Author Topic: Strange behaviour of Siglent SDS1104X-E  (Read 1715 times)

0 Members and 1 Guest are viewing this topic.

Offline Pannenkoek_met_een_lampje

  • Contributor
  • Posts: 21
  • Country: nl
Strange behaviour of Siglent SDS1104X-E
« on: June 09, 2018, 03:13:40 pm »
Hello I just bought a Siglent SDS1104X-E . I try to measure the voltage over a diode used in a DC DC converter. I use channel 3 (blue) at the anode and channel 4(green) at the cathode.


When I use the 2V/div setting I see a delta voltage of 560mV which seems realistic to me. When I switch to 1V/div I get a totally different value. For instance 2.64 volt difference between the probes. which doesnt make sense to me. When I change the input voltage of the DC DC converter the delta V changes a lot when the setting is 1V/div it can even go negative. When I use the 2V/div setting it is always a realistic diode drop voltage.

Other observations / data:
I got a firmware update to 7.0.6.1.25 R2
I hear a relay click sound between the 1V and the 2V/div setting



Can anyone tell me if this is a (known) software bug or is it faulty hardware? Thanks
« Last Edit: June 09, 2018, 08:26:40 pm by Pannenkoek_met_een_lampje »
 

Offline nez

  • Contributor
  • Posts: 11
  • Country: us
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #1 on: June 09, 2018, 08:23:04 pm »
Is the second screenshot accidently a duplicate of the first?  They seem the same.
 

Offline Pannenkoek_met_een_lampje

  • Contributor
  • Posts: 21
  • Country: nl
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #2 on: June 09, 2018, 08:28:46 pm »
Hello Nez, you are right! Thank you for telling, i did change the wrong picture. Please take a look at the delta Y
« Last Edit: June 09, 2018, 08:30:41 pm by Pannenkoek_met_een_lampje »
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 17443
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #3 on: June 09, 2018, 08:32:18 pm »
Neither a bug or a fault, it's user error.
Simply at 1V/div for this waveform you exceed the channel offset limits outlined in the datasheet.

See here:
https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/msg1531490/#msg1531490

And here:
https://www.eevblog.com/forum/testgear/siglent-technical-support-join-in-eevblog/msg1533101/#msg1533101
Avid Rabid Hobbyist
 

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 1072
  • Country: ua
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #4 on: June 09, 2018, 09:42:57 pm »
you-re need to switch 10 V/div or something like that. Because both screenshots shows ADC overflow.
In order to perform measurement, the waveform should not exceed vertical screen limits. It should fit to the screen.
This is common limitation for digital oscilloscopes.
« Last Edit: June 09, 2018, 09:44:44 pm by radiolistener »
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 18870
  • Country: nl
    • NCT Developments
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #5 on: June 09, 2018, 10:16:18 pm »
you-re need to switch 10 V/div or something like that. Because both screenshots shows ADC overflow.
In order to perform measurement, the waveform should not exceed vertical screen limits. It should fit to the screen.
This is common limitation for digital oscilloscopes.
No it isn't There is an offset limit in the datasheet which -more or less- tells you what the full dynamic range is. And this is often way more than the vertical screen limits.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 
The following users thanked this post: Pannenkoek_met_een_lampje

Offline Performa01

  • Frequent Contributor
  • **
  • Posts: 901
  • Country: at
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #6 on: June 10, 2018, 04:56:25 am »
Offset limits in the data sheet have nothing to do with dynamic range. They simply describe the DC offset range covered by the vertical position controls, hence is the maximum DC offset of an input signal that can be compensated by shifting the trace position.

An SDS1000X-E can handle up to +/-500mV at the ADC input, i.e. after the attenuators.

This means that the distortion free input range is about
+/-500mV up to 118mV/div
+/-5V from 120mV/div to 1.18V/div
+/-50V for 1.2V/div up to 10V/div

These numbers are valid for direct connection and both the vertical sensitivity and max. input voltage range have to be multiplied by the probe attenuation factor in use.

With 10x probes:
+/-5V up to 1.18V/div
+/-50V from 1.2V/div to 11.8V/div
+/-500V for 12V/div up to 100V/div

 
The following users thanked this post: rf-loop, Pannenkoek_met_een_lampje

Online nctnico

  • Super Contributor
  • ***
  • Posts: 18870
  • Country: nl
    • NCT Developments
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #7 on: June 10, 2018, 08:26:38 am »
Offset limits in the data sheet have nothing to do with dynamic range.
Yes they do. They tell you the full dynamic range of the analog front end. The ADC can handle only part of that and with the offset control you can shift the 'window' the ADC sees.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10479
  • Country: 00
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #8 on: June 10, 2018, 09:31:54 am »
When I use the 2V/div setting I see a delta voltage of 560mV which seems realistic to me. When I switch to 1V/div I get a totally different value. For instance 2.64 volt difference between the probes. which doesnt make sense to me. When I change the input voltage of the DC DC converter the delta V changes a lot when the setting is 1V/div it can even go negative. When I use the 2V/div setting it is always a realistic diode drop voltage.

Can anyone tell me if this is a (known) software bug or is it faulty hardware? Thanks

Neither.

All op-amps have their limits. Every time you overdrive/saturate one there's a "recovery time" where the output signal can be very distorted.

What this means in DSO terms is that you can't zoom in on signals arbitrarily. The visible trace will distort whenever it goes a long way outside the screen area.

FWIW, topic comes up every couple of weeks in these forums. This thread has some pretty screenshots of the distortions (on a Rigol, but all 'scopes do it).

« Last Edit: June 10, 2018, 09:41:13 am by Fungus »
 
The following users thanked this post: Pannenkoek_met_een_lampje

Offline Pannenkoek_met_een_lampje

  • Contributor
  • Posts: 21
  • Country: nl
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #9 on: June 11, 2018, 06:00:19 pm »
Offset limits in the data sheet have nothing to do with dynamic range. They simply describe the DC offset range covered by the vertical position controls, hence is the maximum DC offset of an input signal that can be compensated by shifting the trace position.

An SDS1000X-E can handle up to +/-500mV at the ADC input, i.e. after the attenuators.

This means that the distortion free input range is about
+/-500mV up to 118mV/div
+/-5V from 120mV/div to 1.18V/div
+/-50V for 1.2V/div up to 10V/div

These numbers are valid for direct connection and both the vertical sensitivity and max. input voltage range have to be multiplied by the probe attenuation factor in use.

With 10x probes:
+/-5V up to 1.18V/div
+/-50V from 1.2V/div to 11.8V/div
+/-500V for 12V/div up to 100V/div

Thank you Performa01, think I get the point. I did take a look at the manual of the scope and did find other numbers though. I did put them in a screen shot.


I Find:
500uV   up to 118mV +/- 2V
120mV  up to 1.18V +/- 20V
1.2V      up to 10V +/- 200V

I now understand that if everything is on screen the results will be okay.
But, one thing is not clear to me though. Lets say my setting is 2V a div with a 10x probe. Then I am in the "120mV  up to 1.18V +/- 20V" range.  If my signal is within the +/- 20V dc offset but the signal disappears of the screen will the opamps need a recovery time and can my result what I read on the screen be still be wrong?

To put it in another way lets say DC offset is 2V:

sure out of range [ -20V ---- Possible dc-offset --- [ (2V)(screen range 8div's*2V = 16V +2V DC offset =18V] ---- Possible dc-offset --- +20V ] sure out of range

So if the signal is below 2V but above -20V or if the signal is above 18V will the signal outside of the screen destroy the values on the screen?

Another question, whats the best option if I want a higher resolution to see whats happening on a signal with a high dc-offset then what are my options? A differential probe? I did take a look at some of them but most of them are 50x and I am more interested in a signal with a high dc-offset which can change between 0.4 to 1.2 Volt.

thanks
 

Offline Pannenkoek_met_een_lampje

  • Contributor
  • Posts: 21
  • Country: nl
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #10 on: June 11, 2018, 06:02:52 pm »
thank you fungus!
You did help a lot! I did make one text in a reply to Performa01, can you please take a look if I understand it?

thanks
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 10479
  • Country: 00
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #11 on: June 11, 2018, 06:22:18 pm »
I now understand that if everything is on screen the results will be okay.
But, one thing is not clear to me though. Lets say my setting is 2V a div with a 10x probe. Then I am in the "120mV  up to 1.18V +/- 20V" range.  If my signal is within the +/- 20V dc offset but the signal disappears of the screen will the opamps need a recovery time

Yes.

can my result what I read on the screen be still be wrong?

Yes.

The offset voltage is completely unrelated to this problem. The offset voltage is simply a voltage that's added to the input signal to position the waveform vertically on screen (ie. it's what you're changing when you twist the vertical position knob).

The op-amp recovery time (and signal distortion) is only related to how far the signal goes off screen.
 

Offline Performa01

  • Frequent Contributor
  • **
  • Posts: 901
  • Country: at
Re: Strange behaviour of Siglent SDS1104X-E
« Reply #12 on: June 12, 2018, 08:39:09 am »
Offset limits in the data sheet have nothing to do with dynamic range. They simply describe the DC offset range covered by the vertical position controls, hence is the maximum DC offset of an input signal that can be compensated by shifting the trace position.

An SDS1000X-E can handle up to +/-500mV at the ADC input, i.e. after the attenuators.

This means that the distortion free input range is about
+/-500mV up to 118mV/div
+/-5V from 120mV/div to 1.18V/div
+/-50V for 1.2V/div up to 10V/div

These numbers are valid for direct connection and both the vertical sensitivity and max. input voltage range have to be multiplied by the probe attenuation factor in use.

With 10x probes:
+/-5V up to 1.18V/div
+/-50V from 1.2V/div to 11.8V/div
+/-500V for 12V/div up to 100V/div

Thank you Performa01, think I get the point. I did take a look at the manual of the scope and did find other numbers though. I did put them in a screen shot.

I Find:
500uV   up to 118mV +/- 2V
120mV  up to 1.18V +/- 20V
1.2V      up to 10V +/- 200V
Regarding input signal amplitude handling in any DSO, there are three parameters to consider.

1. ADC full scale voltage.
This is usually not specified anywhere but is always +/-5 vertical divisions for the SDS1000X-E series. Only ±4 divisions are visible, hence the full dynamic range of the ADC can only be made visible by zooming out vertically in stop mode of the DSO.

Overdriving the ADC always results in a clean clipping and does not affect the visible part of the signal at all.

2. Input buffer signal amplitude limit.
This is not specified in the datasheets as well and is mainly determined by the clamps required for input overvoltage protection. The SDS1000X-E series is specified for 400Vp at the inputs and this high voltage does not cause any harm even in the 500uV/div range. Here the numbers for undistorted input signals I’ve already published apply:

+/-500mV up to 118mV/div
+/-5V from 120mV/div to 1.18V/div
+/-50V from 1.2V/div up to 10V/div

This also means that the input buffer clamps limit the max. undistorted signal amplitude to even less than what the ADC could handle at 118mV/div vertical gain (500mV vs. 590mV), whereas it allows a massive overdrive of the ADC at e.g. 1mV/div (500mV vs. 5mV) for example.

Overdriving the input buffer would normally also only lead to a clean clipping, hence do no harm for the visible part of the waveform, but because of the split-path design of higher bandwidth frontends, there is a symmetrical clipping only for the low frequency portion of the input signal, but it’s a differentiation for the high frequency part. If the two signal portions are combined together again at the output of the split path buffer, severe signal distortions will even affect the visible part of the signal. The effect is frequency dependant and clearly worst near the crossover frequency of the split path input buffer. Distortions are also much more malign if the input overload is asymmetrical, for this causes an offset shift for the HF portion of the signal.

Btw, all this has nothing to do with overload recovery. Recovery times for the various discrete components involved are in the picoseconds or nanosecond range at most, and in the microseconds range for the slow OpAmps used in the DC-path of the input buffer. In contrast, the distortions we see are in the milliseconds range. This is also why it is not obvious when looking at a signal at 500ns/div, because then the error appears to be constant and static. It is the before mentioned offset shift in the HF part of the signal.

3. Offset compensation range.
This is what is specified in the data sheets, e.g. the +/-2V up to 118mV/div. It is simply the range for the vertical position control, which in turn just applies a DC offset to the input buffer in order to alter the trace position and can also be used to compensate for the DC offset of a DC-coupled input signal. While this can be used to center a signal within the permissible signal range of the input buffer and avoid asymmetrical overdrive, it certainly doesn’t tell anything about the clamp level and the signal amplitude where distortions start to emerge.


I hope the explanations above answer your remaining questions as well.


Another question, whats the best option if I want a higher resolution to see whats happening on a signal with a high dc-offset then what are my options? A differential probe? I did take a look at some of them but most of them are 50x and I am more interested in a signal with a high dc-offset which can change between 0.4 to 1.2 Volt.
For high DC offsets, where the signal frequency is above some 3Hz and the absolute voltages aren’t important (like measuring ripple and noise on a supply rail for instance), AC coupling is the best option – and the only one, if the offset compensation range in DC-coupled mode is not sufficient.

Voltages (voltage differences) that aren’t ground related, i.e. floating, are best observed with differential probes indeed. But even this would only enable the measurement of a floating diode forward drop as long as the common mode voltage (you could also call it offset here) does not change faster than the bandwidth of the probe. In your case, you had the flyback voltage ringing from some inductor which is clearly in the MHz range. So the probe needs to be able to handle a violent signal like that in common mode. Please be aware that the quality of a differential probe measurement is closely related to the common mode rejection ratio (CMRR) of the probe and this parameter is frequency dependant. So you need to make sure that the probe in question specifies a high enough CMRR for frequencies of 10MHz for example.
 
The following users thanked this post: bhishmar


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf