Rigol DS1202Z-E Input voltage vs frequency derating?

[TheBay]

A friend has a Rigol DS1202Z-E, he wanted to take some RF voltage readings so used his scope to measure the voltage across his 50R dummy load (Bird 8251).

He had read multiple articles and Youtube videos on doing this and as far as he was aware the max voltage limit on his DS1202Z-E is 300V as per the input labels and documentation.

Using his Apache Labs Anan 7000dleMK2 SDR Transceiver with an Array Solutions Power Master VSWR/Power meter inline he transmitted 10W @ 3.65MHz in to the Bird 8251 dummy load and saw 22.4Vrms on the Rigol DS1202Z-E which is correct.

Happy with his readings and calculations he proceeded to transmit 100W (Expecting around 70.7Vrms) in to the dummy load at 3.65MHz and again measure the voltage across the dummy load with the DS1202Z-E, but something wasn't right it was giving a spurious reading. Thinking it might have been an incorrect setting on the scope and not wanting to reset it to defaults he moved over to CH2 and it gave the 22.4Vrms, so he proceeded to transmit 100W and now CH2 is giving spurious readings.

He then tried 10W again on both channels and still getting spurious readings, so he connected the scope probe back up and went to check the built in squarewave. That was not right, so he reset the scope to factory defaults and all readings are spurious, DC or AC.

Measuring across CH1 with a DMM it seems that has gone really high Z (infinite) and across CH2 it's in a few Mohm ranges. So something has blown in both front ends.

The DS1202Z-E was connected to the dummy load with 50R coax through a BNC T-Piece, as probes were not used this would be at 1X directly to each input.


He's spoken to the dealer (Telonic) where he bought the scope and they told him on the phone after they did some quick calculations that he hasn't done anything wrong and it should cope with that, they then said they will speak to Rigol and get back to him.

Rigol replied with

If there are no signals greater than 300VRMS the customer can’t damage the oscilloscope.

I guess that the generator generates spikes >300VRMS because both input channels are destroyed.

I recommend to use a high voltage probe to absorb any spikes that can occur.

And now the dealer is asking him if he would like a quotation from them to repair the scope, but first of all told him he hadn't done anything wrong and the scope would be fine with this.

He's frustrated as the dealers reply and also Rigol's reply indicate he's done nothing wrong and as far as he's concerned he's not exceded the Max voltage input and has no possible way of generating over 300V, but now has a broken scope!

After he called Telonic back in response to the email from Rigol forwared to him they have asked me for a wiring diagram, which he  has provided, they are going to make further enquiries with Rigol after he stated he was not happy with their catch all 'it is broken, so he must have put more vrms in than he should have' reply.

To exceed the 300V limit quoted by Rigol that would need to be over 1800W in to a 50R load.

I know a lot of test equipment voltage input is derated depending on frequency but I cannot find any information at all on the DS1202Z-E regarding any frequency related derating, there is no mention of this anywhere in the product specifications or the user manual, nor can I find this information on any other Rigol scopes. Does anyone know what this is on this model or series of scope?

[adam4521]

Couldn't find anything for Rigol but this note from Tektronix tends to support your instincts that we should expect a derating of the maximum oscilloscope input voltage with frequency:
https://www.tek.com/en/support/faqs/what-are-maximum-voltage-limits-tds-oscilloscopes-and-how-do-limits-change-over-frequen

If the Rigol input circuits followed the same derating behaviour then while there is not a huge amount of headroom, I calculate that this scenario should still have been ok (assuming no significant harmonics etc). We see a lot of warnings about this for probes, but it is curious that these warnings are not so prominent for oscilloscopes.

[bdunham7]

Predictable (sort of) but not documented.  There are probably lots of things that aren't derated in the specs the way they ought to be and I'm sure this is one of them.  However, I'm puzzled by what he expected to see.  70.7V_RMS would be about 200V_P-P and at 10V/div--the max setting AFAIK--there is only 80 volts on the screen.  So it couldn't have taken the measurement anyway.

[Fungus]

Rigol's manual only says this:


No mention of frequencies but it's in the context of a CAT rating so it's probably at 50/60Hz. Above that? It's anybody's guess.

This video is a similar cautionary tale:

https://youtu.be/4rADgFqFFH8?t=853

[Someone]

Its a mostly ignored specification as people generally use a probe for such high voltages. Rarely does it ever get mentioned:
https://www.tek.com/en/support/faqs/what-are-maximum-voltage-limits-tds-oscilloscopes-and-how-do-limits-change-over-frequen

The maximum input voltage listed in the performance verification procedure manuals for the 1 M Ohms input impedance is: ±400 V (DC + peak AC); derate at 20 dB per decade above 1 MHz.

The 50 Ohms input impedance limit is: 5 VRMS, with peaks less than or equal to ±30 V.

The limits for 1 M Ohms impedance and 50 Ohms impedance converge at 80 MHz. Above that frequency the derating curve continues down at the 20 dB per decade rate for both impedance settings.

Keep in mind that you will usually use a passive probe on the 1 M Ohms input, and that the probe voltage limits will apply. On a TDS754D with a P6139A, the voltage limit is 50 V above 20 MHz out to the probe bandwidth. The recommended active probes also have voltage limits, most less than 50 V.

These voltage limits are imposed by power dissipation limits in the input circuits. The 50 Ohms input impedance is a 1/2 W resistor. This power rating limits the voltage rating to 5 V RMS. On the 1 M Ohms input, as frequencies increase, the capacitive input impedance component decreases and input power is increasingly dissipated in the input capacitance. This heats the whole input attenuator hybrid. When using a probe, the power dissipation in the voltage divider resistors and capacitors in the probe, as well as in the oscilloscope input, limit the voltage. A 10X probe will usually have a 50 V limit at high frequencies because of power dissipated in the divider resistor and capacitor.

Yet the current model scopes dont even mention a frequency range for their input voltage ratings. The above voltage limit from Tek is much lower than you would expect for an 0402 COG capacitor, so there must be some very lossy (or very thermally sensitive) capacitance in the front end.

Lecroy and Siglent at least say their voltage ratings are only valid below 10kHz, but no explicit derating.

[alm]

Yet the current model scopes dont even mention a frequency range for their input voltage ratings.

I clicked on the current scope that was on top on Tek's front page as newly released, and it has this note in the datasheet:

Maximum input voltage:

50 Ω: 5 VRMS, with peaks ≤ ±20 V (DF ≤ 6.25%)

1 MΩ: 300 VRMS , CAT II

For 1 MΩ, derate at 20 dB/decade from 4.5 MHz to 45 MHz;

Derate at 14 dB/decade from 45 MHz to 450 MHz; > 450 MHz, 5.5 VRMS

Is it current model scopes, or is it that some manufacturers do and some don't?

Lecroy and Siglent at least say their voltage ratings are only valid below 10kHz, but no explicit derating.

So the scopes are only rated for signals up to 10 kHz?

[Someone]

Yet the current model scopes dont even mention a frequency range for their input voltage ratings.

I clicked on the https://www.tek.com/en/datasheet/5-series-b-mso-mixed-signal-oscilloscope-datasheet current scope that was on top on Tek's front page as newly released, and it has this note in the datasheet:

Maximum input voltage:
50 Ω: 5 VRMS, with peaks ≤ ±20 V (DF ≤ 6.25%)
1 MΩ: 300 VRMS , CAT II
For 1 MΩ, derate at 20 dB/decade from 4.5 MHz to 45 MHz;
Derate at 14 dB/decade from 45 MHz to 450 MHz; > 450 MHz, 5.5 VRMS

Is it current model scopes, or is it that some manufacturers do and some don't?

I had a look at their lower end models, no such explicit derating.

[Someone]

(Expecting around 70.7Vrms) in to the dummy load at 3.65MHz and again measure the voltage across the dummy load with the DS1202Z-E, but something wasn't right it was giving a spurious reading. Thinking it might have been an incorrect setting on the scope and not wanting to reset it to defaults he moved over to CH2 and it gave the 22.4Vrms, so he proceeded to transmit 100W and now CH2 is giving spurious readings.

Assuming the front end is similar to the 1054, it looks like 70Vrms at 3.6Mhz is close/on the limit of what the 0603 parts would be expected to withstand. Quick estimation from the freely available data suggests the 100 ohm resistor circled below is the likely initial failure point.

[Sighound36]

Hi Bay,

This interests myself on how this could have happened, just call me curious 

Can you show the board an image of the set up that caused the fault that you discribe with the Rigol's possible ADC expiry I feel it would be helpful

Thanks

Sighound

[TheBay]

Hi Bay,

This interests myself on how this could have happened, just call me curious 

Can you show the board an image of the set up that caused the fault that you discribe with the Rigol's possible ADC expiry I feel it would be helpful

Thanks

Sighound

Hi Sighound,

If you mean a block diagram of the setup, attached is what he sent to Rigol/Telonic.

Also my friend just sent me a link to one of many YouTube videos demonstrating this:


https://www.youtube.com/9ilg0JBIgys?t=505 - Time jump to 100W question.

[bdunham7]

Assuming the front end is similar to the 1054, it looks like 70Vrms at 3.6Mhz is close/on the limit of what the 0603 parts would be expected to withstand. Quick estimation from the freely available data suggests the 100 ohm resistor circled below is the likely initial failure point.

Where does the derate chart come from?

[Sighound36]

Hi Bay

 Thank you for the illustration.

The Dummy load from you diagram from your friends is this the same one as in the image I have attached?

Personally I would have used and in series device such as a 30-50dB 100w attenuator so the OUTPUT of the device would then go into the scope front end, from the diagram you are effectively not ultising the dummy load at all with the Tee piece or am I just having a off day, it has been known!

Ideally a spectrum or network analyser with the appropriate set up conditions would be the ideal method, however I appreciate that not everyone has a SA or VNA to hand.

One other thought, grounding of the whole set up and what precautions were taken for this?

SH

[bdunham7]

https://www.youtube.com/9ilg0JBIgys?t=505 - Time jump to 100W question.

Unless I missed it, he didn't actually do it--just said it would be fine.

[Sighound36]

Note the battery in the video

[TheBay]

Hi Bay

 Thank you for the illustration.

The Dummy load from you diagram from your friends is this the same one as in the image I have attached?

Personally I would have used and in series device such as a 30-50dB 100w attenuator so the OUTPUT of the device would then go into the scope front end, from the diagram you are effectively not ultising the dummy load at all with the Tee piece or am I just having a off day, it has been known!

Ideally a spectrum or network analyser with the appropriate set up conditions would be the ideal method, however I appreciate that not everyone has a SA or VNA to hand.

One other thought, grounding of the whole set up and what precautions were taken for this?

SH

Yes the dummy load is the same, I sent your picture to him and he sent back a picture of his.

Having a T there is exactly the same as using a probe or test leads on the load, the scope is just measuring the voltage across the load. The impedance of the scope is higher than the dummy load.

He wanted to measure RF power/voltage with his scope, a VNA doesn't do that and not everyone has a SA at home.
I asked him about the attenuator and he said he does have them and used them. As a ham he wanted to try this method which should work he states, nothing on the scope or paperwork says otherwise and he said as far as he was aware the scope has a limit of 300Vrms which he hasn't exceeded.

[Sighound36]

Hi Bay

What impedance was actually at the scope termination?

How much REAL power was used in this test at the scope termination?

Thanks

SH

[Fungus]

What impedance was actually at the scope termination?

These Rigols don't have switchable 50 Ohm option so it would be 1M Ohm with a direct cable connection.

Having a T there is exactly the same as using a probe or test leads on the load

It's not the same as using a 10x probe on the load 

(which is what you would normally use for anything high voltage/power)

[Fungus]

nothing on the scope or paperwork says otherwise and he said as far as he was aware the scope has a limit of 300Vrms which he hasn't exceeded.

That's a fine line.

You're correct that it doesn't actually say it in the manual, but ... any EE or Ham radio enthusiast could probably predict the outcome just like joe did in the multimeter destruction video I posted above. There's a reason those dummy loads are one giant heatsink.

[TheBay]

nothing on the scope or paperwork says otherwise and he said as far as he was aware the scope has a limit of 300Vrms which he hasn't exceeded.

That's a fine line.

You're correct that it doesn't actually say it in the manual, but ... any EE or Ham radio enthusiast could probably predict the outcome just like joe did in the multimeter destruction video I posted above. There's a reason those dummy loads are one giant heatsink.

Eh, well he isn't an EE and not all ham radio enthusiasts would predect the outcome at all, I'm also a licenced ham and a I don't recall anything remotely related to a scenario like this this being taught with the hobby. In fact it is because of other radio amatuers that he tried to do this as he saw multiple youtube videos and published articles on this specific test which were made by radio amatuers.

His user manual gave no indication of this happening, he has read it cover to cover and in all fairness there really should be a derating guideline in the manual on the basis that this scope does require dertating in relation to frequency, I would assume so but Rigol haven't mentioned this once during any correspondence and informed him multiple times that it is only an issue if has exceeded 300Vrms. If they put something in the manual that mentioned derating even if it was somewhat vague this situation wouldn't have happened. In fact quite these videos and published articles wouldn't exist or would be different if this information was given by the manufacturer. You would expect those who made those videos and articles could predict or assume this, clearly they haven't. Like he couldn't predict or assume as he had no reason to.

The dummy load is that a load and pretty obvious why it requires a heatsink, the scope has a high impedance input! If it was a model that had a 50R termination option which had been enabled then this would be a whole different story. Anything with a 50R input cleary states it's limitations in the manual.

What impedance was actually at the scope termination?

These Rigols don't have switchable 50 Ohm option so it would be 1M Ohm with a direct cable connection.

Having a T there is exactly the same as using a probe or test leads on the load

It's not the same as using a 10x probe on the load 

(which is what you would normally use for anything high voltage/power)

I thought someone would mention 10x in reference to above, I was going to put 1x on my comment above to make it clear.
But comparing apples to apples my point was putting it through a T and putting a 1x probe on across the DL to measure voltage is exactly the same.

My scope probes are all fixed 10X and I never really have reason to use anything other than 10X on my scope apart from when I use differential probes but that's a different story. However I am an EE and he isn't.

[Fungus]

I agree that there's a case.

I'm not a trained EE and I wouldn't have thought about this problem before I saw joe's video a few months ago (thanks joe!), I would have just looked at the voltage. In retrospect it's obvious that derating is needed wherever there's any sort of capacitance in the load.

Judging from Rigol's reply I'm guessing they didn't reach the engineering department.

[Someone]

Assuming the front end is similar to the 1054, it looks like 70Vrms at 3.6Mhz is close/on the limit of what the 0603 parts would be expected to withstand. Quick estimation from the freely available data suggests the 100 ohm resistor circled below is the likely initial failure point.

Where does the derate chart come from?

Simulating the input circuitry of the 1054.

[bdunham7]

Simulating the input circuitry of the 1054.

OK, got it.  Rigol should pay you for that! 

I often wonder why frequency derating is not included in manuals these days--is it because they assume the user knows better or they assume the user is incapable of understanding the concept?

[Someone]

Simulating the input circuitry of the 1054.

OK, got it.  Rigol should pay you for that!

I often wonder why frequency derating is not included in manuals these days--is it because they assume the user knows better or they assume the user is incapable of understanding the concept?

It is only a rough estimate based on the attenuator/compensation and clamping, some of the values/constants could be wrong! There could also be some lower limits further into the signal path in the active buffer (simulation falls apart a bit through there and needs more details) explaining why the OP has seen the input stop working at DC too.

Manufacturers have done worse than not mentioning derating, like the Agilent/Keysight 2000A/3000A/3000T/4000A showing a 300Vrms CAT I rating on the front panel which is (quietly) wound back in the datasheet to only 135 Vrms or 190 Vpk. Even worse on the Keysight 6000 proudly 300Vrms on the front panel, but 30 Vrms or ±40 Vmax in the datasheet. None of those examples provide voltage derating or even suggest it (as Lecroy/Siglent do with their two corner DC-10kHz upper limit), there does remain a possibility those front ends do not need voltage derating. Was there an input schematic from the Agilent 3000 series front end mod developed on this forum?

[Fungus]

I often wonder why frequency derating is not included in manuals these days--is it because they assume the user knows better or they assume the user is incapable of understanding the concept?

It's probably a "marketing" thing.

They could easily include a graph but it would look very bad if it can only take 2V at 100MHz.

(or whatever it works out to be at 100MHz with a 1x probe)

Printing "CAT II 300V" on the front is much better for sales.

[2N3055]

Hi Bay

 Thank you for the illustration.

The Dummy load from you diagram from your friends is this the same one as in the image I have attached?

Personally I would have used and in series device such as a 30-50dB 100w attenuator so the OUTPUT of the device would then go into the scope front end, from the diagram you are effectively not ultising the dummy load at all with the Tee piece or am I just having a off day, it has been known!

Ideally a spectrum or network analyser with the appropriate set up conditions would be the ideal method, however I appreciate that not everyone has a SA or VNA to hand.

One other thought, grounding of the whole set up and what precautions were taken for this?

SH

...........
Having a T there is exactly the same as using a probe or test leads on the load, the scope is just measuring the voltage across the load. The impedance of the scope is higher than the dummy load.
..............

The impedance of the scope is 15-20pF in parallel with 1 MΩ combined (simplified, but good enough for this purpose)..

Damage to the scope from RF will come from heat long before the voltage is too high for the components.

At 1MHz, 15pF will pass as much current as 10kΩ resistor. At  3,6 MHz equ. input resistance is cca 3kΩ.
At 70V RMS that comes down to a 1,6W+ dissipated in just input capacitance. Somewhere there in input circuits.
At higher frequencies impedance drops and power goes up..

And that will be same for all the scopes that have same specified input capacitance. Only thing that will change is that with different input schematics different parts of the circuits will heat up.
Components in input attenuators will not be power components, but smallish SMD types (to have small parasitics.. etc)...

As bdunham7 said, at 10V/div scope will have full range of +/- 40V peak on screen.  Not +/-100V.
300V is "never exceed" peak voltage rating anyways, even at DC. Largest peak voltage (even on DC) you can measure will be +/- 40V... which will be 40V RMS with symmetric squarewave and 28V RMS with sinvewave.
Which would make some level of 16W at 50 Ω that can reasonably be measured and represented on that scope, even without taking derating into account....

Derating curves and or any detailed specifications in that regard might not have helped in this particular case...

With scopes being cheaper than ever in human history, there are many more hobby users than ever before that can afford one. More users, means more mistakes..

But more detailed specifications are always welcome.