[vcf-midatlantic] Open Hardware DSO - 100 MHz/250 Msa/sec - haasoscope

Herb Johnson hjohnson at retrotechnology.info
Mon Nov 20 13:45:42 EST 2017


David Gesswein:

> It all depends on that you want to do with it. This is written assuming you
> are troubleshooting old computers.
> 
> You need sufficient bandwidth
> to view your signal of interest and sufficient sample rate to make it look
> reasonable. Sampling twice the frequency of a sine wave captures the
> information but it doesn't look like [the original signal].
 >
> I think the sample rate will be the biggest limitation. You will need to
> see the character of signals you will want to look at.
> 
> For troubleshooting old equipment the degradation of the signal edges
> is normally not as big an issue. The faults are normally not that subtle
> but I have run across glitches on signals causing trouble that the low sample
> rate can hide. 

I agree with my old friend and colleague, mostly. But I'd put stronger 
emphasis on detecting "glitches". I prefer analog scopes, because they 
can show in some fashion glitches even above their bandwidth. The 
following lecture will be technical. This matters to me, as this is work 
I've done for decades.

"Bandwidth as twice sampling rate" is fine for detection of repeated 
events. It's not fine for reconstructing the signal, or finding one-off 
events. Digital sampling limits aren't the same as the bandwidth rolloff 
of an analog oscilloscope amplifier (or probe). Many 20th C. brand-name 
analog oscilloscopes gave (still give!) good performance well past their 
nameplate bandwidth. also: scope trigger circuits have to respond to 
events far above bandwidth! Triggering is very important.

A "glitch" is a signal where a logic state changes, not because of a 
change in logic condition, but because of a timing event such as 
"propagation" or a "triggering delay". These are very short pulses, on 
the order of several nanoseconds in TTL circuits. Or, they may manifest 
as nanosecond delays in signals. Power supply lines carry glitches, 
caused by literal "spikes" in DC power to logic-switching circuits; 
large currents plus fast changes produce L di/dt or C dv/dt signals.

These matter, because they can trigger logic events. If you can't see 
the trigger, you don't know "why" you have such events. Old TTL designs, 
sometimes generated glitches due to poor design. Aging components may 
cause glitches, by increasing capacitance, or sagging DC voltages. Dead 
or absent bypass caps, fail to filter out power-supply-line glitches.

Look at schematics of the IMSAI front-panel. They actually CREATE 
"glitches" with a series capacitor to turn a logic-level change into a 
triggering pulse! Ugly ugly ugly design.

All that said, any scope beats none. Looking at logic-level signal 
activity (any activity) is a go / no-go test for  logic chips. Many 
faults can be diagnosed with a 20Mhz analog oscilloscope and probe. 
(Probe bandwidth matters. Not today's lecture.) But some faults are 
design problems that finally came due, or non-logic component failures, 
or "tired" chips. TTL logic responds to signals and events of several 
nanosecond duration - that's hundreds of megahertz "bandwidth" 
technically. and even a 20Mhz scope, will trigger on much faster events 
- that at least "detects" them. and you can guess if an analog "bump" is 
actually a pulse above your bandwidth.

End of technical lecture.

In my 20th century view, a $100ish old toolbox-sized HP or Tek 100Mhz 
analog oscilloscope, will show more information about pre 1990's 
computer chips, than most users can appreciate. But a $300 "100Mhz 
sampling" digital oscilloscope you can carry in your pocket, likely 
performs little better than a 30-40Mhz no-brand-I-remember analog 
oscilloscope that you leave behind at a hamfest. Will a 1Ghz modern DSO 
do much better? (shrug) I lack that experience, ask an owner, faster is 
better but.... And a DSO is conveniently tablet-sized, and likely does 
other things too. So the answer for you is: what's in your wallet? Buy 
BOTH if you can afford it, choose otherwise.

Herb "race condition" Johnson
retrotechnology.com


-- 
Herbert R. Johnson, New Jersey in the USA
http://www.retrotechnology.com OR .net
preserve, recover, restore 1970's computing
email: hjohnson AT retrotechnology DOT com
or try later herbjohnson AT retrotechnology DOT info



More information about the vcf-midatlantic mailing list