Author Topic: Rigol DS1000 weird problem  (Read 3977 times)

0 Members and 1 Guest are viewing this topic.

Offline c4757pTopic starter

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Rigol DS1000 weird problem
« on: May 08, 2013, 07:50:47 pm »
(Oh no, not another fix-my-test-equipment from him!)

Actually, I have a pretty good hunch where the problem is. Just wondering if anyone has seen this. Suddenly, I am getting very strong noise on Ch 1 on my DS1052D (converted to 1102D), at almost precisely 100 MHz plus a wide scattering of other frequencies. The exact waveform changes with scope function, so I'm sure it's internal interference from one of the digital circuits. The peak-to-peak amplitude is approx. 0.3 divisions (regardless of vertical scale) when the vertical position is at center, increasing as the position moves upwards, so whatever is causing it is after the attenuator and probably is, or is related to, the DC offset circuit. The less the scope is doing, the rougher it gets. (When I turned on FFT to look at the frequency components, it became almost a perfect sinusoid with just a bit of a second harmonic many dB down.)

Before anybody asks, no, warranty repair is out of the question.

Just wondered if anybody else has seen this same issue and can point me in the right direction. Also - I remember seeing that somebody on here had reverse-engineered the analog frontend and had a schematic, but I'm having a hell of a time finding it. Anybody better with the search tool than I am?

Thankfully, I've still got my trusty Tektronix to help me track this down  :-+
« Last Edit: May 08, 2013, 07:58:51 pm by c4757p »
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline ivan747

  • Super Contributor
  • ***
  • Posts: 2045
  • Country: us
Re: Rigol DS1000 weird problem
« Reply #1 on: May 09, 2013, 12:57:26 am »
Well, just a checklist:
*Ground coupling: is the effect the same?
*Physical inspection, inside and out.
*Bring it back to 1052E see if the problem persists. If it persists, the problem is after the input section.
*Turn on bandwidth limiting, see if the problem persists. If it does, and the BW limit is in hardware, you know where the problem is physically located.
*Self-calibration, after 30 minutes of warm-up (it really needs the warm-up)
*Broken shielding causing it to capture FM radio broadcasts internally? (almost out of question)

That's all what comes into my mind.
 

Offline c4757pTopic starter

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Re: Rigol DS1000 weird problem
« Reply #2 on: May 09, 2013, 01:22:09 am »
Well, just a checklist:
*Ground coupling: is the effect the same?

I am almost certain that ground "coupling" in this scope is done in software, or at least it disconnects the input and then additionally zeroes out the trace. I have never in my life seen such a clean 0V trace - I wouldn't even expect that from a properly functioning scope. It is literally perfectly blanked. (Out of curiosity, to find out whether there is a proper hardware disconnect, sometime I will get around to pumping RF into one channel, setting it 'ground', and seeing if the cross-coupling changes. That's not going to help this, though.)

Quote
*Physical inspection, inside and out.

Done, except under the shielding can. (I'll dig into that next week. I have exams to study for now.)

Quote
*Bring it back to 1052E see if the problem persists. If it persists, the problem is after the input section.

Good idea. Pain in the ass, but I'm sure this will be, whatever the problem is...

Quote
*Turn on bandwidth limiting, see if the problem persists. If it does, and the BW limit is in hardware, you know where the problem is physically located.

Damn. That didn't even occur to me.  |O Thanks! That narrows it down significantly. The BW limit has absolutely no effect on the noise, so it must be a hardware limit and the noise must be coming in after the limiter. The separate software filter does reduce the noise, as expected, so I haven't hosed the digital section or anything.

Quote
*Self-calibration, after 30 minutes of warm-up (it really needs the warm-up)

Done multiple times.

Quote
*Broken shielding causing it to capture FM radio broadcasts internally? (almost out of question)

I can't get this much radio noise on the scope with a foot-long antenna sticking out of the BNC! Also, there's no effect at all when I take off the large, main shielding can around the whole scope, so I doubt it.

Quote
That's all what comes into my mind.

Thank you very much!
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757pTopic starter

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Re: Rigol DS1000 weird problem
« Reply #3 on: May 09, 2013, 01:26:00 am »
Awesome! Here is that analog frontend schematic I mentioned. God, I love this forum...
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757pTopic starter

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Re: Rigol DS1000 weird problem
« Reply #4 on: May 09, 2013, 01:34:55 am »
One more hint - I'm 99.999% sure that it's an internal clock, because it appears to be in phase with the acquisition clock. If I zoom out beyond 200ns/div, it seems to disappear, but switching to peak detect mode reveals that the disappearance is just an aliasing effect that happens to be perfectly synchronized with the acquisition.
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline ivan747

  • Super Contributor
  • ***
  • Posts: 2045
  • Country: us
Re: Rigol DS1000 weird problem
« Reply #5 on: May 09, 2013, 01:56:16 am »
One more hint - I'm 99.999% sure that it's an internal clock, because it appears to be in phase with the acquisition clock. If I zoom out beyond 200ns/div, it seems to disappear, but switching to peak detect mode reveals that the disappearance is just an aliasing effect that happens to be perfectly synchronized with the acquisition.

Something must be wrong inside. By the way, the schematic you see in the first page is not the final, corrected one. Check the notes.
 

Offline c4757pTopic starter

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Re: Rigol DS1000 weird problem
« Reply #6 on: May 09, 2013, 02:10:41 am »
Yep, I saw that. Grabbed the updated schematic as well as the DAC/control schematic.
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757pTopic starter

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Re: Rigol DS1000 weird problem
« Reply #7 on: May 15, 2013, 08:04:20 pm »
Hmm. Finally back to working on this, now that I'm done with my exams. Dug into it and started probing around, and one thing stuck out at me: the amount of EMI in this thing is tremendous. With the shield open, I can't wave a probe in the general direction of the back end of the scope without picking up 50mV or more of noise. Basic power supply ripple measurements that should have been easy required dragging out the old diff probe. Er... is this normal? Or should I be investigating the power supply? I can't even tell if the noise is present or not with a standard probe because of the insane amount of interference the ground lead picks up!

Of course, it does use an open-frame switcher mounted right above where I'm probing. Perhaps it really is normal and I'm barking up the wrong tree? And if it is normal, how do you usually go about working underneath such a hideous EMI radiator?
No longer active here - try the IRC channel if you just can't be without me :)
 

Offline c4757pTopic starter

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Re: Rigol DS1000 weird problem
« Reply #8 on: May 15, 2013, 09:19:56 pm »
Oh dear. Any "I think the noise is here" above must be due to shitty probing technique. I've traced the signal all the way through the analog frontend and it's clean at final output, where it gets shipped off to the ADC. All I can think of at this point is noise on the ADC reference input or power supply rails, which I'm going to investigate. Failing that, I'm at a complete loss for ideas.  :scared: :-BROKE

Edit: Nope, they're clean. So... clean power + clean signal + ADC = garbage. Methinks either a faulty ADC or a faulty connection. (My guess is that because it staggers the ADCs, the "noise" in sync with the system clock frequency is actually a single missing bit, which shows up every time the rotation comes around to that particular ADC chip.) I've already tried resoldering the ADCs to no avail, and the FPGA which receives the data and clocks them will be a bit more "fun" so I'm not doing that right now.

« Last Edit: May 15, 2013, 10:35:12 pm by c4757p »
No longer active here - try the IRC channel if you just can't be without me :)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf