• Please review our updated Terms and Rules here

Cbm 2001 Pet strange boot

Well, if you don't make mistakes you are not going to learn!

Next time you see a horrid 50 Hz sinewave, you will now know to look at the probe GND connection.

And yes, Hogo's throw-away comment about "is it switched on" - been there, done that, got the tee shirt!

Dave
 
Ok.

Make a note of where the timebase setting is (so we can get back to this setting) and then speed up the timebase. What we are going to try and capture is the narrow pulses that should be in the LOW part of the wave that we can't currently see.

What may happen (and I am hoping it won't) is that we may loose our stable wave.

Dave
 
daver2 said:
Ok.

Make a note of where the timebase setting is (so we can get back to this setting) and then speed up the timebase. What we are going to try and capture is the narrow pulses that should be in the LOW part of the wave that we can't currently see.

What may happen (and I am hoping it won't) is that we may loose our stable wave.

Dave
Click to expand...
unfortunately i can't see any low pulses :(
 
No, UD9 is unlikely to be faulty because UD9/20 is the input pin to the ROM (the chip select line). UD2/17 is driving this pin - and we should have checked this out previously. Plus the fact that you could see a HIGH/LOW signal back in post #1,720...

This is, once again, why we take measurements in a methodical and systematic way...

So, is the signal on channel #2 HIGH or LOW and what timebase setting do you have now? I would keep turning the timebase knob clockwise until the pulses appear... You may be in “no man’s” land at the moment - too slow to see the short pulses but too fast to see the high-level structure.

Remember, the NOP generator is accessing all of the 65,535 addresses sequentially, and the UD9 ROM should only be responding to 4,096 of these addresses. So, you have two (2) waves if you like. One ‘wave’ is selecting the UD2 ROM itself for 4096/65536 = 1/16 of the time. So you will observe that UD9/20 is active for one sixteenth of the time. However, there should be 4096 individual pulses within this 1/16 of a wave. There are, therefore, two components to this wave. We have seen the one component, now we are trying to see the other.

Dave
 
Last edited:
Your timebase is still too slow - wind it clockwise more...

A 1 MHz CPU clock = 1,000 ns per cycle. There are 4 clock cycles to each NOP instruction = 1,000 / 4 = 250 ns per NOP instruction - and this is the typical rate that I would be expecting to see the pulses.

Your timebase is set to 0.2 ms/div = 200 us/div = 200,000 ns/div = way too slow...

You have ten (1) horizontal divisions on your oscilloscope, so to see ten (1) pulses (or one pulse per division) you would need to set your timebase to approximately 250 ns/div (=0.25 us/div).

Remember what I stated previously about knowing what you are looking for - or you need to do it by trial-and-error (less successful in my view). But, if you are trying the trial-and-error method, you need to try ALL timebase settings...

Still, it was a very good idea to post a view of the timebase knob setting though - so full marks for that. I forgot to ask for that...

Dave
 
daver2 said:
Your timebase is still too slow - wind it clockwise more...

A 1 MHz CPU clock = 1,000 ns per cycle. There are 4 clock cycles to each NOP instruction = 1,000 / 4 = 250 ns per NOP instruction - and this is the typical rate that I would be expecting to see the pulses.

Your timebase is set to 0.2 ms/div = 200 us/div = 200,000 ns/div = way too slow...

You have ten (1) horizontal divisions on your oscilloscope, so to see ten (1) pulses (or one pulse per division) you would need to set your timebase to approximately 250 ns/div (=0.25 us/div).

Remember what I stated previously about knowing what you are looking for - or you need to do it by trial-and-error (less successful in my view). But, if you are trying the trial-and-error method, you need to try ALL timebase settings...

Still, it was a very good idea to post a view of the timebase knob setting though - so full marks for that. I forgot to ask for that...

Dave
Click to expand...
I am desperate!!! I can't see any pulses with all time/div :(
 
daver2 said:
What is the closest timebase setting you have to 0.25 us/div?

Dave
Click to expand...
From his previous photos showing the scope's panel better it looks like these two top settings are 0.5uS and 0.2us/div at max clockwise of the control.This is about right for a 20MHz rated scope, typically a 50MHz one goes to 0.1uS /div.

(When there are less frequent pulses in a pulse train, this is where the delayed time-base comes in handy to inspect those individually, but that does require more scope skills to master it and a scope that has it like the V-509 I recommend all the time)
 
One interesting thing about being in the position you are in, is that it can seem to be overwhelming, with everyone asking you questions.
Click to expand...

I get involved in the occasional repair thread in other forums but I generally limit myself to one at a time. I don't know how Dave manages to keep all these plates spinning, I would be getting the machines and their stories so far all mixed up.

Watching this thread, as always, with interest.
 
Desperado said:
Dave i am desperate!!! Now how can i do to solve this impossible Jigsaw??
Click to expand...

The solution to the problem is in your own hands.

By now you have figured out how to use your scope and you have done very well with it.

And I think you can understand what the CPU is trying to achieve, either using the NOP generator, or the program in UD9 that is trying to write to the first element of the screen memory display at address $8000.

Therefore you need to put your thinking cap on. What could be broken stopping it from working ?

You have already figured out that the address lines look ok, and the address buffer IC's are ok, and you know the CPU is ok (because you swapped it).

You don't know yet if the data bus buffer IC's E9 and E10 are ok. Or if the UD9 chip select is working properly yet because of the difficulties getting the scope to see it.

It is time to figure this out for yourself, or it could be a very longer road. All the information you need is right in front of you and in this thread.

When it comes to successful repairs, in complex electronic systems, they occur when the person doing it understands how the circuit is supposed to work. Without that it can become a very hit and miss process, guessing at solutions.

In your case, I think you have learnt enough now to be able to figure it out at this point, without being desperate anymore.
 
Last edited:
Hugo Holden said:
The solution to the problem is in your own hands.

By now you have figured out how to use your scope and you have done very well with it.

And I think you can understand what the CPU is trying to achieve, either using the NOP generator, or the program in UD9 that is trying to write to the first element of the screen memory display at address $8000.

Therefore you need to put your thinking cap on. What could be broken stopping it from working ?

You have already figured out that the address lines look ok, and the address buffer IC's are ok, and you know the CPU is ok (because you swapped it).

You don't know yet if the data bus buffer IC's E9 and E10 are ok. Or if the UD9 chip select is working properly yet because of the difficulties getting the scope to see it.

It is time to figure this out for yourself, or it could be a very longer road. All the information you need is right in front of you and in this thread.

When it comes to successful repairs, in complex electronic systems, they occur when the person doing it understands how the circuit is supposed to work. Without that it can become a very hit and miss process, guessing at solutions.

In your case, I think you have learnt enough now to be able to figure it out at this point, without being desperate anymore.
Click to expand...
Please Hugo help me!!
 
We have already given you the next step if you go back and re-read the posts...

See 1730, 1732, 1733 and 1735.

You need to select the correct timebase for the signal you are trying to observe and then adjust the trigger level.

Based on what I see in post #1735, I would set the timebase to 1 us/div as the best compromise. Slower than I would like (0.25 us/div as I have calculated) but you should, nevertheless see some pulses.

Observing NO pulses doesn't mean something is wrong. The questions is, WHAT are you observing with that timebase setting. Only you can determine that - not me and not Hugo.

Dave
 
You must log in or register to reply here.
Back
Top