Rigol MSO 5000 overshoot, or not overshoot?

[kehall]

I'm interested in finding out what other users experiences are with overshoots on the Rigol MSO5000 series...

When probe (10x) is connected to the built-in probe compensation output and trimmed to give a flat response at 500mV/div (10x attenuation set), adjust the vertical to 50mV/div.

To provide the same starting point, press Default and then Auto when the probe is connected.

Which trace best represents what you see on each channel, Ch1 (yellow, with overshoot blip) or Ch2 (blue, flat)?

[Fungus]

"Overshoot" may be due to overloading the opamps when you send the signal so far off screen.

[kehall]

"Overshoot" may be due to overloading the opamps when you send the signal so far off screen.

Not really I don't think, you can see the overshoots at higher settings (100-500mV) but not as clearly, the 'zoom in' just helps show it...

[Fungus]

Either way, it's not an amount that would worry me and I don't think it indicates a problem with the 'scope.

Wait, I think I'm confused about what I'm looking at.

(it happens)

[tautech]

Either way, it's not an amount that would worry me and I don't think it indicates a problem with the 'scope.

It's more than the max 50mV sensitivity of that FNIRSI thing. 

[Verticon]

Just some pictures from my Rigol 5074/5354 for comparison. The first one is for channels 1 and 2 and the second for channels 3 and 4. For both records I used the same probe pair which I adjusted for channels 1 and 2. At least for these settings I definitely see no overshoot and all 4 channels seem to behave similarly.

[Fungus]

Either way, it's not an amount that would worry me and I don't think it indicates a problem with the 'scope.

It's more than the max 50mV sensitivity of that FNIRSI thing. 

Why are you obsessed with the FNIRSI thing?

[Fungus]

"Overshoot" may be due to overloading the opamps when you send the signal so far off screen.

Not really I don't think, you can see the overshoots at higher settings (100-500mV) but not as clearly, the 'zoom in' just helps show it...

Unless all the signal is on screen then you're overloading the opamps and things like this can happen during their recovery time.

[Fungus]

Just some pictures from my Rigol 5074/5354 for comparison. The first one is for channels 1 and 2 and the second for channels 3 and 4. For both records I used the same probe pair which I adjusted for channels 1 and 2. At least for these settings I definitely see no overshoot and all 4 channels seem to behave similarly.

All your signal is on screen. You need to use a signal with a much higher peak-to-peak voltage.

[kehall]

"Overshoot" may be due to overloading the opamps when you send the signal so far off screen.

Not really I don't think, you can see the overshoots at higher settings (100-500mV) but not as clearly, the 'zoom in' just helps show it...

Unless all the signal is on screen then you're overloading the opamps and things like this can happen during their recovery time.

I don't think it matters when the data can be manipulated to produce the screen image - are you saying that the opamps are involved when adjusting offset etc? I doubt it - it'll just be a display routine coded accordingly to show the data according to the required parameters surely?

Anyway overshoots can be seen when all the signal is on screen so it doesn't agree with your suggestion :



Anyway as to it being not significant, if you're investigating a signal that HAS overshoots, how are you supposed to distinguish/make measurements or otherwise trust the display if it can't show a flat response?

Either the scope is factory correct or it's not, would you say this is right?

[kehall]

Just some pictures from my Rigol 5074/5354 for comparison. The first one is for channels 1 and 2 and the second for channels 3 and 4. For both records I used the same probe pair which I adjusted for channels 1 and 2. At least for these settings I definitely see no overshoot and all 4 channels seem to behave similarly.

I think you're in 1x mode there with 10x probes, so you're really at 500mV there as there's no scaling... set attenuation to 10x and go to 50mV (or 5mV staying at 1x)

[Fungus]

are you saying that the opamps are involved when adjusting offset etc?

Yes.

The ADC is 8-bits so apamps are used to scale/offset the signal to fit that range.

[2N3055]

Offset Range
±1 V (500 μV/div~50 mV/div)
±30 V (51 mV/div~260 mV/div)
±100 V (265 mV/div~10 V/div)
that is at 1x probe setting..

But all that is not important.  Analog offset is used only for a signal that has larger DC offset and smaller AC component. So you set offset and channel sensitivity until ALL of the signal is on the screen..
Dynamic range of scope is a bit more than screen.  So if you overdrive input amplifier by signal that would be 20 screens high at that v/div and you're looking at only bottom part of it, you will see overdrive recovery artefacts...

This is how it looks on a very expensive R&S RTM3000:

https://farm1.staticflickr.com/967/40889430405_8f59935293_c.jpg

(from Nctnico's review here : https://www.eevblog.com/forum/testgear/rohde-schwarz-rtm3000-review/msg1604185/#msg1604185)

[kehall]

I think we could be getting a little out of scope here (excuse the pun)

These overshoots are visible when all the trace is on screen per my post above https://www.eevblog.com/forum/testgear/rigol-mso-5000-overshoot-or-not-overshoot/?action=dlattach;attach=1029846;image;

They scale perfectly when altering vertical (yes some signal off screen then but I wouldn't call this particular phenomenon overdriving the amp).

Channel 2 blue trace, does NOT exhibit any overshooting at all right down to the lowest mV/div. The first image in this post shows both channels to the same source, same settings (except offset to compare), yet show completely different visual trace. Ch3 and 4 on that scope were also 'flat'.

... image below with both channels same source, same settings, no offset, overlaid, no difference if probes swapped.



I know there are other MSO5xxx scopes (I had one briefly before it developed another fault) which do not have thes issues. The whole idea of this poll is to determine what other owners MSO5xxx scopes are like, I feel this is down to inadequate factory calibration.

Here's the CH2 blue trace all on screen to compare with CH1 in earlier post.. apologies the scaling wasn't set to 10x so it says 50mV but really 500mV/div

[Fungus]

yes some signal off screen then but I wouldn't call this particular phenomenon overdriving the amp

The calibration signal is 3V peak to peak and you only have 0.2V on screen. It's overloading the amp.

Does it happen when all the signal is on screen?

[kehall]

yes some signal off screen then but I wouldn't call this particular phenomenon overdriving the amp

The calibration signal is 3V peak to peak and you only have 0.2V on screen. It's overloading the amp.

Does it happen when all the signal is on screen?

yes in the post you replied to earlier... https://www.eevblog.com/forum/testgear/rigol-mso-5000-overshoot-or-not-overshoot/msg3151988/#msg3151988

[Sighound36]

It is odd that some of these suffer form this and others do not, both of our 5000 do not have this issue, neither do the 7000 or 8000's.

Maybe a correlation or FW and SW may yield some common ground?

[2N3055]

yes some signal off screen then but I wouldn't call this particular phenomenon overdriving the amp

The calibration signal is 3V peak to peak and you only have 0.2V on screen. It's overloading the amp.

Does it happen when all the signal is on screen?

yes in the post you replied to earlier... https://www.eevblog.com/forum/testgear/rigol-mso-5000-overshoot-or-not-overshoot/msg3151988/#msg3151988

Sorry we are bothering you but photos are screwed up....If I squint on thumbnail i can see it, but enlarged photo is something else...
Try inserting photos inline-full size..

If all the signal is on the screen and there is still over/undershoot and signal is not so fast to prompt Gibbs artefacts, than scope is not right..

[kehall]

Sorry we are bothering you but photos are screwed up....If I squint on thumbnail i can see it, but enlarged photo is something else...
Try inserting photos inline-full size..

If all the signal is on the screen and there is still over/undershoot and signal is not so fast to prompt Gibbs artefacts, than scope is not right..

Interesting.. I was getting the same image issue in another thread from someone else's images (noreply in the 5000 bugs thread)... I think there's probably some web cache/CDN broken somewhere along the line?

I have changed the post to "attachimg=" rather than "attach=", will update the others too. Had to actually re-upload too.. So maybe eevblog needs an ANALYZE / REPAIR TABLES or similar if the images DB is getting messed up

UPDATE: I had to edit the post with the full size ch2 blue trace again and re-upload as I checked on another browser and it was FUBARed so clearly there's something weird going on with image uploading/caching (everything looked ok with original browser having cached the images)

[kehall]

It is odd that some of these suffer form this and others do not, both of our 5000 do not have this issue, neither do the 7000 or 8000's.

Have you voted in the poll?

[Sighound36]

Yes I have now Kehall 

[tv84]

The forum continues with big problems in the attachments but I've seen no msg from Dave about what they are doing to solve it...

(sorry for OT)

[Fungus]

yes some signal off screen then but I wouldn't call this particular phenomenon overdriving the amp

The calibration signal is 3V peak to peak and you only have 0.2V on screen. It's overloading the amp.

Does it happen when all the signal is on screen?

yes in the post you replied to earlier... https://www.eevblog.com/forum/testgear/rigol-mso-5000-overshoot-or-not-overshoot/msg3151988/#msg3151988

That image wasn't there earlier.

Still looks like ordinary probe compensation problems to me. I find it very, very hard to believe that it could overshoot like that on a 1kHz square wave.

nb. Compensation signals aren't usually very good square waves and you won't get a perfect square corner using one. Tweak it so there's no "overshoot" then go look at a better signal.

[kehall]

Still looks like ordinary probe compensation problems to me. I find it very, very hard to believe that it could overshoot like that on a 1kHz square wave.

nb. Compensation signals aren't usually very good square waves and you won't get a perfect square corner using one. Tweak it so there's no "overshoot" then go look at a better signal.

Trust me it's not a compensation issue - the slope is far too abrupt... see this when adjusted...

[Weston]

This looks like a HF compensation issue. Does it change when you swap probes between channels? Have you been adjusting the HF compensation trimmers in addition to the LF compensation trimmer?

Most probes have more than one compensation trimmer, the additional ones determine the wave shape for the first few hundred ns or so of the compensation signal.