LVDT Signal Conditioning using DSP

Nikhil,

Thank you for posting.  That application note is quite old now and many of the newer C2000 parts will do a much better job.  

Just to mention the implementation in the application note used synchronous demodulation, which is easier if the DSP is also generating the excitation signal.  Otherwise you will have to find a way of synchronising to the external oscillator.  As it's a fixed frequency, the overhead for generating the sine wave is not too large.

Anyway, I would be interested in hearing about your progress so please do update the post.  Thanks again.

Regards,

Richard

Hi Richard,

Thanks for the reply, do you have any suggestions as to which other DSPs would work better? Your suggestions are more than welcome. I haven't worked with that particular DSP and we were going to order it in a couple of days, if you can give some advice as to any newer parts, we can go over them and order them instead.

My thought was, looking at the timing diagram in your note, with the method that you used, the processor can handle at most two LVDTs at a time. I am thinking of going for a higher power DSP later, but first try to duplicate what you have done and then proceed from there. The synchronization issue is interesting. Thanks for pointing that out. My main reason for thinking about external excitation was because my LVDTs run at 10 kHz so exact same implementation would not work . So I was going to give the primary winding excitation signal at one of the ADC channels and use a second channel for the return signal and then compare them.

The reason we wanted to explore your approach is, even with traditional analog circuit we have to use ADCs at the end. Cost wise if we can handle multiple LVDTs on a single processor, it would be better as it will eliminate the need for the external ADCs and related circuits and also allow us to have all the advantages mentioned in your paper.

-Nikhil

Hi Nikhil,

If I were doing it today I would base the design on one of the newer floating-point devices, probably the F28377S.  This part will give you a lot of computing performance at a very reasonable price, and hardware support for floating-point will simplify your code.  There is also a cheap "LaunchPad" development board for this device.  There are several similar devices in the same family, so if board space or cost is a concern you could equally well choose a smaller and cheaper part.  

I hope this helps.  Please do share your progress.  Thanks.

Regards,

Richard

I decided to go with 28377S to start and then see if I need to go for duel core. I just received the Launch Pad for 28377S. Still waiting for other things though.