TDS694C works fine but just down to 20ns/DIV, signal distorted from 10ns/DIV and below

Dear all,

our TDS694C began to show what can only be discribed as massive ringing once the set time resolution gets to 10ns/DIV and below. Until down to 20ns/DIV, the signal response just seems fine. After that, it is just a mass with strong oscillations masking the signal up to a point one can only guess its original shape. All four channels show identical behaviour. Trigger still works on all four channel fine, btw.

Is this a common/known problem and what might we do to check for the cause?

Regards,
Christian

Hi Christian,

Is the scope switching to equivalent-time (ET) sampling when you go to 10 ns/div? Those early TDS scopes had that as a feature in their fastest time/div and it could lead to unusual problems if misused (bad triggering usually).

A picture of what you're seeing would also be useful.

Hi Afonso,

thanks for the hint. I'll be back at the lab on monday and check accordingly and will make some screenshots.

Regards,
Christian

Hi Afonso,
I verified today that the oscilloscope remains in real-time mode. I also found the following paragraph in its manual, page 3-17:
 

The general method of sampling are real-time and equivalent-time. The TDS 694C Oscilloscopes use only real-time sampling.

So this does not seem to be the problem. Meanwhile, I created some image dumps from a 10MHz square wave, how it is properly drawn down to 20ns and distorted afterwards. Obviously, triggering fails on anything other than 20ns and above. Hope it helps to track down the origin of the problem.

Regards,
Christian

Hi Christian,

The trigger on those scopes is analog, so it's not obvious at all to me that it should fail when the sampling system fails.

I would suggest trying to use all 4 channels at 20 ns/div. I suspect one of the ADCs has failed (and it only shows up at fast sweeps because it gets muxed into CH1 for higher sample rate).

Hi Afonso,

The failure of a single ADC that distorts the signal in interleaved mode makes sense and I think that such a failure would lead to just the waveform I observed, e.g. what seems to be a jumping between single adjectand sample points of vasty different values. However, I was not able to see a change in behaviour by activating all four channels. I added the trigger signal of the pulse generator to CH2 to see if the distortion is visible on mutliple channels and it certainsly is. Again, up until 20ns/DIV, everything is fine on all channels.

Do you know, how many ADCs are used in the TDS694C and what their single sampling rate is? The distortion taking place when changing from 20ns/DIV to 10ns/DIV corresponds to the change of the sampling rate from 2,5GS/s to 5GS/s. If there are four of them with 2,5GS/S each, what would explain the distortion starting when changing from 2,5GS/s to 5GS/s. Plus, it would explain an additional phenomenon I just observed:

I noticed that despite I am sure I saw the triggered rising edge corresponding with the 50% horizontal time line in the past, it is now somehow delayed, meaning that the rising edge is not on the 50% line despite no horizontal shift is applied. In addition, when I fed the trigger signal to the other channels, I noticed that all four channel exhibit some kind of delay between each other. I created an overlay in an image editor to better visualize what I mean: the rising edge on the pulse signal CH1 (on which I triggered) does not match with any of the trigger edges on the other channels. Plus, the triggered edge on the pulse signal is not where it is supposed to be, namely in the middle of the screen.

Is there some kind of deskew calibration to be performed on the ADCs?

Regards,

Christian

PS: the trigger signal is routed via a 50cm BNC cable, the pulser is attached directly. While the additional BNC cable length will add a certain delay, I would expect it to be around 5ns max and to be constant over all channels.

Channel deskew was set at 0 for all channels.

No nupdates on our side. Any additional ideas?