Lock-in Amplifier Start-Up kit configuration

[obarriel]

I am using Lock-in Amplifier Start-Up kit to measure a small and very noisy signal. Signal is modulated at 50kHz. I would like to know which parameters should I use to make the measurement. Now I am using:

-Samplerate 500kHz (my adquisition board allows up to 25MHz)

-Blocksize: 4096 points

-Time constant: 10ms

-Roll-off: 20dB

-Filter: FIR

The results that I am obtaining are not bad at all, but not as good as I hoped. I don't really know what I should change to get better results without increasing a lot the measurement time (I still need a fast measure).

Thank you very much

[johnsold]

I am not familiar with the Start-Up kit, so my comments are general rather than specific.

Fundamentally the lock-in amplifier trades off time against signal to noise ratio improvement. So, fast and low noise do not happen together.

Make sure your A/D converter has enough dynamic range. The signal you want is buried in the noise, but if the converter resolution is too coarse, quantization may eliminate some of information about the signal.

Perhaps you can do some kind of sliding window or moving average type of measurement where you build up a long block data but update it frequently. I am not sure if this can be done in a lock-in application.

You may need to run some experiments with parameters. There are so many variations in applications that it may be difficult to predict what will work best for your particular situation. The block size and the time constant are of similar duration. I might try varying those to see if it makes a difference. The sample rate gives 10 points per cycle of the modulation. That is probably enough, but I definitely would not go any lower. A little higher might help, but my guess is that the change would be small.

Lynn

[obarriel]

Thank you johnsold!

I have been careful with the resolution of my A/D,and I am not being limited by it.

I have done several experiments with parameters, but I am really unable to understant the lock-in start-up kit behaviour. For example, with the same parameters above, but increasing only the sampling rate  (something like 2 - 5 Mhz) I get much worse results.

Another example, with the same paramters above but with a blocksize of 1024 I also have poor results.

It seems that the samplerate and blocksize parameters strongly affect the lock-in behaviour, so I am trying to optimise the samplerate and blocksize values for my  50 Khz signal in order to improve results.

[johnsold]

I would not have expected the sample rate to have such a strong effect, but I do not know how the kit works internally.

What is your noise bandwidth coming into the A/D? Are there interfering signals or just random noise?

What happens if you change the sample rate slightly, say to 503 kHz? I wonder if the sampling rate being close to a harmonic of the signal has some effect.

Lynn

[obarriel]

There are not interfering signals of other freq. , so I think there is just random noise at all freq.. Signal comes from a photomultiplier detector.

Slighly changing the sampling rate does not affect the result.

It's difficult (for me) to know how the the lock-in really works because the code is protected (non visible) and the manual does not give many details. 

Thank you very much

[obarriel]

I have found this topic related with the lock-in and fast sample rates.

http://digital.ni.com/public.nsf/allkb/0DF72DBB9811E7A586256F80006B9710

This is why I have problems if I increase the sample rate?