LabVIEW

cancel
Showing results for 
Search instead for 
Did you mean: 

FPGA filter design for Lock in Amplifier

Hey guys,

 

I have a problem implementing a working lock in amplifier on my SB-Rio9636. (I have LV 2013 and the add-ons like digital filter design). I will have to sample with at least 50kHz up to 400kHz (therefore I will buy an external ADC).

 

As I am kind of new to FPGA and digital filter design I have a few questions. I am able to synthesize my own sinus signal with the frequency I wish and I am using a PLL to create a phase locked cosinus to this sinus ( I know I would not have to do this, but the PLL also confirms that I am creating the sinus with the right frequency.) For test purpose I am using my reference signal also as measured signal but still my calculated amplitude and phase are oscillating and the amplitude is huge (~10^13).

 

I have attaced my host.vi, RT.vi and FPGA.vi. (As I can only attach 3 items, the SubVis of the fpga.vi can be found here:

https://decibel.ni.com/content/docs/DOC-1762 (LIA_SBRIO2010.zip)

 

1, Is my Gain factor which I use to scale my output right?

2, What is wrong with my demodulation, CIC or moving averager?

3, Can you give me an advice or alternative for the CIC decimation filter and the moving averager filter?

 

 

I hope someone can help me at some points.

 

 

Kind regards

 

Slev1n

 

0 Kudos
Message 1 of 3
(3,177 Views)

Noone has an idea?

 

Maybe someone can tell me what kind of lowpass filter he would choose and how he would implement this filter on the fpga (express vi, self build algorithm or Xilinx Ip nodes). It would be good to be able to change some settings without recompiling.

e.g. cut off frequency/time constant

 

thx

0 Kudos
Message 2 of 3
(3,099 Views)

Hey guys,

 

maybe you can tell me some facts about the attached CIC Filter.

 

#stages = 1?

Changing Rate Factor = decimation-value

Differential Delay = 2?

 

If I want to analyse the frequency response can I use the "plot Mrate NstageFrequency.vi" or would you recommend another?

 

I would like to have an extremly narrow passband lowpass filter (if possible with decimation for less computing effort) followed by an moving average filter, any advice how to implement?

0 Kudos
Message 3 of 3
(3,074 Views)