be UMC. But no idea how legit that is. It wasnt very expensive 😀

CMOS variant (82C84) may require different schematic near crystal:
8284 (your schematic):

82C84:

So 8284 preferable over 82C84

Good point
for 82C84 one would have to replace R11 R12 R15 R16 with 22pf capacitors

clydeRad wrote on 2026-08-30, 08:45:

AAN8009 did not give me absolutely anything on CPU, 82c84a and SN74LS74.

neither one of 82c84s ?

Neither,I tried with Aneng and the scope but nothing changed. In the past I used Aneng as a test on larger frequences, and it showed some low values (as I read because it was otside of its limit), but it showed some signs.

model?

FNIRSI DSO-510
https://www.fnirsi.com/products/dso-510

I disregarded that earlier, but on a second look this actually looks sus. Maybe you DO have short on one of the resistors feeding 22MHz Crystal.
Im guessing those resistors are R11 R12, they do look green-brown-brown in your photo.
Is one of them really shorted to ground? That would be great news.

Only pins closer to the edge of the board are in continuity with ground. But there is a trace from them to the ground so I guess it is fine. Other resistor pins are OK.

Anyway I will see to find 8284 ICs. Locally I was only able to find CMOS ones thinking it would be fine, at least for testing.. And then I was not 100% sure those should go in.

Thanx again to everyne for time and help.

clydeRad wrote on 2026-08-30, 18:18:

In the past I used Aneng as a test on larger frequences, and it showed some low values (as I read because it was otside of its limit), but it showed some signs.

so in theory it should show a false aliased reading - at least confirming there is some clock running

clydeRad wrote on 2026-08-30, 18:18:

FNIRSI DSO-510
https://www.fnirsi.com/products/dso-510

terrible, but 48Msps should still show some weal aliased clock signal

clydeRad wrote on 2026-08-30, 18:18:

Only pins closer to the edge of the board are in continuity with ground.

You wrote "16 crystal which then goes to cpu pin20" and CPU pin 20 is connected to ground, that would mean pin16 is also connected to ground, something not right there. pin 16 should read 510 ohm to ground like the other 8284

clydeRad wrote on 2026-08-30, 18:18:

Anyway I will see to find 8284 ICs. Locally I was only able to find CMOS ones thinking it would be fine, at least for testing.. And then I was not 100% sure those should go in.

replace R11 R12 R15 R16 with 22pf capacitors and those 82C84 should start generating clock

Ok. Regarding crystal going to pin20... That is a stupid mistake on my part. I have probably touched a crystals body when I wrote that. Second crystal pin have a small trace on the other side of the board that goes to the side of the crystal but did not found later where it is going. But it is not shorted to ground.

I will replace the resistors tomorrow after work and hopefully get something different.
I happened to have 4 22pF ceramics. Maybe it is a sign 😀

I made a progress. Board still not working but finaly showing some life.
I replaced the 4 resistors with 22pF capacitors. Now reset LED goes off shortly after powering on (also checked the voltage on CPU pin 21 and it is 0V).
I also get the clock on CPU (4.7MHz and 8MHz with changing the turbo jumper).
But that is all I get.
When powered on I get constant beep, no image on monitor.
I tried the integrated video card and a Hercules ISA card I have and nothing (I adjusted the switches for that).

Tried with and without memory expansion module and then also tried all other RAM switch variations that should not be the case but nothing changed.
So my guess is that either some other components on board are failed. Maybe one of the ram chips on oboard or even BIOS.
I do have the spare new 41256 chips, and also made a simple 4164/41256 dram tester with arduino a while back. But would need to remove all of them from board.
I ordered a t48 eprom flasher, so might try with some opensource BIOS for 8088 board I found on github. If that does not work will have to desolder all ram chips and test them.

👍 Mov AX, 0xDEAD 👍 great spotting that TTL/CMOS difference!

>Hercules ISA

try flashing this one https://github.com/ki3v/xtramtest it does minimal hardware init and runs from ram on CGA/MDA (Hercules is 100% compatible and shoudl work) cards

> I get constant beep

good, means CPU is running now. For the beeping to occur CPU must program few registers - PIT 0x41 0x42, PPI 0x61.
Please upload dumped BIOS when you get the programmer. There might be clues inside as to why its beeping plus Im curious what "Ergo Electronics Co., Ltd." used 😀

T48 arrived 😀
I tried ram test bios and nothing changes.
But beeping part was my another stupid mistake. I used a more modern piezzo buzzer, probably it was the reason it was beeping since I realized that the beeping starts as soon as I turn on the PSU.
With original speaker, which I was too lazy to take out of the case does not make a sound (checked the speaker aside and it works).
Will do more checking. But clock and reset line are still good.

Anyway attached is the original BIOS.

clydeRad wrote on 2026-09-03, 14:11:

Anyway attached is the original BIOS.

Dump is OK, 86Box booting to ms-dos

clydeRad wrote on 2026-09-03, 14:11:

I tried ram test bios and nothing changes.

Custom ROM may works only with true CGA Video card, not EGA/VGA. Usually i use 86Box to check combination of ROM and Video. Perhaps there are ROMs for the XT that output debugging information to a POST card like 286+ BIOS, but I personally don’t know of any.

Mov AX, 0xDEAD wrote on 2026-09-03, 17:23:

Custom ROM may works only with true CGA Video card, not EGA/VGA. Usually i use 86Box to check combination of ROM and Video. Perhaps there are ROMs for the XT that output debugging information to a POST card like 286+ BIOS, but I personally don’t know of any.

xtramtest should work with Hercules

clydeRad wrote on 2026-09-03, 14:11:

But beeping part was my another stupid mistake. I used a more modern piezzo buzzer, probably it was the reason it was beeping since I realized that the beeping starts as soon as I turn on the PSU.

no wonder it was beeping 😀

clydeRad wrote on 2026-09-03, 14:11:

Will do more checking. But clock and reset line are still good.

do you have a logic analyzer? rpi Pico 1/2 board?

rasz wrote:

do you have a logic analyzer? rpi Pico 1/2 board?

No. But I do have an old rpi 2 from some project I never finished 😀 Not sure if that could work since it is pretty old. I see online that sampling on it would be limited to 5-10MHz.

Im surprised xtramtest didnt produce any video 🙁 Its meant for MDA, maybe its not fully Hercules compatible.

full rpi wont do. $5 pico and https://github.com/gusmanb/logicanalyzer can give you
"Make a 100MHz 24 Channel Logic Analyser from your favorite Raspberry Pi Pico" https://www.youtube.com/watch?v=Obd1PiW7RO8
"This 400MHz Logic Analyzer Lets You Spy on Any Circuit!" https://www.youtube.com/watch?v=NaSM0-yAvQs

Imo "Analyzer/Shifter boards" are optional since pico 2 is 5V compatible and XT is 5V so in theory it should work fine bare IO, worth a try for $5.
With LA you will be able to solder probe wires to bios chip or CPU and trace if CPU is even trying to execute something. 24 channels is almost enough to capture whole 8088 CPU bus. Of course debugging this way is more involved, have to fiddle with captures, read what results mean or molest LLM etc.