Making a probe with LDC1000?

[bklein]

I'm curious about the possibilities of making a MD20-like probe using the TI LDC1000 sensor.
What are your recommendations to start with? The probe on the MD20 has an adjustment collar that adjusts the inductance from ~128-143uH. Would it also have a ferrite core? How does this work and is it even applicable to the LDC1000 design? In another thread it was mentioned that the Md20 used 300KHz frequency to gain better differentiation of the audio. I didn't want to hijack that thread but the audio on the MD20 is fairly constant frequency. So I'm confused as beyond 100KHz supposedly skin effect a... takes effect.

[Yeasty]

I just ordered the evaluation board to play with. Apparently this thing can measure conductivity. It would be really cool to come up with a discriminating probe.

[Carl-NC]

I've got the eval board, very unimpressive with the PCB spiral coil but I expected that. I have not yet put a custom coil on it. The Falcon runs at 300kHz just so it can detect fly**** nuggets.

[bklein]

The spiral coil does seem to be able to "see" the little gold spec in Falcon's MD20 card but you have to limit the Rp range. I'm feeling stupid as I've read through their literature several times and still don't know how best to go about using this part. The eval demo software gives Proximity and Inductance real-time graphs, but the spec shows outputs of frequency and proximity. I guess proximity is derived from what the tank experiences with close proximity of the target vs far and frequency is tank frequency which deviates in the presence of the target. So does the frequency go up with ferrous and down with nonferrous? I have never used the MSP430 kits (haven't had hands on with a microcontroller in probably 5 years or so) so not looking forward to it as I don't have that kind of time to dedicate to it. A faster approach may be for me to attempt to interface the kit via SPI to a Labjack and Azeotech DacFactory software we have at work. Making GUI's with that may be easier (I got a power adapter load tester going with it years ago). The MD20's best feature as far as my limited experience shows (just got one last week) is the differentiation between iron and gold with the probe. That's pretty neat. I think it would be imperative to duplicate this feature if doing a probe. For general purpose larger targets my Minelab probe works fine - others work likely just as well so why make another one? I'm curious what you are thinking of trying. There is an analog signal apparently available at the CFB pin...

[bklein]

I played with the eval kit some more last nite and am left encouraged.
I used the Falcon sample card (gold/black sand/pyrite) along with a penny, nickel, and a small chunk of magnetite.
Process was this: set the oscillation frequency, click on autoscale for both graphs, then go to the limits section to set limits just outside of what you saw with the Autoscale. Then take the samples and advance them towards the stock pc coil. It is interesting to see how these response signatures change with the oscillation frequency set by using the eval GUI. There is definitely hope in making a differentiating probe here! This without even trying other coils.
I will be out on vacation for a week or so but upon returning I'll mess with an Aardvark SPI host adapter and Beagle analyser I discovered we have in our lab at work. May be over my head for awhile as I've never used these tools. Fun to experiment... If anyone else has some experiences to add (and bench test suggestions) please discuss.

[bklein]

Jeez how time flies and I don't get anything accomplished. Has anyone got any updates on their experimentation with this eval kit? I haven't spent much more time with the kit but really feel it has potential. I've been to 3 Havasugoldseeker digs so far - last one last Saturday. I use the Falcon MD20 for fine/final detection but the Falcon has been acting up on me though - sounding off all the time or requiring readjustment of the Sensitivity control. I think something's wrong with the unit. I think a better probe could be made with the LDC1000 as it would be trivial to make ferrous and non-ferrous detection sound different. The MD20 detects its own cable, which is a pain. The LDC1000 could be put into something the size of a pen.

[Matt_Rowe]

I was thinking along a line of a PI small enough to work like a MD20.
My Goldmaster MD has a 50khz and can pick up as small as 3 grains.
I think a small unit,that has an iron id, and 300 khz, or a PI would be nice.
Especially a PI.

[bklein]

The ldc1000 would be way better as it can differentiate between ferrous and non-ferrous.