Post Go back to editing

How to drive multiple LVDT primary windings with ADA2200?

Hello all,

I am using the reference design in as described in CN0371 where in an ADA2200 is driving an ADG794 switch. The the reference has 2 of the four outputs of the switch dedicated to driving P+ and 2 dedicated to P-.

So here is my question: I need to drive 8 LVDTs, can i drive them from the one switch with the primary coils in parallel or have a dedicated switch for each LVDT?

Thanks in advance,

-Erik

  • Hi Erik,

    Sorry for the delayed response.  Can you describe what your application is, where you are driving 8 LVDT's?  Do they need to be driven and read simultaneously?  Or are each of these moving sequentially? 

    I don't think tying the primary sides of all the LVDTs together is going to work in that configuration.  Currently the primary winding has 3.3V and 0V on each winding, and the ADA2200 uses a clock to switch the voltage on the winding.  So if you have 8 than you will want to ensure the same basic function.  Now you can put up to 2 LVDT devices on each ADG794 switch because of the other 2 unused outputs.  So you could put 4 of those on your front end to get your 8 channels.  One thing to take note of here would be how to ensure that none of the signal going into the ADA2200 would back feed into the other LVDTs connected.

    Another thing to consider is that the ADA2200 can only measure 1 LVDT at a time.  So that is something else that could limit you in the application.  If you need to do things very fast than it might be a good idea to get multiple ADA2200/AD7192 devices in your signal chain.

    Brandon

  • Hi Brandon, no worries on the late reply. Look at me here now haha... So there will be 8 LVDTs, 8 ada2200s and a 8 channel differential ADC. Each LVDT will have its own signal path, but i want to prevent beat frequencies so i would like to run the primaries on the from the same source. Looking for the best way to do that.

  • Hi Erikg,

    Taking a look at this - Your concern about beat frequencies is valid. In fact, a very powerful debug technique is to purposely force an in-band beat frequency on one of the channels using an external clock generator (multichannel function generator, or two function generators that share the same reference oscillator.)

    That is, set 7 channels to exactly the same excitation frequency (the "offenders") and one channel (the "offended") to a 1Hz offset (or other small offset), and see what you get. Best case - no effect, indicating no crosstalk. Worst case - 1Hz artifacts show up in the offended channel, so synchronization is necessary.

    How much of CN0371 circuit are you implementing? That is, are you using the AD7192 ADC, or some other ADC? CN0371 uses the AD7192's MCLK2 pin to clock the ADA2200 - If you are using 8X AD7192s, the first step would be to drive all MCLK2 from a single master clock source. This can be a crystal oscillator, or one of Analog Devices' silicon oscillators. Note that you will need to configure all AD7192s to use an external clock - see description in the AD7192 datasheet, pages 32-33.

    -Mark