Author Topic: Using arb gen as cheap tracking gen  (Read 819 times)

0 Members and 1 Guest are viewing this topic.

Offline Dan MoosTopic starter

  • Frequent Contributor
  • **
  • Posts: 357
  • Country: us
Using arb gen as cheap tracking gen
« on: August 28, 2020, 12:26:31 am »
I have an HP 8590d SA. I also have a Siglent 2042x arb gen.

The SA has a sweep trigger in. The gen has a sweep trigger out.

I figure I should be able to cobble a poor man's tracking gen.

I set the generator to sweep linearly from one known frequency to another, at a specific sweep rate.

I set the SA to trigger externally, and to sweep at the same speed.

So this kinda works. If I set my start and stop freq for my SA sweep to the same as the generator, I get close. I then fiddle with the center span freq until the level of the trace peaks. My SA has only been on for 1/2 hour, and I've learned it's reference takes a few hours to settle down, so I figure it shouldn't surprise me that the freqs don't match exactly.

What bugs me is that I can't seem to go below 30 KHz RBW before things start looking messed up. I'm currently sweeping from 1 MHz to 1.5MHz on the generator, with a 30 ms sweep. I have the SA going from 1 MHz to 2MHz, with a 60 ms sweep. I'm sweeping the SA twice as long over twice the range so I can see where the signal ends in the middle of the screen.

Is the problem that at lower RBW settings, the poor sync between the sweeps is being exposed? I want to look at the response of filters, and don't want a high RBW to smooth things to the point that I don't see what's actually happening. 

I know if this were completely doable, tracking gens wouldn't be a thing, so I have low expectations. I just want to see how good I can get my janky setup to work.
 

Offline Kean

  • Supporter
  • ****
  • Posts: 2053
  • Country: au
  • Embedded systems & IT consultant
    • Kean Electronics
Re: Using arb gen as cheap tracking gen
« Reply #1 on: August 28, 2020, 12:39:02 am »
I imagine the Siglent arb gen would be sweeping in discrete steps.  No idea how many it can manage in a 30ms sweep, but that would relate to the acheivable RBW setting.
 

Offline hendorog

  • Super Contributor
  • ***
  • Posts: 1617
  • Country: nz
Re: Using arb gen as cheap tracking gen
« Reply #2 on: August 28, 2020, 12:42:06 am »
I have an HP 8590d SA. I also have a Siglent 2042x arb gen.

The SA has a sweep trigger in. The gen has a sweep trigger out.

I figure I should be able to cobble a poor man's tracking gen.

I set the generator to sweep linearly from one known frequency to another, at a specific sweep rate.

I set the SA to trigger externally, and to sweep at the same speed.

So this kinda works. If I set my start and stop freq for my SA sweep to the same as the generator, I get close. I then fiddle with the center span freq until the level of the trace peaks. My SA has only been on for 1/2 hour, and I've learned it's reference takes a few hours to settle down, so I figure it shouldn't surprise me that the freqs don't match exactly.

What bugs me is that I can't seem to go below 30 KHz RBW before things start looking messed up. I'm currently sweeping from 1 MHz to 1.5MHz on the generator, with a 30 ms sweep. I have the SA going from 1 MHz to 2MHz, with a 60 ms sweep. I'm sweeping the SA twice as long over twice the range so I can see where the signal ends in the middle of the screen.

Is the problem that at lower RBW settings, the poor sync between the sweeps is being exposed? I want to look at the response of filters, and don't want a high RBW to smooth things to the point that I don't see what's actually happening. 

I know if this were completely doable, tracking gens wouldn't be a thing, so I have low expectations. I just want to see how good I can get my janky setup to work.

Coupling the references should eliminate the reference drift of both units.

The easiest way to do this is to use Max Hold function on the 8590. Then it should work given enough time - at the extreme you could slow the sweep right down on the 8590 so it does one point for each sweep of your sig gen.

Another option would be to use the LO out on the 8590 if it has one. Then use a mixer and a sig gen to provide the offset and you have a poor mans tracking generator.



 

Offline Dan MoosTopic starter

  • Frequent Contributor
  • **
  • Posts: 357
  • Country: us
Re: Using arb gen as cheap tracking gen
« Reply #3 on: August 28, 2020, 12:50:53 am »
I imagine the Siglent arb gen would be sweeping in discrete steps.  No idea how many it can manage in a 30ms sweep, but that would relate to the acheivable RBW setting.

Does that suggest possibly slower sweeps would achieve a better result? Time to experiment.

If I'm looking at a span of a MHz, what RBW minimum would give me a useful trace? I imagine there must be some rule of thumb ratio of RBW to span that gives the most useful plot.
 

Offline Dan MoosTopic starter

  • Frequent Contributor
  • **
  • Posts: 357
  • Country: us
Re: Using arb gen as cheap tracking gen
« Reply #4 on: August 28, 2020, 01:17:25 am »
Update: At 10 times the sweep time, I can get the RBW down t 3 KHz. If I set the trace to max hold, I can get to the SA's lowest RBW (1k), and in less than 10 seconds the trace is complete.

I would call this a success!
 

Offline Kean

  • Supporter
  • ****
  • Posts: 2053
  • Country: au
  • Embedded systems & IT consultant
    • Kean Electronics
Re: Using arb gen as cheap tracking gen
« Reply #5 on: August 28, 2020, 02:04:45 am »
Good to hear that helped.
To fine tune it, check the docs for the arb gen to see what minimum step or step rate it can acheive.
 

Offline rf-loop

  • Super Contributor
  • ***
  • Posts: 4064
  • Country: fi
  • Born in Finland with DLL21 in hand
Re: Using function gen as cheap tracking gen
« Reply #6 on: August 28, 2020, 05:59:25 am »
Good to hear that helped.
To fine tune it, check the docs for the arb gen to see what minimum step or step rate it can acheive.

Btw, why you and OP talk about "arb generator" when you talk about Siglent SDG2042X in this case.

Yes it have ALSO Arb function for generate arbitrary waveforms. Bit it is just one single function in this multipurpose generator and not even used in this case at all..  Here it is used more like RF signal generator but belongs to category function generators. This Arb function is not at all used here when OP talk about sweeping. Have you never looked SDG2kX function/arbitrary generator sweeping waveform. I have. Also SDG1000X can do it. Do it look like stepping sweep. but if we go to deeply in semantics... well then this story is endless...
Yes I have experience with many kind of signal generators, tens of years. THis generator do not step sweep example like some older HP expensive state of art RF generator. First you select start, then stop and then generator make lookup table for fixed frequencies and start stepping through this internal hidden table where amount of steps depends user defined sweep time and sweep width. It is terrible looking when look with SA or Oscilloscope.

Modern sophisticated generators can do lot of more nice continuous sweep. Well if go deeply to semantics...  but if you look in practice when it do linear sweep it is difficult to find jumps. Even ancient HP3325 synth function generator  can do phase continuous linear sweep.  SDG2kX also can.
« Last Edit: August 28, 2020, 06:08:49 am by rf-loop »
I drive a LEC (low el. consumption) BEV car. Smoke exhaust pipes - go to museum. In Finland quite all electric power is made using nuclear, wind, solar and water.

Wises must compel the mad barbarians to stop their crimes against humanity. Where have the wises gone?
 

Offline Kean

  • Supporter
  • ****
  • Posts: 2053
  • Country: au
  • Embedded systems & IT consultant
    • Kean Electronics
Re: Using function gen as cheap tracking gen
« Reply #7 on: August 28, 2020, 06:28:01 am »
Btw, why you and OP talk about "arb generator" when you talk about Siglent SDG2042X in this case.

Well, I just used the terminology the OP used.  He only gave a partial model number of his equipment.  I did assume that he knew enough to be generating sine waves.

I have no direct experience with the Siglent SDG units, but it appears that my reply helped the OP think about the sweep behaviour, and apparently realised that they were potentially expecting too much with such a fast sweep time.  I'm used to running EMC scans that can take a long time.

If I said anything explicity incorrect, then I am happy to be corrected.  I admit to having a limited knowledge of how some of the equipment works internally.  There is always more to learn.
 

Offline thinkfat

  • Supporter
  • ****
  • Posts: 2150
  • Country: de
  • This is just a hobby I spend too much time on.
    • Matthias' Hackerstübchen
Re: Using arb gen as cheap tracking gen
« Reply #8 on: August 28, 2020, 01:39:34 pm »
It's normal that you need to sweep slower if you reduce the bandwidth. A narrow filter cannot follow fast changes. That's the whole point of it being "narrow".

I imagine the Siglent arb gen would be sweeping in discrete steps.  No idea how many it can manage in a 30ms sweep, but that would relate to the acheivable RBW setting.

Does that suggest possibly slower sweeps would achieve a better result? Time to experiment.

If I'm looking at a span of a MHz, what RBW minimum would give me a useful trace? I imagine there must be some rule of thumb ratio of RBW to span that gives the most useful plot.

I guess if you calculate the step response of an ideal bandpass filter with given bandwidth, that'll give you an idea how long the signal must stay within the filter band.
« Last Edit: August 28, 2020, 01:43:35 pm by thinkfat »
Everybody likes gadgets. Until they try to make them.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf