Garrett Infinium PI mod - fine tuning?

[jtz54]

After building Gary's PI detector and Trevor Hill's GoldPic detector I decided to purchase a Garrett Infinium. So far operation and performance seem on par with what the reports say and the ground balance and reverse discrimination seem to work as advertised (after finding out my cell phone causes issues with the circuitry).

Anyway, I'm a tinkerer and can never leave anything alone without trying to improve it. After seeing what my friend's Minelab GPX3500 can do, I was wondering if screwing around with the Infinium would be worth it (knowing my warranty would go out the door).

Some ideas would be hand selecting the FET (IRF640) and possibly the receive amp transistors. I say this because on the other two PIs I built, hand selecting these components did make a difference in sensitivity.

Also I noted a big difference in charge caps for the coil between different schematics. Some caps seem very small (47uf) while others at 1000uf or more seems to be to get charged and discharged at rates advertised. The Infinium frequency seems to be around 700Hz. So would increasing the cap, increase power to the coil and depth?

I would love to vary the pulse delay but being a uP controlled circuit that would probably require writing some firmware for the and the chip is probably locked or protected.

Anyone care to give some input on these ideas?

[mario]

is there a schematic for this machine?

its suppose to be similar in design to the older
minelab SDs - lots of mods for those.

[mikebg]

Hi jtz54,
You can improve significant the Gary's PI and the GoldPic using separate RX coil placed in induction balance with TX coil.
This is made with Infinium. Note that the 10 x 14” Power DD searchcoil was the standard coil supplied with Infinium. This coil when coupled with the API – Advanced Pulse Technology - of the Infinium LS make it a powerful detector for operation in mineralized soil. Note that Minelab also uses induction balanced coil configuration with PI.

[jon-ecm]

playing with the mosfet seems like the first thing to do, I would even consider upgrading it to something like the FQPF7N65C, it has about a quarter of the capacitance and triple the voltage.

http://www.geotech1.com/forums/showthread.php?t=17073
has some helpful information on what to expect after lifting the cover.

Honestly, The Infinium is a very amazing detector, Its design is pitifully old and the components are bottom of the line dirt compared to what we have now but it still produces very good results.

Jon
http://www.miymd.com

[WM6]

Honestly, The Infinium is a very amazing detector, Its design is pitifully old and the components are bottom of the line but it still produces very good results.

Jon

Hi Jon

What is "old" in Infinium design? Maybe colouring is something unusual, but dont forget that Infinum is underwater detector and underwater optics requirements are not the same as in land. Otherwise all controls all logical and quite simple.

And which "component are bottom of the line dirt compared to what we have now"? It is microcontroled device and I cannot see something way better in hi-tech at competitors.

I cannot say much about its detecting ability as I still dont have one (this can be changed in short) but agree with you that review are quite positive.

[jon-ecm]

I don't want to bring this too much off topic but I will venture to say that the mcu is very old and the adc used is a slow 12 bit type coupled with several noisy op amps.
10 years ago these components would have been "good" but not anymore.
the adc's speed (10 us) limits it to two samples per decay period but I don't belive the mcu could even process that.
Just for reference, On my new detector that I'm working on, I have a 16 bit 1 msps ultra low noise adc coupled with 2 ultra low noise high speed op amps. powered by a 400 mips 32bit eight core mcu and am about half of the resources.
This setup is able to average 10 sample points on the decay curve and do all the dsp required to get a nice smooth threshold.
I have a hard time beliveing that the 20 mhz 8 bit pic in the infinium is capable of doing much,

[WM6]

I have a hard time beliveing that the 20 mhz 8 bit pic in the infinium is capable of doing much,

I know that there is much more Hi-Tech available than those incorporated in existent MD, but for now, there is still important difference between existent applied technology and available technology. First one is proven as working in field, and second one not yet (or maybe in some very expensive devices not accessible to ordinary detectorist).

I am following your great work, Jon and I support you, with best wishes that your work will be confirmed in field.

[wam]

the adc's speed (10 us) limits it to two samples per decay period but I don't belive the mcu could even process that

I do not think it operates on the principals of sampling decay signal. It looks like it is electronic analog hold and compare design and then filter and reading analog output via A/D converter .

Regards,

[jon-ecm]

I do not think it operates on the principals of sampling decay signal. It looks like it is electronic analog hold and compare design and then filter and reading analog output via A/D converter .

Regards,

I think your probably right, Perhaps if we could find those sample averaging capacitors we could make them a little smaller. Stock it's way to slow reacting, It might be deep but you have to sweep the coil so slow that it makes it almost useless for use out in the field.

just my 2 cents worth