• Please review our updated Terms and Rules here

Altair 8800b CPU board issue

W

wstronks

Member
Joined
Jan 16, 2021
Messages
25
Location
Amsterdam, the Netherlands
A 8800b works fine with one CPU board, but swapping the CPU board with another CPU board introduces this strange behavior:
At a valid memory location, using the front panel switches:
- Storing values 00-FE , using deposit works fine
- WAIT, MEMR, MI are ON, all other LEDS are not lit
- Examine shows the right stored value
But
- Storing value FF
- WO and MEMR, WAIT are ON, all others are dimmed ON, A0-A9 and A15 are dimmed ON, A10- A14 are more or less not lit.
- No response anymore on switches DEPOSIT, DEP NEXT. EXAMINE, EX NEXT, etc
- Seems that the Write Operation failed
- Examine the memory location shows 7F was stored instead of FF, bit D7 is OFF instead of ON

Board voltages are OK,
All IC have the right Vcc
Already swapped the 8080 and 8212 without improvement
Checked the states of IC C, M,D,E, F according to troubleshooting section in the manual. They are as specified in the manual
I lack the experience to debug the DEPOSIT cycles to understand what is going wrong on this board (I assume its a CPU board problem)
Who can point me in the right direction to fix this issue?

Thanks,
Wilko
 
Hi Mike, Yes the second board is also 8800B CPU bd rev 0. This behavior also occurs in another 8800b machine/Front panel.
Another aspect that I found was that the good board draws about 630 mA and the board with the issue draw 900 mA.
Regards, Wilko
 
That current consumption variance is definitely worth a first look. Leave it on for a few minutes and look for an unusually warm IC. Most of that extra current is probably being drawn by one failing device.
 
One IC (E), a bi-directional bus driver was a bit hotter (125 F) than the others (90-100 F). I checked its function static and it seemed OK.
However I did replace it and with the new IC E, but the CPU board still has the same behavior.
I do have some more detail on the behavior:

Storing a value of 00-FE in f.i. address 0000h through the front-panel works OK. The WO LED and also the Bus signal SWO (inverted) on 97 stays (static) HIGH.
Quite often, but not always, when storing FF through the front panel the WO LED lights UP and stays lit, and bus signal SWO (inverted) turns low LOW and stays LOW until a next EXAMINE
And in the case when when storing FF is successful without the strange behavior and continue with DEP NEXT, the issue does occurs also after a few DEP NEXT of value FF.
The behavior has never occurred with storing al value other than FF ?

Wilko
 
Try replacing J L S and Y on the CPU board and see if behavior changes at all. If you don’t have enough replacement ICs, try L first, or J and L if you have two replacements.
 
OK, I will do that. But first I will get myself a desoldering station, because I am afraid I will damage the board with my current technique ...
I will get back on this issue after my holiday
Thanks, WIlko
 
wstronks said:
OK, I will do that. But first I will get myself a desoldering station, because I am afraid I will damage the board with my current technique ...
I will get back on this issue after my holiday
Thanks, WIlko
Click to expand...
If they’re not socketed we should probably take a different approach.
 
OK., they are indeed not socketed, except for the 8080 and 8212.
Just to make sure the problem is in this 8800b CPU board, I tested it again in another 8800b system.
This system had all different board than the initial system, also the Front-panel interface board was a different one.
The behavior is a bit different:

Storing a value of 00-FE in f.i. address 0000h through the front-panel also now works OK. The WO LED is not lit.
Similar as testing in the initial system; when storing FF through the front panel the WO LED lights UP and stays lit.
Although in this second system, three aspects are behaving differently:
- When examining the value stored on that address NOW it has stored FF correctly, (although the WO LED lighted UP and stayed lit after pressing DEPOSIT)
- Also, after pressing DEPOSIT of value FF and WO LED is lit you NOW can press DEP NEXT, WO goes out and FF is stored in the next address. In the initial system DEP NEXT had no response in that WO lit state.
- Strangely also the system gives no access (all data bits are lit) on selecting addresses in the address space of 8000h and higher. There is a memory boards and a prom board on the addresses from 8000-BFFF and F800-FFFF. Those addresses are selectable when swapping with an OK CPU board.

Thanks for helping, regards,
Wilko
 
There are like a million DIP switches that control the memory map, the I/O map, ... etc. perhaps one of them is not quite working, before you get to aggressive with chip swapping, check them out for continuity. What I'm thinking is that perhaps you end up with multiple things mapped over top each other. This might also increase power consumption as chip select is tied to that as well. It may be easier to work backwards from the chip selects.
 
Hi Bob, I see what you mean. But I know my way around the address and I/O selections and have a good idea, what is where. Also in a completely different machine with all different board, the CPU board also shows this same weird behavior, but when I swap the board for a 'good' working CPU board, all weird behavior disappears and all functions work well, fi, memory testing using EPROM code. Wilko
 
deramp5113 said:
If they’re not socketed we should probably take a different approach.
Click to expand...
Hi Mike,
I just removed IC J and installed it again using an IC socket. The strange behavior is still the same. I could go ahead with IC L, S and Y, but de-soldering the non socketed IC's is somewhat risky with me.
Would you have another approach to isolate this issue?
Regards, Wilko
 
Hi,

Get your hands on a cheap oscilloscope, my $250 HANTEK DSO2C10 has paid for itself to help me repair boards and figure other things out.

With the good working board, put the scope on each PIN of each chip, hit the <Auto Set> button, make a note of what you see on each PIN.

After you have scoped out the good working board, then install the faulting board and compare notes ... that will lead you to where the fault is, be it a chip or a capacitor near the chip etc.

In the long run, it saves time and money, and then you can share the notes online to help others that do not have a good working board.



.
 
You must log in or register to reply here.
Back
Top