Chicklet pet displays single line only

[Thomas_]

Dear all,

I am (yet another) low-bit enthusiast living in Germany who recently dug out a Commodore Pet 2001 which I always wanted to repair and now the time has come. I'm not overly deep into electronics, though interested.

As you can see on the picture, the symptom is that there is no vertical deflection. The entire picture falls into a single line. Examined closely and with a low intensity setting of the screen, the line rather appears to be slightly dashed, so I would suppose it reflects a character pattern. So it must either be the screen or the board. Or both.

So rather than buying three working Pets I decided to buy one oscilloscope and track down the error by unplugging the video connection and measure the pins. The lower signal on the picture shows the horizontal sync at about 15.62kHz and 1.5V.

That signal can be used as a trigger for the video signal which can be seen on the upper half. The video signal is at about 5V, non-periodic. The oscilloscope was not quite sure about the frequency but it jittered around 1-3MHz which might well be the pixels.

The vsync pins yields a 4.5V DC signal actually with some 15.62 jitter around the 4.5V but I would supposed this is only line cross talk (cross over? - whatever you say).

Since I am doing that for the first time, did I make any false assumptions and if not, how do I get closer to the source of the problem.

Another thing: At the lower left of the board there are some coarsely replaced diodes. The uppermost points into another direction as compared to the lower two. Is this correct and on which circuit diagram can I find the layout of my board?

Hoping for an interesting discussion,

cheers,

Thomas

 

[dabone]

You have a monitor issue.
So look at repairing the monitor board first. Check for any loose connections / bad solder joints on the monitor.


Later,
dabone

 

[dave_m]

The vsync pins yields a 4.5V DC signal

Thomas,
If the Vertical Drive signal is not giving you a skinny negative going pulse every 16.7 mS (about 60 Hz frequency), that would be the place to start looking on the main board. There are many different boards with associated schematics. Provide us the part number on the board such as 32008, 320081, 320132, 320137, 320351 and we'll point you to the correct schematic.

See Zimmers page

 

[Thomas_]

Thanks for pointing into that direction. I saw these schematics before but never mached them with my board number = 320008. I only remember having read somewhere that the video signal is mostly controlled by the CPU itself like on the C64. Is that right and which circuits are still involved in producing the video signal and what are they actually doing?

Thomas

 

[dave_m]

I only remember having read somewhere that the video signal is mostly controlled by the CPU itself like on the C64. Is that right and which circuits are still involved in producing the video signal and what are they actually doing?

No, the video drive timing signals come directly from hardware countdown chain logic from the 8 MHz clock. No CPU involvement at all. This should be easy to fix. Look at this schematic:
Sheet 3 of 32008 board

The Vertical Drive signal going to the video analog board is shown at the very top right. Follow it back to its source, a NAND gate D8 pin 11. If that output is stuck high, there has got to be something wrong with that gate or one of the two inputs. Check the inputs for pulses at D8 pin 12 and 13. If there are pulses there, possibly the gate is bad. If one of the inputs is stuck low, you must follow that signal back to its source. Nothing to it. You will have this PET working in no time. Good luck.
-Dave

 

[Thomas_]

Well, I'd rather say *you* would trace it back in no time. Nevertheless, I'm getting more confident. I followed the signal back to the D8 NAND gate and it seems as if it does not get any input. The next thing to check will be the flip-flops, I already guessed they are used to divide the 8MHz. Actually, one of the NAND gates emits a 61Hz signal but with three consecutive rising edges. From programming the Atari VCS I remember that this is required to trigger a new field.

Will tell you once I found out more.

Thomas_

 

[dave_m]

Will tell you once I found out more.

Just to be sure you understand the package pinouts. Check for the U shape notch at one end of the IC. Pin one is at the left. Count down to 7 (or 8 if a 16 pin package). then around to 8 and back up the package to 14. Here is link to pinout of 7400 NAND.

See photo for dual-in-line (DIP) package.

 

[MikeS]

You have a monitor issue.
So look at repairing the monitor board first.

Really? Not even a "you might have..."? With 4.5V DC on the vertical drive line?

Of course it's possible that he's not familiar with his scope and is missing the vertical pulses, but considering how awkward the monitor is to work on I'd make sure first that the signals from the MB are OK.

 

[MikeS]

Just to be sure you understand the package pinouts.

Another tip: the board is divided into quadrants like a map with letters along the side and numbers along the front and the ICs are numbered accordingly, so it's helpful if you specify the IC number on the board and on the schematic when talking about pins and signals.

Also, although it's probably not relevant to your problem it would be good to know for reference which of the four versions of the 320008 board you have; what are the part numbers of the RAM and ROM ICs?

Welcome, and Good Luck!

 

[dabone]

Really? Not even a "you might have..."? With 4.5V DC on the vertical drive line?

Of course it's possible that he's not familiar with his scope and is missing the vertical pulses, but considering how awkward the monitor is to work on I'd make sure first that the signals from the MB are OK.

Yeah I should have said might had, but 99% of the time, a vertical collapse is a monitor issue, correct?
And I didn't read well enough that he was missing the vertical pulses, my mistake.


Later,
dabone

 

[dave_m]

See the photos at the right of the schematic for an idea of what the waveforms should be for the ~60Hz Vertical Drive (photo#2) and the ~15KHz Horizontal Drive (photo #3).

Schematic

 

[Thomas_]

Thanks for the hint, in particular with the video schematics that includes the signals. Would not be surprised if the monitor turns out to be the next problem to tackle.

I'm not sure about the version of the board, however it contains MPS 6540 (013 4078A) roms and MPS 6550 (4078 ) RAMS, both for main- and video memory.

I went down the path further. BTW: Thanks for the hints about the 'quadrants'. After having followed a few lines manually I figures that out as well. Unfortunately the layout between 320026 and my 320008 differs a bit. D8 e.g., is D7 on my board.

Anyway, I went further down towards the oscillator and got stuck at the 74HS00 NAND gate called D8 (actually D7 here) at the very bottom of the schematics.

The NAND emits a constant 0. On one input it get a constant 1 (3,84V), the other signal (input 4) is a nice oscillation so I would expect an inverted signal rather than a 0.

My first assumption was that I found the problem but looking closer at the signal is does not get down to the very bottom as you can see on the picture. So either the signal is not nice enough or the gate is broken. How far does it have to go down to be considered zero? If it was mistaken as a constant 1, then the gate would be right because 1 NAND 1 = 0.

A dirty signal is a problem because so far I can only think in terms of 1 and 0. The fuzzy thing in between scares me. What if it is a half broken resistor or capacitor?

Thomas_

 

[dave_m]

My first assumption was that I found the problem but looking closer at the signal is does not get down to the very bottom as you can see on the picture. So either the signal is not nice enough or the gate is broken. How far does it have to go down to be considered zero? If it was mistaken as a constant 1, then the gate would be right because 1 NAND 1 = 0.

You are right. That signal does not look like a good TTL level zero (It should be less than 0.2V). D8-3 that drives D8-4 is an important signal that triggers the end of a vertical scan. So if that is not working, you will not get a Vertical Sync signal. You are now an expert troubleshooter! At this point you will have to change out the D8 7400 NAND gate chip. Can you solder?

 

[giobbi]

...Not directly related with your issue, but since you have the 6540/6550 model (that, in case of replacement, require some specific adapters) maybe you would take a look to this interesting page about a PET fixing.

http://www.8bit-homecomputermuseum.at/repair/bluepet/bluepet.html

I already discovered that fixing a PET usually doesn't end at first issue, you fix a problem just to discover there's another behind the corner ;-)

--Giovi

 

[MikeS]

...I'm not sure about the version of the board, however it contains MPS 6540 (013 4078A) roms and MPS 6550 (4078 ) RAMS, both for main- and video memory.

What are the 3-digit numbers, especially on ROMs A2 and H3?

I went down the path further. BTW: Thanks for the hints about the 'quadrants'. After having followed a few lines manually I figures that out as well. Unfortunately the layout between 320026 and my 320008 differs a bit. D8 e.g., is D7 on my board.

Are you sure? I've got a 320008 board in front of me and D8 is a 74LS00 as it is on the schematic, and D6 and D7 are N8281s, equivalent to the 74177s on the schematic...

Anyway, I went further down towards the oscillator and got stuck at the 74HS00 NAND gate called D8 (actually D7 here) at the very bottom of the schematics.

Odd; first time I've heard of a different version of the 320008 board...

 

[giobbi]

the "D" row begins with D2, so maybe this is the mistake. However I have a 320008 and D8 = 74LS00 and D6 and D7 = DM74177.
Fortunately in the further series they started to mark every chip with its own number!

 

[Thomas_]

Last night I meassured D8-3 (feeding D8-4). The signals from D8-1 and 2 look good but what the NAND gate produces from it seems distorted. Interestingly the repair report from Thomas Gutmeier from the 8bit-homecomputermuseum.at Giovanni suggested had an issure with the D8 chip as well so this might be a stressed one in general. Just bought two 74LS00N from ebay with DIP14 sockets. I soldered before, also using one of those tin-solder suction pumps which was never a great joy. The Pet will be a special case: Most of the lines under the plastic look deteriorated. They are rather of the kind you'd not want to touch

Even if this will be a success, I will almost certainly have a video issue then. I remember having seen screen gargabe when I switched the pet on years ago. Anyway, the fun lies in repairing.

I found those conversion sockets for the ROM chips on ebay (260782706654) but they come at a hefty price of about 30$. Has someone ever made one himself and if so, is it complicated?

Thomas_

 

[MikeS]

The Pet will be a special case: Most of the lines under the plastic look deteriorated. They are rather of the kind you'd not want to touch

If it just looks lumpy/bumpy, that's normal.

Even if this will be a success, I will almost certainly have a video issue then. I remember having seen screen gargabe when I switched the pet on years ago. Anyway, the fun lies in repairing.

Right on! If you've got the patience I'm sure you'll get it working.

I found those conversion sockets for the ROM chips on ebay (260782706654) but they come at a hefty price of about 30$. Has someone ever made one himself and if so, is it complicated?

They are indeed a little expensive but they do include the EPROM; the trouble is that you have to know which ROM you need to replace.

If you have the means to program an EPROM the adapters are not hard to make yourself, but I'm a big fan of Jim Brain's goodies:
http://store.go4retro.com/6540-adapter/
$6.49 assembled, $1.50 for the bare PCB

They (and N. Welte's similar adapters from Germany) have the advantage that if you program all the ROMs into one large EPROM you can select any one (or, with some kludging, several) with a jumper or two.

As to the EPROM itself, I'm pretty sure someone on here could be persuaded to program one for you if you don't have the capability; there are also semi-commercial folks out there who would gladly supply and program one for a reasonable price.

 

[dave_m]

I remember having seen screen gargabe when I switched the pet on years ago. Anyway, the fun lies in repairing.

I found those conversion sockets for the ROM chips on ebay (260782706654) but they come at a hefty price of about 30$. Has someone ever made one himself and if so, is it complicated?

A garbage screen on the very old PETs that have a hardwired CRT controller logic rather than the 6545 CRTC chip usually means some bad ROMs and or RAMs that do not allow the PET to start up properly. Since you have the old style RAMs (6550) and ROMs (6540) that are not obtainable, rather than using several adapters for them, consider going with the PETVet from bitfixer. This is mini module that plugs into the CPU socket. It contain a complete replacement ROM set and 32K of ram. It has so much ROM that it has jumpers that will allow you to run BASIC 2 or BASIC 4. You really don't want BASIC 1 (unless you want total authenticity) as it doe not work with disk drives very well. See this link for more info: bitfixer.com. I have one and it works well. The cost is about $76 including shipping.

However if you want to keep your PET as authentic as possible with minimum replacement parts, we'll certainly try to help you troubleshoot down to the bad memory part, but it will be difficult to find the bad part without replacing several parts with adapters in a trial and error basis.

For the present let's get something on the screen besides a straight line, then you can decide on the next step.
-Dave

 

[dabone]

I've got one of these in my 2001, replacing all of those ram/roms, and it runs ALOT cooler, the power draw of those older ram/roms is amazing in a bad way.
I removed all the roms and main ram from the machine, plugged in the petvet, and bingo, upgraded ram and stable ram/roms.
(I did find a replacement 6550 when I first got my pet, but it took several months of looking and they no longer have any for sale.)

Also if you do use a petvet, then you have plently of spare video ram, as the 6550's are used for the main ram and video ram.


http://www.bitfixer.com/bf/petvet

Later,
dabone