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 20th Jul 2023, 6:20 pm   #241
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

OK - resolved now. Waveforms all available from UE12 and they now all flow through to all available ROM chip pin 20 including UD10.

Colin.
ScottishColin is offline   Reply With Quote
Old 20th Jul 2023, 6:43 pm   #242
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

OK. With hindsight then, it looks as though the original UE12 may be OK. I'm glad we decided to do that mundane check-up on the chip enables at the chips otherwise we would have been proceeding for a long time on the assumption that the address decoder was fixed and tested and it would have taken a while to come back around to that area.

Next, still with the NOP generator in place, just try scoping the data lines at the mainboard CPU socket (not on the CPU, because its pins are disconnected from the mainboard with the NOP gen in) or, if it is easier to get to them there, look on UB10 pins 2, 4, 6, 8 and UB9 pins 2, 4, 6, 8.

All we are looking for in this first instance is for any of the signals to look distorted or reduced-height compared to the others. The data itself will just be a stream of random bits, and there may also be periodic 'dead periods' where the currently selected address is not enabling any device, such as when it passes through the address ranges reserved for the empty EPROM sockets.

Do the same on the other side of the data bus buffers, so, UB10 pins 18, 14, 16, 12 and UB9 pins 18, 16, 14, 12.
SiriusHardware is offline   Reply With Quote
Old 20th Jul 2023, 7:28 pm   #243
ajgriff
Nonode
 
ajgriff's Avatar
 
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,576
Default Re: Commodore PET 4032/8050/3022

Quote:
Originally Posted by SiriusHardware View Post
I've just realised one thing though, I have always wondered why the ICs have different second letters A, B, C, D, E, etc. Looking at the layout diagram I've realised that all of the 'A' devices are in the first upper row of chips going across, all the 'B' devices are in the next row down, all the 'C' devices are in the middle row, all the 'D' devices are in the next row down and all the 'E' devices are in the lowest row.

At the same time the numbers in the part reference number ascend as you go from left to right, so it's essentially a way of telling you roughly whereabouts on the PCB the chip is - I've never realised that before. '
What's more the boards themselves are printed with the letters running up the sides and numbers along the bottom edges such that each component has a grid reference.

Alan
ajgriff is offline   Reply With Quote
Old 20th Jul 2023, 7:33 pm   #244
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

..And I didn't know that because I have never ever seen one.
SiriusHardware is offline   Reply With Quote
Old 20th Jul 2023, 10:14 pm   #245
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

Well they're there but they're not pretty.

I attach a screenshot of 6502 pin 39 for reference, then UB10 pin 2 to compare. All the UB10 and UB9 pins below look like this.

I also attach a photo as there seem to be to lines visible to the naked eye on the scope screen - but only one is visible when I pause the scope. I guess it's possible that it's a limitation of the hardware or software but I've not seen this before.

Colin.

Quote:
Originally Posted by SiriusHardware View Post
OK. With hindsight then, it looks as though the original UE12 may be OK. I'm glad we decided to do that mundane check-up on the chip enables at the chips otherwise we would have been proceeding for a long time on the assumption that the address decoder was fixed and tested and it would have taken a while to come back around to that area.

Next, still with the NOP generator in place, just try scoping the data lines at the mainboard CPU socket (not on the CPU, because its pins are disconnected from the mainboard with the NOP gen in) or, if it is easier to get to them there, look on UB10 pins 2, 4, 6, 8 and UB9 pins 2, 4, 6, 8.

All we are looking for in this first instance is for any of the signals to look distorted or reduced-height compared to the others. The data itself will just be a stream of random bits, and there may also be periodic 'dead periods' where the currently selected address is not enabling any device, such as when it passes through the address ranges reserved for the empty EPROM sockets.

Do the same on the other side of the data bus buffers, so, UB10 pins 18, 14, 16, 12 and UB9 pins 18, 16, 14, 12.
Attached Thumbnails
Click image for larger version

Name:	6502 pin 39.jpg
Views:	21
Size:	54.6 KB
ID:	281687   Click image for larger version

Name:	UB10 pin 2.jpg
Views:	27
Size:	50.2 KB
ID:	281688   Click image for larger version

Name:	PXL_20230720_210734514.jpg
Views:	21
Size:	59.2 KB
ID:	281689  
ScottishColin is offline   Reply With Quote
Old 20th Jul 2023, 10:41 pm   #246
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

<Mind Blown>

No, that does not look like what I was expecting. Same on both sides of the buffers, really?

Put your scope into dual channel mode, make sure the settings for both channels (Volts/Div, Probe multiplier setting on probe, probe multiplier setting on scope) match, and put the channel 1 probe on UD7's chip select pin (20), set the trigger for channel 1, falling edge trigger.

Put the channel 2 probe onto UD7's D0 pin (9).

Adjust the time / div until you are capturing one complete low pulse on channel 1 and whatever is happening on UD7 DO (Channel 2) during that same period and show me what that looks like.
SiriusHardware is offline   Reply With Quote
Old 20th Jul 2023, 11:06 pm   #247
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Sorry, I'm going to have to knock off for tonight. Had an exhaustipating day.

I'll check back in tomorrow.
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 9:59 am   #248
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

I suspect you're not going to like this.

Colin.

P.S. No apologies required.
Attached Thumbnails
Click image for larger version

Name:	UD7 pin 20 yellow UD7 pin 9 green.jpg
Views:	28
Size:	61.5 KB
ID:	281694  
ScottishColin is offline   Reply With Quote
Old 21st Jul 2023, 10:17 am   #249
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

I'm not tremendously unhappy about that.

What you are seeing there, coinciding with the low chip select pulse, is about 1K worth of D0 bits all packed closely together so it just looks like a solid wall of noise. I'm not quite sure why you are getting 2 x chip select pulses in quick succession but the address decoder for UD7 is slightly more complicated than it is for most of the others.

Could you do that again, same settings, but this time with UD6 chip select (20) on channel 1 and and D0 (9) on channel 2. Again aiming to get only one low-going chip select pulse across the width of the screen (Maybe turn the time / div up from 2.000 to 1.000 mS).
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 11:56 am   #250
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

How's this?

Colin.


Quote:
Originally Posted by SiriusHardware View Post
I'm not tremendously unhappy about that.

What you are seeing there, coinciding with the low chip select pulse, is about 1K worth of D0 bits all packed closely together so it just looks like a solid wall of noise. I'm not quite sure why you are getting 2 x chip select pulses in quick succession but the address decoder for UD7 is slightly more complicated than it is for most of the others.

Could you do that again, same settings, but this time with UD6 chip select (20) on channel 1 and and D0 (9) on channel 2. Again aiming to get only one low-going chip select pulse across the width of the screen (Maybe turn the time / div up from 2.000 to 1.000 mS).
Attached Thumbnails
Click image for larger version

Name:	UD6 20 yellow UD7 9 green.jpg
Views:	20
Size:	56.7 KB
ID:	281698  
ScottishColin is offline   Reply With Quote
Old 21st Jul 2023, 12:23 pm   #251
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Not bad, although you've gone to 5.000mS rather than to 1.000mS, making the portion of the capture occupied by the chip select shorter rather than longer, as I had hoped. Doesn't matter, let's examine the D0 data in finer detail now.

Staying with those connection points, gradually decrease the time / div value (to 500uS / 200uS, / 100uS, etc). The rising edge of the chip select pulse on the upper trace will disappear off somewhere beyond the RH edge of the display but the falling edge should remain on-screen. Reduce the time / div time value to the point where that thick mess of D0 data on the lower trace separates out into individual pulses, and take a shot of that.

What we are aiming to do now is just to view the first 12 / 16 / 20 or so D0 data bits coming out of the UD6 ROM immediately after the ROM is selected.
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 12:41 pm   #252
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Try to keep the channel 1 and channel 2 settings, like volts / div and scaling settings, matched please. They were OK in the shot in #248 but have somehow become mangled again in the shot in #250, where channel 1 is set to 50V / division and channel 2 is set to 2.0V/ division.
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 1:19 pm   #253
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

Apologies.

How's this at 20ns?

Colin.


Quote:
Originally Posted by SiriusHardware View Post
Try to keep the channel 1 and channel 2 settings, like volts / div and scaling settings, matched please. They were OK in the shot in #248 but have somehow become mangled again in the shot in #250, where channel 1 is set to 50V / division and channel 2 is set to 2.0V/ division.
Attached Thumbnails
Click image for larger version

Name:	UD6 20 yellow UD7 9 green 20ns.jpg
Views:	34
Size:	63.6 KB
ID:	281701  
ScottishColin is offline   Reply With Quote
Old 21st Jul 2023, 1:36 pm   #254
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Now your channel 1 is set to 2KV/ Div and your probe declared to be set to X1000, so the size of the chip select pulse is being displayed as 5.27KV, which is unlikely.

While you are using dual-channel mode please set both channels=: 2V / Div, DC, X10, and the switches on the probes set to X10. Try to keep them that way.

That rising curve on the low section of the upper trace looks odd - you haven't lost your 0V/GND connection between the scope and the PET again, have you?

Now, could you look at the other UD6 data pins in ascending order, still with channel 1 on UD6 chip select, and keeping the time / div at 20nS - no need to post images, but what you are looking for is for all of those Dx pins to look about the same, especially all the same height - the exact pattern of bits coming out of the pin after the falling edge of the chip select signal will differ for each Dx pin. If you find any that are reduced-height or look more distorted than the D0 example, report back.
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 3:22 pm   #255
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Quote:
keeping the time / div at 20nS
Uh, I meant 20uS, of course.
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 3:28 pm   #256
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

I can't be trusted with anything sharper than a butter knife. Apologies.

Here it is again hopefully right this time.

Earth is still there and well connected.

I'll try the other measurements now.

Colin.
Attached Thumbnails
Click image for larger version

Name:	UD6 20 yellow UD7 9 green 20ns.jpg
Views:	22
Size:	60.0 KB
ID:	281719  
ScottishColin is offline   Reply With Quote
Old 21st Jul 2023, 3:37 pm   #257
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Not sure If I've caught you in time but if you haven't started yet, change the time / div to 10uS or 5uS. You'll capture only a half or a quarter as many data pulses but they should look a bit cleaner / squarer due to the increased sampling rate.

If I got that in too late, don't worry about it.
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 3:40 pm   #258
ScottishColin
Octode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,740
Default Re: Commodore PET 4032/8050/3022

Done on all UD6 data pins. Whilst the waves are different to each other, none are obviously 'wrong' or distorted.

Colin.
ScottishColin is offline   Reply With Quote
Old 21st Jul 2023, 3:56 pm   #259
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

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.
SiriusHardware is offline   Reply With Quote
Old 21st Jul 2023, 4:59 pm   #260
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,380
Default Re: Commodore PET 4032/8050/3022

Home. Speaking of sample rates (as I was earlier) I got to wondering whether it would make a difference whether you plugged the Hantek box into a USB 2 or USB 3 port - would plugging it specifically into a USB 3 port let it use a higher sampling rate, I was thinking?

The background to this was my wondering how this constant need to keep an eye on the sample rate never came up before, so I thought possibly you might have had it plugged into a higher speed USB port the first time around.

I've just read the manual and it states a maximum sampling rate of 48MHz which is the sample rate it currently selects for Time / Div = 1uS, so I guess it's already going as fast as it can.
SiriusHardware is offline   Reply With Quote
Reply

Thread Tools



All times are GMT +1. The time now is 6:18 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.