How to get a cleaner sine signal from a waveform generator?

[[email protected]]

Hi all,

I'm using a 2-channel waveform generator (YOKOGAWA FG120) to generate
low frequency(~ 10Hz), low amplitude (40mVpp) sine signals, which
served as differential inputs for measuring a high precision sigma
delta ADC (~100dB SNR).

Now, My test result is not very well. I think it's maybe due to the
impurity of the input signal, whose harmonics and noise will
deteriorate the performance.

Besides buying a more precision generator..., is there way to get a
cleaner sine signal from the generator, say, harmonics decreasing from
-55dBc to -85dBc or lower noise?

Does adding bandpass filter work? or using lock-in amplifer?

I need your opinions. Thanks a lot...

Regards,

Kanjun Shi

 

[Jon Slaughter]

Hi all,

I'm using a 2-channel waveform generator (YOKOGAWA FG120) to generate
low frequency(~ 10Hz), low amplitude (40mVpp) sine signals, which
served as differential inputs for measuring a high precision sigma
delta ADC (~100dB SNR).

Now, My test result is not very well. I think it's maybe due to the
impurity of the input signal, whose harmonics and noise will
deteriorate the performance.

Besides buying a more precision generator..., is there way to get a
cleaner sine signal from the generator, say, harmonics decreasing from
-55dBc to -85dBc or lower noise?

Does adding bandpass filter work? or using lock-in amplifer?

Of course... why wouldn't.

 

[[email protected]]

Any measure you use to improve this is going to be subject to
difficulties due to that extreme SNR that you're measuring against.  So
if you go looking into bandpass filters or lock-in amplifiers you need
to make sure that your follow-on block doesn't generate at least as much
distortion as it is eliminating.  So you can't just sling a solution in
the general direction of your workbench and expect it to improve things.

Have you obtained a model with their "Low Distortion Output" option?

Depending on the nature of their output amplifiers, you may be able to
improve things by using a resistive attenuator and setting the generator
to a higher voltage output -- if they go straight from their amplifier
to their output, and they have some crossover distortion, then the
attenuator may help.  If they _do_ switch in an attenuator, using your
own 'fine tune' attenuator (I'd try steps of 1dB) in conjunction with
increased voltage output may find a sweet spot in their output amplitude
that has lower distortion.

I have to say, if all I ever had to do was this one measurement I'd be
tempted to build a Wien bridge oscillator followed by a long bandpass
chain (and probably powered by batteries at first); it's a
rocks-and-sticks solution, but sometimes those are the solutions that
get you the most bang for your buck when distortion and/or speed are the
highest priority.

--

Tim Wescott
Wescott Design Serviceshttp://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says..
See details athttp://www.wescottdesign.com/actfes/actfes.html- Hide quoted text -

- Show quoted text -

Very thanks for your reply!

I will try the attenuator way...