ELDORADO.VE

Lamar, it might have to do with the cristal structure of the nugget. Check out this site with cristal gold nuggets. http://www.rarities-xls.com/Spec1thru12.htm Eldorado.ve

Hi Reg, On a scale of 1 to 10, how would this coil come out? 375 uH, 0.5A coil current, sample at 5 uS with a gain of 23, Sample at 7 uS with a gain of 100. Eldorado.ve

Hi Reg, thanks for the explanation. This gives me enough info to make my decision. I agree that most of the nuggets are small and if somebody hits a spot with large nuggets deep down it is worthwile for him to get another machine for that purpose. I have looked at the gain of the preamp and it has no more surprises after I found the source of the noise that produced the "sweet spot". I have to finish the coil now and see where the first sampling falls then. What do you think of my idea to use a S+H amplifier to average the sample and eliminate most of the noise and then amplify the now DC current? I can use a sample time of 5us or even shorter for that. ELDORADO.VE

Hi Reg, interesting thread that one at Geotech. My detector is to be used to find gold nuggets. personally I prefer to dig many shallow holes but it seems that many people want "the deep one" ( specially novices) The first question is always "how deep does it go" Since I dont go much out to the goldfields myself, i have to build a detector for the young ones who have the stamina to go and dig. I agree, it is easier to build low power. Question is what do other people want???? Is there maybe a forum where we could put the question and get sensible answers? ELDORADO.VE

Hi Reg, thanks for the good information. Had some thunderstorms here also. The lightning blew up a brand new lightning conductor on the 13800 V line in front of the house. Maybe the lightning conductor was the source of the groundfault that was plaguing me. You are right. the sweet spot at 33 gain was due to 2 opposing kinks cancelling each other. I have looked at Geotech for Erics post on the gain loss but could not find it. I have noticed with some surprise that with a first sample at 6 us and only 33 gain the deflection of the decay curve is very visible for a Nickle. I was used to gains of 1000. So what is happening? For my developing work I have set up a very flexible platform that lets me vary the Tx from 14us to 140us and all other timing accordingly. The coil works well with 12 V from 14us to 140us Tx ( got some smoke from the 3 Watt damping resistors when I pushed the Tx to 200us) and 2 to 12 ohm with only slight changes in the delay time and slight changes in the damping resistors. This puts me in front of the decision now between high power or low power TX. What is your advice? ELDORADO.VE

Hi Reg, some answers to your questions: The 60hz keeps creeping in as fast as I take it out, but this is on the open breadboard I will look for a way to shield the whole system in the final box. The Schottky was the worst, did not expect that. Want to try to get hold of different Germanium diodes to try. After all, 2 us difference at 6 us is 30% difference. The winding is complicated but seems to be good because I used only magnet wire, hope to use Litz when I can get hold of it. have tried many shielding schemes with whatever stuff I can get hold of. Will shield this coil with Graphite paint after potting it with polyurethane foam. I would prefere glass microbaloons with epoxy but cant find the microbaloons. Yes I visit the Geotech forum a lot and much apreciate your frequent and very helpful posts. ELDORADO.VE

Hi Reg, the Tx voltage is 12 volt. I have tested the tx between 14 and 140 us and the coil between 2 Ohm with no additional resistor and with +5 Ohm and + 10 Ohm resistor meaning between 2 and 12 Ohm. the first sample falls between 5.5 to 7 us throughout these ranges at 33 gain. Here is the tricky thing. This is the only gain that this setup works. Kind of a sweet spot. Neither more nor less. From this point I take 7 samples spread at regular distance. Store in Sample and hold amplifyer where I can average the individual samples over n readings. The output of this amplifier is a solid and clean dc voltage that is then amplified in a differential amplifier, or rather several differential amplifiers. the storage and amplification of the seven samples individually should give me the possibility (by pic micro) to pick one specific sample to match with another specific sample, with various gains etc. Each sample pair is developed into a different audio autput. I believe the human ear has the capacity to differentiate a very wide variety of audio signals. I hope this will lead to a higher degree of target ID and ground elimination. The design is still in developing stages and I apreciate all the help I can get. Writing the source code for the PIC 16F817A is one place where I really could use help. I have not worked with microchips or programming before, but this is where the future lies. However, first of all I have to build a real good analog front end that produces a clean good S/N signal before I start shuffling it around digitally. ELDORADO.VE

I was a Garimpero there 40 years ago. Obviously things have changed a lot since then. Wish you the best of luck. ELDORADO.VE

Hi Reg, thanks for your help. My 60hz problem does not have spikes, just all pervading 60Hz noise. I guess if I find a way to control that, my detector will become immune to powerlines. I have made some further tests with diodes. between the 4148, 1N5817 Schottky and a generic Germanium diode, the latter gives the best result by about 2 us. Analysing this, I feel that possibly the reason is that I am using the diodes outside their parameters. These small signal type diodes have a forward current limit. I use a 1k ohm, 3 Watt damping resistor across the coil and another same resistor before the diodes. Both resistors run hot. So far I am quite happy with the coil, but will see what I get after shielding it. This setup works fine with 6 to 7 us first sample delay with pulses from 1 to 6 amps, 14 to 140 us Tx, at 6 Amps obviously I need to beef up the damping resistors and limiting diodes. One more thing I learned: Using the mosfet avalanche diode for clipping the flyback, introduces a lot of noise in the 10 megaherz range. ELDORADO.VE

Hi Garimpo, your reports are very interesting. Hope you find a lot of gold. How do you manage to connect to the Internet out there????? Watch out for those piranhas, both kinds .... All the best ELDORADO.VE

Hi Reg, I use the 5534, in the pix shown, in noninverting mode. The cable is double shielded audio 2 conductor cable (Belden) 17 pf capacitance. I work in a very noisy environment with a 60Hz, 13800 V groundfault noise that is all penetrating. Do you have any advice how to filter this 60Hz noise out? The "Ideal diode" from Linear, LTC4412 is supposed to have a forward voltage drop of 20mV. The way I understand it, using it instead of the other clamping diodes, it would reduce the saturation of the opamp, created by amplifying the 0.7 V forward drop of the diode many times. It also would help in the damping of the coil. However, possibly it will be so noisy that it can not be used for our purpose. ELDORADO.VE

Hi Reg, sorry for using the wrong word. What I meant is CLAMPING diodes, with a C, yes am refering to the limiting diodes before the preamp. The pics are at the exit of the preamp, non inverting. Unfortunately I can not read the numbers on the glass diodes, most of my parts are some 20 years old. The forward voltage of the pair of diodes that give the good result is about 0.35 V or halve that of the other pair, that has about 0.7 V. This is why I thought about trying Schottky diodes that have a low forward voltage drop. There is also something called an ideal diode on the market that gives a forward voltage drop of 20 mV. I wonder if that would work? Since we measure the signal when it is very near 0V. The coil resistance is about 2 Ohm, I added two resistors of 5 Ohm each to reduce the L/R time. I have one more curious thing happening. When I conect the shield of the cable to ground, it reduces the overshoot. Could this possibly mean that the overshoot is due to some external noise? Btw, the coil's self resonant frequency is around 750 Khz with the cable, but the coil is not shielded and pottet yet. ELDORADO.VE

Hi Reg, Attached are two pics of the scope. The divisions are 1us. The coil is 310 uH, 11" diameter and 12 Ohm, including resistor. Tx is 65us. The difference between the 2 pics is only different clamping diodes, all other factors equal. This makes it obvious that by changing to different types of clamping diodes, some faster sampling rates can be obtained. I wonder if you have tried some Schottky diodes? ELDORADO.VE