Checking logic pulses with an oscilloscope (DSO-150)

[analogix]

How can I check for logic high/low pulses with an oscilloscope?
 
Specifically, with a cheap DIY oscilloscope (JYEtech DSO-150). Is it even possible, or do I need a dedicated logic probe for these kind of measurements?

[MikeK]

You mean setting the trigger to "single"?

[Refrigerator]

If your scope has a single trigger function then yes, but these scopes are quite slow.
If you want to analyze logic you can buy a cheap 25MHz logic analyzer for 7 bucks.
I have one and it works reasonably well.

[analogix]

I'm not sure what kind of functions it has as the enclosed "user manual" was rather poor -basically no more than just the assembly instructions 
There is a basic "how to use" section at the bottom of that manual, but it's very basic and isn't much help for a beginner to oscilloscopes.

I don't quite know how to use them, but it has 3 trigger modes:
- auto
- normal
- single

There's also a "hold" function. Is any of this useful for checking for a pulse and counting how many?

I found a cheap logic analyzer (24 MHz) on eBay, but I need to check it now and not in 1 month. Besides the final cost will be much more with import duties and all.
So if my existing scope can be used at all that's a better solution ....for now at least.

[YurkshireLad]

There are a lot of videos on YouTube about the DSO-150. Might be worth watching some as there might be some good tutorials for you.

[Refrigerator]

I'm not sure what kind of functions it has as the enclosed "user manual" was rather poor -basically no more than just the assembly instructions 
There is a basic "how to use" section at the bottom of that manual, but it's very basic and isn't much help for a beginner to oscilloscopes.

I don't quite know how to use them, but it has 3 trigger modes:
- auto
- normal
- single

There's also a "hold" function. Is any of this useful for checking for a pulse and counting how many?

I found a cheap logic analyzer (24 MHz) on eBay, but I need to check it now and not in 1 month. Besides the final cost will be much more with import duties and all.
So if my existing scope can be used at all that's a better solution ....for now at least.

Single mode is what you want, basically the scope will wait for a trigger (you specify the trigger*) and once the trigger happens it captures a single waveform and puts the scope on hold.
* rising edge, falling edge triggers

Also import duties on a 7$ item? Yikes. 

[rstofer]

Nobody can even guess whether the DSO is capable until you tell us what you are measuring and what you want to see.

If the frequency of the square wave is very high, it won't display well.  It claims an analog bandwidth of 250 kHz so the square wave stream better be lower than about 25 kHz.  Anything above that will show significant rounding until, at around 250 kHz, the square wave will look like a sine wave.

Then there is the voltage, the spec is 20V/div max and I assume (without a shred of proof) that there are 10 divisions.  This would imply anything up to 200V peak would be acceptable.  I doubt that but I don't work at those levels anyway.  Around 5V for logic it should be fine.

How long is the sample string you want to see.  If you want to display 10 seconds of trace with 1kHz signal, it's going to be a mess.

There is a 1kHz 3.3V test signal on the scope (according to the docs) so you might want to play with that as you learn the controls.

[analogix]

Nobody can even guess whether the DSO is capable until you tell us what you are measuring and what you want to see.

Sure.
I'm trying to figure the cause of a computer that doesn't boot up and have been asked to check if for a logic pulse (and if present, how many pulses) on the chip enable (CE) pin of the booting ROM (27C010) in order to better trace the source of the problem.

If the frequency of the square wave is very high, it won't display well.  It claims an analog bandwidth of 250 kHz so the square wave stream better be lower than about 25 kHz.  Anything above that will show significant rounding until, at around 250 kHz, the square wave will look like a sine wave.

Then there is the voltage, the spec is 20V/div max and I assume (without a shred of proof) that there are 10 divisions.  This would imply anything up to 200V peak would be acceptable.  I doubt that but I don't work at those levels anyway.  Around 5V for logic it should be fine.

If I'm not mistaken that's it: a normal 5V logic signal.

How long is the sample string you want to see.  If you want to display 10 seconds of trace with 1kHz signal, it's going to be a mess.

There is a 1kHz 3.3V test signal on the scope (according to the docs) so you might want to play with that as you learn the controls.

Yes, I've tried that and I get a nice display of the waveform as I adjust the controls.
I don't think I fully understand the concept of a logic pulse, but an estimated guess (and of course the term gives it partly away): a very short voltage burst at either around +5V (logic high) or close to 0V (IIRC it might actually be anything below +2.5V or so) for a logic low.
In other words, similar to if I was to touch the probe to a constant +5V source, then quickly remove it again (which would be "a single pulse") but of course faster, then briefly touch the +5V source again (for a second pulse). Is this the basic concept?

With this in mind I think I may have figured something out: with the trigger mode to "normal" the display appears to stay at the last reading it took (i.e. the waveform is memorized) until a new reading is taken. Obviously I can't be counting the number of pulses if they're fractions of seconds apart, so I'm guessing the measurement display overlaying the waveform display is what I should look at?
It gives these readings:
- frequency
- cycle
- PW (I assume "pulse width", though I don't exactly know what it means)
- duty

Am I correct to assume the Sec/div control has to be at the scope's fastest (which is 10uS for this particular scope)? And the V/div control needs to be set a 5V (since we're expecting a normal logic signal)?
And I'm assuming the scope is to be set to DC (not AC or GND)?
I also have a probe attached to the scope (instead of using just the crocodile clip originally supplied with it). The GND clip is of course attached to GND from the power supply of the computer. The probe has a switch marked 1x and 10x. Is it correct to assume this is an attenuator function where 10x allows for measuring higher voltages?
Should it be set to 1x or 10x for this particular type of reading?

[rstofer]

Here is a video that describes the difference between Normal and Auto triggering.  As luck would have it, Normal isn't the mode you normally use, it's Auto.

https://youtu.be/HhMrzdJFmDA

You can try probing the CS' pin on the ROMs but my guess is that the rate is in the MHz region, well beyond the frequency capability of the DSO-150.  But I'm just guessing...

I would start with 2V/div and 10us/div.  If it's 5V logic, that will result in a 2-1/2 division display.  It if is 3.0V or 2.4V logic, reduce the V/div.

The memory can access data at 45 ns/byte so it could run around 22 MHz and there is no way in the world the DSI-150 can deal with this frequency.  I don't know what frequency it is actually running but I'd be willing to bet it is MUCH higher than 250 kHz.  You would need a logic analyzer with a maximum frequency of at least 100 MHz.

Speed spec from 1st paragraph here:
http://ww1.microchip.com/downloads/en/DeviceDoc/doc0321.pdf

[rstofer]

You would use 10x as it is intended for higher frequencies and presents less load to the device.  Some scopes allow you to select the probe setting so they display the actual value including the 10x.  I don't know what the DSO-150 does.

If the scope can't account for the probe factor, simply set the V/div to 0.2 instead of 2.

Don't worry too much about being perfect in terms of V/div or t/div.  You can move the trace around and enlarge is with a bit of button twiddling after you get the trace on the screen.

 

[rstofer]

A 'logic pulse' is simply a change of state from one state to another and back again.  For the CE' signal, I would expect it to idle high and pulse low for about 50ns and then return high.  It might stay high for some number of ns and then pulse again to grab the next byte.

When I use CE* or CE' or CEn, I am talking about a negative true or active low signal.  The signal goes idles high and goes low to do its thing which, in this case, is enabling the device.  'Chip Enable not' is one way to read it.

[rstofer]

This is a perfect application for a wide logic analyzer.  You could display the address and the data if you had 25 bits of capture.  So, a 100 MHz 32 channel logic analyzer.  Then you could trace the execution in great detail.  Of course, if you don't have the source code you are in for days of agony trying to figure out what everything means.

In a perfect world, you would have the DIP package and it's pretty easy to buy chip test clips.  The PLCC variant is pretty expensive.

32 pin DIP
https://www.warwickts.com/621/32-Pin-DIL-Test-Clips

32 pin PLCC
https://www.warwickts.com/1408/Pomona-32-Pin-PLCC-Test-Clip-and-Cable-Assembly

There are other sources, these are just the ones that popped up.

[analogix]

Here is a video that describes the difference between Normal and Auto triggering.  As luck would have it, Normal isn't the mode you normally use, it's Auto.

https://youtu.be/HhMrzdJFmDA

Thanks. Yes, you're right: "Auto" is the default mode of the scope.
Since I don't quit eknow what I'm looking for I tried to measure the same pins on a similar computer (which is working). On that one I got a lot of "activity" (spikes and other (seemingly) random looking waveforms) with brief pauses in between. So I assume this means that at least I've got the scope set up more or less right.
So next I measured the non-working computer with the exact same settings at the same CE pins. There it would only go "high" (a horizontal line) and stay there, but without any "spikes" or random waveforms as with the working computer).
Deciphering that into meaningful information on the other hand is a different issue 
But it might be enough for the help I'm getting for repairing it to narrow things down a bit more.

You can try probing the CS' pin on the ROMs but my guess is that the rate is in the MHz region, well beyond the frequency capability of the DSO-150.  But I'm just guessing...

I would start with 2V/div and 10us/div.  If it's 5V logic, that will result in a 2-1/2 division display.  It if is 3.0V or 2.4V logic, reduce the V/div.

Yup. Spot on!
2V and 10uS worked great for a proper display. And the probe switch set to x1.
I wouldn't be surprised if I would need some more advanced scope and a logic analyzer (as well as the knowledge on how to use it) for getting some more meaningful info, but it does appear I've been able to detect a pulse in the working computer, but not from the faulty one.

[rstofer]

2V and 10uS worked great for a proper display. And the probe switch set to x1.
I wouldn't be surprised if I would need some more advanced scope and a logic analyzer (as well as the knowledge on how to use it) for getting some more meaningful info, but it does appear I've been able to detect a pulse in the working computer, but not from the faulty one.

If you're not getting CS' going low, the ROM isn't being read.  That's all there is to it.

About the only way that can happen is if a) the CPU is toast or b) the address decoding logic is toast.  Whatever drives CS' isn't working.  You can try to trace it backwards but that will usually require a schematic and those are often impossible to get.

[analogix]

If you're not getting CS' going low, the ROM isn't being read.  That's all there is to it.

Ah! That helps. Thank you 

About the only way that can happen is if a) the CPU is toast or b) the address decoding logic is toast.  Whatever drives CS' isn't working.  You can try to trace it backwards but that will usually require a schematic and those are often impossible to get.

I've got a schematic, and now I have a place to start!
The same CPU works fine in the other computer so it's likely the problem is elsewhere. I'll trace the CE signal back and check for bad circuit traces, damaged vias, cold soldering etc.