EEVblog #983 - A Shocking Oscilloscope Problem ! aka Whack Triggering ...

[tautech]

My icon is a picture of my scope after taping on the BNC connector!!!   Just kidding, but my TDS220 has a 'microphonic' front end with a 1 Mohm input.  Interesting, but it's not an issue for me.  Of course, the TDS220 doesn't have a touch screen.

The TDS210 and 220 had the BNC's mounted directly to the PCB with no additional fixing to the chassis which seems to dampen the microphonic susceptibility it would seem from the different models of scope that Dave checked.
Tek learnt some lesson from these two models as the BNC's were inclined to bust away from the PCB if subjected to abuse. The TDS1k and 2k did not show anywhere near the same level of problem as their BNC's are fixed with nuts to a steel chassis.

[trophosphere]

I wonder if something like this from Murata would help. Granted it is advertised as the source coming from the capacitor itself it would be logical to assume that it may go both ways?

[Rasz]

the most shocking thing here is a weaksauce clickbait game

[EEVblog]

In case you're not following the R&S thread, the issue pretty much vanishes as soon as there is any kind of probe connected, so pretty much a non-issue.

Nope, it's still a potential issue even with a probe plugged in.
Even with a shorted x10  probe I still get 1-2mV with the poker tap on the screen. Not unexpected as the 9Mohm + low cable cap doesn't present much of a load.
Obviously it's going to be a complex interaction with the source impedance and cable used etc (which are now a "load" for the voltage impulse)
Yes it goes away with a shorted x1 probe (i.e. a low impedance source) but it's not right to say it's not a significant issue when you plug a probe in. I still see it. YMMV?

[Gazza2]

Well I gave my rigol ds2072a some good whacks and found it only triggers on the bnc's. Has Dave started the international slap your scope day?, is this the nerds version of planking?

[Fungus]

the issue pretty much vanishes as soon as there is any kind of probe connected, so pretty much a non-issue.
 
What a scope does with an unconnected input just doesn't matter, as you'll never use it like that.

It's almost as if you missed the part of that video with Dave whacking probes on the bench.

[G7PSK]

What happens in a really noisy environment, such as big diesel engine or gas turbine testing or large amplifiers at concerts etc.

[Brumby]

You wear ear muffs.

[Fungus]

Still waiting for the picoscope results.

[G7PSK]

With a pair on the scope as well 

[mikeselectricstuff]

In case you're not following the R&S thread, the issue pretty much vanishes as soon as there is any kind of probe connected, so pretty much a non-issue.

Nope, it's still a potential issue even with a probe plugged in.
Even with a shorted x10  probe I still get 1-2mV with the poker tap on the screen. Not unexpected as the 9Mohm + low cable cap doesn't present much of a load.
Obviously it's going to be a complex interaction with the source impedance and cable used etc (which are now a "load" for the voltage impulse)
Yes it goes away with a shorted x1 probe (i.e. a low impedance source) but it's not right to say it's not a significant issue when you plug a probe in. I still see it. YMMV?

Probably something they should be checking in production as it will clearly have some dependence on mechanical stress etc. Your unit looks like it's maybe more sensitive.

[bibz]

Maybe yours is more robust? Need more data!

[EPTech]

Hi guys,

I got a Rigol DS4024 here. With probes unplugged and sensitivity on 2mV/div and trigger just above the noise and mode set to normal I have to tap the case quite firmly to get a trigger. On the BNC's I get a spike of about 10mV. Once the probe is connected there is no noticeable microphonic effect on the BNC's, as Mike stated, which is good.

The passive probes that came with the Rigol are very microphonic though. I can even pick up my voice if I talk loud enough. When I took off the tip (hook) the effect went almost completely away. It seems my probe is only microphonic when the probe tip is on there. Any ideas on that? Is the contact inside the probe tip causing this maybe? Is friction between the plastic parts causing minute static potentials?

[IAmBack]

I did another experiment with my 1054z.
I've put 50ohm terminators and then poked the scope (i didn't want to turn off input).
And microphone effect vanished on terminated inputs!
I've some explanation of this phenomena, but I'll wait for experimental confirmation from You.

[mikeselectricstuff]

I wonder if it may also depend on the surface it's sitting on - a hard surface would probably conduct shock away from the scope btter than a soft one.

 

[SeanB]

I tried with the Tenma scope I got from Toploser, and it needs a fair whack to get a response, so I guess pretty much all scopes will have this. But then again, an analogue scope would do the same if tapped, because the deflection plates in the tube move inside the tube supports.

You really had to have pretty good tube design, with massive internal support structures, extra bracing along the edges and at the rear, thicker plates and added vibration mountings, so there were no resonant points inside the aircraft vibration envelope to get a CRT to survive in aircraft. They also had to have much thicker glass faceplates and some scary scanning design, so you got a flat faceplate, though there you were slightly less worried about tube depth in the panel, so you could have a long beam throw making the deflection correction easier in the small scan arc, as opposed to consumer which wanted 110 degrees of scan and more to keep the depth down. Drawback was you really ran the phosphor hard to get good brightness and daylight contrast, and the front ran pretty hot in operation.

[luisprata]

DSO-X2004 results
Para os brasileiros que acompanham o Forum.

[metaxis]

Shock mounts may have unacceptable trade-offs.  Case/bench isolation should be doable, but I suspect it's hard to avoid mechanical coupling from the BNC to the frontend while still providing a fixed, stable  EM environment.

Interesting article from TI on "Stress-induced outbursts: Microphonics in ceramic capacitors" shows this in a DUT.

https://e2e.ti.com/blogs_/b/precisionhub/archive/2014/12/19/stress-induced-outbursts-microphonics-in-ceramic-capacitors-part-1

https://e2e.ti.com/blogs_/b/precisionhub/archive/2014/12/23/stress-induced-outbursts-microphonics-in-ceramic-capacitors-part-2

[EEVblog]

I wonder if it may also depend on the surface it's sitting on - a hard surface would probably conduct shock away from the scope btter than a soft one.

It's a complex resonance mode system thing. Think of the bench as the acoustic equivalent of an RF stub. (they are actually directly equivalent systems, just the mediums and wavelengths differ)

[EEVblog]

And microphone effect vanished on terminated inputs!
I've some explanation of this phenomena, but I'll wait for experimental confirmation from You.

Of course it vanishes. 50ohms is stupidly low impedance compared to a microphonic MLCC cap source impedance.
But of course it depends on where in the front end system chain the offending "singing" cap is located.
If it's after the FET input buffer then you can forget about the input termination doing anything at all.
But by design is almost always likely to be on the input side of the buffer, because the output buffer is going to be low impedance system at that point, so any caps there won't be a problem.

[JPortici]

Still waiting for the picoscope results.

i see if i have time to switch it on later today!
by the way, that tektronix TPS of mine was a real mess: at 20 mV/Div (10x, BW Limit) tapping on scope produced spikes greater than 100 mV, they go outside the screen. but it has never been an issue while working on cars as far as i know it (we don't put it over the engine, but rather on the plastics around the compartment)

edit: it was harder than other scopes. i pretty much had to keep it in one hand while i was beating it

[timgiles]

Interesting papers referenced above. Nice to see some actual real world comparisons?.

I wonder if any one has an idea of the cost increase on a typical frontend if all caps were replaced with tantalums / film caps?

[mikeselectricstuff]

Interesting papers referenced above. Nice to see some actual real world comparisons?.

I wonder if any one has an idea of the cost increase on a typical frontend if all caps were replaced with tantalums / film caps?

It's only going to be something right at the input of the front end, as shorting the input makes it disappear, so will be very small capacitances. I'm not sure there is any alternative to ceramics at these low capacitances. Digikey shows nothing else at 10pf.
If you think how small the required charge is to produce a 1mV spike on a 1megohm input with 10pf capacitance, I doubt it is even possible to eliminate mechanical effects as most components will have some sensitivity to mechanical stresses, and you need to consider a wide range of mechanical frequencies.
The only solution would be careful mechanical isolation, and that would be hard to do without increasing input capacitance. Even something crude like potting the input stage in a suitable damping resin is going to increase stray capacitances. 
 

[madires]

Hameg 1508-2 (CRT): about 2mV pp for a gentle tap on the BNC and about 8mV pp for a proper tap. A tap on the case has a much smaller impact (<0.1mV).

[Stephan_T]

Plugging in a BNC connector can have a huge impact on this problem as mike pointed out. I tested this on my Rigol 1054Z and without even shorting the probe, the impact of tapping on the case and the connector is much smaller compared to a naked BNC socket. Even a BNC T-adapter with nothing else connected dampens the vibrations substantially. So it would be interesting to see the comparison of the different scopes under real world conditions, with probes (open circuit) connected to the inputs.

On my Rigol the Channel 4 seems to be the least effected. Channel 3 is the most microfonic one.

Image 1 shows a tap on the top of the case with
 a T-adapter on channel 1,
 x1 open probes on channel 2 and 3
 and the raw (nothing connected) channel 4.

Image 2 is a similar strike with nothing attached to any connector.