Non-working Commodore PET 3016

[SiriusHardware]

Yes, that one I missed - held high by a 1K resistor - and I am not sure we have ever asked Colin to check that for us, so, before anything else just check that the 'SO' pin (38) is at or close to +5V with the power on. We aren't expecting to see a waveform or frequency on that one, just a steady voltage.

If pin 38 has ~5V on it then try running the CPU with only the following pins connected:

-5V supply (pin 8) - 5V
-0V supply (pins 1 and 21) - 0V
-Clock in (Pin 37) - 1MHz clock
-RDY (Pin 2) (Held high)
-IRQ (pin 4) (Held high)
-NMI (Pin 6) (Held high)
-SO(Pin 38) (Held high)
-Reset (Pin 40) (Momentarily low, then held high. The reset circuit takes care of that.)

The main thing we want to see is activity on address pin A0 (pin 9).

[ScottishColin]

I bought spare 40 pin sockets so I'll do the doings with that. I'll wait a while to see if others chip in

[SiriusHardware]

Originally Posted by Mark1960

It probably is worth setting up the 6502 to run a nop test, but better to do it on a breadboard or protoboard

Only problem is, where does Colin get a clock from then, unless he jumpers it across from the mainboard to the breadboard? That's why I suggested doing it in situ, where there is already a clock and all the required pullups on INT, NMI, RDY, SO, etc, and there is a reset circuit to generate a proper start up reset signal in case the CPU actually needs that.

A NOP test would generate a nice progressive count on the address lines but random / unspecified data should still cause activity on the address lines as the CPU will try to execute it anyway. That's all I'm looking for really, any address line activity at all.

I think Colin did say he bought both types of socket so a conventional socket would be OK for this and the 'unwanted' socket contacts could perhaps be removed altogether by pushing them out from below, to leave only the ones used for the minimum pin set.

[SiriusHardware]

Originally Posted by Mark1960

What I was thinking was to connect pin 36 to 5v, preferably through a resistor somewhere between 3k and 10k. Maybe link pin 36 to pin 2 is easier.

To enlarge on this: Mark is suggesting that your apparently genuine looking 6502 may possibly be some other variant which has been rebadged to look like an original 6502. On an original 6502 some of the pins including pin 36 are marked 'NC' (no connection, or 'do not connect') on the pinout diagram, but some later variants added features to these unused pins, such as 'BE' on pin 36. So Mark is saying try taking that pin 'high' (to +5V either directly or via a resistor) to see if that brings the address lines to life.

If you are going to try that, do it (and also check the voltage on pin 38 for us) before trying to run the CPU with only essential connections.

[ajgriff]

Originally Posted by Mark1960

On later 65C02, I think including the rockwell, pin 36 was BE, bus enable.

The original MOS 6502 and Rockwell 6502 datasheets show pin 36 as NC (not connected). Only the WDC65C02 is shown as BE (bus enable). Is there a more recent Rockwell version of the 65C02?

Alan

[ScottishColin]

Just so I have this right, first thing is the following:

Turn PET on and measure voltage on pin 38.

Then turn PET off

Connect pin 36 to 5v by linking pin 36 to pin 2 with a resistor (between 3k and 10k) in the link.

Then power on.

Then see what frequency and activity there is on pin 9 with my scope.

That sound right?

[SiriusHardware]

Yes please.

[ScottishColin]

OK - pin 38 with power on is 5.173V

Pin 2 connected to pin 36 with a 10K Ohm resistor

Power on

Pin 36 measures 5.08V

Pin 9 no frequency or wave pattern.

https://drive.google.com/file/d/19-w...ew?usp=sharing

[ScottishColin]

Again, for confirmation,

1. remove 10k resistor between pins 2 and 36

2. take an ordinary 40 pin socket, remove al pins except the ones listed below:

pin 8
pins 1 and 21
pin 37
pin 2
pin 4
pin 6
pin 38
pin 40

Then measure frequency and any wave on pin 9.

Have I got that right?

[SiriusHardware]

Thanks for doing that - if you are up for it can you now try the 'minimal CPU pins connected' check either by clipping the unwanted pins from a conventional socket, or by removing the unwanted contacts from a conventional socket, or by bending out the 'unwanted' pins on the CPU so they fall outside the mainboard socket.

The last way is obviously the simplest way to do it, but offers the greater possibility of damage to the IC.

Using a socket requires no modification to or bending of the IC pins but sacrifices a socket - but there are plenty more where that came from.

Edit: Just caught your post #309. Yes, that's right. However you do it be sure there is no connection between any of the other pins and the associated contacts on the mainboard socket. In case it's not clear, we mean for you to plug the modified socket into the mainboard socket and then plug the CPU into the modified socket.

[ScottishColin]

Right. Progress. See attached file.

I took an ordinary 40 pin socket and pushed through all the pins we didn't want and tested with success on pins 9, 10 and 11.

[SiriusHardware]

Again, I can't view that right now (no office suite of any sort here) but I gather from the positive tone that we finally did get something on the address lines?

[ajgriff]

Looks really promising.

Alan

[ScottishColin]

Originally Posted by SiriusHardware

Again, I can't view that right now (no office suite of any sort here) but I gather from the positive tone that we finally did get something on the address lines?

Yes. Definitive square waves and detected frequencies on 9, 10 and 11

[SiriusHardware]

OK. Could you check the voltages on pins 1 and 19 of IC UC3 and pins 1 and 19 of IC UB2 (Both 74LS244). The voltage expected on all four pins is 0V. You can do this either with the CPU plugged in normally or with it still half disconnected, makes no difference.

[ScottishColin]

Pins 1 and 19 of UC3 - 0.102V

Pins 1 and 19 of UB3 (I'm assuming you meant UB3 not UB2) - 0.102V and .2V respectively

[SiriusHardware]

Pins 1 and 19 of UB3 (I'm assuming you meant UB3 not UB2) - 0.102V and .2V respectively

Indeed I did, sorry.

As the pin1 / pin 19 voltages on those buffers are correct,

Power off, CPU removed, please measure...

-The resistance between each of the following CPU socket pins (see below) and 0V, with the black lead on 0V.

-The resistance between each of the same CPU socket pins and +5V, with the red lead on +5V.

Suggest you use your old meter for these resistance measurements. Use whatever range you need to to get a meaningful reading. I would expect the readings from all sixteen pins to 0V to be roughly the same, and the readings from all sixteen pins to +5V to be a different resistance, but again they should all be pretty similar to each other.

9
10
11
12
13
14
15
16
17
18
19
20
..........Not 21!)
22
23
24
25

Let us know how that goes.

[ScottishColin]

Ok. I'll get on with those measurements in the morning.

As a matter of interest are these tests part of a process of elimination, or do you suspect we might find something? Just curious.

[Mark1960]

Hi Colin, as it looks like the 6502 is working when its address lines are not connected to the pcb, next step Sirius described is to try and find out if something on the board is shorting them to 0v or 5v

[SiriusHardware]

Indeed, also to see if there is any significant difference between identical devices UC3 / UB3 which might indicate that one is possibly faulty.