MITS Altair 8800 strange front panel / address bus issue, Need HELP!

[wellswang]

I’m repairing an original MITS Altair 8800 and I’ve encountered a strange front panel/address bus issue. I’d appreciate any troubleshooting suggestions.


Current hardware configuration:

• Right now, only the CPU board is installed in the system.
• I previously tested with a memory board installed as well, but the behavior was exactly the same.
• So the issue does not appear to depend on the RAM board.

Symptoms:

• After power-on and RESET, the address LEDs correctly show all zeros.
• All address/data toggle switches are set to 0.
• Pressing EXAM causes address LEDs A3 and A11 to turn on.
• Pressing DEPOSIT or NEXT does NOT cause A3/A11 to turn on.
• Other address bits behave normally.

Additional testing:

• If I set only A3 switch to 1 (all others 0) and press EXAM, the result is still only A3 and A11 lit.
• If I set only A11 switch to 1 (all others 0) and press EXAM, the result is still only A3 and A11 lit.
• For all other address bits, the LEDs correctly follow the switch settings during EXAM.
• Only A3 and A11 appear stuck high during EXAM operations.
• I tested related 7405 of A3, it works normal.

So currently:

• RESET appears normal
• NEXT and DEPOSIT NEXT appear normal
• The problem seems isolated to EXAM loading addresses from the front panel switches.

Has anyone seen a similar failure mode on an Altair 8800 before?
Any suggestions on which ICs or signal paths I should probe first would be greatly appreciated.

 

[daver2]

There is a complete thread on VCFED (from a while ago) fully testing the front panel out from the ground up.

I was a major poster on the thread, so look for me.

If you can't find it, let me know and I will have a look.

Dave

 

[daver2]

The switch outputs for A3 and A11 share the same signal line on the front panel for D3, along with two (2) further 7405 drivers that do NOT go via switches.

Just checking the 7405 associated with switch A3 may not help. You need to perform a full, and methodical, test of everything...

Dave

 

[wellswang]

Thank you, Daver!


Sorry, I searched the previous threads using keywords like “Altair 8800” and related terms, but I couldn’t find the front panel testing thread you mentioned. Could you please point me to it? Thanks!


I tested the 7405 associated with SA3, SA11, and D3 (the “D” chip in the green box on the schematic), as well as the chips in the blue box directly related to those switches. All of them appear to be working normally.


What’s strange is that after I desoldered the chip D (7405 in the green box) for testing and then soldered it back in, the EXAM behavior remained the same (A3 and A11 still stuck on), but now DEPOSIT NEXT no longer responds — the address no longer increments each time the switch is pressed like it did before.


I carefully checked all the connections and traces, and I don’t think I damaged anything during the work.


At this point I’m still stuck and have no clear direction on what to investigate next.

 

[daver2]

I am away on a business trip this week, so I only have limited access to documents.

What test equipment do you have?

You mentioned about desoldering ICs. Are the ICs in IC sockets or not?

The way to test things is to break the circuitry down into sections and test each section on its own.

If you look at the schematics for the front panel switches: http://dunfield.classiccmp.org/altair/d/88schema.pdf you should note that the 7405 ICs are connected in four (4) groups. Only one group should be active at any one time.

If we can isolate the groups, we can test each group (including the address switches) on their own.

Dave

 

[wellswang]

Thank you, Daver! Sorry for taking up your time. I can do some more testing first, and whenever you have time, I’d really appreciate any suggestions.


At the moment, some of the ICs on the board are socketed, possibly from repairs done by a previous owner. The blue sockets in the photo were already there originally, while the black socket is one I just installed after removing the 7405 for testing.


I have a logic IC tester that can check whether 74-series logic chips are functioning correctly, and I also have a multimeter and an oscilloscope for probing signals in the circuit.


Based on the schematic, I’m wondering if I should trace further upstream from the 7405 and check the 74123 next? Or is it also possible that the Intel 8080 itself could be causing this issue?

 

[daver2]

Until we get some data points, the problem could be anywhere (unfortunately).

If you have an oscilloscope, that is good.

I would check the signals at the 'common' inputs of the 7405's with your oscilloscope and observe what happens when you initiate an EXAMINE, DEPOSIT, etc.

You could also check the D0 to D7 outputs (especially D3) - with all of the address switches set to '0' - under the same conditions and also observe what happens.

You need to divide and conquer. Is it the front panel or the CPU?

Dave

 

[wellswang]

Hi Daver, thank you for your suggestions.


Today I used an oscilloscope to observe the common input line of the 7405, and I found that there was no change (always stay around 0.1V) in the waveform at all when I pressed EXAM (my probe point is marked with a red circle on the schematic).


Also, I’ve ordered two D8080 CPUs, and once they arrive I’d like to try replacing the CPU for testing.

 

[daver2]

Excellent work.

So, if all of those points stay at logic '0', the outputs from the 7405 inverters will all be open circuit and the address switches will not be read.

That indicates we have to look earlier in the circuitry.

However, before we get ahead of ourselves, how did you use the oscilloscope to monitor these points (bearing in mind we will be looking for a pulse)?

Dave

 

[daver2]

I have some ideas on where to go, but I am just going out for an evening meal - so I will post more detailed checks when I get back.

The process is to check from the switches forwards via the logic.

There is something that doesn't look right though feeding the 'supply' rail to the common poles of the control buttons/switches. This goes via what appears to be a 270k resistor (pull down from the 'run' signal) with 1k (pull up) resistors. This doesn't make sense on first looking...

Dave

 

[daver2]

The first thing is to check the RUN/STOP switch and flip-flop equivalent IC 'H'.

With S7 in the 'RUN' position, the following signals should be observed:

IC 'H' pin 9 (H/9) should be oscillating. What is the frequency or period?

H/13 should be '1'.
H/10 should be '0'.
H/11 should be '0'.
H/8 should be '1'.
H/12 should be '1'.

The 'common' point of switches S2, S4, S5 and S6 should be '1'. What is the voltage measurement of this point?

With S7 in the 'STOP' position, the following signals should be observed:

H/9 should be '1'.

H/13 is what? Is it oscillating or a fixed logic level? In this case, is it high or low?

H/10 should be '1'.
H/11 should be '1'.
H/8 should be '0'.
H/12 should be '0'.

The 'common' point of switches S2, S4, S5 and S6 should be '0'. What is the voltage measurement of this point?

If the voltage is not as expected on the 'common' part of the switches, we stand no chance of this ever working!

If the voltage readings are OK, we can move on to the control switches and pulse generators next.

Dave

 

[wellswang]

Thank you, Daver.
During testing, I tried to trigger the oscilloscope on the rising edge to capture the waveform, but in fact it never triggered at all. The oscilloscope consistently showed the voltage staying around 0.1V.

Thank you very much for your testing suggestions. However, I’m currently traveling for work these days, so I won’t be able to continue testing until I return home this weekend. Once I’m back, I’ll follow your suggestions and report the results. Thanks again!

 

[daver2]

No problem. So am I (travelling on business)...

Dave

 

[wellswang]

Hi Dave,


I performed the tests according to your suggestions, and the results were generally consistent with your expectations. The results are as follows:

PIN​	S7-RUN​	S7-STOP​
H/9​	Oscillating / 240.3Hz​	H​
H/13​	H​	H​
H/10​	L​	H​
H/11​	L​	H​
H/8​	H​	L​
H/12​	H (3.76V)​	L (0.08V)​
CommonPoint of S4/S5/S6​	H (2.96V)​	L (0.08V)​
CommonPoint of S2​	H (3.76V)​	L (0.08V)​

It seems that this part of the circuit is probably working normally.

Also, the D8080A CPUs I purchased have arrived. I replaced the original 8080 with one of them, but the symptoms remained exactly the same, so I think we can rule out the CPU as the cause of the problem.


Thank you!


BR,

 

[daver2]

I shall have a more detailed look later on but, as you say, the results initially look good.

I will then post some tests for checking out the 74LS123 devices...

Dave

 

[daver2]

Yep, those readings are fine.

My next set of measurements would be to set S7 to STOP (common points of switches S2, S4, S5, and S6 to 0V) and then measure the logic levels at the inputs to the device pins to which the poles of the switches connect.

The default voltage should be a logic HIGH at the following pins:

M/1 (driven from switch S6/SS).
L/1 (driven from switch S5/EXM).
S/13 (driven from switch S5/EXM NXT).
N/10 (driven from switch S4/DEP).
F/1 (driven from switch S4/DEP NXT).
W/3 (driven from switch S2/PROT).
W/1 (driven from switch S2/UNPROT).

As you operate the associated switch (in the direction of the function indicated) the pin should transition to LOW and then HIGH again when the switch is released.

How is your knowledge of the operation of the 74123 device for the next set of tests after the above?

Dave

 

[wellswang]

Hi Dave,


I tested these pins according to your suggestions, and I found that regardless of whether S7 is switched to “STOP” or released, the logic levels on these pins always remain HIGH.

As for the 74123, my understanding is that an input signal triggers a pulse waveform generated by the RC timing circuit.

I studied the schematic a bit, and it seems to me that when S4/S5/S6 are not being switched, the M/1, L/1, S/13, N/10, and F/1 pins should all be pulled up to +5V HIGH. Simply switching S7 to “STOP” should not affect them. Am I misunderstanding something?

Thank you for your guidance.

 

[daver2]

We may have a bit of confusion between us regarding the testing...

Setting S7 to STOP causes the 'common' signal on the switches S2, S4, S5 and S6 to be set LOW.

The input on (say) M/1 should still be HIGH.

When you now operate S6 to the 'SS' position, M/1 should then go LOW - because switch S6 now pulls the signal level on M/1 down to the 'common' voltage level - which is LOW.

Likewise for all of the input pins verses associated switch positions.

Is this a bit clearer now?

Dave

 

[wellswang]

Hi Dave,


Now I understand exactly what you meant. After setting the mode to STOP using S7, I should operate S2/S4/S5/S6 and observe the logic level changes on the corresponding pins.

I tested the relevant pins accordingly and found that when S4/S5/S6 are activated, the corresponding pins momentarily go LOW and then quickly return HIGH (even if I continue holding the switch in that position). I assume this is the effect of C11, correct?

For S2, however, the signal stays LOW while the switch is held in position, and returns HIGH when the switch is released.

Please see the following table:

PIN	FUNC	SWITCHED	RELEASED
M/1	SS	H->L->H (attachment 1)	H
L/1	EXAM	H->L->H (attachment 1)	H
S/13	EXAM NEXT	H->L->H (attachment 1)	H
N/10	DEP	H->L->H (attachment 1)	H
F/1	DEP NEXT	H->L->H (attachment 1)	H
W/3	PROT	L (attachment 2)	H (attachment 3)
W/1	UNPROT	L (attachment 2)	H (attachment 3)

Thank you.

BR,
Wells.

 

[daver2]

Yep, they are all OK.

The S2 common connection is directly off the active low RUN signal whereas S6, S5 and S4 common connection comes off after the capacitor/resistor network. Sorry, I should have mentioned that. My bad...

This is why the signals that are derived from S2 are a LEVEL whereas the others are low-going pulses.

So far, so good!

I will explain a bit about the 74123 next. It is a little more complicated than what you have described...

Give me a short while to finish off what I am currently doing...

Dave