The point about the "best piece of equipment" is perhaps the thing I like
best about using a logic analyzer for a difficult problem. It's such a pain
to set up, it forces you to think through the analysis before making any
measurements . . . or assumptions. Another nice feature is that you can
rearrange the displayed waveforms without rearranging the probes. If you
choose to see a signal set differently than you previously did, you don't
necessarily have to redefine your trigger equations or any such fun.
Dick
-----Original Message-----
From: Tony Duell <ard_at_p850ug1.demon.co.uk>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Wednesday, June 02, 1999 5:29 PM
Subject: Re: Re[6]: More Bringing up a CPM
>> On the subject of LA/scope/logic probe/voltmeter. I personally
>> consider a LA as the shot gun approach. It will often find
>> problems when one has no idea where to look. The only problem
>
>I can't agree with that. OK, so a logic analyser will monitor a lot of
>points at once, so you can test unrelated sections in the hope of finding
>the fault. But when you've got 64 or 128 or whatever traces, you still
>have to think about the fault.
>
>> is that someone with an oscilloscope can often have the
>> machine repaired by the time one gets the leads connected
>
>I've found that a small logic analyser (say about 3-4 channels),
>intellegently used, is very useful. It doesn't take long to connect it
>up. And it will tell you a lot about the circuit you're investigating.
>
>> I rarely use a logic probe because I don't have one. I
>> can use an oscilloscope just as well with the additional
>> features that an oscilloscope can provide when needed.
>> It is like a logic probe with more added.
>
>You _need_ a LogicDart :-)
>
>> Still the best overall trouble shooting tool I have is
>> an analog voltmeter. I can move quicker with one of these
>
>I agree. I use my old analogue meter a lot more than a digital one. It's
>a lot quicker to read an analogue meter for one thing.
>
>> than many can with other tools. As far as I'm concerned,
>> DVM's are only good for setting power supply voltages.
>
>Oh, they are also useful for setting voltage references on analogue
>subsystems. And checking the exact setting of a variable resistor (when
>you want to change it to (say) 3 time the value as an inital point in a
>modification).
>
>> I have one ( an accurate 5 digit unit ) but the analog meter
>> is still the first thing I reach for.
>
>There are 4 instruments that get most of the use on my bench. In order :
>
>An analogue VOM.
>The LogicDart
>A good bench PSU (30V, 10A). Useful for powering up subsystems
independantly
>A cheap handheld 'scope, audio band only. Useful for checking for PSU
>ripple, looking at motor drive waveforms, SMPSU waveforms, etc.
>
>
>> I find that most don't know how to trouble shoot. The
>> sequence is simple but many still don't get it:
>>
>> 1. Observe symptoms
>> 2. Make predictions
>> 3. validate predictions
>> 4. If predictions do not pan out add this
>> to the observed symptoms and go back to step 2 or 1.
>> 5. Repair bad part.
>>
>> I find that most don't understand the importance of steps
>> 2 and 3. They try to go from 1 to 5 and usually it doesn't
>> work. Step 2 does require that you understand what to
>> expect from each kind of failure.
>
>My method is related to that...
>
>1) Look for silly faults (cables fallen out, etc). Check fuses, etc
>2) Unless there's a very good reason, power up and observe the fault
>3) Do some standard checks. Check PSU rails, master clock, etc
>4) Think of possible causes for the observed symptoms
>5) Do tests (either execute commands or make measurements) to check out
>each possible cause.
>6) If no possible cause is the cause of the fault, you've missed one of
>the possible causes.
>7) You now know the bad part.
>
>The best piece of 'test equipment' you can have is a brain. _Always_
>think about the fault before trying to cure it. Often I make some intial
>measurements and then sit and think for maybe an hour before doing
>anything else.
>
>-tony
>
Received on Wed Jun 02 1999 - 20:23:37 BST