Low Cost PCB's Low Cost Components

Author Topic: Measuring O-scope - Flatness?  (Read 3079 times)

0 Members and 1 Guest are viewing this topic.

Online metrologist

  • Frequent Contributor
  • **
  • Posts: 785
  • Country: us
Measuring O-scope - Flatness?
« on: November 10, 2016, 06:31:07 AM »
I have access to a leveled sweep signal generator so I hooked it up to my Rigol DS1054Z sweeping 0 to 200 MHz.

RF to Channel 1 via BNC
0-10V sweep ramp to channel 2

Set scope to X-Y mode and adjusted the channel sensitivities to maximize the display. I get a sideways funnel type image, but is not smooth. For example, the amplitude rises around 10 MHz then drops around 40 MHz, then rises around 60 MHz, then falls around 100 MHz and is a little flat to about 150 MHz, then gradually falls to 200 MHz. I was expecting something more like the shape of a bullet.

Where do I take my initial reading to determine the 3db rolloff point? Max is around 10 MHz, but 25 to 75 MHz is fairly flat.

What would be a good setup for this and are there any other interesting measurements I can make using the sig gen?
 

Online metrologist

  • Frequent Contributor
  • **
  • Posts: 785
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #1 on: November 10, 2016, 08:13:24 AM »
I changed my setup just to trigger and long sweep. That does not look very flat  :-//

And then I found some modulation...
 

Offline barry14

  • Regular Contributor
  • *
  • Posts: 74
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #2 on: November 10, 2016, 09:32:14 AM »
Did you terminate the cable going to the scope with a 50 ohm feedthru?  At these frequencies, proper impedance levels must be observed if you expect flat responses.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 5565
  • Country: gb
    • Having fun doing more, with less
Re: Measuring O-scope - Flatness?
« Reply #3 on: November 10, 2016, 10:12:55 AM »
Don't forget to plot it on log-log scales, i.e. dB and decades. The same data plotted on lin-lin axes looks very different.

In other words, give numbers, not adjectives :)
There are lies, damned lies, statistics - and ADC/DAC specs.
Gliding 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 newbrain

  • Frequent Contributor
  • **
  • Posts: 366
  • Country: se
Re: Measuring O-scope - Flatness?
« Reply #4 on: November 10, 2016, 10:28:59 AM »
Read the top line on your first screenshot:
5 Ms/s

So, that's good till 2.5MHz, that is to say meaningless (a minor miracle something vaguely resembling the real signal is shown at all)

Is the 2nd screenshot the same signal? That's more difficult to explain.

As for XY display, this being a DSO, you'll have sampling artifacts as in in the 1st picture, depending on the settings.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 5565
  • Country: gb
    • Having fun doing more, with less
Re: Measuring O-scope - Flatness?
« Reply #5 on: November 10, 2016, 07:50:09 PM »
Read the top line on your first screenshot:
5 Ms/s
So, that's good till 2.5MHz, that is to say meaningless (a minor miracle something vaguely resembling the real signal is shown at all)

If there's no modulation, then not necessarily - it is ultimately limited by the analogue front end.

To put it simply, if I have a 1MHz bandwidth signal on a 1GHz carrier, then I only need to sample at >2MS/s. However, the sampling head has to have useful behaviour at 1GHz.
There are lies, damned lies, statistics - and ADC/DAC specs.
Gliding 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 newbrain

  • Frequent Contributor
  • **
  • Posts: 366
  • Country: se
Re: Measuring O-scope - Flatness?
« Reply #6 on: November 10, 2016, 08:09:28 PM »
Read the top line on your first screenshot:
5 Ms/s
So, that's good till 2.5MHz, that is to say meaningless (a minor miracle something vaguely resembling the real signal is shown at all)

If there's no modulation, then not necessarily - it is ultimately limited by the analogue front end.

To put it simply, if I have a 1MHz bandwidth signal on a 1GHz carrier, then I only need to sample at >2MS/s. However, the sampling head has to have useful behaviour at 1GHz.
On paper, yes, no doubts about it, and that's, in fact the principle ETS DSO are based upon, if I'm not mistaken:
if the signal is repetitive, and band limited you can under sample it and obtain a correct reconstruction.

But here we have a signal that varies in time and I can expect the result to depend on interference between the signal period, the waveform update rate and the scan blind time.
So I would not trust this kind of results.

Am I completely off-track? I admit my digital signal theory is a bit rusty, to say the least, so thanks for any clarification.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 5565
  • Country: gb
    • Having fun doing more, with less
Re: Measuring O-scope - Flatness?
« Reply #7 on: November 10, 2016, 08:24:34 PM »
Read the top line on your first screenshot:
5 Ms/s
So, that's good till 2.5MHz, that is to say meaningless (a minor miracle something vaguely resembling the real signal is shown at all)

If there's no modulation, then not necessarily - it is ultimately limited by the analogue front end.

To put it simply, if I have a 1MHz bandwidth signal on a 1GHz carrier, then I only need to sample at >2MS/s. However, the sampling head has to have useful behaviour at 1GHz.
On paper, yes, no doubts about it, and that's, in fact the principle ETS DSO are based upon, if I'm not mistaken:
if the signal is repetitive, and band limited you can under sample it and obtain a correct reconstruction.

That's correct; I first saw it done without a scope, where the output device was an XY pen plotter :)

In addition "subsampling" is often used in "digital" radios with non-repetitive signals, which is a more direct example of the principle.

Quote
But here we have a signal that varies in time and I can expect the result to depend on interference between the signal period, the waveform update rate and the scan blind time.
So I would not trust this kind of results.

Quite possibly, but not necessarily. I would want to play with the scope and signal before deciding what's happening.

If there analogue front end isn't the limiting factor, then the key issues will be the sweep speed and any digital processing of the sampled signal.
There are lies, damned lies, statistics - and ADC/DAC specs.
Gliding 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
 
The following users thanked this post: newbrain

Offline newbrain

  • Frequent Contributor
  • **
  • Posts: 366
  • Country: se
Re: Measuring O-scope - Flatness?
« Reply #8 on: November 10, 2016, 09:07:03 PM »
Quote
But here we have a signal that varies in time and I can expect the result to depend on interference between the signal period, the waveform update rate and the scan blind time.
So I would not trust this kind of results.

Quite possibly, but not necessarily. I would want to play with the scope and signal before deciding what's happening.
ng of the sampled signal.
Ok, then I think we agree: we can not conclude much form that screenshot.

(Admittedly, I was a bit too dry in the first post. I should not answer when I'm >50% asleep... :-[)

As for the analog bandwidth: http://www.eevblog.com/forum/testgear/new-rigol-ds1054z-oscilloscope/msg743080/#msg743080 might be an indication that it should not be so bumpy (though I don't know how it was measured).

I'm still puzzled by the second picture, though: is that an artifact of the persistence + changing frequency + wfm/s ?
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 5565
  • Country: gb
    • Having fun doing more, with less
Re: Measuring O-scope - Flatness?
« Reply #9 on: November 11, 2016, 12:25:07 AM »
(Admittedly, I was a bit too dry in the first post. I should not answer when I'm >50% asleep... :-[)

I've never done that. Oh no. Never ever :)

But is it always difficult to know how many "escape clauses" should be put in a quick statement!
There are lies, damned lies, statistics - and ADC/DAC specs.
Gliding 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 Kalvin

  • Super Contributor
  • ***
  • Posts: 1431
  • Country: fi
  • Embedded SW/HW.
Re: Measuring O-scope - Flatness?
« Reply #10 on: November 11, 2016, 02:03:37 AM »
Did you calibrate the probe using the oscilloscope's internal square wave generator?
 

Online metrologist

  • Frequent Contributor
  • **
  • Posts: 785
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #11 on: November 11, 2016, 04:33:24 AM »
I did not use a 50 ohm term. BNC to BNC. Yeah, I realized I missed my 50 ohm feedthrough termination though.

In the first image, the right side is 200 MHz. I see there could be a problem with a 5MS/s there.

I also tried to take a bunch of screenshots by adjusting the time base for each frequency to give a couple cycles on the display, and then compare p-p voltage at say 10 MHz intervals, but I had not noticed that the scope was telling me that screen save failed! (Aside the point, during this testing, I noticed that my menu flyout's were not happening.  :-BROKE )

The second image is after I found some modulation, AM saw wave...

Anyway, I have access to a sweep signal generator and just wanted to test out my scope.
 

Offline Keysight DanielBogdanoff

  • Frequent Contributor
  • **
  • Posts: 485
  • Country: us
  • ALL THE SCOPES!
    • Keysight Scopes YouTube channel
Re: Measuring O-scope - Flatness?
« Reply #12 on: November 11, 2016, 10:14:18 AM »
Instead of sweeping, you could do it more manually and measure the average Vpk-pk at increasingly higher frequencies and plot it. This allows you to change the time/div scaling to optimize sample rate. It's not uncommon for oscilloscopes to have a non-flat frequency response, as long as it doesn't dip below -3dB it's technically ok. That being said, we design scopes to have as flat of a frequency response as possible up to the bandwidth point.

Online metrologist

  • Frequent Contributor
  • **
  • Posts: 785
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #13 on: November 11, 2016, 11:34:30 AM »
Thanks, Daniel. That's what I tried to do later, but had not noticed the scope was not saving screen grabs. I don't recall if Rigol even specs flatness or if it must be determined from absolute accuracy.
 

Offline vk6zgo

  • Super Contributor
  • ***
  • Posts: 3726
  • Country: au
Re: Measuring O-scope - Flatness?
« Reply #14 on: November 11, 2016, 01:36:38 PM »
Read the top line on your first screenshot:
5 Ms/s
So, that's good till 2.5MHz, that is to say meaningless (a minor miracle something vaguely resembling the real signal is shown at all)

If there's no modulation, then not necessarily - it is ultimately limited by the analogue front end.


To put it simply, if I have a 1MHz bandwidth signal on a 1GHz carrier, then I only need to sample at >2MS/s. However, the sampling head has to have useful behaviour at 1GHz.
On paper, yes, no doubts about it, and that's, in fact the principle ETS DSO are based upon, if I'm not mistaken:
if the signal is repetitive, and band limited you can under sample it and obtain a correct reconstruction.

But here we have a signal that varies in time and I can expect the result to depend on interference between the signal
period, the waveform update rate and the scan blind time.
So I would not trust this kind of results.

Am I completely off-track? I admit my digital signal theory is a bit rusty, to say the least, so thanks for any clarification.

ETS works well for simple signals but for  complex ones.it falls flat on its face.
The classic case,which I quote ad nauseum is that of trying to look at  one field of a PAL video signal using an early '90s DSO.
To see one field,it is necessary to display  20ms worth of signal on the screen.

At the required time/div settings,the sample rate on these devices reduces to one not quite sufficient to reproduce a 15kHz line rate signal,& totally insufficient to reproduce  4.434MHz colour information,or 5MHz HF luma.

The resulting display is  totally unuseable.

In the OP's case,using a far more capable instrument,the DSO sees the signal as a very low frequency one,with a huge range of frequencies superimposed upon it.
It is stretching the bounds of friendship to ask the poor DS1054Z to sample  & store at its native maximum sampling rate  over such a large time.
It will run out of memory!

It does the obvious thing & reduces the sample rate to one which is still quite reasonable for most signals that would be viewed at this time/div setting.
The anti-aliasing filters don't do anything until the top end of the range,so "funnies" happen  after the sweep approaches & then passes 5MHz.

People have checked the frequency response of their DS1054Z's "retro-style" using a series of "spot " frequencies,without the result looking anything like the OP's screenshot.
 
The following users thanked this post: newbrain

Online rstofer

  • Super Contributor
  • ***
  • Posts: 3117
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #15 on: November 11, 2016, 02:14:00 PM »
[
People have checked the frequency response of their DS1054Z's "retro-style" using a series of "spot " frequencies,without the result looking anything like the OP's screenshot.

I think it's important to get this into the thread lest newcomers see the traces and just assume the problem is with the scope.
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 5565
  • Country: gb
    • Having fun doing more, with less
Re: Measuring O-scope - Flatness?
« Reply #16 on: November 11, 2016, 09:31:55 PM »
[
People have checked the frequency response of their DS1054Z's "retro-style" using a series of "spot " frequencies,without the result looking anything like the OP's screenshot.

I think it's important to get this into the thread lest newcomers see the traces and just assume the problem is with the scope.

Well, the problem is with the scope, but it doesn't mean the scope is at fault! Such issues can occur with any measuring instrument, if its operation and limitations aren't understood.

(In general I don't think it is fair to punt the issue and say the "problem is with the operator" - although sometimes that is the inescapable deduction!)
There are lies, damned lies, statistics - and ADC/DAC specs.
Gliding 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 JPortici

  • Super Contributor
  • ***
  • Posts: 1749
  • Country: it
Re: Measuring O-scope - Flatness?
« Reply #17 on: November 11, 2016, 09:41:01 PM »
but had not noticed the scope was not saving screen grabs.

pendrive? check the size. not specified if not in the troubleshooting guide :palm: but size > 2GB will not work.. or not work reliably (i have funny stories about that)
 

Online rstofer

  • Super Contributor
  • ***
  • Posts: 3117
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #18 on: November 12, 2016, 03:43:09 AM »
[
People have checked the frequency response of their DS1054Z's "retro-style" using a series of "spot " frequencies,without the result looking anything like the OP's screenshot.

I think it's important to get this into the thread lest newcomers see the traces and just assume the problem is with the scope.

Well, the problem is with the scope, but it doesn't mean the scope is at fault! Such issues can occur with any measuring instrument, if its operation and limitations aren't understood.

(In general I don't think it is fair to punt the issue and say the "problem is with the operator" - although sometimes that is the inescapable deduction!)

I thought the issue reduced to something bizarre with the signal generator.  Is that not the case?

We know full well that the scope doesn't roll off to -3dB until somewhere around 150 MHz.  This has been documented and there's a video out there somewhere.  If I had a signal generator, I would simply duplicate the test using a simple sine wave and varying the frequency.

I was pretty careful to avoid the term 'operator error'.  I have no idea why the display is showing what it is except that I am not quite ready to throw the scope under the bus given that the signal wasn't even terminated.  And where did the modulation term come from?

I must be missing something.  There is nothing in the discussion that makes me believe it is a scope problem.  But then there is the issue of 'confirmation bias'.  Maybe I really did miss something!

 

Online metrologist

  • Frequent Contributor
  • **
  • Posts: 785
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #19 on: November 12, 2016, 04:29:02 AM »
I have access to a leveled sweep signal generator so I hooked it up to my Rigol DS1054Z sweeping 0 to 200 MHz.

What would be a good setup for this and are there any other interesting measurements I can make using the sig gen?

I made no claim about my screen grabs other than they weren't excatly what I was expecting, so that's why I switched my setup. I just forgot my 50 ohm load, but what was I gonna do, not even hook it up?

The pen drive is 1GB. I think something else happened to stop it from saving. There was a partially written file name with ~ in it. That's probably also when my menu flyout's stopped working. Anyway, I got rid of that file and the scope saved fine after that.

 

Online metrologist

  • Frequent Contributor
  • **
  • Posts: 785
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #20 on: November 12, 2016, 04:37:38 AM »
And where did the modulation term come from?
I must be missing something.  There is nothing in the discussion that makes me believe it is a scope problem.  But then there is the issue of 'confirmation bias'.  Maybe I really did miss something!

Nothing here, really. I just brought my scope to work where we have a very nice sig gen. I had an idea to measure something. Perhaps the method is flawed. Just take it at face value, I hooked the stuff up, set it up this way, and here is what it looked like.

I learned anyway that there was not enough samples the way I had it set up to see enough of the higher frequency signal.

After that, I just thought a saw tooth AM modulation looked cool. I have one of sine wave AM too, but that's been done before.  ???
 

Online tggzzz

  • Super Contributor
  • ***
  • Posts: 5565
  • Country: gb
    • Having fun doing more, with less
Re: Measuring O-scope - Flatness?
« Reply #21 on: November 12, 2016, 06:28:09 AM »
[
People have checked the frequency response of their DS1054Z's "retro-style" using a series of "spot " frequencies,without the result looking anything like the OP's screenshot.

I think it's important to get this into the thread lest newcomers see the traces and just assume the problem is with the scope.

Well, the problem is with the scope, but it doesn't mean the scope is at fault! Such issues can occur with any measuring instrument, if its operation and limitations aren't understood.

(In general I don't think it is fair to punt the issue and say the "problem is with the operator" - although sometimes that is the inescapable deduction!)

I thought the issue reduced to something bizarre with the signal generator.  Is that not the case?

We know full well that the scope doesn't roll off to -3dB until somewhere around 150 MHz.  This has been documented and there's a video out there somewhere.  If I had a signal generator, I would simply duplicate the test using a simple sine wave and varying the frequency.

I was pretty careful to avoid the term 'operator error'.  I have no idea why the display is showing what it is except that I am not quite ready to throw the scope under the bus given that the signal wasn't even terminated.  And where did the modulation term come from?

I must be missing something.  There is nothing in the discussion that makes me believe it is a scope problem.  But then there is the issue of 'confirmation bias'.  Maybe I really did miss something!

My understanding, and I don't think it has been confirmed, is that there is a "missing" 50ohm cable termination and possibly something to do with the way the scope captures/processes the samples. Any such capture/processing problem may well be understandable and predictable; if so there is no fault in the scope, but there is a problem with it. You can get such problems when using any piece of equipment; that is why it is so valuable to "play around" with the settings to see if you can understand what's happening.

An example of another such problem with a scope. With Tek 24x5s and the cal out, when you change the timebase the displayed trace remains unaltered. That's because the scope adjusts the cal ut frequency so there are 5 cycles on the display. The scope isn't faulty, its behaving in an unexpected way is a problem.
There are lies, damned lies, statistics - and ADC/DAC specs.
Gliding 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
 

Online _Wim_

  • Frequent Contributor
  • **
  • Posts: 381
  • Country: be
Re: Measuring O-scope - Flatness?
« Reply #22 on: November 12, 2016, 10:28:09 PM »
You can make a nice frequency response graph by using the graph function on the measurements. Here is a graph I made from my "upgraded" DS1054Z trending Vrms and Frequency using the frequency generator of my CMU200 and the FreRes software. Scope has a 50ohm adaptor on its inputs. For reference, same cable directly to CMU200 input. As can be seen by the plots, -3db point is well over200 MHz! Not bad for a 50Mhz scope...
 

Online rstofer

  • Super Contributor
  • ***
  • Posts: 3117
  • Country: us
Re: Measuring O-scope - Flatness?
« Reply #23 on: November 13, 2016, 02:42:53 AM »
Nice test!

I can see the scope has yet another capability I had never seen.  I really should RTFM.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf