18th Mar 2021, 3:12 pm | #1141 | ||
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
OK - device fitted. Main difference is that the screen is full of characters, and does not clear.
UI10/11 pin 1 - go high and stay there. Checked at 4.30V with a meter. RAM chips - pin 4 has no reading (ie it stays at the same level with and without the scope connected. I checked the voltage at 0.40V on all RAM chips. Pin 15 goes high and stays there - again checked with a meter at 4.30V constant. Scope readings to follow. Colin. Quote:
|
||
18th Mar 2021, 3:26 pm | #1142 | |||
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
OK - scope readings attached - see last page as well.
I'm getting a flickering between the square waves and a roughly 1V reading. I've tried to attach a video so you can see what I can't capture with a screenshot. https://drive.google.com/file/d/1LSZ...ew?usp=sharing ******************************JdbGZ89z5evNXo9z7 Colin. Quote:
|
|||
18th Mar 2021, 3:28 pm | #1143 | |
Nonode
Join Date: May 2004
Location: Halifax, West Yorkshire, UK.
Posts: 2,583
|
Re: Non-working Commodore PET 3016
Quote:
1. The PET might be getting past the character screen stage of the boot process without the PIAs in place. 2. At least one of the PIAs is probably faulty. For what it's worth both 6520 PIAs were faulty in my PET. Failure of PIA2 resulted in the random character screen and PIA1 caused a problem with the internal datasette port. I wouldn't do anything about the PIAs at the moment. Leave that until progress has been made with the RAM issues. Ultimately if one or both PIAs needs replacing there are reasonably priced alternatives to the original 6520s. Perhaps best to cross that bridge when we come to it. Alan |
|
18th Mar 2021, 4:07 pm | #1144 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
Quote:
I expected you would get a screen full of random characters with the NOP device in because instead of executing the PROM code (which includes a routine to clear the screen) the CPU is just executing continual NOP instructions, so never clears the screen RAM. I'll have to wait until this evening before taking a look at the scope traces - duty calls. |
|
18th Mar 2021, 4:50 pm | #1145 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
|
Re: Non-working Commodore PET 3016
In post #1142 you are listing pairs of pin numbers, but only one trace is active in each scope capture. Was it the first or second pin that was on the active trace?
With the NOP generator connected the 6502 is scanning through all 64 K addresses, but the ram is only active during the first 16 K of those, so it does look correct that there would be some flicking of the signal between active and inactive. |
18th Mar 2021, 5:03 pm | #1146 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
|
Re: Non-working Commodore PET 3016
Try to trace /RAS0 back to its source, look for activity on UG1-12 and UG1-13 with the scope.
Also check activity on UH1-3 and UH1-11. |
18th Mar 2021, 5:14 pm | #1147 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
|
Re: Non-working Commodore PET 3016
Tracing the /CAS signals back, check for activity on UH4-12.
Both /CAS and /RAS depend on UH4 so this might be a common fault for both problems. |
18th Mar 2021, 5:30 pm | #1148 | ||
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
Quote:
Quote:
Last edited by SiriusHardware; 18th Mar 2021 at 5:49 pm. |
||
18th Mar 2021, 5:58 pm | #1149 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
These readings are all with the NOP device in place.
UG1-12 - no movement on the scope. UG1-13 - goes high (5V) and stays there) UH1-3 - no movement on scope UH1-11 - nice square waves attached. Colin. |
18th Mar 2021, 6:00 pm | #1150 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
|
18th Mar 2021, 6:04 pm | #1151 |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
All - I have not done anything with the idea in #1132, mainly because I don't know how....
Happy to try it but I'll need more detail I'm afraid. Colin. |
18th Mar 2021, 6:09 pm | #1152 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
I am still most concerned by the lack of any discernible activity on UI10 / UI11 pin 1, without which no data can pass from the system RAM to the buffered data bus.
In addition to Mark's suggestions to chase down the lack of activity on the RAS and CAS lines, could we also, with the NOP test now in place, revisit pins 9 and 11 of UG7 both at the same time so we can see the relationship between the two signals going to those points. I am particularly interested in knowing whether you ever see both of those signals high (~4-5V) at the same time. |
18th Mar 2021, 6:31 pm | #1153 | |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
|
Re: Non-working Commodore PET 3016
Quote:
Lack of outputs from UH4 is common to both, but might be a problem with its inputs. UH5-1 should be 1MHz square wave. UH5-2 should have some activity or held high. UH5-3 should be 1MHz when UH5-2 is high. |
|
18th Mar 2021, 6:40 pm | #1154 | |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
Pin 9 goes high (5v). Pin 11 shifts slightly, but no waves (checked with meter at 0.4V).
Colin. Quote:
|
|
18th Mar 2021, 6:40 pm | #1155 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
I think the BANK SEL signal should be active-high whenever an address in the system RAM address range is being read from or written to, this signal being decoded from the high address lines by one section of UB2, 7425 (input pins 9,10,12,13, output pin 8).
This BANK SEL output is in turn taken to one of the inputs of UG7, 74LS10, (inputs: 9,10,11, output: 8) which combines the BANK SEL signal with the buffered read / write signal 'B R/W' so that read pulses for the system RAM only occur during a read from the system RAM address block. We aren't getting the read enable pulses for the system RAM out of UG7. I note also that the BANK SEL signal feeds back to the main timing chain on sheet 6, and interestingly, it is also one of the initial inputs to the RAS/ CAS generation circuit. BANK SEL and BØ2 (Buffered Φ2) go into UH5, 74LS00, (pins 2,1), output from UH5 pin 3 goes into pin 15 of UH4 which is at the heart of the RAS / CAS generator circuit - so a lack of BANK SEL signal could disable all of this as well as the system RAM read signal. So the first question arising from all of that is: Is there any activity at all on the BANK SEL signal line, UB2 pin 8? - this is with the NOP test still running. Edit: And Colin has just answered that question (pin 11 of UG7 is the same point as UB2 pin 8) - so we don't have a BANK SEL signal. This is a promising lead. You know the jumpers on the mainboard, do they look as they have ever been tampered with or altered from their original settings? My understanding is that they are metal links which were either cut or left linked as required. Last edited by SiriusHardware; 18th Mar 2021 at 6:47 pm. |
18th Mar 2021, 6:43 pm | #1156 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
|
Re: Non-working Commodore PET 3016
If BANK_SEL is always low that would explain both lack of /RAS and /CAS and also lack of signal at UI10-11 and UI11–11.
But that might be jumping ahead a little, lets see what Colin finds. Edit: crossed post with Sirius. |
18th Mar 2021, 6:45 pm | #1157 | ||
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
UH5-1 confirmed 1Mhz waves (screenshot attached).
UH5-2 - no movement off baseline UH5-3 goes high to a measured 4V. Colin. Quote:
|
||
18th Mar 2021, 6:47 pm | #1158 | |
Octode
Join Date: May 2012
Location: Perth, Scotland
Posts: 1,762
|
Re: Non-working Commodore PET 3016
UB2-8 screenshot attached.
Colin. Quote:
|
|
18th Mar 2021, 6:49 pm | #1159 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,264
|
Re: Non-working Commodore PET 3016
For the jumpers on 16K pet, I think M, I and J should be fitted. L and H not fitted.
|
18th Mar 2021, 6:55 pm | #1160 | |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,482
|
Re: Non-working Commodore PET 3016
Quote:
UB2 pin 8 and UG7 pin 11 and UB2 pin 8 and UH5 pin 2 |
|