• Please review our updated Terms and Rules here

KIM-1 Repairs continue

I might have goofed here - I used a six inch long wire connected to the ground pin of the port where power connects.. maybe that was too long? There's not a lot of great places to connect close by the CPU, the ground clip is too big to grab the ground pin of that, and i hate soldering all over the board to take measurements.. hmm.
 
falter said:
I'm going to post a google drive link with my results, I took some video to show was was on the scope. https://drive.google.com/drive/folders/1tpW--wA_6FQCVBroazurQbY7CxYkYY0I?usp=sharing
Click to expand...

Ah, just put 'em on youtube. I hate that google drive crap.

You've basically got it right. I can't see the rest of the front panel, so there are some details
that would probably clean up the display a little, particularly in the twitchy triggering. First,
if you have a vertical amplifier bandwidth limiter (typically 20MHz or something), turn it on
and you'll see a little less of that noisy stuff that doesn't matter much in this system. And
you should be DC-coupled for these measurements (well, AC if you want to scope the crystal,
but there isn't much reason to want to do that). Your vertical attenuator and horizontal sweep
time are perfectly appropriate for what you're looking at, but make sure that your triggering
is also DC coupled on the falling edge (which is cleaner than the rising edge), and you may be
able to bandwidth-limit it too. And you're pretty much there - those things should stabilize
the clock signal view considerably. Oh - and make sure you're triggering from the same
channel as you're feeding the signal into.

And the A0 signal you're getting off the RAM looks just as expected.

falter said:
Before trying this, I watched this excellent 'for dummies' video by David Murray's brother. I'm still not fully sure if I have the time base, position etc set right but I see something that looks like a square wave kinda. For fun I probed the RAM starting at A0.. as you can see the frequency, at least as far as my scope sees it, is changing and I think it is lower than what Dave said it should be at to start (was it 500hz or 250hz)? I'm wondering if I've got it dialed in right here. Probe is set to X10.
Click to expand...

I'll take a look at the vid, but have no idea who David Murray is.

As for the probe, the reason they're usually 10x is to reduce the loading on
the circuit under test. The probes usually have a little pin on the BNC connector
where it connects to the scope, and the scope senses it, knows that it means the
signal is actually 10x the amplitude of the signal it's getting, and adjusts your
readout accordingly.
 
Last edited:
falter said:
I might have goofed here - I used a six inch long wire connected to the ground pin of the port where power connects.. maybe that was too long? There's not a lot of great places to connect close by the CPU, the ground clip is too big to grab the ground pin of that, and i hate soldering all over the board to take measurements.. hmm.
Click to expand...

No goof there. Check the schematic and find a suitably convenient ground, maybe
on one of the cassette interface's discrete componets (resistor or capacitor). And
get yourself a 40p DIP clip.
 
falter said:
I'm going to post a google drive link with my results, I took some video to show was was on the scope. https://drive.google.com/drive/folders/1tpW--wA_6FQCVBroazurQbY7CxYkYY0I?usp=sharing
Click to expand...
You need to also adjust the trigger threshold. You have it set to 0v so its triggering off that noise burst. Set it to 2v or so and it should stabilize the square wave.

You will see overshoot, ringing, and noise pickup with long ground leads. To learn what that looks like try to pick up a ground close to the chip to get a reasonably clean signal then use longer ground leads and see how it changes. You will learn what is likely just probing artifacts so can mostly ignore them. Another trick I use to check noise pickup is put the probe on the same ground your using for probe. Anything you see then is noise pickup.

With troubleshooting faults you can occasionally get odd spikes from faults that you may have difficulty telling from ringing/noise pick up and have to ground better but most of the time we aren't talking subtle problems so longer ground leads are OK.
 
See what I mean? There's no always-right answer. I suggested triggering on the falling
edge (because it's cleaner) until you get a handle on the adjustments, @djg (whose avatar
I really love) suggests sticking with the rising edge and adjusting the threshold. Play with
both, see what happens.
 
Thanks guys. My main thing is just making sure I'm looking at valid information and not random nonsense because I have the thing set wrong. What sort of timebase is usually recommended? LIke 50ns? I'm not sure what the T-> thing at bottom right is.

One other thing I noticed while probing the 2102s.. not all address pins were active. 0-4 were, but then 5-7 were not... is that normal?
 
Last edited:
falter said:
Thanks guys. My main thing is just making sure I'm looking at valid information and not random nonsense because I have the thing set wrong. What sort of timebase is usually recommended? LIke 50ns? I'm not sure what the T-> thing at bottom right is.
Click to expand...

No, you're running a 1MHz clock, so the 1uS/div setting you have in your video is perfect.
The 40nS thing you're seeing there appears to be a trigger delay - I don't know whether
you did that on purpose or the scope is doing it for you. That lets you move your window
of visibility back and forth in time relative to the trigger point. There's going to be a
control that allows you to make that adjustment, and watch what happens to the trace -
and that delay indicator at the top of the screen - when you play with it.

falter said:
One other thing I noticed while probing the 2102s.. not all address pins were active. 0-4 were, but then 5-7 were not... is that normal?
Click to expand...

I won't answer that one now because I don't have the schematic in front of me and don't
remember the 2102 pinning...
 
Many thanks. Yeah still getting the hang of the controls here. Actually it turns out all the address lines on the 2102s are active.

I figured out how to set the trigger threshold to 2.0V, I think, however when I do this, it makes the waveform so short it's hard to read. If I try to zoom in with the vertical control the threshold drops and won't allow me to reach 2.0.

There is a way to redefine the size of the squares/divisions and zoom in while maintaining that 2.0v threshold?
 
falter said:
The highest the trigger level seems to let me go is 1.21v..
Click to expand...

That doesn't sound right. No reason for it to limit you if you want it to be higher.
But that's a peculiarity of your scope that you'll have to sort out via RTFM or similar.
 
falter said:
Many thanks. Yeah still getting the hang of the controls here. Actually it turns out all the address lines on the 2102s are active.

I figured out how to set the trigger threshold to 2.0V, I think, however when I do this, it makes the waveform so short it's hard to read. If I try to zoom in with the vertical control the threshold drops and won't allow me to reach 2.0.

There is a way to redefine the size of the squares/divisions and zoom in while maintaining that 2.0v threshold?
Click to expand...

Changing the trigger threshold should not affect the vertical display - that's set by the
vertical attenuator - in volts/div - and the horizontal sweep (including trigger level) have
no business there. Doesn't sound at all right.
 
circa77 said:
Changing the trigger threshold should not affect the vertical display - that's set by the
vertical attenuator - in volts/div - and the horizontal sweep (including trigger level) have
no business there. Doesn't sound at all right.
Click to expand...
Maybe tomorrow I'll do a video showing my adjustments.. I could be describing it wrong.. still getting used to the terminology. But I think I've made a bit of a breakthrough today thanks to you guys. I've been so intimidated by this piece of equipment.
 
It can be a steep learning curve using a scope. The trigger issue is always the vexed setting.

(everything else is Child's-play, the sweep rate to see a few cycles of what you are interested in, the vertical amplifier gain so its not too small and not off screen, all that is intuitive as you adjust the controls).

So why is the trigger always the stumbling block ?

The reason is that what you see on the scope's screen, represents the result of plotting a repetitive signal.

In the dream case it is something like a regular repetitive square or sine wave or some other wave, but it is regular. This is often the case in the world of analog electronics where I mainly live, TV's radios etc, and in those fields, the Engineers could pick up the use of the scope in 20 minutes. You are faced with a harder task.

When the scope's trigger circuit locks to those regular waves (much like an oscillator can lock to a sync pulse) what you see is the same wave plotted over an over again, on top of itself on the scope's screen, all stable and very lovely, and you can examine the wave shape and even write an essay why it looks as it does.

The only problem is though, not all waves we are interested in are that regular.

This is the case in computers, because there are many pulses with variable length timings (depending on what code is running and what the CPU is doing).

In these cases, the best you can hope for ( tweaking the trigger level control for best lock, be it on the rising or falling wave), is to get a "pseudo stable" lock, but in the display you may well see various pulses coming and going.

There is a "sort of solution": One trick is that you can use CH1 of your scope and lock that to the master crystal oscillator in the computer (because all other signals are ultimately derived from that). This oscillator has a very stable and regular output and is easy for the scope's time-base, to lock to.

Then use channel 2 and its probe to look at various circuit points.

In many cases, in the "using the scope to repair a computer", you don't have to bother with anything that elaborate though, because, most of the time, what you are looking for is a line that is stuck high, stuck low, or has an indeterminate logic level on it. So pay attention to the high & low voltage levels of the circuit points you are examining. The other utility of it is to check that various chips, FF's, counters and Logic Gates, and other IC's are obeying their truth tables.

So you can be "brave & bold" with the scope, it works for you, not you for it.

It is a lot like the line on the first Rambo Movie:

Trautman: I don't think you understand. I didn't come to rescue Rambo from you. I came here to rescue you from him.
 
Last edited:
First check should have been the eight data lines (D0 through D7).

You should observe stable DC voltage levels. What are the voltage levels you observe?

Note: you need to know where the 0V/GND level is on the oscilloscope screen when you measure the voltage! I set the probe to GND and adjust the Y position to something sensible on the graticule, then switch the probe/oscilloscope channel back to DC coupling.

Dave
 
Hugo Holden said:
There is a "sort of solution": One trick is that you can use CH1 of your scope and lock that to the master crystal oscillator in the computer (because all other signals are ultimately derived from that). This oscillator has a very stable and regular output and is easy for the scope's time-base, to lock to.
Click to expand...

Good explanation there, Oz. Another way to skin that cat is to feed the system clock (phase 2 in 6502 and 68xx systems)
into the external trigger input, then you still have all of your vertical channels available rather than "wasting" one just on
the trigger. But mom, then I'd have to go get another probe, and that's suuuuuuuuuuuuuuuuch a hassle!

Hugo Holden said:
In many cases, in the "using the scope to repair a computer", you don't have to bother with anything that elaborate though, because, most of the time, what you are looking for is a line that is stuck high, stuck low, or has an indeterminate logic level on it. So pay attention to the high & low voltage levels of the circuit points you are examining. The other utility of it is to check that various chips, FF's, counters and Logic Gates, and other IC's are obeying their truth tables.
Click to expand...

And as I said in the beginning, in time you'll get a feel for what "doesn't look right".

This quote has been broadly attributed (usually to Asimov), but it bears a strong similarity to serious debugging:

The classic phrase of discovery in science isn’t “Eureka!” It’s more “That's odd...”

Hugo Holden said:
So you can be "brave & bold" with the scope, it works for you, not you for it.

It is a lot like the line on the first Rambo Movie:

Trautman: I don't think you understand. I didn't come to rescue Rambo from you. I came here to rescue you from him.
Click to expand...

As long as we're doing movie quotes, here's perhaps my favourite of all time, from "Wag the Dog":

Nice enough people. They just haven't thought it all the way through.
 
falter said:
The highest the trigger level seems to let me go is 1.21v..
Click to expand...
I should have been clearer. I was referring to 2V at probe tip. You are using x10 probes so that is 200mv at scope. To avoid that confusion I normally go under the channel menu and set the input to x10 so all the readouts on the scope will read the voltage at the probe tip.
 
daver2 said:
First check should have been the eight data lines (D0 through D7).

You should observe stable DC voltage levels. What are the voltage levels you observe?

Note: you need to know where the 0V/GND level is on the oscilloscope screen when you measure the voltage! I set the probe to GND and adjust the Y position to something sensible on the graticule, then switch the probe/oscilloscope channel back to DC coupling.

Dave
Click to expand...
So here is video of my attempt at probing all of the data pins on the 6502. I think I managed to set the base (?) onscreen at 0V, and at 500mv per division(?) if I'm reading right, each data pin is fluctuating between 0 and a little under half a volt? I note D0 looks a bit more.. er.. sparse?

Or have I mixed up my scale and it's actually x10, so fluctuating between 0v and just under 5v?
 
Actually.. yeah I goofed here.. I found the channel menu setting to set the input to X10. Retesting.. standby
 
You must log in or register to reply here.
Back
Top