Oscilloscope waveform averaging

[Greg Bredthauer]

Hi all,

I'm using a Tektronix 3034 scope to acquire a 10000 point waveform. I'm
sampling at 250MS/sec. The waveform repeats at 5kHz.

This is all fine and good... what confuses me is the behavior of the
averaging function. If I set the scope to average 16 waveforms to cut down
on random noise, it takes 3-4 seconds for the displayed waveform to settle
down to its final values when I change the input signal. Why is this? The
Tektronix user manual claims that the scope can acquire 450 10000-point
waveforms per second. If this is true, I'd expect the scope to settle on a
new waveform in about 1/30th of a second, not 3 or 4 seconds.

Is this just a limitation of the scope's processing power? Or is the scope
using a decaying average instead of a simple boxcar?

Is there anything I can do to speed this up on my current scope? I've
found that switching to fast-trigger mode greatly speeds up the average's
settling time, but I need my 10000 samples. Will a scope like the TDS5000
series help, or perhaps a PC-based scope add-in card?

Thanks in advance for any help or advice!

-Greg

 

[John Miles]

Hi all,

I'm using a Tektronix 3034 scope to acquire a 10000 point waveform. I'm
sampling at 250MS/sec. The waveform repeats at 5kHz.

This is all fine and good... what confuses me is the behavior of the
averaging function. If I set the scope to average 16 waveforms to cut down
on random noise, it takes 3-4 seconds for the displayed waveform to settle
down to its final values when I change the input signal. Why is this? The
Tektronix user manual claims that the scope can acquire 450 10000-point
waveforms per second. ...

It is probably doing averaging in the DPO accumulation buffer rather
than the acquisition memory itself. (The DPO feature inserts an
accumulation buffer between sample storage and video memory; the video
buffer itself is built from information appearing in the accumulation
buffer over time.)

Is there anything I can do to speed this up on my current scope? I've
found that switching to fast-trigger mode greatly speeds up the average's
settling time, but I need my 10000 samples. Will a scope like the TDS5000
series help, or perhaps a PC-based scope add-in card?

I imagine that yes, this is one of the features that they give you when
you drop the big bucks for a TDS5000. The TDS3000s are nice scopes but
they do have their limitations, including the excessive front-end noise
that necessitates video averaging in the first place.

-- jm

 

[TekMan]

Hi all,

I'm using a Tektronix 3034 scope to acquire a 10000 point waveform. I'm
sampling at 250MS/sec. The waveform repeats at 5kHz.

This is all fine and good... what confuses me is the behavior of the
averaging function. If I set the scope to average 16 waveforms to cut down
on random noise, it takes 3-4 seconds for the displayed waveform to settle
down to its final values when I change the input signal. Why is this? The
Tektronix user manual claims that the scope can acquire 450 10000-point
waveforms per second. If this is true, I'd expect the scope to settle on a
new waveform in about 1/30th of a second, not 3 or 4 seconds.

Is this just a limitation of the scope's processing power? Or is the scope
using a decaying average instead of a simple boxcar?

Is there anything I can do to speed this up on my current scope? I've
found that switching to fast-trigger mode greatly speeds up the average's
settling time, but I need my 10000 samples. Will a scope like the TDS5000
series help, or perhaps a PC-based scope add-in card?

Thanks in advance for any help or advice!

-Greg

shure it only 5 KHz?

Reduce the sampling rate - for a 5 kHz signal go for 0.1 ms/division
timebase. 250MS/s is grossly overkill.

Tell us what happens at 0.1 ms/div.

And don't forget that "acquire rate" is not equal to "screen update
rate"
450 acquisitions, approx. 30 to 100 updates is not uncommon (all "per
second").


A TDS 3xxx should be fine for 5kHz. Even an old TDS420 would do (15 k
points; 150MHz; 100 MS/s) for that application.


hth,
Andreas

 

[AliasFan]

[email protected] (TekMan) wrote in

shure it only 5 KHz?



hth,
Andreas

The signal itself is from an ultrasound hyrophone, with interesting signal
info in the 0-20MHz range, which is why I'm sampling at 250MS/sec. The
ultrasound pulse (and trigger) come at 5kHz rate.

-Greg

 

[Greg Bredthauer]

Hi all,

I'm using a Tektronix 3034 scope to acquire a 10000 point waveform.
I'm sampling at 250MS/sec. The waveform repeats at 5kHz.

This is all fine and good... what confuses me is the behavior of the
averaging function. If I set the scope to average 16 waveforms to cut
down on random noise, it takes 3-4 seconds for the displayed waveform
to settle down to its final values when I change the input signal.
Why is this?

I've solved my own problem... two things:

1. The scope captures and processes waveforms much more slowly at 250MS/s
than at 100MS/s for an unknown reason.

2. There is some kind of persistence beyond just averaging (some kind of
decaying averaging?) In any case, I can get around it by stopping and
restarting acquisition.

Making these 2 changes lets me get 128 averages done in under 1 second. I
hope this helps someone else out there

-Greg

 

[Ed Price]

[email protected] (TekMan) wrote in


The signal itself is from an ultrasound hyrophone, with interesting signal
info in the 0-20MHz range, which is why I'm sampling at 250MS/sec. The
ultrasound pulse (and trigger) come at 5kHz rate.

-Greg

20 MHz seems incredibly high for mechanical transfer of energy in water.
What are you doing, and do you really need this bandwidth? Could you need be
related to viewing high-order artifacts of the transducer rather than an
actual propagated signal?

Ed
wb6wsn