Geotech

Yes, that's exactly the plot I was thinking about, thanks!

Nice plot, GB. Pimento, I was looking on Geotech for the very plot we're probably both thinking of, can't find it. Locator, that illustrates it quite nicely... a big coil will go deeper for big objects than a small coil. Ferinstance, if you want to detect a deep ammo box full of silver, use an 18" coil, not a 6" coil.

In a nutshell: Big coil = more depth on big targets, less depth on small targets, worse separation Small coil = more sensitivity to small targets, less sensitivity on big targets, better separation Concentric coil = Slightly better depth, easier pinpointing, worse ground handling DD coil = Slightly worse depth, harder pinpointing, better ground handling There is way way more to it than this, I would need to write 3 chapters to do it justice.

What Phrunt talks about is EMI interference and can be used to trace an active power line. The V3 has a frequency offset control that you normally adjust to make it quiet to EMI; instead, you want to get near a powerline and adjust it for maximum interference. Try it in MF mode and each SF mode to see which is most sensitive, my guess would be the 2.5kHz mode. If you have access to an end of a pipe or cable you can inject your own signal and trace it, again using interference. The signal generator needs to have fairly precise control because you want to tune it to the operating frequency of the detector (say, 2.5kHz) and then move it a few Hertz off for maximum interference. Connect the (+) signal to the pipe/cable, and connect the (-) signal to a ground probe that's 10-20 feet perpendicular to the pipe/cable. Obviously, never use this method on a live cable. If you need to trace a PVC pipe, it alone can't be detected but sometimes smart installers lay a tracer wire in with the pipe. Where the pipe comes out of the ground, look for an unconnected wire sticking out. If you find one, go thank the guy who installed it.

It's an example of exceptional marketing. MF has an inherent disadvantage in that it is, by necessity, a wideband system which is more susceptible to EMI. But with both SF & MF it depends on how well it's designed.

As mentioned 100 or so times on these forums, BBS & FBS are 2-frequency detectors, not 17 or 28.

The SL board is a newer design and has some changes to make it quieter. It was intended that the TDI would switch to the SL board but it never did. The SL board has more options for mods. Yes, lowering the cap values makes the demods faster with a slight loss of depth. I would leave them alone, I'm not even sure why Reg lowered them on the SL as it also tends to be noisier.

Compare your Minelab curve to the price of gold over the same time span and you'll have your answer. A very large percentage of ML revenue is in African gold prospecting and sales there are very dependent on the price of gold.

I'm not sure what you mean by "turntables." In the TDI, C59-C60 (same as C20-C21) are already 0.22uF but you could decrease them to 0.1uF to equal the SL speed improvement. The 2 caps need to be matched to 1% or better. The TDI has no C56 or D20-21, and C42 is C50 in the TDI and can be shorted. I'm not familiar with the other mods in Jim's post. I think the preamp replacement was the LME49990 but without some other changes just replacing the opamp won't do much.

Jeff, it's likely that something in the transmitter popped, maybe the TX FET overheated. Components should be able to handle up to a 20V battery but there is a chance that a high battery voltage will cause the TX to run hot. Unlike the TDI, the FET on the SL does not have a heat sink, rather it uses the PCB as a heat sink. I'm surprised that 14V caused a problem. Let me know what you find out from Todd.

Yes, 3 positions, on-off-on.

With PI the induced eddy currents spike and then decay back to zero. All MF designs induce eddy currents that exponentially rise to some value before being driven in the opposite direction, also exponentially. In both case you can use time-domain sampling to look at the exponentials, and the exponential time constant is directly related to target phase. With SMF it is easier to use channel bandpass filters to separate the frequencies into respective sinusoids and use freq domain analysis. You can't do this with sequential MF (FBS) so it is left in the time domain. IMO, yes, the differentiator is that PI looks at the RX signal during a TX dead zone. The TX current could be a sawtooth (traditional PI), a square wave (GPZ), or a half-sine (Barringer). But you can have a hybrid VLF/PI. GPX leans slightly this way when using a DD coil to look at the active TX region for iron ID. At White's I built a half-sine detector where a true VLF-like phase angle was extracted during the half-sine pulse and a PI response was extracted during the dead time.

Yes, to a large degree. With GPX models you can use a balanced DD coil and look at the target response during the TX pulse where there is a reactive signal and do some iron discrimination. The TX transition in the GPZ is very fast and therefore creates a very large signal even in an IB coil so it's tough to look into that event and see the reactive signal. But not saying it can't be done.

On the Pro the Coarse/Fine pots are wired in series. The coarse is sill 100k, the fine is 10k and includes the GEB on/off switch. Both pots are wired as variable resistors, i.e. the center lug is shorted to one of the outer lugs to make a 2-terminal pot. You will need to figure out which outer lug to short it to so that counterclockwise gives a minimal value.

Correct, a Bigfoot for the Classic won't work on a DFX.

It's a fairly complicated topic that really requires a whole book chapter and a lot of diagrams to explain*. In the old days all VLF designs were frequency domain (or, more accurately, phase domain) and all PI designs were time domain. But BBS/FBS/FBS2 (all really the same basic tech) created sequentially-transmitted frequencies where freq-domain processing doesn't work because the dead time between frequencies would screw up the channel filters. So ML used time-domain processing which, honestly, looks somewhat similar to normal freq-domain sampling. But it's done on ramped-exponential decays instead of sinusoids. From the time-domain sampling you can still get the equivalent of a target phase. PI is also time-domain (and always will be) and also samples a decaying exponential but it does so during a TX "quiet time" where the TX signal is not changing. With FBS the TX signal is always changing, just like VLF. And with PI, you can never extract a target phase. This was a good differentiator until the GPZ came along. In traditional PI the TX current is turned completely off during the RX sampling. In the GPZ it is not, but it is also not changing (the TX current is a bipolar square wave) so the result is exactly the same: there is no changing TX field during RX sampling. So now my definition of PI is a system that receives during a TX dead time, whether current is zero or a DC value. Put another way, during sampling there is no reactive signal. *Inside the Metal Detector, 3rd ed, Ch 2; pub. 2023 (I hope)

FBS2 is VLF with time-domain sampling & processing. There is no PI going on.

All pots are mono. The pin connections are all linear, that is, outside-middle-outside on the pot (or switch) is the same on the connector. This means you will have to figure out which outside connection goes where but that's pretty easy once you fire up the circuit. If the pot operation is backwards, swap the outside pins. Sorry, I don't have a diagram or a pic.

Pulse delay & Frequency are 10k, although 100k will also work. Gain is 50k. Ground balance is 100k. Threshold is 100k. As I recall the PCB is marked where everything plugs in.

If I read it right, someone posted that the single F modes were the same as the Eq. But the Eq MF modes use different frequencies than the SF modes (2.6, 7.8, 39) and I suspect that 2.6 is for ground analysis. If I were to guess then I would say it's expanded on the low end to favor deep silver. Maybe there's a new boosted 1.56kHz component. That would also explain the incompatible coil. But then, I would also expect to see a lower SF mode, maybe 2kHz.

Most any coil that will work on a TDI/GPX will work on the AQ. A potential problem is that a lot of those coils may not work down to 7us. I once put an AQ coil on a TDI and could then run the TDI down to 7us, but with the stock TDI coil 10us was the limit.

Generally whatever you can do in time-domain system you can do in a frequency-domain system, just differently. You could determine target ferrousity in MIQ by looking at phase linearity through the response so I wouldn't be surprised to see that the Manticore does FE/CO with much the same waveform as MIQ. Also, it's difficult to mix simultaneous MF with sequential MF due to signal discontinuities that will screw up the channelization filters in the simultaneous side. Probably you could hook up a BFO coil to the Manticore and it wouldn't physically burn it up. I took Tom's comment to be more tongue-in-cheek than literal. A typical VLF TX coil current might be in the 100mA range yet battery current may only be 25mA. This happens because the coil is resonated with a capacitor and the TX energy is mostly recycled, with the periodic energy loss due mostly to coil resistance. This is obviously true with a sinusoidal TX but it is also true with a square-wave-ish MF TX. With an ideal coil (R=0) the power delivered to the coil is zero even if the TX current is 100mA. So when someone says "this TX has 50% more power" I don't know what that even means unless it's a total-loss PI transmitter. I assume ML's "50% power" statement actually means 50% more current which also means a 50% higher voltage drive. I think Tom mentioned this results in only a 7% depth increase which is right assuming a 6th power signal loss, which is close. But, hey, 7% is 7%. The V3 had a TX Boost option which kicked the TX drive from 10V to 30V, a 3x increase. This added about 1" of depth to most coins but cut battery life in about half because the TX coil resistance wasn't all that good. And in an MF detector (and really, any detector) TX coil resistance determines the lowest practical frequency you can run due to power losses. This could be why the Manticore's coils are different, they focused on a lower R to run a lower frequency and/or higher drive voltage. Although MIQ runs at 2.6/7.8/39kHz the 2.6kHz component appears to me to be weak and is probably mostly used for ground analysis, not target detection. That's why BBS/FBS was better for deep coins, it was dominated by 3kHz. So if Manticore can run a much stronger 2.6kHz you'll see deeper coins. On the display... It appears to me that the manual brightness goes down to 1 but not 0. This suggests that the display is a transmissive type which (like most cell phones) always requires some level of backlight or you can't read it at all. The CTX and V3 used a transflective display which was sunlight-readable even with the backlight completely off. But very few companies make transflective displays because, in a showroom where sales are made, they tend to look washed out next to an eye-popping transmissive display. Transmissive displays obviously chew up more battery power and any color display is far worse than the near-zero power of a segmented LCD such as used in most detectors. So the Manticore is going to have a higher batter draw just due to the display.

Brightness is now fixed but you can turn it off.

I would start with Todd at Centerville-NW. As I recall White's ran out of the exact displays several years ago so if Todd doesn't have them poke around DigiKey, Mouser, or eBay. There are several compatible displays that more-or-less should be an exact fit. They run $10-15 each.

To clarify on Aureous' post, I modified a TDI to test out an idea I had, it was a lot easier than trying to mod an Impulse. To do this I had to write new micro code from scratch. The idea worked but needs more effort. The code includes a switch to exactly replicate the original TDI timing and I sent a few chips to someone with a bunch of dead TDIs but he says the chips were acting squirrely so, again, needs more work.