Hi guys, I need a circuit which will delay a varying frequency sine wave by 90 degrees constantly (for use in a lock-in amplifier). I have a dsPIC controlling everything which I can use to control this quadrature phase shift circuit. Does anyone have any ideas on how this can be implemented? Thankyou

 

A lock-in amplifier is in basic a PLL and synchronous demodulator followed by a low-pass filter. Then you use a lock-in amplifier with in internal reference it functions as a synchronous demodulator. What will your application be? A true lock-in amplifier or a synchronous demodulator. And what kind of frequencies do you want to use

 

Use an NE567 chip.

 

Why do you need a 90 degree shift for the lock-in amplifier?

 

I'm sorry guys, I have figured out a much better implementation of what i needed to do. I basically needed a two reference signals going into two AD630's. One signal is in phase with the signal to be measured, one is 90 degrees out of phase. I am going to use the pic to generate the square wave reference signals and use software to offset the second reference signal by 90 degrees for a given frequency. pretty obvious now that this is a far superior method than trying to get a digitally controlled analogue circuit to achieve 90 degrees at a given frequency.

I'm a little confused by that IC dave, what could I have used it for?

Thanks for the replys

 

Hmm what are you trying to make. Do not be shy just tell us. I am not so sure using square waves is the best thing to do at all. I have used the AD630 as a a synchronous demodulator but with sine waves. If I remember correct the AD630 is very expensive. If you are going to use a dsPIC. You will probably have more than enough power do do the synchronous demodulation in hardware. You just multiply the measured signal with the two reference signals. And then apply the lowpass filter

 

Yeah it definitely can be implemented in the dsPIC, but I am afraid I am not capable of that level of programming yet (http://ww1.microchip.com/downloads/en/AppNotes/01115A.pdf). The end goal is to have most of this system built into the dsPIC (that is really what these chips are for!).

For now, I am designing an analogue solution. I am building an impedance spectroscopy system involving a potentiostat. I am feeding a DDS-generated sine wave into the potentiostat (after setting the correct amplitudes and offset using DACs and two stages of JFET attenuation (http://www.tij.co.jp/jp/lit/an/snoa620/snoa620.pdf top of page 5). the current from the potentiostat (expressed as a voltage) is then fed into this lock in amplifier (implemented using two AD630's like these guys have done http://www.patarnott.com/atms360/pptATMS360/LockInAmplifier2010.ppt).

The DDS is going to be sweeping frequences, hence the need for this frequency independent phase shift for the quadrature demodulator... Or more sensibly just clock out another reference signal which is 90 degrees out of phase with the original.

The system is then going to have some digitally controlled low pass filters to obtain the magnitude and phase readings as fast as possible depending on what frequency is being passed.

Hope this clears up a few queries (yeah right lol)

 

First of all that ppt file do no describe a lock-in amplifier. But a synchronous demodulator followed by a low-pass filter But do not worry about that. Second what frequency range are you looking for? and what will the impedance range be. Is it for skin conductance measurements? Making a digital synchronous demodulator is not that hard. If you have continuous stream of samples you just multiply sample by sample. I have done this a lot. Using a sound card and Labview. But also the AD630 of course

 

Ahh right OK, I am rather new to the whole field! Would you mind explaining to me the difference between a synchronous demodulator and a lock-in amplifier? The application is for reading capacitative biosensors. Some really good background information here:

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2174792/

Figure 4 outlines pretty much the responses I will be after when the system is working, 0.1Hz to about 100KHz for our experiments (they go up to 1MHz). Impedance ranges I am really not sure about yet, just going to have to experiment and find out!

 

This sound to me like a job for a proper commercial lock-in amplifier. For that frequency range anything not made by professionals will in the best case be mediocre. So get a commercial lock-in. And use say Labview to control the lock-in to make the test automated.

Ahh right OK, I am rather new to the whole field! Would you mind explaining to me the difference between a synchronous demodulator and a lock-in amplifier?

The heart in every lock-in amplifier is a synchronous demodulator followed by a low-pass filter. Another name for this is a "Phase-Sensitive Detector(PSD)" And then you use a commercial lock-in with internal reference it works as a PSD. However then a external reference signal is provided to the lock-in. The lock-in use a PLL in the lock-in amplifier to lock the internal reference oscillator to this external reference. And hence the name "lock-in" If you need more info take a look here http://www.thinksrs.com/support/app.htm and also here http://forum.allaboutcircuits.com/showthread.php?t=74293 post 3
The power point you pointed me to. Do have kind of the principals OK. But it is something with it that is somewhat strange

 

Why do you need a 90 degree shift for the lock-in amplifier?

See post #10 and the documents about lock-in basics

 

Well this is a development job so it will be worked on until the system reaches enough accuracy to be able to read these sensors. The results I have obtained from the prototyping have been very promising in fact, with proper PCB layout and the proper use of the dsPIC the results will only get better. Every stage of the build I have tested thoroughly for enough accuracy. There will be problems but I intend to tackle every one head on.

Right I understand, thanks for clearing this up. So the lock-in aspect of the amplifier refers to the generation of the reference signals (via a PLL) within itself based on the external Vref signal you provide it. Just wondering, why is the whole PLL thing necessary if the reference signal has already been generated and send to the amplifier? Perhaps it would be more accurate to this a phase sensitive detector then!

Thanks again for your help

 

Yes a lock-in amplifier is more accurate then does not have to "lock-in" to an external reference signal. But in some measurement setup this is not possible.
For your setup I would strongly recommend using sinusoidal refence waves. Since you have signals up to 100Khz. You will for a full digital solution have to use at least 1MHz samplerate. What may be kind of hard to implement and you will need a lot of MIPS to achieve that. So I would suggest a two way approach. Then you are in the lower frequency range. Your PSD is fully digital, but then you come to the limit then this become impractical you switch from a digital PSD to an analog approach using the AD630. Ofcourse using the same AD converts as for the digital lock-in

 

Well, my thinking is this. The DDS which is supplying the original signal through the whole system is clocked by the dsPIC itself, using the reference clock module. This means that the DDS and the PIC are both clocked by the same source. The DDS simply outputs a sine wave, the frequency of which is a division of the clock source you send to it. This means that using a timer module on the PIC I can simply toggle a pin and send it to the PSD and it uses this signal as a reference source (since it will be exactly the same frequency if care is taken when choosing the divisors on the DDS and the timer module. Also, I can just preload a value into the timer(s) which are being used to generate a reference signal to vary the phase of the references to get the system running very accurately.

As I mentioned, hopefully all of this will be taken care of digitally eventually, but for the time being I think this will work.