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 11th Mar 2021, 8:19 pm   #941
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

Mark, I think your suggestion, which is 'non invasive', is a good one and there is no harm in trying it before bringing the focus back onto the video RAM / output circuit if that proves necessary.

Just keep going for now and guide Colin through that process of following R/W. Maybe also look at the 'TV RAM R/W' signal emanating from UF6 pin 7, since that line only ever activates when the CPU is trying to write data into UF7 / UF8 through UE7 / UE8.
SiriusHardware is offline  
Old 11th Mar 2021, 8:29 pm   #942
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

Alan(AJ), George, good leads, I will track through those areas of the circuit.
SiriusHardware is offline  
Old 11th Mar 2021, 8:40 pm   #943
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by buggies View Post
I am following this massive thread with interest but will not be interfering - just to remark that the "@" display reminded me of this page:-

http://www.dasarodesigns.com/project...dore-pet-2001/
(I realise that it is a different Pet and the CPU has been replaced - but might the clock phase circuitry be similar?)

They say:-
"This screen is usually the result of a failed CPU. Because the CPU contains the clock divider circuitry that produces the clock-phases used by the video circuitry, if it is not present the characters stored in the video RAM (random noise) will not be displayed and instead the character represented by the code 0x00 is displayed."

Switching back to lurking mode...
Thanks George, We’ll need to wait for the replacement UF9 but that might be a useful link. I think we verified the control inputs before removing the suspect UF9.

Don’t stay too quiet, even just occasional encouragement is helpful.
Mark1960 is offline  
Old 11th Mar 2021, 8:41 pm   #944
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
Alan(AJ), George, good leads, I will track through those areas of the circuit.
For the avoidance of confusion I deleted my post when I realised George had already covered the clock-phase issue although I should say that I found Matthew D'Asaro's troubleshooting guide very helpful when working on my 2001-8, something I mentioned a couple of times in the other thread. Care needed though because the two PETs are completely different in many respects. It just so happens that the video RAM/ROM circuitry is very similar.

Alan

Last edited by ajgriff; 11th Mar 2021 at 8:49 pm.
ajgriff is online now  
Old 11th Mar 2021, 9:19 pm   #945
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

The Φ1 signal from the CPU is officially 'NC' on the PET so we don't have to follow that.

Φ2 on the other hand does get about a bit. After passing through buffer A10 pins 9/8 it becomes signal 'BØ2' which we can assume stands for 'Buffered Φ2'.

'BØ2' goes to UC6 (6520) on sheet 2 (IEE 488 interface), It goes to UC5 (6522) and UC7 (6250) in the keyboard scanning circuit, and it does go to the video logic area (sheet 7) where it arrives, or should, as BØ2 on UA5 pin 1.

So it may be worth looking for the Φ2 signal on
-The CPU pin 39, which I am sure we have checked before
-UA10 pin 8
-UA5 pin 1

Also look for activity specifically on the 'TV RAM R/W' line during the post-reset phase to see if the CPU is attempting to clear the video RAM.
-UF6 pin 7
-UE7, UE8 pin 19s

Does anyone recall if the cursor on the PET's initial screen blinks or is steady-state?
SiriusHardware is offline  
Old 11th Mar 2021, 9:23 pm   #946
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

It blinks. Mine's doing it at this very moment.

Alan
ajgriff is online now  
Old 11th Mar 2021, 9:25 pm   #947
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

OK - slight change now in that I still get the signal on UC4/34 but only for a second or two after power up.

I get the same signal on UA3/12.

I get a similar signal on UA3/8, but only half the width.

I get no waves on UA10/4 - the signal goes high and stays there.

Colin.


Quote:
Originally Posted by Mark1960 View Post
Quote:
Originally Posted by ScottishColin View Post
Mark - I get the attached on UC4/34 after power on. It does not go away even minutes later.
That looks OK, you should see something similar on UA10-4, UA3-12, UA3-8.
ScottishColin is offline  
Old 11th Mar 2021, 9:37 pm   #948
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

CPU pin 39, UA10 pin 8, UA5 pin 1

All attached - all the same

UF6 pin 7, UE7, UE8 pin 19s

No activity even straight after pwoer on any any of these three pins. The go straight high.

Colin.



Quote:
Originally Posted by SiriusHardware View Post
The Φ1 signal from the CPU is officially 'NC' on the PET so we don't have to follow that.

Φ2 on the other hand does get about a bit. After passing through buffer A10 pins 9/8 it becomes signal 'BØ2' which we can assume stands for 'Buffered Φ2'.

'BØ2' goes to UC6 (6520) on sheet 2 (IEE 488 interface), It goes to UC5 (6522) and UC7 (6250) in the keyboard scanning circuit, and it does go to the video logic area (sheet 7) where it arrives, or should, as BØ2 on UA5 pin 1.

So it may be worth looking for the Φ2 signal on
-The CPU pin 39, which I am sure we have checked before
-UA10 pin 8
-UA5 pin 1

Also look for activity specifically on the 'TV RAM R/W' line during the post-reset phase to see if the CPU is attempting to clear the video RAM.
-UF6 pin 7
-UE7, UE8 pin 19s

Does anyone recall if the cursor on the PET's initial screen blinks or is steady-state?
Attached Thumbnails
Click image for larger version

Name:	UC4 pin 39 - 20210311.jpg
Views:	26
Size:	44.0 KB
ID:	228938   Click image for larger version

Name:	UA10 pin 8 - 20210311.jpg
Views:	28
Size:	43.3 KB
ID:	228939   Click image for larger version

Name:	UA5 pin 1 20210311.jpg
Views:	29
Size:	43.8 KB
ID:	228940  
ScottishColin is offline  
Old 11th Mar 2021, 9:37 pm   #949
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

Are you really sure there is no R/W signal on UA10 pin 4? Because the input to UA3 13->12 comes from UA10 pin 4, and you say you do have a signal on UA3 pin 12.

Regarding there only being R/W activity for a second or so, can you check the CPU A0 pin (9) to see if the CPU is still running continually, or if it is stopping after a second or so.

Last edited by SiriusHardware; 11th Mar 2021 at 9:44 pm.
SiriusHardware is offline  
Old 11th Mar 2021, 9:48 pm   #950
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

Just caught your latest, your scope captures show BØ2 getting as far as the video logic circuit.

Your other observation regarding these pins, which are all on the TV RAM R/W line...

Quote:
UF6 pin 7, UE7, UE8 pin 19s
...confirms that the system is not trying to write anything to the video RAM during initialisation.
SiriusHardware is offline  
Old 11th Mar 2021, 10:01 pm   #951
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

Quote:
It blinks. Mine's doing it at this very moment.
So all being well, in normal circumstances, with the machine sitting in the initial 'waiting for user input' loop, we would expect to see regular, brief writes to the video RAM as the cursor character is written / deleted / written / deleted / written / deleted.
SiriusHardware is offline  
Old 11th Mar 2021, 11:18 pm   #952
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Might be worth noting that even when the 2001-8 is displaying the correct opening screen there's no cursor when the two PIAs are removed. Can't immediately recall which of the two determines the appearance of the cursor and haven't looked into the nature of the rôle played by the relevant PIA.

Alan
ajgriff is online now  
Old 11th Mar 2021, 11:45 pm   #953
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Here's UA10 pin 4 - only present for a few seconds after power up.

Also attached is pin9 from the 6502. It does not go away, although the voltage is a little lower than the others.

Colin.


Quote:
Originally Posted by SiriusHardware View Post
Are you really sure there is no R/W signal on UA10 pin 4? Because the input to UA3 13->12 comes from UA10 pin 4, and you say you do have a signal on UA3 pin 12.

Regarding there only being R/W activity for a second or so, can you check the CPU A0 pin (9) to see if the CPU is still running continually, or if it is stopping after a second or so.
Attached Thumbnails
Click image for larger version

Name:	UA10 pin 4 20210311.jpg
Views:	41
Size:	45.2 KB
ID:	228958   Click image for larger version

Name:	6502 Pin 9 20210311.jpg
Views:	43
Size:	43.6 KB
ID:	228960  
ScottishColin is offline  
Old 12th Mar 2021, 12:19 am   #954
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

OK, so really the situation is that you have read / write activity in the places you were asked to look for it, but only for a few seconds after power-up. In normal circumstances I would expect there would be a high level of writes for a very short time as the system does things like write configuration commands to the large peripheral ICs and also writes to the screen RAM to clear it. It probably also clears the entire system memory as well, so all of those things would create a big burst of write-activity just after switch-on. After that I would still expect to see continuous writes at a less high rate, for example to the IC which drives the keypad rows.

A few things to look at tomorrow:-

-Look for activity on UD9 chip select (pin 20) - whether there is any, and if there is, is it momentary (just for a couple of seconds after switch-on) or continuous.

-Look for activity on UC7 pin 2,3,4,5. They are part of the drive signal for the keyboard key scan circuit. Do you see continuous activity on there? If not, what state are they in (High? Low?)

-What state do you see on UC7 pin 18? 'Video On'?

-What state do you see on UC7 pin 19?

-What state or activity (if any) do you see on UC7 pin 9?
SiriusHardware is offline  
Old 12th Mar 2021, 2:24 am   #955
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
Default Re: Non-working Commodore PET 3016

Colin, when looking for activity on R/W and related signals switch your scope to trigger on falling edge.

After the initial burst of writes there will only be brief intermittent negative pulses on the R/W signals and your scope trigger on positive edge might not display the negative pulse.

Edit: Also change the trigger level back to 1.4v, your last set of traces look like they were triggering on negative overshoot, and there might not be enough overshoot on all the signals.
Mark1960 is offline  
Old 12th Mar 2021, 11:24 am   #956
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Non-working Commodore PET 3016

Is there a disassembly of the UD9 ROM that anyone knows of? It would be useful to know what the CPU should be trying to do and in what order. (I'm assuming that's where the CPU goes to when it comes out of reset - could be wrong of course).
SiriusHardware is offline  
Old 12th Mar 2021, 1:02 pm   #957
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by SiriusHardware View Post
Is there a disassembly of the UD9 ROM that anyone knows of?
Isn't it on Bo Zimmerman's site somewhere? Great resource but not always easy to find what you're looking for, especially for someone like me who doesn't understand half the terminology!

Alan
ajgriff is online now  
Old 12th Mar 2021, 1:19 pm   #958
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

UD9 pin 20 waves attached.

UC7 pins 2-5 go to 4V for about 1 second then back to 0.

UC7 pin 18 pulses between 5V and 0V and does not stop pulsing

UC7 pin 19 goes 'high, but only to 2.7V and stays there

UC7 pin 9 goes high to 4.9V (tested with a meter) and stays there.

Colin.


Quote:
Originally Posted by SiriusHardware View Post
OK, so really the situation is that you have read / write activity in the places you were asked to look for it, but only for a few seconds after power-up. In normal circumstances I would expect there would be a high level of writes for a very short time as the system does things like write configuration commands to the large peripheral ICs and also writes to the screen RAM to clear it. It probably also clears the entire system memory as well, so all of those things would create a big burst of write-activity just after switch-on. After that I would still expect to see continuous writes at a less high rate, for example to the IC which drives the keypad rows.

A few things to look at tomorrow:-

-Look for activity on UD9 chip select (pin 20) - whether there is any, and if there is, is it momentary (just for a couple of seconds after switch-on) or continuous.

-Look for activity on UC7 pin 2,3,4,5. They are part of the drive signal for the keyboard key scan circuit. Do you see continuous activity on there? If not, what state are they in (High? Low?)

-What state do you see on UC7 pin 18? 'Video On'?

-What state do you see on UC7 pin 19?

-What state or activity (if any) do you see on UC7 pin 9?
Attached Thumbnails
Click image for larger version

Name:	UD9 pin 20 20210321.jpg
Views:	42
Size:	45.0 KB
ID:	228992  
ScottishColin is offline  
Old 12th Mar 2021, 1:20 pm   #959
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Done - apologies. It is possible I'll get the hang of this one day.

Do any of my previous ones need repeating?

Colin.


Quote:
Originally Posted by Mark1960 View Post
Colin, when looking for activity on R/W and related signals switch your scope to trigger on falling edge.

After the initial burst of writes there will only be brief intermittent negative pulses on the R/W signals and your scope trigger on positive edge might not display the negative pulse.

Edit: Also change the trigger level back to 1.4v, your last set of traces look like they were triggering on negative overshoot, and there might not be enough overshoot on all the signals.
ScottishColin is offline  
Old 12th Mar 2021, 1:27 pm   #960
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
Default Re: Non-working Commodore PET 3016

Does this page help?

http://www.6502.org/users/andre/petindex/roms.html

Colin.


Quote:
Originally Posted by SiriusHardware View Post
Is there a disassembly of the UD9 ROM that anyone knows of? It would be useful to know what the CPU should be trying to do and in what order. (I'm assuming that's where the CPU goes to when it comes out of reset - could be wrong of course).
ScottishColin is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 3:41 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.