Author Topic: DS1054Z distortion issue?  (Read 16871 times)

0 Members and 1 Guest are viewing this topic.

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 9655
  • Country: 00
Re: DS1054Z distortion issue?
« Reply #75 on: April 30, 2016, 10:50:24 pm »
You are not the first person to run into such problems.

Famously Jim Williams was very interested in equivalent problems, and developed ways to circumvent them. His writings are widely available on the web. Look out for his measurements of opamp's settling time.

You will note, and DSO proponents will rail against, his opinions of modern scope's performance in this respect.

People who agree with that opinion are free to find old oscilloscopes and use them.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 16943
  • Country: nl
    • NCT Developments
Re: DS1054Z distortion issue?
« Reply #76 on: April 30, 2016, 10:54:30 pm »
It is probably also a niche thing what I am trying to do (scopes have 3% vertical accuracy any way) and so probably an application for higher-resolution ADC instruments like higher-end bench scopes, USB scopes etc. In that case I don't have to overdrive inputs to be able to capture the minuscule details. I don't have a 10-bit or better scope either, so instead I built a solution my self. Building test equipment is both fun and also an learning experience for a whole range of skill sets.
One solution to your problem is to create a buffer amplifier which can also the input signal. IIRC Analog Devices has amplifiers with signal clipping inputs. Not exactly cheap but still affordable. Otherwise a couple of zener diodes (after a buffer amplifier) may do the job.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline hans

  • Super Contributor
  • ***
  • Posts: 1032
  • Country: nl
Re: DS1054Z distortion issue?
« Reply #77 on: April 30, 2016, 11:40:38 pm »
Having to fiddle 6 knobs to make 1 measurement, and then doing it all over again 2 minute later after a change is not productive. That is my view on it anyway - I'm fine turning all the knobs a hundred times on an oscilloscope to find, lock and explain the details I am looking for (I disgust auto set), but not when I'm measuring 1 quantity repeatably in a very limited changing environment while still needing to make adjustments back and forth for the scope to work. This level of interface is maybe acceptable for some, but not for me.

The reality of manufacturing/marketing is that you're an edge case. This problem could be solved in entry level DSOs but at the risk of going out of business because the resulting instruments would be unsalable to most people (they'd cost a lot more).

If you need this function a lot and all the fiddling is costing you money then you either need a more specialized oscilloscope or some sort of front-end adapter to clamp the signals before they even reach the oscilloscope (a fun project?)

I can relate to that. As pointed out, what I'm trying to do is quite niche and can be argued over if this capability should be supported or not (I certainly wish it was). However it likely won't because if you were to support every little niche feature customers require you will end up adding menu buttons, options and knobs at every corner of the menu/screen. It quickly becomes a mess. Hardware specifications get out of hand, product becomes too hard/expensive to engineer or produce, etc.

The place where I used to work every menu button change at 1st or 2nd level was heavily discussed and criticized by service & product engineers, which I think is a good thing. This resulted in some frustration of other colleagues which already promised some features and changes to customers, that did not happen.

I think nctnico also suggests a similar thing, like a fast-signal clamp board that you can safely overdrive but will recover in e.g. <1us or faster so influence on measurements is minimal. That sounds like a nice project idea, maybe something I could work on to solve this once and for all :)
« Last Edit: May 01, 2016, 12:24:54 am by hans »
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 9189
  • Country: gb
    • Having fun doing more, with less
Re: DS1054Z distortion issue?
« Reply #78 on: May 01, 2016, 12:13:45 am »
But that isn't what's being discussed here. This discussion was about a problem that potentially affects all DSOs.
There's a couple of people who try to hijack every oscilloscope related thread and turn it into a DS1054Z-hating thread. I don't see any need for it.

Overload recovery affects more than just DSOs; most analogue scopes have similar problems.

Hence it is a poor argument against the DS1054Z.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 9655
  • Country: 00
Re: DS1054Z distortion issue?
« Reply #79 on: May 01, 2016, 12:48:40 am »
Overload recovery affects more than just DSOs; most analogue scopes have similar problems.

It's not just oscilloscopes either, it's every op-amp ever made.

FWIW the DS1054Z says "300V RMS CAT I" on the front.

That suggests it's been designed for up to 300V RMS on every input range. It should also withstand very low energy 1500V transients.



Would anybody disagree with that, based on Dave's schematic?
« Last Edit: May 01, 2016, 12:53:31 am by Fungus »
 

Offline Fungus

  • Super Contributor
  • ***
  • Posts: 9655
  • Country: 00
Re: DS1054Z distortion issue?
« Reply #80 on: May 01, 2016, 01:13:22 am »
Here's the front end from the DS1054Z.

The only part that's switchable is the part labelled "Attenuator" in the schematic. This part can be completely switched out with a relay. When switched out, the input will go directly to point I've labelled "A".

It follows that everything beyond point "A" can therefore withstand 300V:



So ... the only question is whether or not the capacitors in the Attenuator are rated for a 300V input. That doesn't seem expensive to do and I don't see why they wouldn't be.

« Last Edit: May 01, 2016, 01:25:52 am by Fungus »
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 9189
  • Country: gb
    • Having fun doing more, with less
Re: DS1054Z distortion issue?
« Reply #81 on: May 01, 2016, 02:59:47 am »
I think nctnico also suggests a similar thing, like a fast-signal clamp board that you can safely overdrive but will recover in e.g. <1us or faster so influence on measurements is minimal. That sounds like a nice project idea, maybe something I could work on to solve this once and for all :)

What will be the clamp voltages? And will that prevent problems with internal nodes in the scope entering saturation?
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline pxl

  • Regular Contributor
  • *
  • Posts: 126
  • Country: hu
Re: DS1054Z distortion issue?
« Reply #82 on: May 01, 2016, 07:55:25 am »
I've made a small investigation, now from the opposite viewpoint: I checked the max voltage for every (well, most of) vertical divisions, where the smallest distortions starts to appear (I stopped and then zoomed in to check the details, HiRes ;)). The signal is 10 Hz, square wave from 0 to the noted voltage:

vertical div.max voltage
1mV60mV
2mV60mV
5mV60mV
10mV60mV
20mV1.38V
50mV1.37V
100mV1.31V
200mV1.18V
500mV>10V

from 500mV and above I cannot spot any distortions, I have 10V square wave max. With keeping this table in mind, it is very easy to find vertical ranges, where it is almost perfect, and the ranges where it is hardly usable for this purpose.
« Last Edit: May 01, 2016, 07:57:49 am by pxl »
 

Offline pxl

  • Regular Contributor
  • *
  • Posts: 126
  • Country: hu
Re: DS1054Z distortion issue?
« Reply #83 on: May 01, 2016, 08:04:46 am »
If you need this function a lot and all the fiddling is costing you money then you either need a more specialized oscilloscope or some sort of front-end adapter to clamp the signals before they even reach the oscilloscope (a fun project?)

Seems logical, but in reality it would be very impractical. The first step, when measuring this kind of small parts of largest signals, is to view the whole and then zoom in the interested part. This way, we would need a switchable level, which should be always kept synchronized with the internal vertical settings. :scared:

This limiters should be in the scope, I am afraid.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 16943
  • Country: nl
    • NCT Developments
Re: DS1054Z distortion issue?
« Reply #84 on: May 01, 2016, 08:30:23 am »
If you need this function a lot and all the fiddling is costing you money then you either need a more specialized oscilloscope or some sort of front-end adapter to clamp the signals before they even reach the oscilloscope (a fun project?)

Seems logical, but in reality it would be very impractical. The first step, when measuring this kind of small parts of largest signals, is to view the whole and then zoom in the interested part. This way, we would need a switchable level, which should be always kept synchronized with the internal vertical settings. :scared:

This limiters should be in the scope, I am afraid.
No, use 2 channels (one clipped externally). Alternatively an external trigger channel can also be used.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 15093
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: DS1054Z distortion issue?
« Reply #85 on: May 01, 2016, 07:01:55 pm »
I've made a small investigation, now from the opposite viewpoint: I checked the max voltage for every (well, most of) vertical divisions, where the smallest distortions starts to appear (I stopped and then zoomed in to check the details, HiRes ;)). The signal is 10 Hz, square wave from 0 to the noted voltage:

vertical div.max voltage
1mV60mV
2mV60mV
5mV60mV
10mV60mV
20mV1.38V
50mV1.37V
100mV1.31V
200mV1.18V
500mV>10V

from 500mV and above I cannot spot any distortions, I have 10V square wave max. With keeping this table in mind, it is very easy to find vertical ranges, where it is almost perfect, and the ranges where it is hardly usable for this purpose.
@ pxl
Simple checks with the SDS2304
10 Hz square wave (source HiZ SDG1010)
Normal acquisition HiZ

No distortion of the fundamental waveform @ 3V p-p down to 1mV/div (max sensitivity)
No distortion of the fundamental waveform @ 10V p-p down to 200mV/div
Distortion of the fundamental waveform was present @ 4V+ p-p @ 100mV/div and lower

BTW, SDS2304 does not show the clipped off-screen portion of the waveform on the display, there are just the rising edges tracing off the edge of the display.
Avid Rabid Hobbyist
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 3043
  • Country: fi
  • Starting with DLL21
Re: DS1054Z distortion issue?
« Reply #86 on: May 01, 2016, 08:52:54 pm »

BTW, SDS2304 does not show the clipped off-screen portion of the waveform on the display, there are just the rising edges tracing off the edge of the display.

Naturally. ADC is 8bit. It do not clip display top or bottom and so it not also draw false line there like some others.
ADC is 8bit. Display vertical division is 25 ADC levels. Display height is not ADC full scale.
8div (what is Siglent vertical display range isheight) show only 200 from full 256 range.)
Stop scope. Move vertical position up and down, there can see these clips (and straight line where it cut)
But what I hope is that Siglent do small improvement and same as example even in -80 era old HP digital scope. It give warning on the screen about signal clipping. (and also warning about possible aliasing situation)



« Last Edit: May 01, 2016, 09:02:26 pm by rf-loop »
If practice and theory is not equal it tells that used application of theory  is wrong or the theory itself is wrong.
It is much easier to think an apple fall to the ground than to think that the earth and the apple will begin to move toward each other and collide.
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 15093
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: DS1054Z distortion issue?
« Reply #87 on: May 01, 2016, 09:27:35 pm »
But what I hope is that Siglent do small improvement and same as example even in -80 era old HP digital scope. It give warning on the screen about signal clipping. (and also warning about possible aliasing situation)
Where there was confusion early in this thread was with the Rigol displaying this clipped portion on the display, early SDS2000 FW did this too but now this clipped portion is missing in the 2000 series it is quite  (should be) obvious to the user the V/div need be adjusted to display the full waveform. For the 2000 series IMO a warning is not required..... but maybe for the blind.  :palm:
« Last Edit: May 01, 2016, 09:40:34 pm by tautech »
Avid Rabid Hobbyist
 

Offline pxl

  • Regular Contributor
  • *
  • Posts: 126
  • Country: hu
Re: DS1054Z distortion issue?
« Reply #88 on: May 02, 2016, 05:08:24 am »
I've made a small investigation, now from the opposite viewpoint: I checked the max voltage for every (well, most of) vertical divisions, where the smallest distortions starts to appear (I stopped and then zoomed in to check the details, HiRes ;)). The signal is 10 Hz, square wave from 0 to the noted voltage:

vertical div.max voltage
1mV60mV
2mV60mV
5mV60mV
10mV60mV
20mV1.38V
50mV1.37V
100mV1.31V
200mV1.18V
500mV>10V

from 500mV and above I cannot spot any distortions, I have 10V square wave max. With keeping this table in mind, it is very easy to find vertical ranges, where it is almost perfect, and the ranges where it is hardly usable for this purpose.
@ pxl
Simple checks with the SDS2304
10 Hz square wave (source HiZ SDG1010)
Normal acquisition HiZ

No distortion of the fundamental waveform @ 3V p-p down to 1mV/div (max sensitivity)
No distortion of the fundamental waveform @ 10V p-p down to 200mV/div
Distortion of the fundamental waveform was present @ 4V+ p-p @ 100mV/div and lower

Nice, thanks. And after all this long thread, it turned out that there is a scope which seems to be immune to this problem (in the range of our use cases, e.g. with 1:100 probes). Where is the trade-off?
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 15093
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: DS1054Z distortion issue?
« Reply #89 on: May 02, 2016, 06:25:50 am »
And after all this long thread,
Let's make it longer.  ;)

Quote
it turned out that there is a scope which seems to be immune to this problem (in the range of our use cases, e.g. with 1:100 probes).
Problem  :-// 
Isn't it quite clear this style of measurement is outside normal DSO intended functionality and therefore outside what manufacturers might consider incorporating into their designs.

There be a few more as more members contribute. But remember the OP had an old Tek DSO that didn't clip for his measurement needs of what this thread is based on.



Quote
Where is the trade-off?
IMHO, product and FW maturity and front end design.
Avid Rabid Hobbyist
 

Online MarkL

  • Supporter
  • ****
  • Posts: 1484
  • Country: us
Re: DS1054Z distortion issue?
« Reply #90 on: May 02, 2016, 07:12:15 am »
...
I think nctnico also suggests a similar thing, like a fast-signal clamp board that you can safely overdrive but will recover in e.g. <1us or faster so influence on measurements is minimal. That sounds like a nice project idea, maybe something I could work on to solve this once and for all :)
Jim Williams has been referenced several times in this thread with his approach to the overdrive problem.

Here's another couple of solutions from him that are fairly simple.  They are in the venerable AN47, starting on page 39:

  http://www.linear.com/docs/4138

(Although a bit dated, it's worth reading the whole AN47.)
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 16943
  • Country: nl
    • NCT Developments
Re: DS1054Z distortion issue?
« Reply #91 on: May 02, 2016, 07:19:54 am »
I've made a small investigation, now from the opposite viewpoint: I checked the max voltage for every (well, most of) vertical divisions, where the smallest distortions starts to appear (I stopped and then zoomed in to check the details, HiRes ;)). The signal is 10 Hz, square wave from 0 to the noted voltage:

vertical div.max voltage
1mV60mV
2mV60mV
5mV60mV
10mV60mV
20mV1.38V
50mV1.37V
100mV1.31V
200mV1.18V
500mV>10V

from 500mV and above I cannot spot any distortions, I have 10V square wave max. With keeping this table in mind, it is very easy to find vertical ranges, where it is almost perfect, and the ranges where it is hardly usable for this purpose.
Did you set the offset so you can see the flat piece of the square wave or did you look at the edges? In case of the latter you are not likely to find any distortions because when the opamp is driving the edge it has already recoverd from the overdrive. You need to look at the flat parts of the square waves!
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline pxl

  • Regular Contributor
  • *
  • Posts: 126
  • Country: hu
Re: DS1054Z distortion issue?
« Reply #92 on: May 02, 2016, 07:30:08 am »
I've made a small investigation, now from the opposite viewpoint: I checked the max voltage for every (well, most of) vertical divisions, where the smallest distortions starts to appear (I stopped and then zoomed in to check the details, HiRes ;)). The signal is 10 Hz, square wave from 0 to the noted voltage:

vertical div.max voltage
1mV60mV
2mV60mV
5mV60mV
10mV60mV
20mV1.38V
50mV1.37V
100mV1.31V
200mV1.18V
500mV>10V

from 500mV and above I cannot spot any distortions, I have 10V square wave max. With keeping this table in mind, it is very easy to find vertical ranges, where it is almost perfect, and the ranges where it is hardly usable for this purpose.
Did you set the offset so you can see the flat piece of the square wave or did you look at the edges? In case of the latter you are not likely to find any distortions because when the opamp is driving the edge it has already recoverd from the overdrive. You need to look at the flat parts of the square waves!

Yep, this is. From this point up there is no distortions. So 500 mV and 10 Vpp is ok. (1:1 BNC)

 

Offline pxl

  • Regular Contributor
  • *
  • Posts: 126
  • Country: hu
Re: DS1054Z distortion issue?
« Reply #93 on: May 02, 2016, 07:46:10 am »
Quote
Where is the trade-off?
IMHO, product and FW maturity and front end design.

I mean that I can spot 3 amp steps in my table: 1mV-10mV, 20mV-200mV and from 500mV. From your data I suspect, there is only two steps in siglent: one is from 1mV to 100 mV and one step from 200mV. Why the difference? What are the benefits having 3 steps if there is any?
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 16943
  • Country: nl
    • NCT Developments
Re: DS1054Z distortion issue?
« Reply #94 on: May 02, 2016, 07:51:46 am »
I just did a similar test on my GDS2204E. With 10Vpp input it can display the signal correctly up to 50mV/div. IMHO an easy test is to input a 100kHz square wave to channel 1 & 2 (terminate properly!), set the vertical position of both channels to 0 and set the trigger level to 0. Set channel 2 as the trigger source and adjust the volt/div so channel 2 displays the signal. Adjust the timebase so that channel 2 shows a slope instead of a sharp edge. Now increase the sensitivity of channel 1 until the trace of channel 1 shows a significant (>0.2 div) horizontal offset from channel 2. Once that happens you are in overdrive territory. Another good indicator of when an oscilloscope will overdrive is the horizontal offset range. On the GDS2204E the 50mV/div has a +/-5V offset range so it seems it has been designed to not overdrive at 50mV/div when driven with a 10Vpp input.
« Last Edit: May 02, 2016, 07:53:17 am by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 15093
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: DS1054Z distortion issue?
« Reply #95 on: May 02, 2016, 08:13:22 am »
Quote
Where is the trade-off?
IMHO, product and FW maturity and front end design.

I mean that I can spot 3 amp steps in my table: 1mV-10mV, 20mV-200mV and from 500mV. From your data I suspect, there is only two steps in siglent: one is from 1mV to 100 mV and one step from 200mV. Why the difference? What are the benefits having 3 steps if there is any?
Much TE will differ in this respect as overdriven input measurements are not part of normal DSO use.
So any perceived benefits you imagine are just design idiosyncrasies.

There are so many other features to consider when buying a DSO, so don't sell yourself short by focussing entirely on this issue.

I might add, my tests/checks were done using a square waveform with 0V offset ie. all amplitude was above the DSO ch 0V baseline.

Avid Rabid Hobbyist
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 16943
  • Country: nl
    • NCT Developments
Re: DS1054Z distortion issue?
« Reply #96 on: May 02, 2016, 08:17:06 am »
Quote
Where is the trade-off?
IMHO, product and FW maturity and front end design.

I mean that I can spot 3 amp steps in my table: 1mV-10mV, 20mV-200mV and from 500mV. From your data I suspect, there is only two steps in siglent: one is from 1mV to 100 mV and one step from 200mV. Why the difference? What are the benefits having 3 steps if there is any?
Much TE will differ in this respect as overdriven input measurements are not part of normal DSO use.
I beg to differ. IMHO when the input signal is within the offset adjust range there should not be any distortions because the oscilloscope has been designed to be able to handle such signals.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 16943
  • Country: nl
    • NCT Developments
Re: DS1054Z distortion issue?
« Reply #97 on: May 02, 2016, 08:45:14 am »
I think nctnico also suggests a similar thing, like a fast-signal clamp board that you can safely overdrive but will recover in e.g. <1us or faster so influence on measurements is minimal. That sounds like a nice project idea, maybe something I could work on to solve this once and for all :)
What will be the clamp voltages? And will that prevent problems with internal nodes in the scope entering saturation?
a) whatever you want them to be
b) make sure that the clamped signal doesn't exceed the scope's specifications (the answer to a).
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online tautech

  • Super Contributor
  • ***
  • Posts: 15093
  • Country: nz
  • Taupaki Technologies Ltd. NZ Siglent Distributor
    • Taupaki Technologies Ltd.
Re: DS1054Z distortion issue?
« Reply #98 on: May 02, 2016, 05:30:26 pm »
Another good indicator of when an oscilloscope will overdrive is the horizontal offset range. On the GDS2204E the 50mV/div has a +/-5V offset range so it seems it has been designed to not overdrive at 50mV/div when driven with a 10Vpp input.
A check of the SDS2000 manual shows a table with these vertical offset ranges:

Volt Scale Range of Vertical Position

2 mV/div - 100 mV/div   ±1V
102 mV/div - 1 V/div      ±10 V
1.02 V/div - 10 V/div      ±100 V
« Last Edit: May 02, 2016, 06:31:04 pm by tautech »
Avid Rabid Hobbyist
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 16943
  • Country: nl
    • NCT Developments
Re: DS1054Z distortion issue?
« Reply #99 on: May 02, 2016, 06:44:44 pm »
GW Instek GDS2000E offset ranges:
1mV/div ~ 20mV/div : ±0.5V
50mV/div ~ 200mV/div : ±5V
500mV/div ~ 2V/div : ±25V
5V/div ~ 10V/div : ±250V
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf