# Thread: PET 2001-32N - Diag works in UD9, but with Kernal it fails to start

1. Senior Member
Join Date
Oct 2012
Location
New York
Posts
504
Originally Posted by daver2
First things first.

The address lines are OK going into UB3 and UC3? Yes?

What is most likely happening is that when one address line is LOW it is dragging the other address line down with it. You should be able to work out which address line it is interacting with by looking at the frequency of how often the 'funny' bit occurs. Measure the time from when the 'funny' bit starts to occur to when it starts to occur again and relate that to a frequency (f=1/t). This frequency should equate to one of the other address lines.

This is the first step.

Dave
AH HAH! It's actually pin 3 on UB3 that is interfering with pin 7 on UC3. On the chip side of the 244s everything is fine. The problem is only on the other side of the buffers.

I'll poke around a bit to see if those traces ever come close to one another or what else they both enter while you think about it more.

2. >>> Wow, that's so obvious and so clever, I'd never have thought of it. I'll do some math and work it out. I do wonder, though, why it's not also pulling the same lines high.

I try. Very trying according to my wife...

The way TTL logic works is that never happens!

There is an effective internal resistor to +5V within the output stage of the IC and a transistor pulls the level down to 0V. If you put a short circuit (say) between two TTL outputs, the only time you will see a logic HIGH is when both output transistors of both gate outputs are OFF. If either transistor is ON, the output voltage will always be low.

This is why you can get away by pulling a TTL signal low (with a piece of wire for debugging), but if you connect the output of a TTL gate directly to VCC (and the output transistor is ON) that will probably be the last thing you do with that gate or IC as there is now no current limiting resistor to prevent the transistor from going pop and joining the great hoard of damaged ICs in the sky .

Dave

3. So, we are talking about BA5 (UC3/7) and BA9 (UB3/3).

I can't see any TTL logic where both of these address lines come into contact (via a single IC that is).

Did you say that you have physically removed all of the ROMS?

The only other option (which you are now chasing down) is a solder bridge or contamination somewhere on the PCB.

Dave

4. Senior Member
Join Date
Oct 2012
Location
New York
Posts
504
Originally Posted by daver2
So, we are talking about BA5 (UC3/7) and BA9 (UB3/3).

I can't see any TTL logic where both of these address lines come into contact (via a single IC that is).

Did you say that you have physically removed all of the ROMS?

The only other option (which you are now chasing down) is a solder bridge or contamination somewhere on the PCB.

Dave
Yes, physically removed. I've followed the traces all over the top side of the board; they do run adjacent, but I can't find any bridges anywhere. I'm tempted to get out this "Leak Seeker" and try it out, though I've never had any success with it. I read 15-20 ohms between pin 3 and pin 7 on those two chips.

5. Senior Member
Join Date
Oct 2012
Location
New York
Posts
504
Originally Posted by dhoelzer
Yes, physically removed. I've followed the traces all over the top side of the board; they do run adjacent, but I can't find any bridges anywhere. I'm tempted to get out this "Leak Seeker" and try it out, though I've never had any success with it. I read 15-20 ohms between pin 3 and pin 7 on those two chips.
Oh my GOD!

Well, Dave, you told me so. Despite all of my careful looking and cleaning, there is a via that passes from the top to the bottom of the board next to pin 3 of UD4. There was, apparently, FLUX bridging them under the EPROM socket. The leak seeker didn't really help much.. I just used a really sensitive old-school needle style ohmmeter to figure out where the resistance was lowest and started looking carefully... which lead the the via, which had the lowest resistance at 18.7 ohms.

At this point I have your tests running. I'll resocket the PIA/CIAs and let it run for a while to see if anything else crops up. Fingers crossed!

6. Senior Member
Join Date
Oct 2012
Location
New York
Posts
504
Originally Posted by dhoelzer
Oh my GOD!

Well, Dave, you told me so. Despite all of my careful looking and cleaning, there is a via that passes from the top to the bottom of the board next to pin 3 of UD4. There was, apparently, FLUX bridging them under the EPROM socket. The leak seeker didn't really help much.. I just used a really sensitive old-school needle style ohmmeter to figure out where the resistance was lowest and started looking carefully... which lead the the via, which had the lowest resistance at 18.7 ohms.

At this point I have your tests running. I'll resocket the PIA/CIAs and let it run for a while to see if anything else crops up. Fingers crossed!
Ah, well, nope. Good for about 2 minutes, now they are coupling again. I think I'll just be done with it and remove all of the eprom sockets. I don't know who put them on... was it me? I don't remember, I've had this thing too many years... Let me pull all of the sockets and give those neighboring traces a good cleaning... I hesitate to share that the thought just crossed my mind to cut the traces where they enter the UD section, but that's just crazy talk.

7. Senior Member
Join Date
Oct 2012
Location
New York
Posts
504
Originally Posted by dhoelzer
Ah, well, nope. Good for about 2 minutes, now they are coupling again. I think I'll just be done with it and remove all of the eprom sockets. I don't know who put them on... was it me? I don't remember, I've had this thing too many years... Let me pull all of the sockets and give those neighboring traces a good cleaning... I hesitate to share that the thought just crossed my mind to cut the traces where they enter the UD section, but that's just crazy talk.
Sorry for so many posts. All sockets removed (except the top two, since I use them both with testing)...

I'm now certain that I'm not the one who put in those sockets... what makes me so sure is the distinctive, acrid smell of acid-based flux/acid-core solder. Someone used acid core solder on this stinkin board! No wonder there are all of these weird transient signals!

Anyway, sockets removed, traces thoroughly cleaned... it's now sitting burning in with the NOP generator watching the two address lines for signs of trouble.

UPDATE: It's much improved, but the TIM behavior coupled with Dave's EDIT ROM diags not running persists. Given this discovery, however, and not knowing/remembering who put the rest of the sockets in under the PIA, CIA, and CPU, I think I will be best served to remove those sockets and give the traces the same treatment since there was a coupling behavior visible there as well.
Last edited by dhoelzer; October 29th, 2019 at 12:08 PM.

8. I think you have it on the run...

I came across a tiny bit of solder once just lodged under an IC socket. It was the 'metal mouse'! As it moved about under the socket we got all sorts of strange behaviour. And the behaviour changed as well depending upon which pins it decided to short out !

Dave

9. Senior Member
Join Date
Oct 2012
Location
New York
Posts
504
Originally Posted by daver2
I think you have it on the run...

I came across a tiny bit of solder once just lodged under an IC socket. It was the 'metal mouse'! As it moved about under the socket we got all sorts of strange behaviour. And the behaviour changed as well depending upon which pins it decided to short out !

Dave
Just an update for today. I removed the rest of the sockets that had been installed by a former owner (the UE chips were mostly socketed)... Pin 17 on the PIA is now rock solid after 30 minutes with no signs of any coupling. I've been looking at a Basic prompt with a normal cursor for the same amount of time.

Unfortunately, I can't test further because I'm still waiting for the replacement 74145 for UC9 to arrive and the keyboard will not work without it. Otherwise, it looks like I'm about ready to start testing the cassette port and then the IEE488.

Dave: For some reason, it seems that your DIAG won't start.. I get a screen of PETSCII chart but nothing changes after that. Any thoughts?

10. Ah, so if you get a screen of PETSCII characters and it doesn’t change, are the PETSCII characters that are displayed actually correct as per the screenshots in my documentation?

My tester writes ascending characters into the video RAM and then performs a read back and compare. If that is going wrong then it will ‘jam’ at that test.

You should see the screen ram (8xxx) and the Exxx ROM being constantly accessed and nothing else.

It may mean you have a video RAM read problem if the characters look OK on the screen.

Presumably, then you had some low resistance ‘stuff’ between the two address lines somewhere. Did you ever find anything?

Dave