Tweet
There was a discussion on another forum about digital metal detectors. Suppose you have the receive coil fed directly into an ADC and everything is processed digitally after that. Therefore there is no analog circuitry to speak of. How much variation in depth (detector to detector) would you expect in such a design? They claimed 1/2 of 1%, but never explained why.
-
-
While the digital circuitry will behave identically unit-to-unit, the ADC, coils, and TX driver might not. Coils will have some variation unless they are PCB coils, which can be done in PI-land but not practical in VLF. The ADC performance can vary, typically a few percent. And the TX coil drive can vary, depending on the design, maybe several percent.
But a digital detector will have far better reproducibility, and each unit will probably "respond" the same (audio, disc, etc).
- Carl -
I believe that you must have at least one analog amplifier between coil and ADC. I think that it is problem to convert a signal of few mV or Hundreds uV direct to digital.Comment
-
It is probably very possible now. You can get ADC's that are 16 bits that run 1Mhz. For a 5 volt range this is 75uV LSB which might be enough for IB. Has anyone tried this? ...BillComment
-
Digital Pi
bmx79,
Another aspect of the ADC as the 'immediate' front end that must be thought about is actually protecting the ADC itself. You would also normally need to match or lower the input impedance to suit the ADC inputs and using a buffer Opamp with a bit of gain (or unity gain) is a reliable way of doing it. Check out some of the input specs of a few ADCs including MPUs with ADC inputs and you will see that PIC MPUs for instance require <6k input impedance or there abouts. Also, I would rather replace a $1 Opamp than a $10 ADC anytime. Although you are correct in thinking 'direct' to digital is possible, the real world application of it requires a few protection measures to be reliable. Just some food for thought.
Cheers
MuntariComment
Comment