AnsweredAssumed Answered

Sweep Arbitrary Waveforms with 33500 series

Question asked by jpb on May 16, 2013
Latest reply on May 17, 2013 by jpb

Like • Show 0 Likes0
Comment • 5

The current data sheets for the 33500A and B series clearly states that frequency (or sample rate) sweeps cannot be applied to arbitrary waveforms.

But the 33500A series (recently discontinued) service/user manual states that arbitrary waveforms can be frequency swept.

The data sheet also lists a settling time of <200 nsecs for arbitrary waveforms. Settling time for a step say make sense but I don't understand what this means in terms of an arbitrary and constantly changing waveform.

If arbitrary waveforms can't be swept or frequency modulated, then doesn't some of the advantages of TrueForm technology become lost in the sense that the sample rate could just be a fixed 250MS/sec and the user could load different arbitrary waveforms for different frequencies.

I'm trying to compare the 33500 series with cheaper simple DDS systems and also more expensive variable clock systems. Not being able to sweep or modulate the frequency seems to be a bit of a drawback.

Can someone from Agilent confirm that this is the case and also explain what is meant by settling time in the context of an arbitrary time varying waveform.

No one else has this question

Outcomes

lhornburg

May 16, 2013 10:32 AM

The datasheet is wrong.

Yes, you can sweep and modulate arbitrary waveforms, in both the 33500A and 33500B. (Assuming you have the arb option in the B)

Can you provide the page numbers with this incorrect information? I will make sure it gets fixed.

I'll look into the 200 ns characteristic and get back to you.

Actions

jpb @ lhornburg on May 16, 2013 12:50 PM

Thanks for the quick response, it is good to hear that sweeps are allowed for Arbs.

In both data sheets (for the A and for the B series) there is no dot next to Arb under sweep and FM in the table showing which modulations are allowed.

The table is at the top of page 17 of the data sheet for the newer B series.

The same table in the manual does have a dot next to Arb.

It would be a good idea to correct this as people (like me) who are thinking of buying look at the data sheet rather than searching the manual.

I am still curious to understand what the

Settling time (typ) < 200 ns to 0.5% of final value

is in terms of time varying arbitrary waveforms.

Actions

jpb May 17, 2013 7:54 AM

Further to my last post.

The user manual for the 33500 series on page 31 for Frequency Sweep states:

"For sweeps, you can select sine, square, ramp, pulse, triangle, or PRBS waveforms (arbitrary waveforms, noise, and DC are not allowed). "

This, presumably is also incorrect?

Actions

lhornburg

@ jpb on May 17, 2013 9:00 AM

Yes, this is incorrect as well. When we originally introduced the 33500A, we did not allow you to define your arbitrary waveform in terms of frequency, so you were not allowed to use an Arb for FM, or a frequency sweep. In firmware release 1.14 and all later releases, we added the ability to define an arb in terms of frequency/period. Along with this, we enabled the instrument to use FM and SWEEP for arb waveforms.

It looks like the documentation was not updated to reflect these changes, and they copy and pasted them into the 33500B manuals.

200 ns Settling time:

The settling time spec only applies with “Filter Off”. Then if a step change in the waveform is initiated at some point in time, the output will be within 0.5% of its final value 200 ns later. Any sudden change in voltage is a step, and the 0.5% is relative to the size of the step. So, for example, a 1-volt step would be considered “settled” when the output is within 5 mV of its final value. A 10-volt step is “settled” when the output is within 50 mV of the final value. When one of the filters is on, this characteristic does not apply and you don't have to worry about settling time.

I took a look at the 33220A Characterstics since it is a DDS box, it has a longer settling time (250 ns) when the filter is off.

Actions

jpb @ lhornburg on May 17, 2013 10:50 AM

Thanks for the answers to my questions.

So the settling time is a measure of how long ringing/overshoot etc on steps will last if not damped by the use of the filter?

What still slightly puzzles me is that the no-filter case has a bandwidth of 40MHz whilst the normal filter is 0.27 x Sampling rate which at the maximum 250MS/s comes to 67.5MHz which is greater than the no-filter bandwidth, but I guess any overshoot will be at frequencies much above this.

Actions

5 Replies

lhornburg May 16, 2013 10:32 AM
Mark Correct
Correct Answer
The datasheet is wrong.

Yes, you can sweep and modulate arbitrary waveforms, in both the 33500A and 33500B. (Assuming you have the arb option in the B)

Can you provide the page numbers with this incorrect information? I will make sure it gets fixed.

I'll look into the 200 ns characteristic and get back to you.
Like • Show 0 Likes0
Actions
- jpb @ lhornburg on May 16, 2013 12:50 PM
  Mark Correct
  Correct Answer
  Thanks for the quick response, it is good to hear that sweeps are allowed for Arbs.
  
  In both data sheets (for the A and for the B series) there is no dot next to Arb under sweep and FM in the table showing which modulations are allowed.
  
  The table is at the top of page 17 of the data sheet for the newer B series.
  
  The same table in the manual does have a dot next to Arb.
  
  It would be a good idea to correct this as people (like me) who are thinking of buying look at the data sheet rather than searching the manual.
  
  I am still curious to understand what the
  
  Settling time (typ) < 200 ns to 0.5% of final value
  
  is in terms of time varying arbitrary waveforms.
  Like • Show 0 Likes0
  Actions
jpb May 17, 2013 7:54 AM
Mark Correct
Correct Answer
Further to my last post.

The user manual for the 33500 series on page 31 for Frequency Sweep states:

"For sweeps, you can select sine, square, ramp, pulse, triangle, or PRBS waveforms (arbitrary waveforms, noise, and DC are not allowed). "

This, presumably is also incorrect?
Like • Show 0 Likes0
Actions
- lhornburg @ jpb on May 17, 2013 9:00 AM
  Mark Correct
  Correct Answer
  Yes, this is incorrect as well. When we originally introduced the 33500A, we did not allow you to define your arbitrary waveform in terms of frequency, so you were not allowed to use an Arb for FM, or a frequency sweep. In firmware release 1.14 and all later releases, we added the ability to define an arb in terms of frequency/period. Along with this, we enabled the instrument to use FM and SWEEP for arb waveforms.
  
  It looks like the documentation was not updated to reflect these changes, and they copy and pasted them into the 33500B manuals.
  
  200 ns Settling time:
  
  The settling time spec only applies with “Filter Off”. Then if a step change in the waveform is initiated at some point in time, the output will be within 0.5% of its final value 200 ns later. Any sudden change in voltage is a step, and the 0.5% is relative to the size of the step. So, for example, a 1-volt step would be considered “settled” when the output is within 5 mV of its final value. A 10-volt step is “settled” when the output is within 50 mV of the final value. When one of the filters is on, this characteristic does not apply and you don't have to worry about settling time.
  
  I took a look at the 33220A Characterstics since it is a DDS box, it has a longer settling time (250 ns) when the filter is off.
  Like • Show 0 Likes0
  Actions
  - jpb @ lhornburg on May 17, 2013 10:50 AM
    Mark Correct
    Correct Answer
    Thanks for the answers to my questions.
    
    So the settling time is a measure of how long ringing/overshoot etc on steps will last if not damped by the use of the filter?
    
    What still slightly puzzles me is that the no-filter case has a bandwidth of 40MHz whilst the normal filter is 0.27 x Sampling rate which at the maximum 250MS/s comes to 67.5MHz which is greater than the no-filter bandwidth, but I guess any overshoot will be at frequencies much above this.
    Like • Show 0 Likes0
    Actions

Sweep Arbitrary Waveforms with 33500 series

Outcomes

Related Content

Recommended Content