Dwell time for HP4195A and HP8563E

[Larry Martell]

I am trying to determine what the dwell time is for these instruments.
By dwell time I mean the amount of time it samples the signal for each
data point. This does not appear to be a configurable value, nor does
there seem to be a way to query this value.

For these instruments, is this parameter fixed or variable? If it is
fixed how
do I find out what it is? If it is variable, is it calcualted by simply
dividing the
sweep time by the number of data points?

TIA,
-larry

 

[qrk]

I am trying to determine what the dwell time is for these instruments.
By dwell time I mean the amount of time it samples the signal for each
data point. This does not appear to be a configurable value, nor does
there seem to be a way to query this value.

For these instruments, is this parameter fixed or variable? If it is
fixed how
do I find out what it is? If it is variable, is it calcualted by simply
dividing the
sweep time by the number of data points?

TIA,
-larry

It is indirectly adjustable, at least on the 4195.
The measurement time depends on your resolution bandwidth (Res BW
button) setting. Lower bandwidth setting will take longer to make a
measurement. You can divide the total time by number of points to come
up with an estimate if you use the same bandwidth for all points
(non-auto mode). If you use the Auto Res BW, the resolution bandwidth
will be different for different frequency bands which can speed up the
total measurement time (slow at low freqs, faster at the higher
freqs). If you use the video filter, this will add more time to the
measurement process. On the 4195, you can manually change the
bandwidth in the middle of the sweep if you need to hurry things up or
reduce noise in critical areas.

 

[Glenn]

Larry
I'm not certain about the 856X series but I suspect it is the same as
for the newer analyzers. If so, the way the detector works depends on a
number of factors:
o how fast one is sweeping.
0 detection mode (peak,sample, neg. peak)
0 resolution bandwidth

As the analyzer's LO traverses the spectrum between measurement bins
(each bin of width = span/#measurment points), depending on how the
detector is selected, the A/D takes 1-to-many measurements. Maximum
sweep rate is limited by the need to let the detected signal settle (the
group delay of the RBW filter) and the ability of the LO itself to move
quickly and accurately. For the ESA series, multiple measurements may
be made within each bin. Peak Detect modes just save the max or min
value for the bin. Sample Detect (probably) measures the value at the
center(or end, I 'm not sure) of the bin and throws away any other
measurment. For some analyzers there is a non-sweeping mode wherein the
A/D runs as fast as it can. I think that is a 50 ns measurement time for
ESA.

While there may be no dwell time selection, you may be able to influence
the time spent by judicious selection of span, RBW and sweep time. But
then, the question arises "why do you want to do that?". The SA should
already manage things to avoid over-sweeping, trying to take data faster
than the if filters can handle, and the detection mode selection should
allow you some fleximbility as to whether you're looking at a peak or
sampled mode.

Glenn

 

[Larry Martell]

It is indirectly adjustable, at least on the 4195.
The measurement time depends on your resolution bandwidth (Res BW
button) setting. Lower bandwidth setting will take longer to make a
measurement. You can divide the total time by number of points to come
up with an estimate if you use the same bandwidth for all points
(non-auto mode). If you use the Auto Res BW, the resolution bandwidth
will be different for different frequency bands which can speed up the
total measurement time (slow at low freqs, faster at the higher
freqs). If you use the video filter, this will add more time to the
measurement process. On the 4195, you can manually change the
bandwidth in the middle of the sweep if you need to hurry things up or
reduce noise in critical areas.

Thanks much for the reply Mark. This is very helpful.

-larry

 

[Larry Martell]

Larry
I'm not certain about the 856X series but I suspect it is the same as
for the newer analyzers. If so, the way the detector works depends on a
number of factors:
o how fast one is sweeping.
0 detection mode (peak,sample, neg. peak)
0 resolution bandwidth

As the analyzer's LO traverses the spectrum between measurement bins
(each bin of width = span/#measurment points), depending on how the
detector is selected, the A/D takes 1-to-many measurements. Maximum
sweep rate is limited by the need to let the detected signal settle (the
group delay of the RBW filter) and the ability of the LO itself to move
quickly and accurately. For the ESA series, multiple measurements may
be made within each bin. Peak Detect modes just save the max or min
value for the bin. Sample Detect (probably) measures the value at the
center(or end, I 'm not sure) of the bin and throws away any other
measurment. For some analyzers there is a non-sweeping mode wherein the
A/D runs as fast as it can. I think that is a 50 ns measurement time for
ESA.

Thanks very much for the info Glenn, it is very hlepful.

While there may be no dwell time selection, you may be able to influence
the time spent by judicious selection of span, RBW and sweep time. But
then, the question arises "why do you want to do that?".

I have to perform some EMI emissions tests that have specific dwell
time requirments
(>= 0.15 sec), and I will need to prove to my customer that I am
meeting these.

Thanks!
-larry

 

[Glenn]

Larry

I have to perform some EMI emissions tests that have specific dwell
time requirments
(>= 0.15 sec), and I will need to prove to my customer that I am
meeting these.

OK, that helps. In that case you'll want to use the peak detect mode
(max) and slow the sweep time down to (.15sec/bin * #bins/sweep). That
should get you to .15 seconds within each bin and a measured level
corresponding to the peak within that time. To be thorough, you'll
probably want to make sure that the RBW is set to be at least as wide as
a bin.

Glenn

 

[Larry Martell]

Larry


OK, that helps. In that case you'll want to use the peak detect mode
(max) and slow the sweep time down to (.15sec/bin * #bins/sweep). That
should get you to .15 seconds within each bin and a measured level
corresponding to the peak within that time. To be thorough, you'll
probably want to make sure that the RBW is set to be at least as wide as
a bin.

Glenn, thanks very much for the info!

-larry