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 Feb 2021, 10:11 pm   #701
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by ScottishColin View Post
OK - so if each 'large' horizontal line is 50ms, then I reckon about 1.35 x 50ms = 67.5ms between falling edges.

And the active CS pulse is about 0.1 * 50ms = 5ms.
Have another look. There are over two divisions between each falling edge, each division = 50mS so that's at least 100mS and more between falling edges. Try again.

The smaller ticks divide the large divisions into fifths, so one subdivision is one fifth of 50mS (=10mS). So how long is the active-low _CE pulse?

As to active low / active high, in the retro computer world chip select signals are usually (but not always) active low, that is they hang around up at logic 1 when inactive and only perform their designated function when taken low, to 0V.

There is a handy way to tell, when looking at circuit diagrams, whether an IC pin or a signal line is active low. If it is active low the pin or line name will have an 'overline' (as opposed to an underline) over the name. The attached close up of UD9 shows two signal or pin names which are active low, and one which is not.

In plain text like this, active-low signals are often indicated by putting an underscore in front of the signal name like this '_CS', or a forward slash, like this, '/CS'.
Attached Thumbnails
Click image for larger version

Name:	ActiveHigh_ActiveLow.jpg
Views:	39
Size:	52.7 KB
ID:	227747  
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 1:15 pm   #702
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

Apologies for the delay - we had a nice day yesterday so went for a walk at The Hermitage...

So - I reckon there's c. 2.6 'major' divisions, so 2.6 * 50ms = 130ms. Am I closer?

And thanks for the other information - it's great for me to keep learning.

Colin.


Quote:
Originally Posted by SiriusHardware View Post
Quote:
Originally Posted by ScottishColin View Post
OK - so if each 'large' horizontal line is 50ms, then I reckon about 1.35 x 50ms = 67.5ms between falling edges.

And the active CS pulse is about 0.1 * 50ms = 5ms.
Have another look. There are over two divisions between each falling edge, each division = 50mS so that's at least 100mS and more between falling edges. Try again.

The smaller ticks divide the large divisions into fifths, so one subdivision is one fifth of 50mS (=10mS). So how long is the active-low _CE pulse?

As to active low / active high, in the retro computer world chip select signals are usually (but not always) active low, that is they hang around up at logic 1 when inactive and only perform their designated function when taken low, to 0V.

There is a handy way to tell, when looking at circuit diagrams, whether an IC pin or a signal line is active low. If it is active low the pin or line name will have an 'overline' (as opposed to an underline) over the name. The attached close up of UD9 shows two signal or pin names which are active low, and one which is not.

In plain text like this, active-low signals are often indicated by putting an underscore in front of the signal name like this '_CS', or a forward slash, like this, '/CS'.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 1:17 pm   #703
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

https://en.wikipedia.org/wiki/Killer_poke#Commodore_PET

Colin.


Quote:
Originally Posted by Mark1960 View Post
Quote:
Originally Posted by SiriusHardware View Post
Mark, are you thinking that PROM _CS is held low for the entire time that the address is within the address range of the PROM? It does look that way - 4 inputs from BA12 to BA15 to the 74154 4-line-to-16 line decoder UD2. No other signal involved in the generation of the chip select. If so this is somewhat reminiscent of the situation with the MK14, you could intentionally damage something by deliberately writing to the PROM address area.
That does look like a possible problem, but when the PET is running NOPs it shouldn’t be a risk as there are no write operations.

I have heard it was possible to damage PETs by poking to the wrong memory addresses, maybe the PROM decoding was responsible for that.

It would also mean that PETs could be damaged if the software crashes, seems like a very poor design, especially with all those unused chip select pins on the proms.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 1:21 pm   #704
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

Just done this - UD9 pin 9 is only getting to 2V as well.

I'll go through one of the PROMs and send the link to the document through this afternoon.

Colin.

Quote:
Originally Posted by Mark1960 View Post
Quote:
Originally Posted by ScottishColin View Post
When you say "Instead of UE9-2 use the pin for the same data line on UD9." does that mean pin 9
Yes, if UE9-2 is D0, check UD9-9.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 2:56 pm   #705
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

OK - here's the complete UD9 test.

https://drive.google.com/file/d/1JMz...ew?usp=sharing

Colin.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 2:56 pm   #706
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

We had a great day here as well yesterday, certainly not a day to be hunched over a scope.

One thing I should have done while I had your PROMs to read them was to scope the data outputs as they were being read in order to see what logic levels they would normally be at, but I was more intent on turning them around quickly so I didn't do that, unfortunately.

However, you could do that if you still have your Arduino Mega reader rigged up - no need to connect the data output lines, just connect the address lines, strap the CS pins into the necessary states (CS1 low, CS3 high), run the reader sketch and while it is reading, scope the data outputs on the PROM to see what data levels would normally come out of the PROMs. You don't have to, it's just something you can try if you have doubts about the logic levels coming out of the PROMs.

Last edited by SiriusHardware; 28th Feb 2021 at 3:02 pm.
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 3:04 pm   #707
Timbucus
Heptode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 860
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by ScottishColin View Post
https://en.wikipedia.org/wiki/Killer_poke#Commodore_PET

Colin.


Quote:
Originally Posted by Mark1960 View Post
Quote:
Originally Posted by SiriusHardware View Post
Mark, are you thinking that PROM _CS is held low for the entire time that the address is within the address range of the PROM? It does look that way - 4 inputs from BA12 to BA15 to the 74154 4-line-to-16 line decoder UD2. No other signal involved in the generation of the chip select. If so this is somewhat reminiscent of the situation with the MK14, you could intentionally damage something by deliberately writing to the PROM address area.
That does look like a possible problem, but when the PET is running NOPs it shouldn’t be a risk as there are no write operations.

I have heard it was possible to damage PETs by poking to the wrong memory addresses, maybe the PROM decoding was responsible for that.

It would also mean that PETs could be damaged if the software crashes, seems like a very poor design, especially with all those unused chip select pins on the proms.

Interesting video from TechTangents - I think I suspect the LS08 output is not just weak but, perhaps has a damaged output circuit - anyway draw your own conclusions...

https://www.youtube.com/watch?v=7bMJ0NIuWU0
Timbucus is online now   Reply With Quote
Old 28th Feb 2021, 3:11 pm   #708
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

Ref: #705, for DA2 and DA3 you didn't catch the edge of the CS pulse on the upper trace so it's hard for us to know exactly what stage of the cycle we are looking at.

DA0 looks decidedly feeble (half of the size of the others) so I'm wondering if a fault on something else on DA0 is keeping that line clamped low.

We'll know whether that is the case if you also look at DA0-DA7 on UD6, UD7, UD8 and find that the DA0 signal is significantly lower on all of those too.
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 4:04 pm   #709
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

With the machine set up the way it is, (NOP test still in place, but everything else still connected as it should be) could you please, with power off, measure the resistances of the DA0 to DA7 lines to 0V and to +5V?
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 4:17 pm   #710
Mark1960
Hexode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 410
Default Re: Non-working Commodore PET 3016

For UD9 it might be worth trying to measure DA2 and DA3 again, just in case of a bad scope probe connection. I wouldn’t expect to see no activity on those data pins.

DA0 could possibly try adding 4.7k pull up to 5v. That might show if its a weak output instead of something pulling it down. Try measuring DA0 from the other proms first.
Mark1960 is offline   Reply With Quote
Old 28th Feb 2021, 4:35 pm   #711
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

Quote:
Originally Posted by ScottishColin View Post

So - I reckon there's c. 2.6 'major' divisions, so 2.6 * 50ms = 130ms. Am I closer?
Yep. Actually a smidge more than 130mS, call it 131mS.

Now you've got the hang of measuring voltage and time, you can try your hand at another measurement which is traditionally possible with scopes - frequency measurement. I know your scope actually indicates the frequency of the signal which is a nice feature, but you can have a stab at it anyway.

The chip select signal is repetitive, it does the same thing over and over again so it can be said to have a steady frequency. 'Frequency' is the number of repetitions, or 'cycles', in one second.

You are working / measuring in milliseconds so let's stick with those units. There are a thousand milliseconds in one second.

You've already established how long one waveform cycle is - 131mS - by measuring from the start of one waveform cycle to the same point on the next waveform cycle (falling edge to falling edge).

Armed with this information, and given that there are 1000 milliseconds in a second, can you say how many cycles there are in one second? Whatever answer you get, you can check by using the frequency measurement feature of your scope to see how close you were.
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 4:40 pm   #712
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

Here's a re-do of DA2 and DA3 of UD9.

https://drive.google.com/file/d/1Pwv...ew?usp=sharing

I'll go to UD8-6 next.

Colin.


Quote:
Originally Posted by SiriusHardware View Post
Ref: #705, for DA2 and DA3 you didn't catch the edge of the CS pulse on the upper trace so it's hard for us to know exactly what stage of the cycle we are looking at.

DA0 looks decidedly feeble (half of the size of the others) so I'm wondering if a fault on something else on DA0 is keeping that line clamped low.

We'll know whether that is the case if you also look at DA0-DA7 on UD6, UD7, UD8 and find that the DA0 signal is significantly lower on all of those too.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 4:42 pm   #713
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

This seems too easy - 1000 / 131 = 7.633

Colin.


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

So - I reckon there's c. 2.6 'major' divisions, so 2.6 * 50ms = 130ms. Am I closer?
Yep. Actually a smidge more than 130mS, call it 131mS.

Now you've got the hang of measuring voltage and time, you can try your hand at another measurement which is traditionally possible with scopes - frequency measurement. I know your scope actually indicates the frequency of the signal which is a nice feature, but you can have a stab at it anyway.

The chip select signal is repetitive, it does the same thing over and over again so it can be said to have a steady frequency. 'Frequency' is the number of repetitions, or 'cycles', in one second.

You are working / measuring in milliseconds so let's stick with those units. There are a thousand milliseconds in one second.

You've already established how long one waveform cycle is - 131mS - by measuring from the start of one waveform cycle to the same point on the next waveform cycle (falling edge to falling edge).

Armed with this information, and given that there are 1000 milliseconds in a second, can you say how many cycles there are in one second? Whatever answer you get, you can check by using the frequency measurement feature of your scope to see how close you were.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 5:11 pm   #714
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

Quote:
This seems too easy - 1000 / 131 = 7.633
Well, as we now know the UDx chip select pins are being activated once per complete run through the address range.

Logically, we know that the A15 line is also running through one complete cycle on each run through the address range because it is low for the first 32K addresses and high for the second 32K addresses, so the frequency on A15 should be the same as your calculated figure for the _CS signal.

Let's see what frequency you measured on the A15 line after you got the NOP test running....

Quote:
A15 25 7.632Hz
Pretty close.

The small difference is due to the fact that the length of a cycle won't be exactly 131.000mS, but it is difficult to read the length of a pulse off a scope screen with much greater accuracy than that.

So now you know how to (roughly) measure frequency using a scope. Unfortunately this marvellous ability will be somewhat wasted since your scope is so posh that it just tells you the frequency of the signal anyway. Most scopes, certainly most CRT scopes, don't have that feature so it is still worth knowing how to do it.
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 5:12 pm   #715
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

Meter set to 2k

0V taken from ground screw

+5v taken from pin24 of UD9

Code:
                OV           +5V
DA0          0.502       0.403
DA1          0.502       0.409
DA2          Varied      0.2
DA3          0.508       0.412
DA4          0.511       0.413
DA5          0.523       0.413
DA6          0.509       0.412
DA7          0.508       0.410

DA2 went all over the place on the 0V check from 0 to 0.45 to 0.508 - I could not get a steady reading from it.

Colin.

Quote:
Originally Posted by SiriusHardware View Post
With the machine set up the way it is, (NOP test still in place, but everything else still connected as it should be) could you please, with power off, measure the resistances of the DA0 to DA7 lines to 0V and to +5V?
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 5:30 pm   #716
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

Well, that's certainly interesting as it looks as though there is a distinct difference between DA2 and the rest, which is also borne out by your scope observations as well.

Remove all of UD6-UD9, re-measure DA2 resistance to +5V. If that 200R (0.2K) resistance is still there, remove all of UC5, UC6, UC7 and see if it is still there. Note removing devices will probably change all of the resistances, so what you are looking for is the moment when DA2 resistance to +5V is no longer different to DA0-1 and DA3-7 resistance to +5V.

If there is still a difference between DA2 and the other data lines with all of those devices removed, that will point the finger back at the data bus buffers, specifically UE9 which is on the DA2 line.
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 6:41 pm   #717
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

Does heat affect resistance?

I ask as I have just checked again before I removed any ICs and all resistance readings on all pins are c 1.13 at the 20k setting on my meter. the last set of readings were just after I had switched off the PET which had been on for quite some time.

the long and the short of it is that I cannot get Pin 11 UD9 to mis-behave at all now with ICs still inserted.

Colin.


Quote:
Originally Posted by SiriusHardware View Post
Well, that's certainly interesting as it looks as though there is a distinct difference between DA2 and the rest, which is also borne out by your scope observations as well.

Remove all of UD6-UD9, re-measure DA2 resistance to +5V. If that 200R (0.2K) resistance is still there, remove all of UC5, UC6, UC7 and see if it is still there. Note removing devices will probably change all of the resistances, so what you are looking for is the moment when DA2 resistance to +5V is no longer different to DA0-1 and DA3-7 resistance to +5V.

If there is still a difference between DA2 and the other data lines with all of those devices removed, that will point the finger back at the data bus buffers, specifically UE9 which is on the DA2 line.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 7:19 pm   #718
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

There certainly are such things as thermal faults.

Keep it simple for now: If the resistance difference on DA2 has disappeared, what do you now get when you scope DA2? Still no signal, or do you have one now? If you do, is the signal on DA0 still small compared to the others?
SiriusHardware is online now   Reply With Quote
Old 28th Feb 2021, 7:57 pm   #719
ScottishColin
Heptode
 
Join Date: May 2012
Location: Perth, Scotland
Posts: 573
Default Re: Non-working Commodore PET 3016

OK. Meter set to 20k

Red probe on earth screw for 0V readings.

Red probe on pin 24 UD9 for +5V readings


Code:
                        OV           +5V
DA0 (pin9 UD9)           1.26       1.10
DA1 (pin10 UD9)          1.29      1.12
DA2 (pin11 UD9)          1.29      1.13
DA3 (pin13 UD9)          1.30      1.14
DA4 (pin14 UD9)          1.30      1.14
DA5 (pin15 UD9)          1.31      1.14
DA6 (pin16 UD9)          1.30      1.14
DA7 (pin17 UD9)          1.29      1.13
Colin.
ScottishColin is offline   Reply With Quote
Old 28th Feb 2021, 8:02 pm   #720
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 6,309
Default Re: Non-working Commodore PET 3016

Resistances certainly look more uniform now although DA0 is a tad lower than the rest in both directiions.

What about DA0 and DA2 compared to the others on scope?
DA0 still half the size of the others?
DA2 still flatlined?
SiriusHardware is online now   Reply With Quote
Reply

Thread Tools



All times are GMT. The time now is 4:34 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 - 2021, vBulletin Solutions, Inc.
Copyright ©2002 - 2021, Paul Stenning.