UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Computers

Notices

Vintage Computers Any vintage computer systems, calculators, video games etc., but with an emphasis on 1980s and earlier equipment.

Closed Thread
 
Thread Tools
Old 29th Jan 2021, 7:53 pm   #341
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
As was explained earlier in the thread, resistance meters use different amounts of current on different ranges, when applied to a non-linear item like a semiconductor device or junction you will get a different reading from one range to the next.

We'll go with the 2K setting for the first run, what we are mainly interested in is whether this value is approximately the same for all of the pins measured to +5V or only some of them.

To consider Refugee's point, after measuring the resistance of all 16 address pins to +5V with the red probe on socket pin 8, could you then make the same measurements again, but keep the black probe on pin 8 and move the red probe from address line to address line.

For the second lot with the probes reversed, dial up a range or two if the reading is infinite / overflow.

Pin 1 testing

PET off, Red pin in Pin 1, meter set to 2k:
9 - 0.628
10 - 0.629
11 - 0.629
12 - 0.630
13 - 0.630
14 - 0.629
15 - 0.630
16 - 0.626
17 - 0.631
18 - 0.630
19 - 0.630
20 - 0.631
22 - 0.630
23 - 0.630
24 - 0.630
25 - 0.556

So Pin 25 looks like an outlier to me?

PET off, Black pin in Pin 1, meter set to 2k:
All readings were 1.

I tried all meter settings and the readings stayed at 1. no matter what
ScottishColin is offline  
Old 29th Jan 2021, 7:54 pm   #342
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: Non-working Commodore PET 3016

I watch a lot of those by Jerry Walker and he seems to find a lot of machines with those chips faulty.
Timbucus is offline  
Old 29th Jan 2021, 7:55 pm   #343
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
OK, that's enough to suggest to me that, as all of those buffer inputs read the same, they are probably OK.

Now: You know the contacts you removed from the socket? Do you still have them? If so, replace the address pin contacts (9-25 inclusive - leave 21 in place) only, but not the data pin contacts (pins 26-33). Let's see what happens to the address pins when you have the address pins connected, but the data pins left disconnected.

I am beginning to wonder if the 6502 is running into a HALT instruction (or whatever the 6502 equivalent is) when it is connected to the data bus.
So the same Pin 1 and pin 8 tests as before with red and black each inserted into pins 1 and 8, with the above pins reinserted in the socket and the power off - have I got that right?

Colin.
ScottishColin is offline  
Old 29th Jan 2021, 7:58 pm   #344
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

That slight difference on 25 (A15) isn't enough to alarm me unduly. Although the resistance looks rather low on 2K range it rises to over 10K on your next range, so I understand, so I think this is OK.
SiriusHardware is online now  
Old 29th Jan 2021, 8:03 pm   #345
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by ScottishColin View Post

So the same Pin 1 and pin 8 tests as before with red and black each inserted into pins 1 and 8, with the above pins reinserted in the socket and the power off - have I got that right?

Colin.
No, sorry, I think you've done all the resistance measurements we need on the address lines for now. They have told us that they are all very similar so the likelihood is that they are OK.

What I'd like now is for you to reinstate the address pin contacts ie, pins 9-20 and pins 22-25 in your modified socket so that you can insert the 6502 with all the address pins connected, but the data pins (26-33) still disconnected. With the modified socket fitted in the CPU socket and the CPU fitted in the modified socket, power on and use your scope to look for activity on the address lines starting with pin 9, 10, 11....
SiriusHardware is online now  
Old 29th Jan 2021, 8:09 pm   #346
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by Timbucus View Post
I watch a lot of those by Jerry Walker and he seems to find a lot of machines with those chips faulty.
They could still be, hence asking Colin to reconnect the address pins (only) to see if that kills the address activity again.
SiriusHardware is online now  
Old 29th Jan 2021, 8:32 pm   #347
Slothie
Octode
 
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
I am beginning to wonder if the 6502 is running into a HALT instruction (or whatever the 6502 equivalent is) when it is connected to the data bus.
The 6502 has no HALT instruction - it's no quitter!
Some variants of the 6502 (like the Sally used in some Atari 8 bits) have a /HALT input, and the 65C02 has a BE input to disable the busses but the original 6502 used in the PET and VIC20 does not.
Slothie is offline  
Old 29th Jan 2021, 8:50 pm   #348
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Non-working Commodore PET 3016

With power off the resistance between 5v and 0v might be quite low, so red on 5v and black on address lines may just be measuring the parasitic diode from 0v to the inputs of the buffer, which would not be a fault condition.

I would suggest putting a 1k resistor from A0 to 0v, without the 6502 connected to the address lines, then measure the voltage across that resistor with the board powered up. The voltage measured should be less than 0.8v.
Mark1960 is online now  
Old 29th Jan 2021, 8:54 pm   #349
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
Quote:
Originally Posted by ScottishColin View Post

So the same Pin 1 and pin 8 tests as before with red and black each inserted into pins 1 and 8, with the above pins reinserted in the socket and the power off - have I got that right?

Colin.
No, sorry, I think you've done all the resistance measurements we need on the address lines for now. They have told us that they are all very similar so the likelihood is that they are OK.

What I'd like now is for you to reinstate the address pin contacts ie, pins 9-20 and pins 22-25 in your modified socket so that you can insert the 6502 with all the address pins connected, but the data pins (26-33) still disconnected. With the modified socket fitted in the CPU socket and the CPU fitted in the modified socket, power on and use your scope to look for activity on the address lines starting with pin 9, 10, 11....
Confirmed that I am still getting scope readings on 9, 10 and 11 with the above pins reinstated.
ScottishColin is offline  
Old 29th Jan 2021, 9:13 pm   #350
dave cox
Nonode
 
dave cox's Avatar
 
Join Date: Jan 2009
Location: Bristol, UK.
Posts: 2,059
Default Re: Non-working Commodore PET 3016

Would Mark1960's pull-down idea be better as a pull-up ?

NMOS is resistive pull-up and active pull-down.

dc
dave cox is offline  
Old 29th Jan 2021, 9:16 pm   #351
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Non-working Commodore PET 3016

Check the outputs of C3 for activity with the scope, pins 18, 3 and 16, should be the same as 9, 10 and 11 from the 6502.

If that looks OK then verify all the outputs of both 74ls244.

If that checks out then it seems to be pointing to ram or rom problems. I think next step would be to connect the data lines again, to see if activity on 9, 10 and 11 stops, just in case you fixed something without realising.

Is the ram in your pet in sockets or only the rom?
Mark1960 is online now  
Old 29th Jan 2021, 9:19 pm   #352
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by dave cox View Post
Would Mark1960's pull-down idea be better as a pull-up ?

NMOS is resistive pull-up and active pull-down.

dc
That was an idea when I thought faulty buffers might be driving the address lines of the 6502 high due to a fault in the buffer inputs, latest from Colin seems to eliminate that.
Mark1960 is online now  
Old 29th Jan 2021, 9:30 pm   #353
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by Slothie View Post
The 6502 has no HALT instruction - it's no quitter!
Ah, but... see here and scroll down to 'The KIL Opcodes'

https://www.pagetable.com/?p=39

If the system is broken to the extent that the CPU is not receiving valid opcodes to execute it may very well be ingesting one of these undocumented halt instructions from the databus. This could be the reason why the CPU appears catatonic, but only when it is reading the actual data bus.
SiriusHardware is online now  
Old 29th Jan 2021, 9:38 pm   #354
Slothie
Octode
 
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
Quote:
Originally Posted by Slothie View Post
The 6502 has no HALT instruction - it's no quitter!
Ah, but... see here and scroll down to 'The KIL Opcodes'

https://www.pagetable.com/?p=39

If the system is broken to the extent that the CPU is not receiving valid opcodes to execute it may very well be ingesting one of these undocumented halt instructions from the databus. This could be the reason why the CPU appears catatonic, but only when it is reading the actual data bus.
I never knew this... this sounds plausible. A corrupt rom could feed any old rubbish to the cpu.
Slothie is offline  
Old 29th Jan 2021, 9:54 pm   #355
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

Again building on what Mark suggested, if you feel like enhancing your Scope-Fu, Colin, with the databus pins still disconnected, try using two probes and both channels of your scope to look at both of these pairs of pins at the same time: (You'll need the second channel turned on to try this, and both probes set to x10 with the dashboard settings also at x10).

The purpose of this is to verify that the buffers UB3 / UC3 are working. If they are you should get the same 'picture' on both channels of your scope, because you are looking at each address line first before, then after it has passed through the buffer. There may be some differences in height / shape but the highs and lows should be in the same places on both traces. As you work higher up through the address pins the pulses will become longer and less frequent, if so keep dialling down the horizontal Time / Div setting on the scope so you can always see at least several pulses.

6502 pin 9, UC3 pin 18
6502 pin 10, UC3 pin 3
6502 pin 11, UC3 pin 16
6502 pin 12, UC3 pin 5
6502 pin 13, UC3 pin 14
6502 pin 14, UC3 pin 7
6502 pin 15, UC3 pin 12
6502 pin 16, UC3 pin 9

6502 pin 17, UB3 pin 18
6502 pin 18, UB3 pin 3
6502 pin 19, UB3 pin 16
6502 pin 20, UB3 pin 5
6502 pin 22, UB3 pin 14 ; Note the jump from CPU pin 20 to 22
6502 pin 23, UB3 pin 7
6502 pin 24, UB3 pin 12
6502 pin 25, UB3 pin 9

You need not necessarily post images of all these checks, just satisfy yourself that each address line looks the same on the output side of the buffer as it does at the CPU. If any don't, then of course we'll want more detail.

Last edited by SiriusHardware; 29th Jan 2021 at 10:04 pm.
SiriusHardware is online now  
Old 29th Jan 2021, 10:46 pm   #356
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by Mark1960 View Post
Check the outputs of C3 for activity with the scope, pins 18, 3 and 16, should be the same as 9, 10 and 11 from the 6502.

If that looks OK then verify all the outputs of both 74ls244.

If that checks out then it seems to be pointing to ram or rom problems. I think next step would be to connect the data lines again, to see if activity on 9, 10 and 11 stops, just in case you fixed something without realising.

Is the ram in your pet in sockets or only the rom?
All RAM chips are soldered to the motherbard.

Colin.
ScottishColin is offline  
Old 29th Jan 2021, 10:49 pm   #357
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
Again building on what Mark suggested, if you feel like enhancing your Scope-Fu, Colin, with the databus pins still disconnected, try using two probes and both channels of your scope to look at both of these pairs of pins at the same time: (You'll need the second channel turned on to try this, and both probes set to x10 with the dashboard settings also at x10).

The purpose of this is to verify that the buffers UB3 / UC3 are working. If they are you should get the same 'picture' on both channels of your scope, because you are looking at each address line first before, then after it has passed through the buffer. There may be some differences in height / shape but the highs and lows should be in the same places on both traces. As you work higher up through the address pins the pulses will become longer and less frequent, if so keep dialling down the horizontal Time / Div setting on the scope so you can always see at least several pulses.

6502 pin 9, UC3 pin 18
6502 pin 10, UC3 pin 3
6502 pin 11, UC3 pin 16
6502 pin 12, UC3 pin 5
6502 pin 13, UC3 pin 14
6502 pin 14, UC3 pin 7
6502 pin 15, UC3 pin 12
6502 pin 16, UC3 pin 9

6502 pin 17, UB3 pin 18
6502 pin 18, UB3 pin 3
6502 pin 19, UB3 pin 16
6502 pin 20, UB3 pin 5
6502 pin 22, UB3 pin 14 ; Note the jump from CPU pin 20 to 22
6502 pin 23, UB3 pin 7
6502 pin 24, UB3 pin 12
6502 pin 25, UB3 pin 9

You need not necessarily post images of all these checks, just satisfy yourself that each address line looks the same on the output side of the buffer as it does at the CPU. If any don't, then of course we'll want more detail.
Just to confirm again - I'll leave pins 9-20, 22-25 in the socket, plug it back in with the CPU and whilst powered on, compare the pairs above to each other. Have I got that right?

Colin.
ScottishColin is offline  
Old 29th Jan 2021, 10:55 pm   #358
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

Exactly so.
SiriusHardware is online now  
Old 29th Jan 2021, 11:00 pm   #359
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Thanks - tomorrow.

Incidentally, I found this Televideo service manual which has the first three pages talking about 6502 testing. In particular, page 3 shows the wave forms and which pins should have similar waveforms to the others on the 6502.

I know it's not what you are asking me to test but I just thought it was interesting - looks like it should be applicable to the PET doesn't it?
Attached Files
File Type: pdf TVI955_Service_Bulletins.pdf (1.19 MB, 39 views)
ScottishColin is offline  
Old 29th Jan 2021, 11:17 pm   #360
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
Default Re: Non-working Commodore PET 3016

I've just had a quick look - it's a cross between what you are already doing and something Mark referred to before, a 'NOP' test. 'NOP' (No Operation) is a microprocessor instruction which doesn't do anything, it just uses up one unit of microprocessor instruction time.

Normally the opcodes / data which the CPU fetches and executes come from either PROM or RAM but there is a trick (which is described in your document) where you disconnect the CPU data pins from the system and hardwire the CPU data pins so that the CPU always reads a 'NOP' instruction - in the case of the 6502 that would be the hex code 'EA' - as mentioned in your document.

Why would you do that? Well, because if you keep the CPU fed on a diet of NOP instructions it runs smoothly up through the address range producing nice regular patterns on all of the address lines, and while it's doing that it is also generating chip select signals and doing other generally useful things which can be looked at. So yes, it's a valid technique and we may well ask you to do something like that a little way down the line. You are almost doing it now, except that you aren't imposing a NOP instruction on the data pins at the moment so the CPU is just executing whatever code it sees there over and over again, probably hex FF.
SiriusHardware is online now  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 6:28 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.