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 21st Jul 2023, 5:31 pm   #261
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

Here's a sample one at 5ns. Is this the kind of thing you're looking for?

Just so you know how I'm doing this, I pause a trace at 1ms then I can zoom in to the right setting using the paused waveform. I can't get a capture manually at much lower than 1ms

Colin.

Quote:
Originally Posted by SiriusHardware View Post
Just about to head off home, while I'm doing that try switching the time/ div to 10uS, or 5uS - that will increase the sampling rate and although you'll capture fewer data pulses they should look better. Have another look at D0 through D7, still with CH1 triggering from UD6 chip select.
Attached Thumbnails
Click image for larger version

Name:	UD6 20 yellow UD7 9 green 5ns.jpg
Views:	28
Size:	53.9 KB
ID:	281723  
ScottishColin is offline   Reply With Quote
Old 21st Jul 2023, 5:54 pm   #262
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

Well, I don't think it is going to get any better, stick with 5uS or with 10uS, whichever you think gives a better 'picture' of the data.

Now, more boring stuff I'm afraid. Do the same thing at the same settings for all of the fitted ROMs except for the character generator ROM - scope the DO-D7 pins on each ROM in turn, triggering from the falling edge of that ROM's chip select pin, and make sure the data coming out of each data pin of each ROM looks reasonable / similar to what you're seeing coming out of all the other data pins but as before, expect different bit patterns coming out of each data bit of the ROM.
SiriusHardware is online now   Reply With Quote
Old 21st Jul 2023, 5:56 pm   #263
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,252
Default Re: Commodore PET 4032/8050/3022

Looks like 5us not 5ns, through that looks to be the full sweep time.

Sample rate shown is 16MHz. If the scope is quoted at 48MHz sample rate that is possibly only when running single channel. You might be able to get the dual channel sample rate up to 24MHz, maybe try capturing signal at 0.5ms sweep rate before zooming in.

If the scope has an external trigger mode, or you can trigger off channel 1 while only channel 2 is captured and displayed, then you might be able to get the sample rate up to 48MHz.
Mark1960 is online now   Reply With Quote
Old 21st Jul 2023, 6:00 pm   #264
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

It does go to 48MHz at 1uS. The sample rate varies automatically with the time / div setting.

I don't think it does have an external trigger (would be very useful if it did). Would it be usual to be able to trigger from a channel which is de-activated?
SiriusHardware is online now   Reply With Quote
Old 21st Jul 2023, 6:07 pm   #265
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,252
Default Re: Commodore PET 4032/8050/3022

Quote:
Originally Posted by SiriusHardware View Post
It does go to 48MHz at 1uS. The sample rate varies automatically with the time / div setting.

I don't think it does have an external trigger (would be very useful if it did). Would it be usual to be able to trigger from a channel which is de-activated?
I don’t know if it makes a difference on the sample rate if a channel is deactivated on the Hantek, but I think most scope’s I’ve used can still be triggered off a channel that is not actively displayed.
Mark1960 is online now   Reply With Quote
Old 21st Jul 2023, 6:13 pm   #266
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

Could be worth a try?

Quote:
5us not 5ns, through that looks to be the full sweep time.
5uS, yes, that was just a typo but it IS the time / div value and not the full sweep time (look again at the value boxes next to the 'Horizontal' time / div control in the image in #261).
SiriusHardware is online now   Reply With Quote
Old 21st Jul 2023, 6:50 pm   #267
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

Want to try that, Colin? Switch off channel 1, but leave the trigger set to trigger on the falling edge of channel 1 and see if you get a steady lock on the UD6 D0 signal regardless.
SiriusHardware is online now   Reply With Quote
Old 21st Jul 2023, 7:09 pm   #268
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

I'll give that a try. And also try to be more precise in my testing and responses. Apologies. It irritates me when others aren't accurate and I need to pay more attention.

I will be busy tonight so will probably pick this up in the morning.

Colin.

Quote:
Originally Posted by SiriusHardware View Post
Want to try that, Colin? Switch off channel 1, but leave the trigger set to trigger on the falling edge of channel 1 and see if you get a steady lock on the UD6 D0 signal regardless.
ScottishColin is offline   Reply With Quote
Old 21st Jul 2023, 7:43 pm   #269
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

Scroll back not more than 20 posts to find several avoidable errors by me, and they are just the ones I didn't edit in time.

The thinking (Mark's thinking, really) is that when you have two channels enabled, maybe the (up to) 48MHz sample rate is shared between the two active inputs, so, put crudely, instead of sampling one channel at the 48Mhz rate it may sample the two channels alternately at only 24Mhz each.

By switching off one channel we may improve the sampling resolution of the active channel, but the question is whether, with channel 1 switched off, the scope will still continue to look at the CH1 input for trigger activity?

Personally, I suspect that the channel 1 input sampler still has to be working in order to detect the change in level that the trigger circuit is looking for and I doubt that will be the case but I would be glad to be wrong about that.

You may find that as soon as you turn channel 1 off the 'trigger from channel 1' control option becomes greyed out, if so, that would be telling you that you can't use channel 1 as a trigger source when channel 1 is turned off.

Slightly higher end scopes often have an 'EXT TRIGGER' BNC connector and you can connect a third signal to that which, although it won't be shown on the display, can be used to trigger the sweep. This is useful when you want to see what two other signals are doing when the trigger signal activates. If yours had it, you could for example use it to trigger from the chip select signal while the scope showed the data coming out of D0 AND D1 simultaneously. Unfortunately your scope doesn't seem to have an EXT TRIG input.

Last edited by SiriusHardware; 21st Jul 2023 at 7:52 pm.
SiriusHardware is online now   Reply With Quote
Old 22nd Jul 2023, 5:02 pm   #270
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

I can't get anything to perform any differently if I deselect CH1 or undo the probe from the back of the scope and deselect it.

Shall I go back to post 262?

Colin.
ScottishColin is offline   Reply With Quote
Old 22nd Jul 2023, 5:55 pm   #271
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

Yes please Colin - I'm just off out shortly for the evening with SWMBO so I'm afraid I won't be around. To recap, look at all the D0-D7s on all the fitted main ROMs using each ROM's own chip select signal as the trigger for CH1. And make sure that the general level / look of the pulses coming out of each ROM Dx pin is about the same.

Then, power off, remove the NOP gen and check for continuity between the D0 of each fitted ROM and the mainboard CPU socket DA0, and for continuity between the D1 of each fitted ROM to the mainboard CPU socket's DA1, and so on all the way up to D7 / DA7.

The ROM data signals don't pass through any buffers on the way to the CPU, all should be directly connected.
SiriusHardware is online now   Reply With Quote
Old 23rd Jul 2023, 1:14 pm   #272
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

OK - that took me a lot longer than it should have.

All ROMs have repeating waveforms on all the data pins - I can get pictures if required.

All data pins continuity from the 6502 to the ROMs all check out fine.

Colin.
ScottishColin is offline   Reply With Quote
Old 23rd Jul 2023, 1:51 pm   #273
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

OK, you've got address signals going all the way to the ROMs and data signals coming all the way back from the ROMs and the address decoder seems to be working OK so I think we need to try harder to work out whether the CPU is able to run code resident in the ROMs and if not, why not. Let me have a think about that (suggestions from others about how to proceed are always welcome). One thing we haven't looked at is the behaviour of the R/W signal chain, mainly because it stays in the 'Read' direction when we are using the NOP tester.

This machine contains even more of the sensitive electronic fuses also known as 74LS244 buffers and the complex nature of the video circuitry will make it quite hard to analyse their operation particularly because they are used in a bidirectional mode so we have to find ways of observing the data as it passes through them in each direction.

There is one 'one weird trick' we can use to test them all without necessarily replacing them or buying replacements, and that is to remove them in pairs, and temporarily fit them in the UD13 and UD14 positions, fit the NOP generator and observe that we have the expected squarewave outputs on the buffered side of the address bus, because if we do that will prove that every gate in the 74LS244s is working. Of course this would put the onus on you to remove and socket and test all of those buffers one pair at a time, although of course you could stop and retest the machine if you find and replace any more broken buffers during the process. I'll leave that up to you, but you seem so skilled now at removing ICs and sockets that you could always, when all faultfinding is done, remove the sockets and refit the ICs directly to conserve the original appearance.

Last edited by Radio Wrangler; 24th Jul 2023 at 4:48 am. Reason: IC designator error fixed in "One weird trick..."
SiriusHardware is online now   Reply With Quote
Old 23rd Jul 2023, 2:40 pm   #274
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

I'll leave the creative juices simmering for a little while. I'm happy to work on this motherboard - it's a lot easier than the 3016.

In the meantime, I'm going to replace the white UD7 socket as there was a bit of pressin required to get continuity. I'll let you know if anything improves.

Colin.
ScottishColin is offline   Reply With Quote
Old 23rd Jul 2023, 3:43 pm   #275
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

No change after replacing the UD7 white socket, although it fell apart quite easily.

Colin.
ScottishColin is offline   Reply With Quote
Old 23rd Jul 2023, 5:11 pm   #276
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,419
Default Re: Commodore PET 4032/8050/3022

And when you say no change after changing that socket, you mean no change with the machine configured normally.. what about with the Daver2 EPROM fitted in UD7 position? Or did you already try that as well?
SiriusHardware is online now   Reply With Quote
Old 23rd Jul 2023, 5:39 pm   #277
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

Tried every combination. 6502 straight in, Tynemouth board and 6052 with Dave's EPROM fitted.

Colin.
ScottishColin is offline   Reply With Quote
Old 23rd Jul 2023, 5:48 pm   #278
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,751
Default Re: Commodore PET 4032/8050/3022

If I have the Tynemouth board fitted with it using ROM and RAM on itself, what would stop that initialising?

Colin.
ScottishColin is offline   Reply With Quote
Old 23rd Jul 2023, 7:09 pm   #279
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,354
Default Re: Commodore PET 4032/8050/3022

Quote:
Originally Posted by SiriusHardware View Post
OK, you've got address signals going all the way to the ROMs and data signals coming all the way back from the ROMs and the address decoder seems to be working OK so I think we need to try harder to work out whether the CPU is able to run code resident in the ROMs and if not, why not. Let me have a think about that (suggestions from others about how to proceed are always welcome). One thing we haven't looked at is the behaviour of the R/W signal chain, mainly because it stays in the 'Read' direction when we are using the NOP tester.
>>
>>
I've probably forgotten how much has been verified so far / if the NOP-Tester is running OK, but if the NOP-Tester is running OK and a NOP takes two E-cycles (rather than theoretical only one needed, due to design-simplifications in the original 6502 which result in a minimum of two address fetches?) then presumably:

1. Address line A0 should be a squarewave at half the E-Clock (1MHz?) frequency, if a NOP takes 2 fetch / execute cycles, but PC address is only implemented every 2 cycles?
2. Each subsequent increases in address line number, should result in that being half the frequency of the previous one.
3. The Data-lines on the 6502 should always be C9h (D70 = 11001001), whenever it does a read. So can check each of the data-lines is at the required state, whenever 6502 R/nW is High (Presumably always with NOP's) and Sync is High? (To indicate Fetch cycle)

I was a little puzzled how the NOP generator copes with Reset vector. But I presume it causes an initial jump to address C9C9h if that's the only data it outputs, before cycling through the memory space.

I found these webpages useful:
https://wilsonminesco.com/6502primer/MysteryPins.html
https://ist.uwaterloo.ca/~schepers/MJK/pics/6502.gif (Also attached, for reference)
And a simulator, that lets you put C9 etc into memory and run through a simulation what the processor does next:
http://visual6502.org/JSSim/expert.html

Things to check on the 6502, to make sure it runs OK are:
1. Power to it is +5V +/-0.25V, with ground also close to 0.0V relative to a good ground from PSU.
2. Clock input and output pins 37, 39 & 3 all have 1MHz clock on them (especially of output pns are used to feed other parts of the h/w)
3. Reset goes low for a few hundred ms at power-up reset, then remains high
- which I think have mostly been done, but also check:
4. nNMI and nIRQ pins are not stuck low, but mostly high
Attached Thumbnails
Click image for larger version

Name:	6502.jpg
Views:	12
Size:	50.4 KB
ID:	281883  
ortek_service is online now   Reply With Quote
Old 23rd Jul 2023, 7:09 pm   #280
ortek_service
Octode
 
ortek_service's Avatar
 
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,354
Default Re: Commodore PET 4032/8050/3022

Quote:
Originally Posted by SiriusHardware View Post
>>
>>
This machine contains even more of the sensitive electronic fuses also known as 74LS244 buffers and the complex nature of the video circuitry will make it quite hard to analyse their operation particularly because they are used in a bidirectional mode so we have to find ways of observing the data as it passes through them in each direction.

There is one 'one weird trick' we can use to test them all without necessarily replacing them or buying replacements, and that is to remove them in pairs, temporarily fit them in the UB10 and UB9 positions, fit the NOP generator and observe that we have the expected squarewave outputs on the buffered side of the address bus, because if we do that will prove that every gate in the 74LS244s is working. Of course this would put the onus on you to remove and socket and test all of those buffers one pair at a time, although of course you could stop and retest the machine if you find and replace any more broken buffers during the process. I'll leave that up to you, but you seem so skilled now at removing ICs and sockets that you could always, when all faultfinding is done, remove the sockets and refit the ICs directly to conserve the original appearance.

Well if you're going to the trouble of socketing any 74LS244's on this PET, as many had failed on the previous PET, then you could use the 74LS244 sockets in that one to check any removed 74LS244 are OK - Although always a bit of chance of breaking that again, and probably preferable to test them on breadboard (ideally with the assistance of an Arduino-based Logic-IC tester program, to save a lot of playing with links / switches etc) / or a cheap IC Programmer that can tests logic IC's (IIRC the TL866 can, if you could get hold of an old cheap one, or an alternative).

However, testing the 74LS244's in-circuit would save a lot of (de)soldering, but you probably really need at least a 3 channel 'scope, to be able to see that the output follows the input whenever it is enable / goes tri-state when not being enabled. Although you'd probably be able to get a fair idea of these functioning by 'scoping these separately / each In & out pair, without the enable.
ortek_service is online now   Reply With Quote
Reply

Thread Tools



All times are GMT +1. The time now is 1:56 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.