Inductive halfbridge calipers

[branadic]

Hi folks,

I found those ±1 mm calipers on ebay and thought they are cheap to get my hands on them for a project I'm working on.
Unfortunately, they turned out to be inductive halfbridge calipers, but I was expecting them to be LVDTs in the first place.

There are these old panel meters to display the position, but even used they cost an arm and a leg and are purely analog without access to the electrical position information. I've searched the web in the hope to find some ideas or solution how modern technics can be used to excitate and read out the signals, such as the digi 3plus, which costs a fortune compared to the calipers.
But I couldn't find anything suitable. Anyone here has an idea how they work or can come up with a good suggestion or even a DIY / OpenSource solution?

What I could find out so far is, that beside the changing amplitude with position, they seem to show a position dependent phase shift for frequencies <10 kHz.  At 10 kHz and larger only the amplitude changes with position.

Every suggestion is welcome, be it an RPi solution with an appropriate shield or microcontroller solution. I'd like to have a resolution down to ~100 nm, just for completeness.

-branadic-

[Henrik_V]

I would start with a soundcard.
Drive it with a sine , one mic channel and divider as reference, the other for the signal. Apply a sine fit to get amplitude and phase ...

If you hook it to a scope and a sine generator you will find it's not too complicated

A DAC (rude: one bit -> rec) and one well timed sample is the simple method.

[branadic]

Thanks Henrik_V. A soundcard is not really an option for me as I don't wanted to put a computer next to the DUT, but if you were refering to ADC/DAC shields for e.g. a RPi, that's more of an option.

If you hook it to a scope and a sine generator you will find it's not too complicated

Did that yesterday already, but a scope with 12 bit resolution (15 bit with oversampling) is still different to a standalone solution. I looked into AD630, as I found this publication: https://www.mdpi.com/1424-8220/23/7/3703 which is a pretty ancient piece of silicon. But that is true for AD698 too.

I was also thinking about simultanious sampling of excitation voltage and signal and doing the rest in software (devision and lowpass filtering), but that requires quite a fast ADC with sufficient resolution (24 bit) given that the excitation frequency is 13 kHz, which is what most of the halfbridges seem to be used with.

-branadic-

[Terry Bites]

I remembered this note from LT on LVDT signal conditioning.
There's another handy note on an exitation source.
It all looks sane enough.

PGA970, if only....

[Henrik_V]

Good analog multiplier and all these analog sync demodulators/LVDT/AC-Bridge IC for low freq are obsolete nowadays.
Digital DAC and ADC is the way to go.

No real need for a second channel. A DAC (or 8bit DO  with resistors, as it's been done in the past) with filter & stable amp/buffer feeded by the same Vref than the ADC is fine. Enshure a stable amplitude at known frequency.
Recent µC have a 16bit ADC buildin,  64 sample per periode is fine, take 4+x periodes apply a 3 parameter sinefit .. will beat any 24 bit peak detection.  LVDTs (or halfbridge) of the size you have will measure temperature anyway for values with nm index

[branadic]

Received a working meter today and indeed the halfbridges work with the meter.
Measurement ranges down to ±300 nm, sum, difference... insane.
-branadic-