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.

Reply
 
Thread Tools
Old 26th Jul 2023, 3:07 pm   #341
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

OK - new probes in use for a test. This is repeatable - I get 18 double blips - see screenshots on UE13 pin 12 after Reset goes high. I have taken two screenshots of the same waveform but using different time/div settings.

Is this in the right direction?

Colin.
Attached Thumbnails
Click image for larger version

Name:	CH2 green pin 40 6502 CH1 yellow pin 12 UE13 200 us.jpg
Views:	21
Size:	60.3 KB
ID:	282097   Click image for larger version

Name:	CH2 green pin 40 6502 CH1 yellow pin 12 UE13 50us closeup.jpg
Views:	23
Size:	62.4 KB
ID:	282098  
ScottishColin is offline   Reply With Quote
Old 26th Jul 2023, 3:34 pm   #342
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

It's a start. That's still the Tynemouth board, yes? It's interesting that it does exactly what Daver2's EPROM does at startup, accesses the CRTC IC 36 times in total. This is a nice bit of behavioural knowledge worth scribbling down somewhere for the future.

Now - keeping all settings the same, move the trigger channel probe to UE13 pin 12 and put the other channel probe on the R/W pin of UB13 (pin 22) or if you can't get to that pin easily, pin 8 of UD15 which is where that signal comes from.

This is to see which of those 36 accesses coincide with write operations and which of them coincide with read operations - also to check that the B R/W signal is present and correct since it is one of the ones we haven't looked at yet.

If possible try to catch it immediately after reset because we want to see those same 18 pairs of chip access pulses lined up with whatever activity there is on the BR/W line - this is where an EXT TRIG input, or indeed the analyser feature of your scope, would have come in handy.
SiriusHardware is offline   Reply With Quote
Old 26th Jul 2023, 4:53 pm   #343
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

Still with the Tynemouth board set to ROM and RAM enabled.

Here's the screenshot - shows activity beforehand on UD15 and the 36 waves on UE13 pin 12. Did you want it more zoomed in?

Colin.

P.S. It's nice to finally start to get some consistent results - that's what was doing my head in last night. I've also started to understand how to do this with this software now, although it needs a lot of force-closing and restarting.

Quote:
Originally Posted by SiriusHardware View Post
It's a start. That's still the Tynemouth board, yes? It's interesting that it does exactly what Daver2's EPROM does at startup, accesses the CRTC IC 36 times in total. This is a nice bit of behavioural knowledge worth scribbling down somewhere for the future.

Now - keeping all settings the same, move the trigger channel probe to UE13 pin 12 and put the other channel probe on the R/W pin of UB13 (pin 22) or if you can't get to that pin easily, pin 8 of UD15 which is where that signal comes from.

This is to see which of those 36 accesses coincide with write operations and which of them coincide with read operations - also to check that the B R/W signal is present and correct since it is one of the ones we haven't looked at yet.

If possible try to catch it immediately after reset because we want to see those same 18 pairs of chip access pulses lined up with whatever activity there is on the BR/W line - this is where an EXT TRIG input, or indeed the analyser feature of your scope, would have come in handy.
Attached Thumbnails
Click image for larger version

Name:	CH2 green UE13 pin 12 CH1 yellow UD15 pin 8 200us.jpg
Views:	16
Size:	63.3 KB
ID:	282104  
ScottishColin is offline   Reply With Quote
Old 26th Jul 2023, 5:06 pm   #344
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,241
Default Re: Commodore PET 4032/8050/3022

Quote:
Originally Posted by ortek_service View Post
You may also be able to use multiple 8ch ones together, as 16ch ones tend to be rather more (although often much-higher sampling rates)
I tried a clone 8 channel one but couldn’t get it to work, I think it was showing as a usb blaster, not sure if I had the right drivers installed. The 16 channel ones are not expensive and it was easy to get it working, search for CY7C68013A-56 EZ-USB FX2LP USB2.0.
Mark1960 is offline   Reply With Quote
Old 26th Jul 2023, 6:03 pm   #345
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Quote:
Did you want it more zoomed in?
Is that possible? I'd like to see the 36 individual CS pulses spread as far as possible across the screen so that we can associate individual R/W pulses with particular read / write pulses - oh, and also change to falling edge trigger on channel 2, the channel you are using to monitor / trigger from the CRTC CS pulses.

It's OK that there is activity on R/W before the 'conversation' with the CRTC because the CPU chats to other devices before it gets to the CRTC.
SiriusHardware is offline   Reply With Quote
Old 26th Jul 2023, 6:17 pm   #346
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

One of the good things about the software is that once I get a trace like this, I can alter the time/div setting and it will apply that to the stored data.

Here's the 36 as requested, along with a much more zoomed in version to show detail - I'll do it again with falling instead of rising.

Colin.
Attached Thumbnails
Click image for larger version

Name:	CH2 green UE13 pin 12 CH1 yellow UD15 pin 8 50us.jpg
Views:	31
Size:	67.3 KB
ID:	282122   Click image for larger version

Name:	CH2 green UE13 pin 12 CH1 yellow UD15 pin 8 10us.jpg
Views:	23
Size:	55.2 KB
ID:	282123  
ScottishColin is offline   Reply With Quote
Old 26th Jul 2023, 6:20 pm   #347
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

Set the trigger to falling and I get the same waveforms as inn post 346. I'm at least happy it's consistent and reproducible.

Colin.


Quote:
Originally Posted by SiriusHardware View Post
Quote:
Did you want it more zoomed in?
Is that possible? I'd like to see the 36 individual CS pulses spread as far as possible across the screen so that we can associate individual R/W pulses with particular read / write pulses - oh, and also change to falling edge trigger on channel 2, the channel you are using to monitor / trigger from the CRTC CS pulses.

It's OK that there is activity on R/W before the 'conversation' with the CRTC because the CPU chats to other devices before it gets to the CRTC.
ScottishColin is offline   Reply With Quote
Old 26th Jul 2023, 7:03 pm   #348
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Yes, I suppose it was just triggering on the rising edge of the first CS pulse, so hardly any difference in the start up time.

Anyway, good job, we can see that all 36 of those initial accesses to the CRTC are writes, (when R/W = high, CPU is reading, when R/W = low, CPU is writing), so that's useful to know.

Now can you do the same thing for the pins / signals below, still using the CRTC CS signal as the trigger...

CRTC 'RS' pin (24), signal (AKA BA0, also available at UD13 pin 3)

CRTC 'E' pin (23), signal (AKA BA2, also available at UD13 pin 7)

One at a time, CRTC data pins D7 through D0 (Pins 26-33 inclusive) same signals also available on the D0 to D7 pins of any of the main ROMs.

Finally if you can get to them, show the waveforms and frequencies on pins 39 and 40 of the CRTC - this can be in single channel mode and with AUTO trigger as they should be regular, continuous signals.
SiriusHardware is offline   Reply With Quote
Old 27th Jul 2023, 9:07 am   #349
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

OK - images of results from UD13 pin 3 and UD 13 pin 7 attached - two images for each to show detail.

D7->D0 to follow in a while along with 39/40 of the CTRC.

Colin.
Attached Thumbnails
Click image for larger version

Name:	CH2 green UE13 pin 12 CH1 yellow UD13 pin 3 50us.jpg
Views:	17
Size:	69.7 KB
ID:	282151   Click image for larger version

Name:	CH2 green UE13 pin 12 CH1 yellow UD13 pin 3 20us.jpg
Views:	14
Size:	66.8 KB
ID:	282152   Click image for larger version

Name:	CH2 green UE13 pin 12 CH1 yellow UD13 pin 7 50us.jpg
Views:	15
Size:	68.3 KB
ID:	282153   Click image for larger version

Name:	CH2 green UE13 pin 12 CH1 yellow UD13 pin 7 20us.jpg
Views:	12
Size:	63.3 KB
ID:	282154  
ScottishColin is offline   Reply With Quote
Old 27th Jul 2023, 9:18 am   #350
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Good so far - this is very useful information you are collecting because on a working system with the motherboard ROMs enabled we should expect to see exactly the same patterns because the same code resident at the same ROM addresses will be running around the same loop as it sends instructions to the CRTC - provided the code in the ROMs in your board and the code in the ROMs in the Tynemouth board are identical.

I'll let you continue and check back in when you've had to time to capture everything.
SiriusHardware is offline   Reply With Quote
Old 27th Jul 2023, 9:46 am   #351
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

I forgot you asked for ROM chip numbers. Here they are.

UD6 - 901465-22
UD7 - 901498-01
UD8 - 901465-21
UD9 - 901465-20
UD10 - 901465-23

Colin.
ScottishColin is offline   Reply With Quote
Old 27th Jul 2023, 10:05 am   #352
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

I'm in danger of guessing that you want the following readings:

"One at a time, CRTC data pins D7 through D0 (Pins 26-33 inclusive) same signals also available on the D0 to D7 pins of any of the main ROMs." with the Tynemouth board ROMs enabled - is that correct?

And then again with the Tynemouth ROMs disabled?

Colin.
ScottishColin is offline   Reply With Quote
Old 27th Jul 2023, 10:12 am   #353
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

I'll lay hands on file copies of those and disassemble a bit of the code to try to see what happens / what gets initialised in what order. I have a feeling not much happens in the initial code in UD6, the CPU is soon told to jump to a large subroutine in UD7 where all the 'Magic Happens'. That's probably where the 'Chirp' code is and certainly where the IC initialisation / clear the screen code will be. Also, someone back in the past pointed to a disasssembly of the 5-ROM set, can anyone remember where that was?

Once you've done all the captures asked for earlier, now that you are quite expert at this I will ask you to put the Daver2 EPROM in the UD7 and try again to observe those 36 CRTC CS pulses shortly after the system comes out of reset. If we still don't see them then UD6 might possibly be faulty and we (you) may have to desolder it and read it and verify it against the matching code file from zimmers. It would be useful to have that in a socket anyway because then we would be able to try running Slothie's test code - although - I'm not sure if that test code tries to initialise the CRTC. (Slothie may be able to tell us, if he is around).
SiriusHardware is offline   Reply With Quote
Old 27th Jul 2023, 10:16 am   #354
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Quote:
And then again with the Tynemouth ROMs disabled?
Not for the moment, but it is a diagnostic tool we can put on the shelf and come back to later if necessary.

As mentioned, once you've captured the activity on the data lines I want to go back to the Daver2 EPROM now that you have mastered the art of capturing specific events using SINGLE sweep, to see once and for all if that EPROM is (or is not) trying to initialise the CRTC. The giveaway would be those 36 CRTC CS pulses shortly after the system comes up out of reset.
SiriusHardware is offline   Reply With Quote
Old 27th Jul 2023, 10:23 am   #355
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,338
Default Re: Commodore PET 4032/8050/3022

Quote:
Originally Posted by ScottishColin View Post
As a matter of interest, my scope is a 6022BL which seems to have a logic analyzer function available - never used it. Would it be of more use for what we are trying to do here (section 5.1 of the manual link below).

http://www.hantek.com/Product/Hantek...2BL_Manual.pdf

Colin.
I had seen a few PC 'scopes that also had logic-analyzer inputs, but hadn't realised that Colin's one had - Or that it has 16 rather than just 8 channels, so in theory even more useful.

However, from what I found here there may be some issues with the provided software and there are some alternative (but maybe Linux only?) ones that they did get to work OL: https://www.eevblog.com/forum/testge...ll-16-channels

It looks like there may also be some helpful info in the review comments on here: https://www.amazon.co.uk/Hantek-Digi.../dp/B01M74UKKC

- I do notice from the spec's on there, that the Logic-Analyzer bandwidth is 10MHz (rather than 20MHz max. for 'scope mode - Although maybe only single-ch mode). So it could be sample rate is only 24M in logic-analyzer mode, rather than 48M max for (single-channel?) 'scope mode. However, for these logic-analyzer digital inputs, you don't need to multiplex-share a single ADC, so you do get 24M sample rate on all 16 channels simultaneously.

It also seems that if you use the original Saleae software for the 8ch 24M sampling? that those cheap 8ch - as well as your Hantek 6022BL? are apparently compatible with, then that can be a restriction on number of channels / speed. But some of the alternatives they suggest to use with the 6022BL may give use of all 16 channels / higher-speed.


Quote:
Originally Posted by SiriusHardware View Post
Quite useful if you can work out how to use it - as far as I can see, you run a specific 'BL' version of the software and maybe press a button to put it into logic analyser mode. You can trigger from one channel with the usual ability to specify rising edge / falling edge etc and you can monitor what is happening on up to 15 other lines. I could wish it had 32 inputs rather than just 16 but that would be churlish.

The main thing is - do you have the necessary means of connecting the analyser inputs on the scope box to any IC pins of interest?
Well it seems they do provide all the 'dupont' receptacle cables and 'mini-grabber' IC pin test clips, from what's shown in the Accessories, here: http://www.hantek.com/products/detail/153
- I assume these are provided with it, rather than having to buy extra.
(They are shown in the picture, on the Amazon listing I found (above). But maybe there are slightly-cheaper options that exclude these - although they are a common-standard on this type of thing, so easy & cheap to buy separately).
ortek_service is offline   Reply With Quote
Old 27th Jul 2023, 10:31 am   #356
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,338
Default Re: Commodore PET 4032/8050/3022

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

Keep an eye out for 'IC test clips' going cheap, as they can be a good time saver when trying to monitor signals on IC pins. Problem is, you ideally need as many different sizes of test clips as there are sizes of DIP IC.

The blue connectors under the link below (from CPC) are the kind of thing I mean -

https://cpc.farnell.com/c/cable-lead.../ic-test-clips

As you can see they aren't exactly cheap from mainstream sources but you may be able to find them cheaper elsewhere. Ideally you would want a 40-pin one as well, but one intended to fit a 24 pin or 28 pin wide bodied IC could possibly be used on a 40-pin IC by placing it at the high end or the low end of the device under investigation. Anyway, these only really come under the heading of 'would be nice to have'.

Edit: Crossed with Owen, who I think was talking about something similar.
No, the problem with those IC Test clips (which although useful for making lots of connections in one go / In-circuit reading/(re)programming of soldered-in EEPROM Memory etc. and I do have various ones (Inc. built in logic probe LEDS on some), is that they stick up too high to get access to IC's under the Tynemouth etc. board (without jacking that up on lots of IC sockets, which could make it a bit wobbly).

Whereas I had in mine the individual 'twin grabber hooks' type test clips, commonly used with logic analyzers etc and also shown on Hantek accessories, here: http://www.hantek.com/products/detail/153
ortek_service is offline   Reply With Quote
Old 27th Jul 2023, 10:35 am   #357
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,338
Default Re: Commodore PET 4032/8050/3022

Quote:
Originally Posted by Mark1960 View Post
Quote:
Originally Posted by ortek_service View Post
You may also be able to use multiple 8ch ones together, as 16ch ones tend to be rather more (although often much-higher sampling rates)
I tried a clone 8 channel one but couldn’t get it to work, I think it was showing as a usb blaster, not sure if I had the right drivers installed. The 16 channel ones are not expensive and it was easy to get it working, search for CY7C68013A-56 EZ-USB FX2LP USB2.0.
Well it turns out Colin already has a 16ch one built into his Hantek 6022BL 'scope. But might also have some 'fun' trying to find software that works propery for it - Although I have posted some links above to findings other people had as to what software did work correctly (seemingly more open-source community ones, that are better supported than something the manufacturer may have provided but not updated)
ortek_service is offline   Reply With Quote
Old 27th Jul 2023, 10:47 am   #358
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Another precautionary measure - as I understand it you (Colin) have been taking these captures from alternatives to the CRTC pins because they are difficult to get to with the Tynemouth board fitted (this is one reason I'd like to get the Daver2 test EPROM up and running instead, as it doesn't block access to other ICs).

That being the case, can you also (with power off and Tynemouth board removed) check continuity between those pins on the CRTC controller and the pins you have been measuring from - so for example you've been getting the CRTC CS pulses from UE13 pin 12, so check continuity between UE13 pin 12 and the actual CS pin (25) on the CRTC controller and do the same for all the other CRTC pins which you've been scoping 'remotely' because access to the actual CRTC was difficult.

It would be unwise to assume that just because those signals are present at their source or elsewhere on the PCB, they must also be present at the actual pins of the CRTC. As long as we make sure that each point of measurement is directly connected to the CRTC pin, we can be reasonably sure that the signal we saw on the measurement point was also present at the CRTC pin.
SiriusHardware is offline   Reply With Quote
Old 27th Jul 2023, 2:35 pm   #359
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

Continuity between UE13 pin 12 and CTRC pin 25 is ok by me (0.5 ohms).

I've not posted as I've been suffering from a lack of consistency in my readings again.

I'm going to replace the 6502 socket, socket UD6 ad solder a wire to the underneath of UE13 pin 12 to allow for easier testing.

I'll be back.

Colin.
ScottishColin is offline   Reply With Quote
Old 27th Jul 2023, 2:42 pm   #360
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

OK Colin, good luck.

Found a disassembly of the 5-ROM set on, where else but Zimmers:-

http://www.zimmers.net/anonftp/pub/c...isassembly.txt

This will be a great time saver when it comes to trying to understand what should happen when, but suffice to say that on reset, the CPU works through only several bytes in UD6 before being told to jump to the first address in UD7, E000. There, the first instruction it encounters is a jump to E04B where there is an immediate Jump To Subroutine at E06F in which initialisation of the I/O devices is carried out. On return from that subroutine There is another Jump To Subroutine at E07A which sets up the CRTC controller, and then on return from that subroutine it clears the screen, the clear screen routine is at E051-E079.

At first glance it looks like the Chirp is generated towards the end of the I/O initialisation subroutine (E06F) but I will have to look at it more closely.
SiriusHardware is offline   Reply With Quote
Reply

Thread Tools



All times are GMT +1. The time now is 5:59 am.


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.