Deunan wrote:

Anyway, either I'm going blind or the original IBM 5150 did not have a reset switch.

That came much later, after the AT era, when nobody cared anymore what IBM tried to push as "standards" (think PS half and EISA and all that deadborn stuff), and IBM finally succumbed.
Seems IBM was afraid of questions "why does it need a reset switch?", which could put its reliability into perspective 😀

retardware wrote:

Deunan wrote:

Oops. When I wrote "that" '138 I meant the one you've traced the connection to from the ROM socket. But it's actually helpful that you did all of them, since this mobo has 4 and the schematic has 5 and the U numbers are different. Helps me map which might be which.

It is possible that this mobo does not have a demultiplexer for the ROS chip selection, as the 4-input NAND goes low when the high address bits are all high. Could well be possible that this directly goes to !OE on all ROM sockets. This can be verified by checking if both (small and large) ROS sockets share the same !OE line.

retardware wrote:

@HanJammer: just one suggestion: look up "74LSxxx pinout" etc in the web. This will help you understanding what you are measuring (and let you learn a lot) and save us from translating pin numbers to signal names.

Yes. I did it every time (hence I'm also pretty much aware about which pin is GND on these TTLs - regarding the previous comment 😀 ).

Deunan wrote:

Pull-ups are quite common even in TTL systems which default to logic level high, due to the relatively low voltage on output which gets worse the more inputs of other gates are connected.
When you look for one, just check the resistance (on unpowered system, obviously) between the signal you are measuring and +5V. This system would have pull-ups in range of 1k to 10k. In general for a TTL a value less than 100 ohms or greater than 47k is definitelly not a pull-up.

Anyway, either I'm going blind or the original IBM 5150 did not have a reset switch. And it bothers me that you didn't see any reaction to it on the '138. I need to think a bit about your results so I'll make another post later but for now can you check what the reset switch connector pins are actually connected to. I would expect one pin to be ground, and the other to go to pin 11 of 8248 (on the left, beside ISA slots). Possibly it's tied along with the PWR GOOD signal.

Regarding the resistors - from what I understood - it will be sufficient to measure between VCC pin and the other pins on the TTL chips, right?

I will check the reset pins when I'm back at home. Also reset switch connector is in the most peculiar place (next to the keyboard connector).

retardware wrote:

Deunan wrote:

Anyway, either I'm going blind or the original IBM 5150 did not have a reset switch.

That came much later, after the AT era, when nobody cared anymore what IBM tried to push as "standards" (think PS half and EISA and all that deadborn stuff), and IBM finally succumbed.
Seems IBM was afraid of questions "why does it need a reset switch?", which could put its reliability into perspective 😀

EISA was introduced by the Gang of Nine not by the IBM - did you mean MCA?

Yes, it's enough to measure resistance between VCC and the pin. But don't do all of them - for now just check if there is a pull-up on the U50 pin 5.

Other than reset here's a next round of test:

Ok, so the first address latch we need is U24 '373. I was hoping the pins would go in ascending or descending order but apparently not, so you'll have to map them all for me.
For this test only check U24 and the CPU socket. You've already did pins 2,3 now also find out where on the '373 are the following CPU pins: 39, 3, 4, 5, 6, 7 and 8.
If you can't find a connection, then also check the other '373 nearby (U25). Perhaps it was easier to route the signals by swapping them between the chips.

After you're done with that, check CPU pins 35, 36, 37, and 38. Those should go to either the U25 or some other '373, '244 or '245, most likely near the CPU.

Last, measure the voltage on pins 1 and 11 of U24 and U25 with both CPU and 8288 chips removed from sockets. Let's see if we can test those chips in the board.

retardware wrote:

It is possible that this mobo does not have a demultiplexer for the ROS chip selection, as the 4-input NAND goes low when the high address bits are all high. Could well be possible that this directly goes to !OE on all ROM sockets. This can be verified by checking if both (small and large) ROS sockets share the same !OE line.

They don't share the same /OE line.

Deunan wrote:

Anyway, either I'm going blind or the original IBM 5150 did not have a reset switch. And it bothers me that you didn't see any reaction to it on the '138. I need to think a bit about your results so I'll make another post later but for now can you check what the reset switch connector pins are actually connected to. I would expect one pin to be ground, and the other to go to pin 11 of 8248 (on the left, beside ISA slots). Possibly it's tied along with the PWR GOOD signal.

One pin is the ground, the other pin is tied with pin 11 of the 8284 clock oscillator chip (-RES).

Deunan wrote:

Yes, it's enough to measure resistance between VCC and the pin. But don't do all of them - for now just check if there is a pull-up on the U50 pin 5.

U50 pin 5 vs VCC = solid 4,7kOhm, pin 5 vs GND = 5kOhm but it slowly goes up (stopped measuring at 5,5kOhm)...

Deunan wrote:

Ok, so the first address latch we need is U24 '373. I was hoping the pins would go in ascending or descending order but apparently not, so you'll have to map them all for me.
For this test only check U24 and the CPU socket. You've already did pins 2,3 now also find out where on the '373 are the following CPU pins: 39, 3, 4, 5, 6, 7 and 8.
If you can't find a connection, then also check the other '373 nearby (U25). Perhaps it was easier to route the signals by swapping them between the chips.

I thought I might as well check all address lines...

CPU pin 39 = U24 '373 pin 14
CPU pin 2 = U24 '373 pin 7
CPU pin 3 = U24 '373 pin 13
CPU pin 4 = U24 '373 pin 8
CPU pin 5 - no connection to U24 or U25 / but connection to U15 '244 pin 8
CPU pin 6 - no connection to U24 or U25 / but connection to U15 '244 pin 6
CPU pin 7 - no connection to U24 or U25 / but connection to U15 '244 pin 4
CPU pin 8 - no connection to U24 or U25 / but connection to U15 '244 pin 2
CPU pin 9 = U25 '373 pin 3
CPU pin 10 = U25 '373 pin 18
CPU pin 11 = U25 '373 pin 17
CPU pin 12 = U25 '373 pin 4
CPU pin 13 = U25 '373 pin 7
CPU pin 14 = U25 '373 pin 14
CPU pin 15 = U25 '373 pin 8
CPU pin 16 = U25 '373 pin 13

Deunan wrote:

After you're done with that, check CPU pins 35, 36, 37, and 38. Those should go to either the U25 or some other '373, '244 or '245, most likely near the CPU.

CPU pin 35 = U24 '373 pin 18
CPU pin 36 = U24 '373 pin 3
CPU pin 37 = U24 '373 pin 17
CPU pin 38 = U24 '373 pin 4

Deunan wrote:

Last, measure the voltage on pins 1 and 11 of U24 and U25 with both CPU and 8288 chips removed from sockets. Let's see if we can test those chips in the board.

U24 '373 pin 1 = 0,23V
U24 '373 pin 11 = around 2,5V hard to say because it slowly raises over time - around 0,003V each second.

U25 is exactly the same as above:
U25 '373 pin 1 = 0,23V
U25 '373 pin 11 = around 2,5V hard to say because it slowly raises over time - around 0,003V each second.

HanJammer wrote:

They don't share the same /OE line.

Good, then there is some more decoding logic.

Btw, this seems to be the manual, in case you are interested in DIP/jumper setting info:
http://www.retroprograms.com/IBMPC/PIM-TB10-Z … oard_Manual.pdf

retardware wrote:

Good, then there is some more decoding logic. […]
Show full quote

HanJammer wrote:

They don't share the same /OE line.

Good, then there is some more decoding logic.

Btw, this seems to be the manual, in case you are interested in DIP/jumper setting info:
http://www.retroprograms.com/IBMPC/PIM-TB10-Z … oard_Manual.pdf

Yup, I linked this one and another one in one of the posts above… Unfortunatelly these manuals don't contain electrical diagrams like some other manuals from this era I've seen 🙁

It's nice to know I had the same idea on how to connect a reset switch to this system as the engineers a few decades earlier. All the tests below are to be run with no CPU and no 8288.

First lets check that reset signal before we start tracing it any further. Use a meter rather than probe, it's slow changing and fully controlled.
1) Check that the voltage on pin 11 of the 8284 clock generator (U91?) is above 3.5V at all times - I mean, just watch it for some 30s or so to make sure the PSU is outputting a stable PWR_GOOD.
2) Then check that it drops to near 0V when you press the reset switch. To make sure the switch works.
3) Check pin 10 of the 8284, it should be the opposite of pin 11. In other words, close to 0V at all times except when the reset switch is pressed, then it should be at least 3.5V

Assuming that's all good, find out where the ISA slot pin B02 is connected. It should be a pin on one of the '04 chips, but that can be anywhere on the mobo. Pin 10 of the 8284 should also be connected to the same '04 chip. Once you find those connections, also test the voltage on those '04 pins with and without reset being applied.

Now, it should be possible to test the '373 chips. The rising voltage you've observed is most likely the chip heating up a bit, that's normal. The chip itself is a transparent latch and in this particular state it should just pass the logic level from inputs to outputs. There are 8 pairs, mapped like this (input->output):
3->2, 4->5, 7->6, 8->9, 13->12, 14->15, 17->16, 18->19
If in doubt, remember that inputs are tied to CPU socket and outputs are not. Also, preferably use a 100ohm (up to some 330 should work as well) resistor rather than a wire to short the inputs to ground. No need to connect them to 5V, TLL chip should default to high logic level by itself.

So, first check the voltage on all outputs. The actual value is not important as long as it represents a correct logic level. So lets add some margins and say below 0.6V is L, and above 3.5V is H. I'll use L/H instead of 0/1 as to not confuse logic level and voltage. Unless you see one output being considerably different voltage than others, then report that. Small differences (up to 0.2V or so) are OK.
In default state all output should be H. Then just use the resistor (or a wire) to bring each input to GND, and check if the output also went L. Test both '373 (U24, U25) that way.

Next, remember that '20 gate (U21)? It's pins 9,10,12 and 13 should be connected to U24 pins 19,16,5,2. Except I don't know the exact order but it's not critical. Verify that pin 8 on U21 is L in default powered state. Then bring each of the following U24 pins 18,17,4,3 to GND with resistor, as above, and make sure the pin 8 on U21 goes H each time you do that. It's enough to test one by one, no need to test pin combinations or anything like that.

Now move to U50 '138 chip. It should be connected like this: U50#1 -> U24#12, U50#2 -> U24#6, U50#3 -> U24#15. If that is not the case, then see if these 3 pins (1,2 and 3) are connected to any other pins on U24, U25 or U15. Just to make sure I didn't get the wrong '138 chip. Again if not, then I guess there is another buffer chip in between and we'll have to find it.

Deunan wrote:

First lets check that reset signal before we start tracing it any further. Use a meter rather than probe, it's slow changing and fully controlled.
1) Check that the voltage on pin 11 of the 8284 clock generator (U91?) is above 3.5V at all times - I mean, just watch it for some 30s or so to make sure the PSU is outputting a stable PWR_GOOD.

Stable 4,7V

Deunan wrote:

2) Then check that it drops to near 0V when you press the reset switch. To make sure the switch works.

Drops to 0,01V

Deunan wrote:

3) Check pin 10 of the 8284, it should be the opposite of pin 11. In other words, close to 0V at all times except when the reset switch is pressed, then it should be at least 3.5V

It's 0,17V and then 4,18V with Reset depressed.

Deunan wrote:

Assuming that's all good, find out where the ISA slot pin B02 is connected. It should be a pin on one of the '04 chips, but that can be anywhere on the mobo. Pin 10 of the 8284 should also be connected to the same '04 chip. Once you find those connections, also test the voltage on those '04 pins with and without reset being applied.

ISA B02 is connected to U20 '04 pin 12. It's 0,14V and then 3,75V with Reset depressed.

8284 pin 10 is not connected to U20 '04 pin 1. It's 0,17V and then 4,16V with Reset depressed.

Deunan wrote:

Now, it should be possible to test the '373 chips. The rising voltage you've observed is most likely the chip heating up a bit, […]
Show full quote

Now, it should be possible to test the '373 chips. The rising voltage you've observed is most likely the chip heating up a bit, that's normal. The chip itself is a transparent latch and in this particular state it should just pass the logic level from inputs to outputs. There are 8 pairs, mapped like this (input->output):
3->2, 4->5, 7->6, 8->9, 13->12, 14->15, 17->16, 18->19
If in doubt, remember that inputs are tied to CPU socket and outputs are not. Also, preferably use a 100ohm (up to some 330 should work as well) resistor rather than a wire to short the inputs to ground. No need to connect them to 5V, TLL chip should default to high logic level by itself.

So, first check the voltage on all outputs. The actual value is not important as long as it represents a correct logic level. So lets add some margins and say below 0.6V is L, and above 3.5V is H. I'll use L/H instead of 0/1 as to not confuse logic level and voltage. Unless you see one output being considerably different voltage than others, then report that. Small differences (up to 0.2V or so) are OK.
In default state all output should be H. Then just use the resistor (or a wire) to bring each input to GND, and check if the output also went L. Test both '373 (U24, U25) that way.

Well, this part was tricky as they are crumped between ISA slot and CPU socket...

[#pinnumber] [default output voltage] -> [output voltage with input grounded with 100Ohm resistor]

U24
#19 3,82V -> 0,14V
#16 3,80V -> 0,14V
#15 3,81V -> 0,14V
#12 3,82V -> 0,13V
#9 3,83V -> 0,13V
#6 3,83V -> 0,13V
#5 3,80V -> 0,14V
#2 3,80V -> 0,14V

U25
#19 3,80V -> 0,14V
#16 3,80V -> 0,14V
#15 3,80V -> 0,14V
#12 3,85V -> 0,13V
#9 3,85V -> 0,13V
#6 3,80V -> 0,14V
#5 3,80V -> 0,14V
#2 3,80V -> 0,14V

Deunan wrote:

Next, remember that '20 gate (U21)? It's pins 9,10,12 and 13 should be connected to U24 pins 19,16,5,2. Except I don't know the exact order but it's not critical. Verify that pin 8 on U21 is L in default powered state. Then bring each of the following U24 pins 18,17,4,3 to GND with resistor, as above, and make sure the pin 8 on U21 goes H each time you do that. It's enough to test one by one, no need to test pin combinations or anything like that.

There is no connection between U21 or U86 (which is '20 as well) and U24 or U25.
U21#8 is green (L) and turns red (H) when pins 18,17,4,3 on U24 are grounded through resistor (one by one).

Deunan wrote:

Now move to U50 '138 chip. It should be connected like this: U50#1 -> U24#12, U50#2 -> U24#6, U50#3 -> U24#15. If that is not the case, then see if these 3 pins (1,2 and 3) are connected to any other pins on U24, U25 or U15. Just to make sure I didn't get the wrong '138 chip. Again if not, then I guess there is another buffer chip in between and we'll have to find it.

No connections between U50#1/2/3 and U24 or U25 or U15 or U23 or U17 or U8(any pin) detected.

However there are connections like this to U48 '244 chip:
U50#1 -> U48#7
U50#2 -> U48#5
U50#3 -> U48#3

U24#6 -> U48#15
U24#9 -> U48#11
U24#10 -> U48#1 & #10
U24#12 -> U48#13
U24#15 -> U48#17
U24#20 -> U48#20

Also some pins on U25 '373 are connected to U43 '244

Also there is another '373 chip below U48 (marked U54) which is connected to U58 '245 below it

All these chips are in front of the ISA slots.

OK then, reset is properly applied by the 8284, and it seems to reach ISA slots as well. But I don't understand what you mean by "8284 pin 10 is not connected to U20 '04 pin 1". Did you mean pin 13 of U20 here? If you only tested 8284 #10 connection to U20 #1 then also test all the other pins on the '04. there should be a connection there somewhere. Possibly there is more than 1 NOT gate used here to split the reset signal cleanly beteen ISA and mobo chips. To find that out also check if the newly found connection to ISA B02 (U20 #12) is also connected to U50 #6.

'373 tested OK. We didn't test the latching but that's not possible in the mobo, but that would be a very uncommon failure. The output drivers work.

Unfortunately since there is no connection between U24/U25 and the '20 gate U21 there has to be another chip in between. So you'll need to hunt it down. Use the meter beep and sweeping over pins. The connection we want is to the U21 pins 9,10,12,13 - pick one and test. If you find a connection then confirm the rest is also connected to the same chip. In general it should be another 373, 244 or 245. However near the CPU is U26 chip, a '670 - start with that one. Maybe it's some sort of weird DMA bypass.

In general the test of this gate turned out good but I have to know what is driving it - I guess even though BIOS ROM is part of the mobo, they've connected it through ISA bus. Your U50 findings seem to confirm it.

I need to draw myself quick diagram of the address connections because I'm starting to get lost, for now though do this: U48 pins 1 and 19 - are those connected to GND? If not, what is their voltage?

Deunan wrote:

OK then, reset is properly applied by the 8284, and it seems to reach ISA slots as well. But I don't understand what you mean by "8284 pin 10 is not connected to U20 '04 pin 1". Did you mean pin 13 of U20 here?

Hm… to be honest I don't know either. Now When I tested it 8284 pin 10 seems to be definitely connected to U20 '04 pin 1 - those tests take time and I try to make notes every each one of them but maybe I've made some mistake in the notes.

Deunan wrote:

To find that out also check if the newly found connection to ISA B02 (U20 #12) is also connected to U50 #6.

Nope, there's no connection between B02 to U50#6 or between U20#12 to U50#6.

Deunan wrote:

Unfortunately since there is no connection between U24/U25 and the '20 gate U21 there has to be another chip in between. So you'll need to hunt it down. Use the meter beep and sweeping over pins. The connection we want is to the U21 pins 9,10,12,13 - pick one and test. If you find a connection then confirm the rest is also connected to the same chip. In general it should be another 373, 244 or 245. However near the CPU is U26 chip, a '670 - start with that one. Maybe it's some sort of weird DMA bypass.

It seems it's U17 '244 buffer.

Beside common GNDs and VCCs:

U21#9 -> U17#18
U21#10 -> U17#16
U21#12 -> U17#14
U21#13 -> U17#12

U24#1 -> U17#17
U24#2 -> U17#6
U24#5 -> U17#2
U24#10 -> U17#1 & #10 & #19 (GND but on both pins)
U24#16 -> U17#4
U24#19 -> U17#8

Deunan wrote:

I need to draw myself quick diagram of the address connections because I'm starting to get lost, for now though do this: U48 pins 1 and 19 - are those connected to GND? If not, what is their voltage?

That's right. Both connected to GND.

Edited: Corrected pins on U24 vs U17 and the picture.

So it is connected to the ISA bus. I mean, it's the same signals, but separated by these buffers. So we need to be sure the buffers are OK or not just the slots but also the BIOS ROM will be not be addressed properly.

Let's try and speed this up, I'm going to skip a step or two in hopes my assumptions are correct. With the CPU and 8288 out of sockets it should be possible to test all the address lines all the way to ISA slots.
1) First check that in the default state all address pins on ISA, A12-A31 contacts, are level H. As before anything above 3.5V is good.
2) Then short each of the CPU address pins: 35-39,2-16 one by one with 100 ohm resistor to GND. You can use CPU pin 1 or 20 as GND.
Shorting each pin should bring the ISA address line to level L. The order between the pins is easy here, on CPU it's 35-39,2-16 and on ISA it's A12-A31.

Regarding reset, there must be some connections between U20 '04 pins then. Just to confirm, the 8284 reset signal goes into pin 1, and pin 12 to ISA B02? Then there should be at least one more NOT gate used because two would not negate the signal. Also, we still need to find which pin of this '04 is actually connected to pin 6 of U50.

If you can't find any connection between U50 pin 6 and U20 then maybe pin 6 is actually connected to +5V permanently? Check with ohm meter.

Just saw the nice picture. Are you sure of this connection: U24#18 -> U17#20? It doesn't seem right. Maybe a pin miscount?

Deunan wrote:

Just saw the nice picture. Are you sure of this connection: U24#18 -> U17#20? It doesn't seem right. Maybe a pin miscount?

You are right. I must have touched the probe to the capacitor sitting next to U24#18 pin. I corrected the picture and the pin mappings.

I will perform rest of the tests later today...

Deunan wrote:

1) First check that in the default state all address pins on ISA, A12-A31 contacts, are level H. As before anything above 3.5V is good.

Yeah, they are as high as Jim Morisson at Mardi Gras.

All around 3.8-3.9V.

Deunan wrote:

2) Then short each of the CPU address pins: 35-39,2-16 one by one with 100 ohm resistor to GND. You can use CPU pin 1 or 20 as GND.
Shorting each pin should bring the ISA address line to level L. The order between the pins is easy here, on CPU it's 35-39,2-16 and on ISA it's A12-A31.

A19 OK
A18 OK
A17 OK
A16 OK
AD15 OK
AD14 OK
AD13 OK
AD12 OK
AD11 OK
AD10 OK
AD9 OK
AD8 OK
AD7 OK
AD6 OK
AD5 OK
AD4 OK
AD3 OK
AD2 OK
AD1 OK
AD0 OK

Deunan wrote:

Regarding reset, there must be some connections between U20 '04 pins then. Just to confirm, the 8284 reset signal goes into pin 1, and pin 12 to ISA B02? Then there should be at least one more NOT gate used because two would not negate the signal.

Reset switch grounds pin 11 of the 8284.
U20 '04 pin 1 is connected to 8284 pin 10 and CPU RESET (pin 21)
U20 '04 pin 2 is connected to U17 '244 pin 11
U20 '04 pin 12 is connected to ISA B02.
U20 '04 pin 13 is connected to U17 '244 pin 11
U17 '244 pin 9 is connected to U3 '175 pin 1
There is no direct connection between 8284 and ISA B02.

So, yeah it seems it's routed through two of the gates in U20.

Deunan wrote:

Also, we still need to find which pin of this '04 is actually connected to pin 6 of U50. If you can't find any connection between U50 pin 6 and U20 then maybe pin 6 is actually connected to +5V permanently? Check with ohm meter.

Yeah, U50 pin 6 is permanently connected to +5V (Vcc on U20, CPU and other chips).

Hm. That is good, but also disappointing. I was rather hoping we'd find a faulty 74LS transciver, replce it and fix the mobo. So far all the address lines seem to work and both the ROM selector '20 gate and the '138 decoder work as well.

So what's left. The data bus and the '245 bi-dir transciver - we'll check that next. Or it could be some annoying dynamic glitch but I hope not. These are very rare and also major PITA to find. Since we now know the ROM decoding is tied to ISA signals, so the address lines should be as well. Test the address pins on BIOS socket vs ISA slot pins - there should be a connection between each pair.

Next we need to know which '245 is the data bus buffer. Could be U23 or U8 next to CPU. In fact it could be both if the bus is split between mobo and ISA slots. See if you can find which of these chips is connected to CPU data lines AD0-AD7.
Lastly, check which one of the '245 is connected to data lines on the BIOS socket. Again it could be one more chip in between so if it's not U8/U23, test other '244 and '373 as well.

It might turn out you'll have to run some test programs from the BIOS socket after all so try to find cheap AT28C256-150 , or AT28C64 - or another electrically erasable part.

Deunan wrote:

Hm. That is good, but also disappointing. I was rather hoping we'd find a faulty 74LS transciver, replce it and fix the mobo. So […]
Show full quote

Hm. That is good, but also disappointing. I was rather hoping we'd find a faulty 74LS transciver, replce it and fix the mobo. So far all the address lines seem to work and both the ROM selector '20 gate and the '138 decoder work as well.

So what's left. The data bus and the '245 bi-dir transciver - we'll check that next. Or it could be some annoying dynamic glitch but I hope not. These are very rare and also major PITA to find. Since we now know the ROM decoding is tied to ISA signals, so the address lines should be as well. Test the address pins on BIOS socket vs ISA slot pins - there should be a connection between each pair.

Next we need to know which '245 is the data bus buffer. Could be U23 or U8 next to CPU. In fact it could be both if the bus is split between mobo and ISA slots. See if you can find which of these chips is connected to CPU data lines AD0-AD7.
Lastly, check which one of the '245 is connected to data lines on the BIOS socket. Again it could be one more chip in between so if it's not U8/U23, test other '244 and '373 as well.

It might turn out you'll have to run some test programs from the BIOS socket after all so try to find cheap AT28C256-150 , or AT28C64 - or another electrically erasable part.

I will do that, but first let me know if schematics like the one below are acceptable for you as I really need some tool to quickly sketch connections I find - writing it down while holding probes is not very convenient…

I can deal with text description, but if it helps you as well then these schematics are fine. Very "electric" / old American style but I can read them just fine. Modern style is to just cross the wires directly and without those hops. If you need to indicate a connection you make a dot. No dot = no connection. You can use any style that suits you - I will adapt.

Deunan wrote:

I can deal with text description, but if it helps you as well then these schematics are fine. Very "electric" / old American style but I can read them just fine. Modern style is to just cross the wires directly and without those hops. If you need to indicate a connection you make a dot. No dot = no connection. You can use any style that suits you - I will adapt.

AD0 -> U23#9
AD1 -> U23#8
AD2 -> U23#7
AD3 -> U23#6
AD4 -> U23#5
AD5 -> U23#4
AD6 -> U23#3
AD7 -> U23#2

U38#11 -> U54#3 -> U58#11
U38#12 -> U54#4 -> U58#12
U38#13 -> U54#7 -> U58#13
U38#15 -> U54#8 -> U58#14
U38#16 -> U54#13 -> U58#15
U38#17 -> U54#14 -> U58#16
U38#18 -> U54#17 -> U58#17
U38#19 -> U54#18 -> U58#18

U58#9 -> ISA A9 DATA 0
U58#8 -> ISA A8 DATA 1
U58#7 -> ISA A7 DATA 2
U58#6 -> ISA A6 DATA 3
U58#5 -> ISA A5 DATA 4
U58#4 -> ISA A4 DATA 5
U58#3 -> ISA A3 DATA 6
U58#2 -> ISA A2 DATA 7

U54#2 -> U15#18
U54#5 -> U15#16
U54#6 -> U15#14
U54#9 -> U15#12
U54#12 -> U24#9
U54#15 -> U24#12
U54#16 -> U24#6
U54#19 -> U24#15

U15#2 -> CPU#8
U15#4 -> CPU#7
U15#6 -> CPU#6
U15#8 -> CPU#5

U24#3 -> CPU#36
U24#4 -> CPU#38
U24#7 -> CPU#2
U24#8 -> CPU#4
U24#13 -> CPU#3
U24#14 -> CPU#39
U24#17 -> CPU#37
U24#18 -> CPU#35

U8 is connected to 8288... (not in picture)

U25 is connected to other CPU pins… (not in picture)

It seems the BIOS ROM data bus is connected to ISA as well. We can try to test it all the way in one go, except we need to have it set to different direction. As before, all test are to be done with CPU and 8288 removed.

First, check the voltage on pins 1 and 19 of U23. And confirm that pin 1 of U23 is connected to pin 4 of 8288.
Then check voltage on pins 1 and 19 of U58. And finally check if any of these 2 pins on both U23 and U58 are connected to U17.

Assuming you find that U23 and U58 have pin 11 driven L, and pin 19 in H state, try this: Put 100 ohm resistor in 8288 socket between GND (pin 1 or 10) and pin 4. This should cause pin 19 on both U28 and U58 go to L state. Hopefully it won't take more to reverse the bus direction.

Deunan wrote:

First, check the voltage on pins 1 and 19 of U23. And confirm that pin 1 of U23 is connected to pin 4 of 8288.

U23#1 - 1,69V
U23#19 - 0,14V

Deunan wrote:

Then check voltage on pins 1 and 19 of U58. And finally check if any of these 2 pins on both U23 and U58 are connected to U17.

U58#1 - 3,85V
U58#19 - 0,23V

Deunan wrote:

Assuming you find that U23 and U58 have pin 11 driven L, and pin 19 in H state, try this: Put 100 ohm resistor in 8288 socket between GND (pin 1 or 10) and pin 4. This should cause pin 19 on both U28 and U58 go to L state. Hopefully it won't take more to reverse the bus direction.

Pin 11?
Pin 11 is H on both of them.
Pin 1 is 'meh' on U23 and H on U58.
Pin 19 is L on both of them.

Resistor between 8288#4 and GND didn't changed anything. Both U23#19 and U58#19 are still in L state.