VOGONS


First post, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie

So In the "stuff I bought" thread I let you all know about this XT clone machine I found in neighbourgh's garage…

First thing I noticed is that this machine's PSU is shot.

So I dismantled it and started testing it out part by part. I'm most interested in the motherboard and ISA cards…

Unfortunatelly motherboard (designed KT 10 M/B, Made in Japan) won't start.

WP_20190406_21_39_04_Rich.jpg
Filename
WP_20190406_21_39_04_Rich.jpg
File size
819.74 KiB
Views
1760 views
File license
Fair use/fair dealing exception

---------------------------

Things we know:
- BIOS EPROM is fine - I dumped the bin file: https://mega.nz/#F!JZcy1A7T!SxCYxYLulEkQZM7dIohowQ

- There are two BIOS sockets - 27256 (empty) and 2764 (with EPROM).

- Dip switches on this board are set correctly (I think) - according to specs on TULARC (I found two very similar XT board there: mobo 1, mobo 2). There are no additional jumpers and LED/buton connectors are in places shown in these specs.

- CPU is gettin quite warm (almost hot) after longer period of activity.

- UMC and Siemens chips in the middle of the board are getting slightly warm.

- On the U53 resistor block one of the resistors had too long pin on the back side of the PCB and it was bent and touching the pin of the resistor next to it - I removed the pin, but it didn't changed anything.

- I visually inspected the board and I didn't see any physical dammage to the traces or missing components. Traces I suspected may be bad (ie. I've seen some black spots on them) tested good.

- I removed all of the chips (but DRAM) from their sockets and inspected them for corrosion and/or bent pins - none found.

- I observed that IRDY (I/O ready) LED on the diag card tried to come on up (at first it was very faint, then it was blinking with low frequency, then it was lit constantly and brightly) - it did it at first, but it doesn't do it now.

- PC Beeper tried to beep several times but it was like the voice from the grave with a lot of screeching/crackling noise and so on...

- Most of the time nothing shows up on the diagnostic board, but from time to time it will throw some POST codes. Also I noticed it goes through these codes at extreme rates unlike my 286 motherboards. Looks like this (codes shown around 2:00):
https://youtu.be/pJtouhw7gmI

- I've tested the orange tantalum caps with multimeter and some of them show around 4,7 uF (I guess it's close to their rating) and rest behaves erratic (my cat agrees) but I have no idea how believable these reading are when the cap is mounted in the circuit (for spare caps I have they seem to be pretty fine). Looks like this:
https://youtu.be/RXXFbYE1Pnw

- I've tested some of the ceramic and tantalum caps on the rest of the board and got mixed responses (ie the cap over the 8087 socket show the same erratic behaviour as some of the orange ones).

---------------------------

Do you guys have any suggestions what to do next (I would like to exhaust all possible diagnostic options before I start replacing parts). Also I don't have 5uF tantalum caps at the moment - can I safely replace them with 10uF caps?

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg

Reply 1 of 139, by retardware

User metadata
Rank Oldbie
Rank
Oldbie

I would advise to never test components in circuit, except you know for sure that the meter never introduces voltages above, say, 0.2V. Some multimeters have so high metering voltages that they can do actual damage.
For component testing *always* lift off one leg from the board/circuit.

Regarding upsizing tantalums from 5 to 10uF, there should be no problem if there are not too much of them (I did that a few times when that was necessary to stabilize supply voltages sufficiently for overclocking).

Do you have an oscilloscope or frequency counter to measure the 14.318 clock?
I ask because I find odd how bent the 14.318 crystal is. Does not match the otherwise impression of careful component placing. Maybe some damage that makes it resonate incorrectly?

Reply 2 of 139, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie
retardware wrote:

Do you have an oscilloscope or frequency counter to measure the 14.318 clock?
I ask because I find odd how bent the 14.318 crystal is. Does not match the otherwise impression of careful component placing. Maybe some damage that makes it resonate incorrectly?

I don't have oscilloscope or frequency counter. It is bent, but I don't see any dammage - I guess it was installed like this.

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg

Reply 3 of 139, by retardware

User metadata
Rank Oldbie
Rank
Oldbie

Crystals are extremely sensitive and can get damaged without any optical hint.
As these cost only a few cents, in your place I would swap in a new one, at least if you swap the tantalums on suspicion only as well.

Reply 5 of 139, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie
Predator99 wrote:

There are no diagnostic codes on most XTs - POST card nearly useless. Which lights are on?

Do you have a TL866? Make a Supersoft ROM 😉

Yeah, I have TL866. Where would I find this Supersoft ROM? Are you referring to this: http://www.minuszerodegrees.net/supersoft_lan … dmark%20ROM.htm
?
Can I burn it on 27256 (motherboard has socket for it - I guess it can be used instead of currently occupied 2764?). I'm asking because I don't have any 2764s at hand right now...

On the diag board all of the voltage LEDs are on (-12V, +12V, +5V, 3.3V), CLK LED and FRAME LED.

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg

Reply 6 of 139, by Predator99

User metadata
Rank l33t
Rank
l33t

Yes, put in on the 27256. I think the image will not fit 2764 anyway. As far as I remember its 32kb. If it smaller than your IC size copy it multiple times behind each other until filled.

Reply 7 of 139, by Predator99

User metadata
Rank l33t
Rank
l33t

Not 100% if this was the correct one, but you see its the same data after each 4000h bytes. Fill your Eprom with it and should be OK.
Of course, from minuszerodegrees 😉

Filename
Supersoft xt.rar
File size
5.37 KiB
Downloads
53 downloads
File license
Fair use/fair dealing exception

Reply 8 of 139, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie

I burned and used this one "U33 image for a 27256 EPROM (or 27C256) in a MK36000/27256 adapter".

I used CGA (from the same machine as the motherboard) multi i/o card - I tested it in another PC (I was surprised quite good quality for composite output acutally!):

56426998_2644757562206249_2953707506913771520_o.jpg
Filename
56426998_2644757562206249_2953707506913771520_o.jpg
File size
232.31 KiB
Views
1682 views
File license
Fair use/fair dealing exception
56182455_2644757608872911_1600797438648516608_o.jpg
Filename
56182455_2644757608872911_1600797438648516608_o.jpg
File size
166.31 KiB
Views
1682 views
File license
Fair use/fair dealing exception

I installed the EPROM and CGA and tried to turn it on... but this doesn't changed much - motherboard behaves the same way as before… Take a look - I think it has some serious issues with demons (and by demons I don't mean smtpd or bind 😉 ).
https://www.youtube.com/watch?v=sUf6swK328c

CAUTION: Use low volume or your brain may melt in the last few seconds.

I guess the next step is cap replacement...

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg

Reply 9 of 139, by Predator99

User metadata
Rank l33t
Rank
l33t

You are lucky and all ICs are on socket. Turn on without BIOS and CPU. What happens? Then remove all other ICs one after one and see if the Sound disappears. Did you check for shorts on the bottom side?

Reply 10 of 139, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie
Predator99 wrote:

You are lucky and all ICs are on socket. Turn on without BIOS and CPU. What happens? Then remove all other ICs one after one and see if the Sound disappears. Did you check for shorts on the bottom side?

Like I said before did visual check and found one bent pin (resistor's pin) which was touching the pin next to it.

The problem is the behaviour shown in the youtube clip is intermittent - like one time per 30-40 tries (or more)… most of the time it just sits there and nothing happens.

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg

Reply 11 of 139, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

I'd say the mobo is running but executing random data rather than BIOS image. This is mostly TTL logic there, it's stable without input signals (basically all inputs are pulled-up). So it should not "go crazy" - though it just might if the CPU is CMOS and clock line is cut.

I can't tell from the photo but I'm assuming the CPU here is 8088 (is it 8088 or 80C88?) below that empty 8087 socket? If so, it has 8-bit data bus and that makes interfacing with other 8-bit chips trivial. So there should be 1:1 connection between the CPU and the BIOS chip. If not, it must be through the '244/'245 chips nearby. No need for chipset, everything can be traced here and most stuff can be replaced as well.

I wonder why there are 2 ROM sockets here - I mean there's just 2 pins different between 2764 and 27256, maybe one if you consider the N/C pin 26. A jumper or two would do. So maybe there's something missing after all?

Reply 12 of 139, by Predator99

User metadata
Rank l33t
Rank
l33t

Regarding random data on the Post-card: It looks exactly like in my working XTs. Updating very fast. So I think its executing BIOS Code in the Video. But the Sound is weird and the fact its not reproducable makes it not a good point to start with.

Reply 13 of 139, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Just had a quick look at that BIOS, it does indeed seem OK. The checksum is zero, as it should be in this case.
But then I noticed the card used is a CGA? Well then, the dip switches are set to monochrome so that's wrong. Try SW1/6 to ON (or 1/5 to ON).

Reply 14 of 139, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie

OK, I indeed set the output on the card to colour composite and left the dip switches alone. I will try that. I didn't found any jumpers on the board or missing components but I will inspect again and also post better quality photos. CPU is Siemens 8088-2 (8MHz).

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg

Reply 16 of 139, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie

It seems that with Ruud's diag rom it does this crazy thing with POST codes everytime I start it. This behaviour is different than observed with the original AWARD BIOS or with Supersoft diag ROM (when it only did it once in a while).

It will do the same thing without BIOS ROM too.

It won't do this without CPU.

It still does this without UM82C55A (I guess it's keyboard controller)

It won't do this without SAB8237A5P (DMA controller)

It still does this without SAB8259A2P (Interrupt controller)

Last edited by HanJammer on 2019-04-08, 20:50. Edited 1 time in total.

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg

Reply 17 of 139, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

I still think it's not reading the BIOS ROM correctly. Since you own EPROM programmer, do you have any electrically erasable 5V parts (FLASH or EEPROM) in 64-512 size? I can prepare some test images for you but I expect at least a few attempts will be needed to figure things out. Having to erase a standard EPROM each time will be major PITA.

Reply 18 of 139, by HanJammer

User metadata
Rank Oldbie
Rank
Oldbie

At the moment I don't have any EEPROMs or FLASH. I only have OTP and UV EPROMs. Only 27256.

Also I noticed that now it does this crazy thing almost every time I start it (currently with the AWARD ROM).

Edit: nevermind, it won't do this again… I guess it's recap time. I will replace 4,7uF tantalum caps with 10uF one I have.
BTW - CPU voltage is ~4,7V all the time.

New items (October/November 2022) -> My Items for Sale
I8v8PGb.jpg