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

[egonotto]

Hallo,

@flodins:
flodins wrote:
"I recently got the R&S RTO2024 2GHz 10GS/s 1Mwfms/s 16bit ADC".

Are you sure that has a 16bit ADC?


I tab with a blue ball pen on the BNC of a picoscope 5243A and a picoscope 4262

It is no problem for normal work.

Best regards
egonotto

 

[David Hess]

I played with some other oscilloscopes (7A18, 7A26, 2445B, 2247A, and 2232) and got the same results at about the same levels of roughly 4 to 10 millivolts peak-to-peak.  The impulse in all cases appears to be coming from the BNC itself and I ran other tests using different adapters and coaxial patch cords which seemed to show the same thing.

I tested the 2445B because it thought maybe a high impedance input built as part of a hybrid would perform better or at least different but the results were the same.

Another source that I have not seen mentioned is the printed circuit board or substrate that the high impedance circuits and buffer are built on.  Tektronix had enough problems with hook on FR4 laminates that they used other substrates for their early printed circuit high impedance circuits.  The 7A18 (polysulfone I think) and 7A26 (something white) that I tested use special substrates but normal printed circuit construction.

[KedasProbe]

The good thing is there is a very good measuring device attracted to it to locate it    (your scope)
Just a few ideas:
- based on the time difference on each channel you can find the vibration travel speed (or multiple)
(in an ideal case you could do like some opposite triangulation, like you know the source but want to find the 3 sensors)
- Pressing something close to the vibration sensor will reduce the signal, this may help you locate it (can be hard if multiple sensors are present per channel) the problem here will be a constant vibration source, attaching some small mechanical oscillator (speaker...) to your coax can make this constant. Or getting closer with a source to the sensor will increase the amplitude,

[Stephan_T]

Hi Dave,

at the end of #938 part 2 (@8:03min) you explicitly talk about high impedance loads/sources, but you only showed shorted probes. In my previous post I already presented some  high impedance results for my Rigol DS1054Z.
Just the mechanical presence of some mass on the BNC connector seam to have a significant damping effect alone.

For high impedance circuits capacitive coupling (of 50Hz fields) and EMF from moving charges has a much larger impact on real measurements.

For an other test I mounted 3 different probes side by side on the back of a chair as far away from the desk as possible.
channel 1 = open BNC (for comparison of microphonic effect)
channel 2 = 1x probe
channel 3 = 10x probe
channel 4 = 100x probe

Image 3 shows the effect of a bang on the top of the case compared to the pick up of 50Hz just by the tiny pins of the probes (nothing attached to the probe) The nearest mains cable was at least 1m away from the probes. You can see how the open BNC (yellow channel 1) is showing the noise of the bang, but only channel 3 (magenta) is slightly effected. In my previous tests I already found, that channel 3 is the weakest (most microphonic) one.

But compare this now to the EMF impact of an electrostatic charge on a PVC pipe waved 50cm away from all the three probes shown in Image-4. And remember: only the 5mm long tiny pins of the probes are acting here as "antennas" to pick this up.

Image-5 demonstrates how much larger the EMF and 50Hz signals become, when the standard spring hooks are attached to the probes.


BTW. with a 1m cable on a DMM I can pick up a "Harry Potter" like waving of the (electro-statically charged) magic PVC wand even from 3m distance.   

[EEVblog]

Hi Dave,
at the end of #938 part 2 (@8:03min) you explicitly talk about high impedance loads/sources, but you only showed shorted probes.

The shorted x10 probe was representative of a high impedance source. Effectively the 9M resistor in the probe.

[joeqsmith]

Thanks for running this. 

At 1:50 "which it does by the way", does this mean you tried the other scopes used in the first video and the effect of tapping the probes base vs directly on the BNC eliminates or greatly reduces the effect?   If so, did you see much of a difference between scope probe combos?

[joeqsmith]

Just as a reference, my old LeCroy with touch screen.  With and without the probe. 
https://www.youtube.com/watch?v=q3Itb5TCjLE&feature=youtu.be

[BravoV]

Just as a reference, my old LeCroy with touch screen.  With and without the probe. 
https://www.youtube.com/watch?v=q3Itb5TCjLE&feature=youtu.be

Looks like LeCroy did their homework pretty well, I can see while without the probe connected, you tapped the screen quite hard and its not affected. 

[ataradov]

Looks like LeCroy did their homework pretty well, I can see while without the probe connected, you tapped the screen quite hard and its not affected. 

That LeCroy is heavy. Transferring energy from the screen to the PCB is much harder in that case. Attach a chunk of led to R&S and you will get similar result.

[mtdoc]

I just watched the "part 2" video.  Dave should be arrested for oscilloscope abuse. Every time he stabbed the screen with his plastic poker I cringed. Seriously, who treats an oscilloscope like that? It's a touch screen not a "stab with a plastic poker" screen.

Maybe it's just me but I was taught that when you are measuring small signals you don't touch anything during the measurement. Not only is this "not a big issue in everyday use" as Dave says but IHMO it's not an issue at all in any proper use of an oscilloscope.  Really!

[joeqsmith]

Just as a reference, my old LeCroy with touch screen.  With and without the probe. 
https://www.youtube.com/watch?v=q3Itb5TCjLE&feature=youtu.be

Looks like LeCroy did their homework pretty well, I can see while without the probe connected, you tapped the screen quite hard and its not affected. 

After a fair amount of TLC, I've been fairly happy with it.    I do play around with some projects that emit enough RF to cause the LCD to flicker but other than that and it's poor case, it's solid.  The probe I show was a LeCroy PP002A.

My scope includes a plastic poker stick to stab it with.  I've never used it and normally keep the touch screen turned off.  I like the mouse and keyboard.   

[Bud]

The impulse in all cases appears to be coming from the BNC itself and I ran other tests using different adapters and coaxial patch cords which seemed to show the same thing.

Yes I tried tapping a open BNC connector on the other end of the patch cord, same thing. An even tapping an unterminated coax cable with no connector on the far end does the same. I fact it seems tapping the coax has a more severe effect then tapping the connector. So yes, no escape 

[BravoV]

Looks like LeCroy did their homework pretty well, I can see while without the probe connected, you tapped the screen quite hard and its not affected. 

That LeCroy is heavy. Transferring energy from the screen to the PCB is much harder in that case. Attach a chunk of led to R&S and you will get similar result.

Exactly my point, and that is not a rocket science isn't it ? Either using a piece of heavy lead .. or other "more advanced & creative" mechanical solution which I believe R&S engineers capable of doing it. 

We are not talking a sub $500 scope here, isn't that normal to have "a bit" of higher expectation ?

[Fungus]

Looks like LeCroy did their homework pretty well,

Either that or they got lucky. Or their scopes simply weigh more so you need to hit them a lot harder.

"Never attribute to cleverness that which can be adequately explained by dumb luck" - Fungus.

[gnif]

For some reason hitting people triggers them too 

[Fungus]

We are not talking a sub $500 scope here, isn't that normal to have "a bit" of higher expectation ?

Maybe it's just that nobody ever saw it as a real problem in the 60+ years they've been making oscilloscopes.

Oscilloscope probes are usually held in soft squishy supports or rigidly attached to the DUT. I've never seen a qualified engineer try to take a reading by bashing the circuit with the probes.

[ataradov]

I've never seen a qualified engineer try to take a reading by bashing the circuit with the probes.

Not a day goes by without me punching a scope while doing measurements on the mV scale

I agree, the whole thing is a cool party trick, but no further action is required on the side of manufacturers.

[Fungus]

Not a day goes by without me punching a scope while doing measurements on the mV scale

Maybe you could try cleaning the trigger adjustment potentiometer.

[RoGeorge]

For some reason hitting people triggers them too 

Brilliant!
 

You, Sir, win the Internet for today!

[KedasProbe]

I've never seen a qualified engineer try to take a reading by bashing the circuit with the probes.

Not a day goes by without me punching a scope while doing measurements on the mV scale

I agree, the whole thing is a cool party trick, but no further action is required on the side of manufacturers.

It's not that important but would be nice.
Could be annoying if your environment generates much vibration or if the vibration is the source your DUT is measuring.
The next generation scopes will have "vibration proof" in the feature list

[EEVblog]

Maybe it's just me but I was taught that when you are measuring small signals you don't touch anything during the measurement. Not only is this "not a big issue in everyday use" as Dave says but IHMO it's not an issue at all in any proper use of an oscilloscope.  Really!

If you know about the effect, and that is the point of the videos. I thought that was obvious, but it seems not, as quite a few have complained the videos were pointless.
There would be a hell of a lot of people who would have no idea it was even possible.
And BTW, it's not uncommon to accidentally bump your scope probe etc (which are very susceptible), and I have seen accidental impulses in real use. Luckily I knew what caused it, others might think it's a real signal, or an accidental trigger could send you up the garden path etc.

[bktemp]

Maybe it's just me but I was taught that when you are measuring small signals you don't touch anything during the measurement. Not only is this "not a big issue in everyday use" as Dave says but IHMO it's not an issue at all in any proper use of an oscilloscope.  Really!

If you know about the effect, and that is the point of the videos. I thought that was obvious, but it seems not, as quite a few have complained the videos were pointless.
There would be a hell of a lot of people who would have no idea it was even possible.

Maybe you should do a video on how to deal with high impedance sources / low level signals & currents.
There are many possible topics like how to deal with noise from mains (both 50/60Hz + harmonics and all the other high frequency noise from SMPSs) or other error sources like triboelectric effect or errors caused by thermoelectric voltages.
You have mentioned some of them in other videos but I can't remember a full video on those topics.

[Fungus]

Could be annoying if your environment generates much vibration

Only if nobody ever invents a thing called "foam rubber" and decides to sell it all over the place at really low prices.

[KedasProbe]

Could be annoying if your environment generates much vibration

Only if nobody ever invents a thing called "foam rubber" and decides to sell it all over the place at really low prices.

There are almost always ways around it but it would be nice if the way around was build in the scope. (if it cost very little)
I hope someone would give it a try to prove that with a spare scope.

[mtdoc]

There are almost always ways around it but it would be nice if the way around was build in the scope. (if it cost very little)

The fact that no one has is testament to the fact that this is a complete non-issue in real world use.