Sample Jitter for AD2 Scope When Averaging

[Adam K]

 

I have a trigger that is syncronized to an amplifier output that I'd like to measure with the AD2 scope. From other experiments I know this trigger signal is synchronized to within 10-20ns to the amplifier ouput.  I'm using the Record feature and capturing 48 repeat 100-ms measurements of this waveform at 500kHz sampling (2us period).  I also have averaging turned on.

The plot below shows a small segment of the waveforms with all 48 measurements overlaid.  As you can see there is noticeable timing jitter that covers 2us across all the measurements.  I'm inferring this comes from the AD2 averaging 2-us blocks in a continuous running mode and then just providing me the nearest sample following the detected trigger. As a result the samples may be provided anywhere from 0 to 2us after the trigger signal is detected. Is this correct? 

Are there any options I can use to preserve processing gain (by averaging) but give me samples without this timing jitter?  I'm not quite sure what the oversampling option does.  Brute force I could turn the sampling up much higher (say 20MHz) and then just average after to 500kHz samples but wondering if there is a better option.

Thanks! 

 

[attila]

Hi @Adam K

For jitter measurement one continuous capture is required.
With Analog Discovery you could analyze lower frequency signals by recording at up to 1-2MHz.
With ADP3450, EclypseZ7-Zmod you can capture up to 128/256Mi samples at up to 125MHz, with ADP5250 1M samples up to 1GHz.

 

 

[Adam K]

Thanks @Attila for the reply but I'm still not sure about the Discovery2 operation when averaging is turned on and an external trigger is used. 

Channel 1 - measuring a repeated signal S1

• 500kHz sampling
• Averaging mode - blocks of 200 samples are averaged together to go from 100MHz to 500kHz sampling

Channel 2 - TTL trigger that is produced at the onset of the signal S1

• Edge detection

My question re: AD2 averaging sampling mode is whether block averaging begins with the detected trigger OR is block averaging continuously being performed and the next block averaged sample is output after the trigger is detected.  In the first case the delay from the trigger to the output of the first block average will be 200 samples or 2us, while in the second case the output of the first block average will be anywere from 1 to 200 samples: e.g. 10ns to 2000ns. 

The data I presented implies that it is the second case above, but I'd like to know if the first case is possible.  

Thanks!

[attila]

Hi @Adam K

The averaging is continuous before, through and after the trigger; respecting the set 100MHz/N sampling rate.