Geotech

Detector coils are not antennae, they are designed as a transformer. There is no EM propagation, only a local magnetic field. The coils don't have to be impedance-matched to anything. Most RX coils are just connected to a high-impedance preamp, with some token load resistor (typically 10k). Some designs throw in a few extra components to reduce EMI.

The color scheme for Mixed Mode is mine, it's the Garnet & Gold of Florida State, my alma mater. It looked surprisingly good so I used it to demonstrate the custom color abilities of the V. BTW, Geotech shares a similar color theme. Go Noles!

I was fortunate to work closely with Dan for a few years, he taught me a lot. But, no, he & I agreed the TDI design had gone about as far as it could.

1974 or 75, Jetco Goldstar TR. Then a White's 6000/d when I realized how lousy the Jetco was. I get a chuckle from the irony that I ended up at White's for a while, and now I work at Jetco. Or, what used to be Jetco.

Not in a VLF, the "ground zone" (ferrite) is a pure magnetic response (no eddy response) and has a 0° phase shift. Everything from there up to 90° represents ferrous content (includes some amount of magnetic response), and from 90° to 180° represents non-ferrous. By definition, there are no eddy targets in the "ground zone." A traditional PI is different, as it doesn't give a reactive response. The "ground zone" is usually a GB point set for magnetically viscous material, which can mimic a certain eddy response. There seems to be some confusion over whether you're describing a VLF or PI.

Fisher CZ and White's DFX/V3 all simultaneously transmit & receive 2 (or 3) frequencies at the same time, and use continuous-time (so-called frequency domain) demodulators. What I call "Concurrent MultiFrequency," or CMF. Minelab BBS/FBS/FBS2 all sequentially transmit & receive 2 frequencies, and use discrete-time demodulators. What I call "Sequential MultiFrequency," or SMF.

Hahahaha I ask David that every day.

I'm not feeling near as pissy as I was when I learned that my wireless pinpointer patent was dropped and XP would own that technology, so I won't comment on all the behind-the-scenes stuff other than to say, official statements rarely reflected reality. One bit of truth was that the Cypress module has a low transfer rate (62kbits/sec as I recall) which is good enough for audio packets, but transferring programs turned out to be slow and prone to drops. New modules are 15-30x faster and can easily do the job. And it's only with Bluetooth 4 that BT audio has gotten fast enough to be usable for detectors, but still tends to be more power-hungry than using a proprietary protocol.

Deus uses a PIC24F processor, which is simple to hack and almost impossible to code-protect. By taking the Deus to the African market XP was begging to be hacked.

This idea has been batted around for years but, to my knowledge, has ever been implemented. Several reasons why come to mind: 1. Target responses don't vary a whole lot until you get beyond a 2:1 frequency ratio. That is, the difference between 10k and 12k isn't enough to make both of those frequencies worthwhile. So we could sweep the following: 1k, 2k, 5k, 10k, 20k, 50k, & 100kHz. 7 frequencies would give you all the information you need, maybe even as little as 4. 2. Most detectors use zero-IF demodulators and for good SNR require multiple cycles of the frequency. The more frequencies you have, the less time the demods have to accumulate data. Beyond 3 or 4 frequencies you'll see this as a lag in responsivity. On the upside, it would be a good method for pinpointing where lag is not important, and you could even do full non-motion (static) disc. In fact, White's has a patent on a "zero motion 3F disc method" that was supposed to end up in the V3 but did not. Ideally, you would want to use direct sampling in a sweep-frequency detector, but this requires a really fast high-precision ADC. Direct sampling detectors (X-Terra, Go-Find, Prizm 6T, Deus) currently max out at less than 20kHz. 3. Sweeping the TX across a wide range of frequencies is easy, a DDS can do it. But as the frequency increases the coil current decreases, so a 100kHz TX signal would be very weak. You can see this in Minelab BBS/FBS detectors, where the 25kHz TX signal has 1/8th the strength of the 3.125kHz signal; they go deep on silver, but don't do well on small gold. Solving this means increasing the TX voltage with frequency which sounds easy but it's not. Designing a coil to work well enough over a 100:1 frequency range that you can easily swap coils without having to recalibrate the detector is also a huge challenge. 4. Until about 10 years ago the processing horsepower just wasn't there. Now it is, but it takes a while for detector companies to catch up to what the rest of the mass electronics market is widely doing. So a sweep-frequency detector is doable, but I'm not sure it would offer the benefits people imagine it would. Probably a wide-band 3F detector would offer 99% of the target ID benefits while being overall more usable and deeper.

Nothing was ever changed for EMI issues. The PC interface was done and working, really nicely. I pushed hard to challenge the ML patent but was alone in that effort. I expect XP will win.

Well, it's more complicated than that. Instead of a sinusoid, you could run a square wave drive which generates a triangle wave current, which does have a fundamental and harmonics. This is exactly what a Fisher CZ does. A SF detector could demodulate only the fundamental (ignore the harmonics), and vary the square wave frequency to whatever it wants, the advantage of which no coil capacitor needs to be selected. The V3 does it this way (in SF mode), and I'd guess (but don't know) that the Alter 71 does it this way, too. But I expect most of the switchable frequency detectors on your list run sinusoidal transmitters.

Good article Steve, but wrong on the harmonic info. Switchable frequency machines switch to different fundamental frequencies, not harmonics of the base frequency. Most of them run sinusoidal transmitters so there are no harmonics, and they switch in different capacitors to do so. Also, the DFX doesn't really run in single frequency mode. It always transmits/receives 2 frequencies, just ignores one of them in "single" mode.

Yes, nuggets often have multi-domain eddy responses.

Both polar and cartesian plots can have XY axes, what they represent depends on what they represent. Technically, for a metal detector they should be the IQ axes, for the "in-phase" and "quadrature" signals. Sometimes they are called the XR axes, for reactive (X) and resistive (R). Sometimes the ferrite is on the left, sometimes on the right, and sometimes at the 12 o'clock position, depending n the preference of the engineer. Yes, kind of. Bent and twisted metal ends up having eddy responses with different phase angles (the metal thickness appears to vary), so the response is not homogeneous. A bent nail, or can slaw, are good examples. Jewelry can also be all over the place.

Deus looks to be the same as V3, maybe a little more real time. The V3 attempts to plot a single sweep over the target, which is hard to figure out without a motion sensor in the coil. A perfectly straight tight line represents a strong single-domain eddy response. The tilt of the line is the raw phase. Close to 3 o'clock (180°) is high silver; 2 o'clock might be copper cent; 1 o'clock nickel; 12 o'clock would be salt; between 9 and 12 o'clock is iron; 9 o'clock (0°) is pure ferrite. The distance from any point on the line to the origin is the signal strength. Yes, this is a simple vector response. Most (all?) detectors map the raw phase (0-180°) into a VDI range (i.e. 0-100); there is no standard for this. As the response balloons out into a tilted ellipse, the ballooning represents a changing phase as the coil sweeps over. Often this is simply due to a weak signal. So you can read the tightness of the response as a confidence level. The wild responses can be due to 2 things: multi-domain eddy responses, as in the silver chains; and combined magnetic & eddy responses, as with the bottle caps. As a bottle cap enters the coil field it initially looks ferrous due to the iron content. As it goes past the center of the coil the eddy response of the flat steel dominates and so the phase at the peak amplitude of the response looks non-ferrous. Many detectors report target VDI as read at the signal peak, which is why bottle caps fool them. With the XY plot you can visually see the whole response, so bottle caps become easy to discern.

X=reactive and Y=resistive; together they give you the phase. So for a straight line response, the angle of the line is the phase, and the length of the line is the amplitude.

That was pointed out by customers over & over on the Prizms. Before I left, the MX Sport package had reached proto stage, and I tried to convince mgmt that the h/p jack was on the wrong side, and why. I think my effort actually persuaded them to leave it where it was. In a nutshell, that's why there were so many missed opportunities.

The wireless pinpointer connect was something that the White's TRX had designed into it, but it was never implemented. I even filed a patent application on the technique about 4 years ago. Because the V3 wireless module was going obsolete, everything was designed around a new module that was supposed to have debuted with the first new 'Sport design. The pinpointer would tell the detector to disable its transmitter, and pinpointer audio would be sent to the headphones. Since the TRX is a full VLF design, it could even send VDI info to be displayed on the detector screen. User setting changes could be made from the detector or a cell phone app. However, White's mgmt didn't think much of the idea from the beginning*, it was entirely driven by Engineering. After I left they let the patent app expire, and apparently even wireless audio got dropped from the MX Sport. As far as I know, the socket for the wireless module is still being put in the TRX. It was intended that early pinpointers could easily be upgraded to wireless, I'm guessing something that will never happen. - Carl * The verbatim quote: "I don't like that idea at all." That was the day I first asked myself, "Why am I here?"

It works, I've done it. The ink used in paper money has iron in it, any good BFO will detect it but only an inch or so away.

The short answer is no. And no. And no. And no. A patent forbids you from even making a personal device that infringes, with one exception: you can build the device for the purpose of evaluating the patent claims. If you build a device for normal use, that's infringement whether you sell it or not. If you build a device and give it to someone else to use, you both are infringing. If you sell a kit that, when built, infringes, then you are guilty of inducement to infringe. Etc etc.

From 2012, Carl's 7 rules of PUD: http://www.geotech1.com/forums/showthread.php?19376-QED&p=152608#post152608 For those with the dream of being a one-man metal detector manufacturing company, post #129 suggests: Here's a challenge for you... build a production run of 20 units. Just 20! See how long it takes you. See how much money it costs you. Keep very accurate records. Figure in labor, overhead, taxes, marketing, and eventual service. Then see how much you need to charge for those detectors to make it worth your while. Those 20 units shoulda been ready to ship upon announcement.

If I don't get back within a year, then a follow-up is probably in order. Send me your email, I'll set you up right away. carl@geotech1.com

All ya have to do is read the "Can't Register?" post. It tells you why, and how.

Yep, Geotech ain't much different than the rest of the internet, but it is interesting. Lots of good ideas pop up from time to time, very few ever get developed into something usable. A lot of the guys don't have the complete know-how, or don't have the time, to get it to the end. This feeds back into MD's "kickstarter" thread, where the kind of people needed to turn a metal detector idea into a metal detector are people who have done it before and know all the nuances and gotchas of what it takes to get to the end. The people who say they can do it are plentiful, the people who can do it are exceptionally rare. As for Funfinder's thread, I think his frustration is that people on Geotech have been build plain ol' PI detectors for, what, 16 years now? And there doesn't seem to be much forward progress in what they've done. I would agree with that. I think he'd like to see more VLF and multifrequency developments. But, compared to PI, those are really hard for DIY folks, so most of them stick with simple PI circuits. I get his frustration.