How did Tony blow up his scope?

[metrologist]

Tony makes YT videos, mostly on machining/shop topics. In his recent video he mentions he blew up his scope when measuring a 40khz Piezo Ultrasonic Transducer from an inexpensive ultrasonic cleaner. He said it was high voltage, but even so, I do not see how he damaged his scope.

I can only see a couple ways to connect a probe to this device, either ground and probe to each transducer leads (what he did), or ground to the driver and probe to either lead.

Any guesses on what went wrong (high voltage, really?)? I have a similar cleaner (4 transducers) and was going to take it apart and make sure it's working right, cuz, you know, it ain't broke yet. 

*** WARNING *** no actual scope vapors were released in video, although they were mentioned...

[TheDane]

I feel your pain. 

My bet is that it was primarily back-EMF that killed the scope; the end result from the drivers turning off.

In the video it is mentioned that the head is 60W, and the driver seems like it is mains driven indirectly through isolation transformers.
That is potentially a lot of (reverse)power, and since the scope is rated at max 400V (it is written clearly on the front), it and/or the probe will take a massive hit when exposed to high(er) levels, especially at high frequencies.
A 100x probe is a must for high voltage measurements, especially equipment connected to mains - Is my experience.
They can be found pretty cheap on eBay, and is much less costly than a repair/new scope
1:100 High Voltage Oscilloscope Scope Probe - https://www.ebay.co.uk/itm/P4060-1-100-High-Voltage-2000V-2KV-60MHz-Oscilloscope-Scope-Probe-100X/282780450437

https://hackaday.io/project/4689-improve-the-haber-process/log/16986-analysis-of-the-ebay-ultrasonic-power-supply
The hackaday project is pretty good at explaining and illustrating voltage levels, at resonance and off-resonance.

So, yeah - Hertz hurts. (DC hurts as well, so having those exposed wires hanging around is a no-go in my book - heat shrink tubing works  )
Happy to see nobody got washe'd in the face during the recording.

Good luck getting the welder to function.

[SeanB]

Consider the typical ultrasonic driver is a autotransformer connected to one side to the 300v power rail, gotten from half wave redctification of the mains and using a voltage doubler. This transformer has a tappeed primary connected at one end to the mid point of the 2 voltage doubler caps, and the one end with few windings is the drive, connected to the power half bridge driver. The other is a lot more turns, and the transducer is connected to the centre tap and the open end, and then there is a small toroid used as the standard self oscillating driver for the circuit. The coil inductance and the transducer are the frequency setting element, and as it is a parallel resonant circuit the voltages are very high, over 2kV with almost any transducer that is not broken, and thus a very high very fast voltage waveform. The probe used, being a regular 10x probe, probably only has a voltage rating of 400VAC at 50/60 Hz, and is meant to be derated at higher frequencies, probably being limited to only 30V at 40kHz. Thus with 2kV applied to the probe, the capacitive currents in the probe would be very high, even if the ground probe was not connected. This then would allow enough current through to burn out the lower compensation resistors in the probe itself, and also allow very high voltages to be exposed to the scope front end, and with more than enough to exceed the 40VDC rating of the front end, and with more than enough current capacitively coupled to the input to burn out pretty much every semiconductor in the front end and also probably also most of the attenuator resistors and most likely the ADC input as well.

[ogden]

https://hackaday.io/project/4689-improve-the-haber-process/log/16986-analysis-of-the-ebay-ultrasonic-power-supply
The hackaday project is pretty good at explaining and illustrating voltage levels, at resonance and off-resonance.

Right. Tony is running transducer w/o nominal load (water in the tank), thus amplitude of piezo insert is much higher, voltage as well. Probe divider was blown first, then high voltage came straight into scope frontend which had no chances. I would say this is partially probe's fault as well. That thing is designed to fail so users have to buy next scope.

RTFM: "Do not allow the solution to drop more than 3/8 inch below the operating level line with heat or ultrasonics on. Failure to comply may cause transducer and/or heater damage and will void your warranty.". In this case scope was voided

p.s. I Love Tony's videos - contents and visuals are perfect.

[David Hess]

That is potentially a lot of (reverse)power, and since the scope is rated at max 400V (it is written clearly on the front), it and/or the probe will take a massive hit when exposed to high(er) levels, especially at high frequencies.

The maximum input voltage of an oscilloscope decreases at higher frequencies starting at about 1 kHz.  Probes start derating at a much higher frequency.

[TheDane]

Thanks David Hess,
That was news to me   

I found a datasheet (1:10 600V CAT I - 300V CAT II, DC-500MHz probe), with a derating curve.
http://www.pintech.com.cn/en/uploadfiles/201386112758359.pdf
It seems to start derating quickly from 600V above 27kHz, dropping to 140V at 1MHz. 


Hertz hurts  - it's not a wise idea to poke around high voltage with hands/fingers, which I guess is why the 'real' 1:1000 high-voltage probes are so huge.
I remember using those when fixing tube televisions back in the day - it gave such a good feeling using them.
http://www.farnell.com/datasheets/1725284.pdf
I wish they would include the info in all the datasheets, not omit the potentially damaging stuff.

[Vtile]

That is potentially a lot of (reverse)power, and since the scope is rated at max 400V (it is written clearly on the front), it and/or the probe will take a massive hit when exposed to high(er) levels, especially at high frequencies.

The maximum input voltage of an oscilloscope decreases at higher frequencies starting at about 1 kHz.  Probes start derating at a much higher frequency.

Interesting piece of information... Again one of the quirks of the scopes and especially what is the de facto standard how the typical markings goes. So it seems that I can not safely assume that my 400V max input (1 meg) 100 MHz scope is guaranteed 400V max in whole range, but only at certain fly droppings sized portion near DC. 

I wish they would include the info in all the datasheets, not omit the potentially damaging stuff.

Indeed also what is often not told is the behavior against high energy, in example those HV probes are at some cases told not to use high-energy circuits and sometimes there is nothing in the "manuals". IE. good example is the differential probe 'manuals', I have been interested on those lately as I have been mapping my options to measure one 90kW max. peak VFD braking resistor waveforms, where obviously is more energy involved that in typical CRT. It seems for me the only way to know if some product is really good for you is to buy two and toast one in controlled manner and see.  (..Or cook my own with way over engineered parts)

[T3sl4co1l]

I've ran a generic Tektronix 10x probe (I don't remember which number, a classic basic one) over 500V before, which was risky.  All it takes is the internal element to arc over, and poof, now you've got high voltage on the scope input.

I wouldn't trust a cheap Chinese probe to handle more than it's rated, and even that's probably risky... consider, using a single 1/4W resistor to hold off 400V.

I did once make my own 100x probe, which was fine until the 10M resistor arced over and zapped that channel.   That was on my Tek 475, which instead of buying a replacement attenuator module, I just shoved some resistors into the socket... now that channel has some 'hook' that can't be trimmed out of the >5V/div range.

Tim

[Vtile]

I've ran a generic Tektronix 10x probe (I don't remember which number, a classic basic one) over 500V before, which was risky.  All it takes is the internal element to arc over, and poof, now you've got high voltage on the scope input.

I wouldn't trust a cheap Chinese probe to handle more than it's rated, and even that's probably risky... consider, using a single 1/4W resistor to hold off 400V.

I did once make my own 100x probe, which was fine until the 10M resistor arced over and zapped that channel.   That was on my Tek 475, which instead of buying a replacement attenuator module, I just shoved some resistors into the socket... now that channel has some 'hook' that can't be trimmed out of the >5V/div range.

Tim

Yep. I'm also worried about the flash over as if there is enough power behind the flash is not just unnoticed *click* that might fry some opamp, but a serious boom. ...And I'm not even worried about personal safety aspects (can be arranged with measuring place and procedure) nor measuring equipment, but the damage for the DUT.

I assume your 10Meg resistor weren't something like this http://www.royalohm.com/pdf/product2017/PowerDissipationPDM.pdf
I might need to get a bucket of those and some meters of shielded HV-powercable as a compensation capacitor.

[Jwillis]

I'm still learning stuff.I did a little search on those cleaning transducer power boards and they can have any where from 400VAC peak to peak to as high as 1200VAC peak to peak. Wouldn't that be a bit much for his oscilloscope?

[David Hess]

I'm still learning stuff.I did a little search on those cleaning transducer power boards and they can have any where from 400VAC peak to peak to as high as 1200VAC peak to peak. Wouldn't that be a bit much for his oscilloscope?

There is some margin to the published maximum voltage specification but not that much.  If have had oscilloscopes survive repeated breakdown of the input AC coupling capacitor but most will not and when you reach that level of overload, that is not the only part which can breakdown or fail.

Old designs tended to use axle leaded 1/8th and 1/4 watt carbon composition resistors where overload tolerance was required and they are very tolerant of excessive voltage and tended to fail open.  There are modern small axial leaded fusible high voltage film resistors which replaced carbon composition resistors in safety and protection applications but I have never found them in surface mount packages.

The input protection structure is a high value series resistor, 470 kilohms is typical, between the 1 megohm shunt and the FET gate which is protected by one or two low leakage low capacitance shunt diodes; JFETs really only need one external shunt diode for protection as they have the other built in.  That high value of series resistance would provide unacceptable frequency response into the capacitance at the FET gate so it is bypassed with about 1000 picofarads of capacitance which is where the derating of the maximum AC input comes from.  As the frequency is increased, the bypass capacitor passes more current which eventually reaches the point where it damages the shunt protection diodes or worse.  I have done the calculations based on real circuits and the advertised maximum voltages and derating, at least on old instruments, is about right.

Any high impedance input attenuators are sized for about the same protection level so high voltage breakdown there tends to cause breakdown further down the chain of circuits.  They should improve the derating with frequency however most new DSOs only have one which is engaged at a relatively low sensitivity.  The same applies to external probes with the exception that practically all x10 probes (1) and most x100 "high voltage" probes (2) lack shunt attenuation relying on the oscilloscopes 1 megohm input resistance as a shunt so they provide no attenuation to DC if AC coupling is used.

(1) Some old x10 probes include an internal shunt but I have never seen a current production one which does.  LeCroy might have some for their differential amplifiers as they allow for adjustment of probe attenuation.
(2) I know of one (!) currently produced x100 probe with an internal shunt and it actually looks more like a real high voltage probe.

[Jwillis]

That scope is a Owon sds6260v and it says in the manual that the  max input voltage is 400V (PK-PK) (DC + AC PK-PK) .What ever does DC+AC PK-PK mean?
Anyway this is a good lesson to be learned but I was wondering if theirs such a thing as probes that work by inductance instead of direct contact?By this I mean a clamp that goes on the shielded wire instead of the metal. 

[David Hess]

That scope is a Owon sds6260v and it says in the manual that the  max input voltage is 400V (PK-PK) (DC + AC PK-PK) .What ever does DC+AC PK-PK mean?

That means that the combined peak of the AC and DC voltage must not exceed 400 volts.  So if you have a 240 volt AC signal which produces about 340 volts peak, you can apply another 60 volts of DC to it before reaching the maximum input specification.

Anyway this is a good lesson to be learned but I was wondering if theirs such a thing as probes that work by inductance instead of direct contact?By this I mean a clamp that goes on the shielded wire instead of the metal.

He was using a 40 kHz current clamp in earlier videos but that would not be fast enough for this and it would not reveal the voltage anyway.  This would be a good application for a high voltage differential probe.

[tautech]

If one's to 'play' with electronics outside their field of experience , and like most, their scope was a significant investment.......well as much as a cellphone, then they should invest in a 100x probe at the very least. They cost no more than a night out and if you add a 1000x probe too, the total cost (cheap DSO included) is less than a good cellphone !

What's more, the scope and probes will outlast the cellphone !

[metrologist]

What's more, the scope and probes will outlast the cellphone !

I bet they are working on "smart" probes to take care of that problem...

This has been a good thread, exposing more for me to consider. Thanks all 

[tautech]

What's more, the scope and probes will outlast the cellphone !

I bet they are working on "smart" probes to take care of that problem...

To quote an old sales pitch; You can have anything you like Sir/Madam, providing you can pay for it !

Therein lies the issue, how much is the average punter prepared to spend ? ?