Analog versus digital oscilloscopes 2

[bdunham7]

So, the best way to avoid a post being closed is focus only on the Technical Arguments instead of commenting about other people's posts using terms like: "idiot and a troll", "incoherent babbling", "stupidity", "you argue is silly".

I didn't say that.  There's nothing wrong with using qualitative terms.  However, if you prefer technical terms, the foundation of any technical debate is agreement on the meaning of the terms.  If you use the term 'AVERAGE' to mean a rarely found low value, 500MHz to refer to a scope with a -3db point of 350Mhz, etc, it makes it difficult to know exactly what you are saying.  I'm still wondering what you think 'DPO' means...and I don't mean just the words behind the acronym.

As for the 2465A that can display a signal beyond its specified BW, most scopes will do that if you overdrive them, the limiting factor is usually the trigger.  A Rigol DS1054Z can be made to trigger on and display a 400MHz signal, but that doesn't make it a 400MHz scope and isn't all that useful in real life.  Not useless, just limited to an occasional crude display of whether a signal is present.

[bdunham7]

Most of the things that benefit from single shot capture are digital signals which are best captured, filtered and processed by digital domain tools such as logic analysers and protocol analysers.

Two examples of when single shot actually solves unusual, non digital problems.

1.  Uncorrelated signals on multiple channels.  Modern DSOs have dropped the 'ALT' trigger function, but a single shot can yield a clear view of uncorrelated multiple channels.

2.  Noisy or unknown signals that you can't seem to trigger properly on--just slow the timebase down a bit and take a random (or triggered in some cases) single shot and then zoom in and see what is going on.  When you see it, it makes it easier to figure out how to trigger on it.

[Trader]

Most of the things that benefit from single shot capture are digital signals which are best captured, filtered and processed by digital domain tools such as logic analysers and protocol analysers.

Two examples of when single shot actually solves unusual, non digital problems.

1.  Uncorrelated signals on multiple channels.  Modern DSOs have dropped the 'ALT' trigger function, but a single shot can yield a clear view of uncorrelated multiple channels.

2.  Noisy or unknown signals that you can't seem to trigger properly on--just slow the timebase down a bit and take a random (or triggered in some cases) single shot and then zoom in and see what is going on.  When you see it, it makes it easier to figure out how to trigger on it.

What about this?

[bdunham7]

See my earlier reply to David Hess.  Those infrequent aberrations are captured quite well by any decent modern DSO, and then usually can be observed separately using advanced trigger techniques.  So the 2467B, a very impressive and expensive instrument in its time, can merely show you occasional glimpses of your rogue signals, whereas a proper DSO can display them quite clearly in living color.

[Trader]

See my earlier reply to David Hess.  Those infrequent aberrations are captured quite well by any decent modern DSO, and then usually can be observed separately using advanced trigger techniques.  So the 2467B, a very impressive and expensive instrument in its time, can merely show you occasional glimpses of your rogue signals, whereas a proper DSO can display them quite clearly in living color.

Did you see the videos I posted before? ("#11: Tektronix Oscilloscope Triggering controls and their usage" and "Tektronix 2465A oscilloscope demonstration"), that CROs have a lot of triggering and zoom features.

Do you know the 80-20 rule? You are highlighting specific situations, but most time (maybe 99%) the Waveform and Amplitude are the most important info to measure, and complex debug are hard even for a professional.

A Rigol 1054z can read up to 125MHz inside the -3db and barely anything after that.  A CRO with good sensitivity up to 300-500 MHz is not so expensive.

I'll repeat one more time, I'm NOT saying that a CRO is better than a DSO (I also prefer a DSO), I'm saying that a Good CRO can do almost all the necessary work for a hobbyist and even a professional, and they cost much less than a DSO of the same bandwidth.

[bdunham7]

A Rigol 1054z can read up to 125MHz inside the -3db and barely anything after that.  A CRO with good sensitivity up to 300-500 MHz is not so expensive.

I was able to test my garage door opener at 315MHz with mine, using the FFT function.  And I've demonstrated it previously displaying a stable signal at 400MHz.  And that's a pretty low-end scope-- sold mine and moved up.  Where are you seeing these inexpensive CROs in good working order?  I think a good portion of the novices buying these don't even realize that they have problems--I rarely find one that is 100% tip-top, they all seem to need some work.  The only reason these might seem cheap (and they don't seem cheap to me these days) is because they are very old and completely depreciated.

I'll repeat one more time, I'm NOT saying that a CRO is better than a DSO (I also prefer a DSO), I'm saying that a Good CRO can do almost all the necessary work for a hobbyist and even a professional, and they cost much less than a DSO of the same bandwidth...that CROs have a lot of triggering and zoom features.

I have a small collection of CROs and I routinely repair, calibrate and even sell them.  I'm well aware of the triggers, B-sweeps, expansion, etc.  The only way you could believe that this would suffice for any advanced use or that a 'good CRO' can do most of what a DSO can is if you have no clue about what the capabilities of even a low-end decent modern DSO are.  If you have an old-school cheapo DSO like a DS1052E, then you don't have much of an advanced feature set to compare to.  Or perhaps you don't know what other people are looking at with their scopes--one clue for you is that BW isn't usually all that big of a deal.  If 200MHz @3dB isn't good enough, it's not all that likely that 350MHz is going to be the answer.  With exceptions, of course.  If a CRO works for you, great.  Don't assume it will work well for most other people, that's not helping them at all.

[james_s]

Most of the things that benefit from single shot capture are digital signals which are best captured, filtered and processed by digital domain tools such as logic analysers and protocol analysers.

Two examples of when single shot actually solves unusual, non digital problems.

1.  Uncorrelated signals on multiple channels.  Modern DSOs have dropped the 'ALT' trigger function, but a single shot can yield a clear view of uncorrelated multiple channels.

2.  Noisy or unknown signals that you can't seem to trigger properly on--just slow the timebase down a bit and take a random (or triggered in some cases) single shot and then zoom in and see what is going on.  When you see it, it makes it easier to figure out how to trigger on it.

I very frequently use single shot to capture things like startup and transient response of power supplies. I also frequently capture and save waveforms for later comparison. It's also not uncommon that I want to look at the analog aspect of a digital signal.

[Siwastaja]

I very frequently use single shot to capture things like startup and transient response of power supplies. I also frequently capture and save waveforms for later comparison. It's also not uncommon that I want to look at the analog aspect of a digital signal.

Exactly this, and much more. Power sequencing, RC circuits...

Problems with three-stating, pull-up values, speed, interference... Analog levels like current measurements, comparator setpoints. Whatever. Most control aspects in modern electronics are non-repetitive, or irregularly repeating (i.e., not constant frequency) events, and can't be sampled as binary signals either, or doing that loses too much information. So both analog scope and logic analyser are out, often.

Logic analyser is more optimal, possibly fast to use, for some large datasets, but if you go the recommended-by-some "buy an analog scope and a cheap logic analyser" way, you very likely find out you need to still buy a DSO.

A beginner usually does not have unlimited budget and bench space.

A modern low-cost DSO can alone and completely replace these three instruments (DSO, analog scope, logic analyser) no problem whatsoever, and to a limited but often quite sufficient extent, also a spectrum analyser. I do exactly that, all the time, both professionally and for hobby, no problem. If I required very large amounts of bus decoding, a logic analyser would be a timesaver, making life a bit easier. A process optimization, really. I can see no use for analog scope at all, it would be "just for fun".

You can buy a massive set of different kitchen knives and lie to yourself that now you have "right tool for every job" but as every professional chef and kitchen enthusiast know, a good universal kitchen knife that is kept sharp does most of the jobs and all those specialized ones see no use.

I have had all three instruments mentioned at my disposal and I always end up working with the DSO alone.

[tggzzz]

As for the 2465A that can display a signal beyond its specified BW, most scopes will do that if you overdrive them, the limiting factor is usually the trigger.

Amusingly, my Tek485 (nominally 350MHz) will trigger on sine waves >1GHz. The limiting factor appears to be the amplitude, because that has fallen to ~0.2 divisions. Tek made good triggering circuits

[tggzzz]

My digital domain tools are on my DSO. 

I can't afford an 8/16/32/64 channel scope

[tggzzz]

Most of the things that benefit from single shot capture are digital signals which are best captured, filtered and processed by digital domain tools such as logic analysers and protocol analysers.

Two examples of when single shot actually solves unusual, non digital problems.

1.  Uncorrelated signals on multiple channels.  Modern DSOs have dropped the 'ALT' trigger function, but a single shot can yield a clear view of uncorrelated multiple channels.

2.  Noisy or unknown signals that you can't seem to trigger properly on--just slow the timebase down a bit and take a random (or triggered in some cases) single shot and then zoom in and see what is going on.  When you see it, it makes it easier to figure out how to trigger on it.

There are use cases where digitising scopes are very valuable. Anybody that has needed to use an analogue storage scope will be in no doubt of that!

Nonetheless, an attitude that only digitising scopes are useful is incorrect. The aphorism in my .sig indicates my views.

[bdunham7]

Amusingly, my Tek485 (nominally 350MHz) will trigger on sine waves >1GHz. The limiting factor appears to be the amplitude, because that has fallen to ~0.2 divisions. Tek made good triggering circuits

My 485 is as of yet unrestored and has an aberration on the very fastest timebase, so I can't test that.  On most scopes if you just keep increasing the input amplitude, you can still get a vertical amplifier response well after the trigger has given up.  Perhaps the tunnel diode triggers of the 485 are exceptionally good. Trigger response on sine waves well above the specified bandwidth is not an unusual feature on scopes, but rather a necessity to ensure a stable trigger in normal operation. 

Nonetheless, an attitude that only digitising scopes are useful is incorrect. The aphorism in my .sig indicates my views.

OK, then compare scopes to aircraft.  Gliders do fly, and certainly build skills like coordination, energy management, etc.  If you just want to demonstrate that they can be made to work, you can.  However, if you actually need to get somewhere, you select a conventional powered aircraft --unless you happen to live on the edge of a huge cliff overlooking your destination.  I can demonstrate the abilities of an analog scope all day long, but when I need to see something there's a 99% chance I'm grabbing a DSO of some sort.

[Trader]

About the Single Shot Capture, yes I totally agree this is big disadvantage of the CROs, solved by the storage scopes.

I'm not recommending buying a CRO instead of a DSO, I just think, finding a good ~$50 CRO will help a lot a hobbyist with a restricted budget for a DSO. (but today this is changing with a lot of chinese cheap scopes)

As I showed in the previous video ("DSO Oscilloscope vs Analog - When I use which one and why"), some Real-Time events are much better in a CRO and only good in an expensive DSO.  So the CROs could be interesting as a 2nd or 3rd option.

With an Analog Oscilloscope, some events can be perceptible by the eye, enough to find the problem, other events, using a camera give a pretty nice result.

https://www.eevblog.com/forum/beginners/single-sequence-on-analog-scope/msg165889/#msg165889

http://www.fadstoobsessions.com/Downloads-Products/Webcam-Oscilloscope.php

[Trader]

This "Output Channel" could be a very interesting feature to add Storage Capability to an Analog Scope.

Maybe a device that gets the output signal, Store it, and output as a continuous/repetitive signal (to be injected in the scope's channel 2?).

Of course, this is just a thought exercise, this Problem is Already Solved by the DSOs.

[rstofer]

There's nothing wrong with 2 channel CROs!  We used them to get to the Moon back in 1969.  But that was a long time ago and we have moved on.  We don't use pay phones anymore either.

The thing about triggering, whether DSO or CRO, is that it is sometimes necessary to create a hardware signal to trigger the trace.  This happens all the time in FPGA work.  In fact, when looking at SPI, the first thing we trigger on is CS'.  I can't find it at the moment but somewhere around here I have a circuit for detecting the start condition for I2C.  Very handy when trying to synch up to the data stream.

Not only can we trigger on these conditions but with a DSO we also get to see the signals before the trigger.  That is a very powerful feature!

[Trader]

This "Output Channel" could be a very interesting feature to add Storage Capability to an Analog Scope.

This also reminds me the RAW option, you can get the raw data, send to PC and do a lot of good work, e.g.: in the DS1054Z, the FFT is much better calculated using the PC power processing, you can even calculate a good accurate THD.

Be able to use a CRO to get the signal and sending it to a PC (like a USB scope) will be very interesting, since the high quality of some CROs (Teks).

https://groups.google.com/g/hermes-lite/c/zpvR1FV1JZw/m/EaDyjHMyCQAJ

[tggzzz]

Amusingly, my Tek485 (nominally 350MHz) will trigger on sine waves >1GHz. The limiting factor appears to be the amplitude, because that has fallen to ~0.2 divisions. Tek made good triggering circuits

My 485 is as of yet unrestored and has an aberration on the very fastest timebase, so I can't test that.  On most scopes if you just keep increasing the input amplitude, you can still get a vertical amplifier response well after the trigger has given up.  Perhaps the tunnel diode triggers of the 485 are exceptionally good. Trigger response on sine waves well above the specified bandwidth is not an unusual feature on scopes, but rather a necessity to ensure a stable trigger in normal operation. 

Yes.

Nonetheless, an attitude that only digitising scopes are useful is incorrect. The aphorism in my .sig indicates my views.

OK, then compare scopes to aircraft.  Gliders do fly, and certainly build skills like coordination, energy management, etc.  If you just want to demonstrate that they can be made to work, you can.  However, if you actually need to get somewhere, you select a conventional powered aircraft --unless you happen to live on the edge of a huge cliff overlooking your destination.  I can demonstrate the abilities of an analog scope all day long, but when I need to see something there's a 99% chance I'm grabbing a DSO of some sort.

Of course, but that completely misses the point. If you don't have skill and imagination, then you need to be rich (and vice versa).

(Apart from that, gliders can fly higher than commercial airliners (except Concorde), often fly higher and faster than light aircraft, can travel further than you imagine, and can stay up for longer than their pilots can stay awake. But that is irrelevant to this thread)

[tggzzz]

This also reminds me the RAW option, you can get the raw data, send to PC and do a lot of good work, e.g.: in the DS1054Z, the FFT is much better calculated using the PC power processing, you can even calculate a good accurate THD.

That cheap DSO is, I believe, limited to post processing what's on it screen - not everything captured in its memory. Post-processing in the PC removes that problem.

As for measuring distortion, the ENOB becomes important.

You clearly understand that the art is in knowing the strengths and limitations of each of your tools, and using the appropriate combination. Not everybody grasps that!

[TimFox]

When using good DSOs at work, I preferred to download the raw data as a .csv file and do my processing and graphing on my office computer.  I especially liked the modern DSOs that would load such files onto a USB thumb drive.
Even for trivial math processing, I would put the data into Excel, then into "Grapher" (from Golden Software, a high-powered graphic-presentation software package) that allowed me to limit, scale, set axis limits, set axis titles, change point symbols, change line width and color, etc. before inserting the graph into a report.
Unfortunately, most analog CROs don't have a useful analog output that could be externally digitized (some do).

[David Hess]

CROs seem to have reached collectable status. Every online auction I've seen in years has resulted in competition and greatly inflated prices for what you're getting as a tool.

I recommended against purchasing analog oscilloscopes even a couple years ago before that was the case.  With rare exceptions, the same people who have such a limited budget that a used analog oscilloscope is all that they can afford are also unsuited to recognizing good deals or being able to repair and maintain such an old instrument.

As far as collectable status, I suspect something more general is going on because used test equipment of all types has become much rarer and more expensive over the past couple of years.

Single shot capture is a must have nowadays.

It was a must 20 years ago yet somehow we got by, or at least I did, through various clever tricks.

Unfortunately their intensity grading is always or almost always horridly implemented.  It neither duplicates the intensity curve of a CRT, so it cannot be used for tangential measurements, and it does not look nearly as good.

Some are bad, some newer ones are better but they definitely have a different brightness/repetition rate curve, with the DSO being a lot flatter.  That means you can't tell as much by how it looks, but you can spot more infrequent signals.  But it is always problematic to judge new technology by how well it replicates old--a decent entry-level DSO these days will have intensity color grading.  With some setup care, the feature is very useful.  Here's a picture of a pulse that repeats at 20MHz but every millionth pulse has some extra amplitude.

While a DSO should be superior in displaying signal anomalies, this is highly dependent on the specifics; levels of parasitic oscillation or glitching which would be apparently on an analog display can be all but invisible on a DSO.  I have run into this problem several times now with 8-bit DSOs, although oddly enough my Tektronix 7854 which is 10-bits and has no index grading at all could see them easily despite being a digitizing oscilloscope rather than a digital storage oscilloscope.

That is why I gave an objective test although most people are not aware of its significance.  Can a DSO with intensity grading make tangential measurements?  No, or at least I have never found one which could.

What good would this be?  Not all DSOs can apply standard deviation and RMS measurements to every parameter, and tangential measurements are excellent for verifying that the standard deviation or RMS measurement is accurate.  When I was testing DSOs to see if they could make accurate standard deviation and RMS measurements of Gaussian noise, tangential measurement on an analog oscilloscope served to verify the signal source to better than 2%.  The alternative would have been something like a sampling RMS voltmeter.

The DSOs which have problems with this are the ones which make measurements on the display record, because the display processing destroys much of the information contained in the acquisition waveform.

[bdunham7]

(Apart from that, gliders can fly higher than commercial airliners (except Concorde), often fly higher and faster than light aircraft, can travel further than you imagine, and can stay up for longer than their pilots can stay awake. But that is irrelevant to this thread)

It's not irrelevant, it's directly on point.  Despite your glider theoretically having all those attributes on a good day, it has almost zero practical applications other than recreation.  If I suddenly needed to go to Las Vegas quickly (a 200-mile straight line for me, ~300 mile/4-6 hour drive), I could go to a local airport and get a beat up rental Skyhawk (or a T182 Skylane, perhaps) and reliably be there in under two hours.  That's going to be difficult to match with any glider no matter how skilled you may be.

Also, I don't know the economics angle very well, but I suspect that gliding is overall a tad more expensive then general aviation, at least here in the US. 

[tggzzz]

As far as collectable status, I suspect something more general is going on because used test equipment of all types has become much rarer and more expensive over the past couple of years.

I, and I suspect others, would also like to understand that phenomenon.

Single shot capture is a must have nowadays.

It was a must 20 years ago yet somehow we got by, or at least I did, through various clever tricks.

Make that 42 years, in my case.

[TimFox]

50 years ago, we survived by taking Polaroid exposures of single-sweep waveforms.  CROs, cameras, and films were rated for "writing speed", in cm/sec, for that purpose.
Polaroid film was available with a nominal (not well-controlled) speed of ASA 10,000 for high-speed capture.
This was cumbersome, but more reliable than analog storage oscilloscopes.
Nostalgia isn't what it used to be...

[bdunham7]

That is why I gave an objective test although most people are not aware of its significance.  Can a DSO with intensity grading make tangential measurements?  No, or at least I have never found one which could.

What good would this be?  Not all DSOs can apply standard deviation and RMS measurements to every parameter, and tangential measurements are excellent for verifying that the standard deviation or RMS measurement is accurate.  When I was testing DSOs to see if they could make accurate standard deviation and RMS measurements of Gaussian noise, tangential measurement on an analog oscilloscope served to verify the signal source to better than 2%.  The alternative would have been something like a sampling RMS voltmeter.

By 'tangential measurements' are you referring to tangential noise measurement, similar to this but applied more generally to things other than amplifier input noise?

https://w140.com/tekwiki/images/c/c8/Measuring_conventional_oscilloscope_noise_garuts_samuel.pdf

I'm not sure why this technique would not work on any DSO, intensity grading or not.  It appears to at least nominally work on mine.  As for the accuracy of the DSOs built in RMS measurements, there is a significant limitation when measuring input noise because the RMS measurement will include any DC offset, which shows up at very low levels (500uV/div, e.g.) even when AC coupled.  At higher levels of Gaussian noise, the RMS and StDEV match what my AWG says they are.

[tggzzz]

(Apart from that, gliders can fly higher than commercial airliners (except Concorde), often fly higher and faster than light aircraft, can travel further than you imagine, and can stay up for longer than their pilots can stay awake. But that is irrelevant to this thread)

It's not irrelevant, it's directly on point.  Despite your glider theoretically having all those attributes on a good day, it has almost zero practical applications other than recreation.  If I suddenly needed to go to Las Vegas quickly (a 200-mile straight line for me, ~300 mile/4-6 hour drive), I could go to a local airport and get a beat up rental Skyhawk (or a T182 Skylane, perhaps) and reliably be there in under two hours.  That's going to be difficult to match with any glider no matter how skilled you may be.

The point I was making is in the sentence you snipped. Maybe you missed it, so here it is again:
"Of course, but that completely misses the point. If you don't have skill and imagination, then you need to be rich (and vice versa)."

Also, I don't know the economics angle very well, but I suspect that gliding is overall a tad more expensive then general aviation, at least here in the US.

In the region of £12/launch, plus £30/hour including instructor. More importantly it is fun, e.g. +3g/-1g and formation flying and aerobatics all before going solo, flying alongside raptors, no falling asleep because flying straight and level is so boring.