Multi-Frequency

[Davor]

The other point is that it MAY be very fuel-efficient if you let it. I guess the marketing leader needed to do something different than Fisher simply to avoid patent quarrel. With nowadays MOSFETs, square wave excitation can be more efficient than a classic free running oscillator using a high Q tank, but with a freedom to assume just about any frequency you need - within reason.
Now that both patents are expired, this approach is a fair game for everyone.

[Skippy]

Anyone had any more thoughts on this? Carl says BBS isn't as complicated as it first appears, so I'm happy to consider it (and ignore my Fisher CZ-based idea). Has anyone reverse-engineered the BBS machines? I don't recall seeing anything myself, other than coil pre-amp stuff. But I haven't actually searched for any details.
I rather like the idea of BBS running at a higher frequency. Firstly if you want the standard frequencies (3K/25K) you might as well just go and buy a ML. Secondly: many single-freq machines are favouring freqs around 13KHz. Now some of that is just "compromise", but my gut feeling is that a "BBS" machine that ran at 8K/64K would be a pretty good all-rounder.
I'm a little concerned about the 64K, I admit, it is a bit high. Which made me wonder if a 4:1 or 6:1 frequency ratio would be possible? If the details of how BBS worked were understood, then we would know if lower ratios would be possible. Then maybe an 8K/48K machine would be a possibility.

[Davor]

The first thing to understand with multifrequency is that it deals more with subtraction than summation. What you subtract are the components that are found at both frequencies that are resulting from soil. What remains are the components that are different at each frequency, and that belong to the target. These are increasingly different with increased frequency ratio, but there are differences even with ratios as small as 1:2. However, going too close you lose sensitivity. Going too far makes for other complications, so it is a compromise that will get you farther.
"Normal" soils will have fairly equal reaction (power) for a wide range of frequencies, while a sea water will have reaction proportional with frequency. Hence, with a 2 frequency rig you may have either soil or sea elimination using frequencies subtraction only, but not both at a same time. You may instead have two pairs of frequencies to do both soil and sea, each pair eliminating sea, and one pair against the other eliminating soil. Also 2 pairs may have a common frequency, which makes for a more practical 3 frequency solution. So frequencies 1 and 2 make the first pair, and frequencies 2 and 3 make the other. Say, 1:2:4 or 1:3:9
Also with a single frequency pair you may expect the infamous hole in response, so going for a 3rd frequency is a better choice.
The beauty of BBS Tx timing approach is serialisation of a Rx process. BBS transmits a sequence of pulses belonging to a single frequency at any given moment, and the clever part is only in transitions between frequencies that happen at middle of a pulse. You need only one Rx channel without complex fixed filtering, simply because there is no interference between frequencies, as they are not active at the same time. If only there were micros out there with 16 or more bits ADC, and you could crunch everything in there with only a simple low noise preamp. Frequencies are processed one after the other, their values are updated and stored, and in a separate process you do all the arithmetics, as if they are happening concurrently. Easy.
A little less obvious thing is that BBS timing provides for a relaxed choice of frequencies. You can have any kind or ratio, even 1:π if you like. However, you need to take care that each frequency takes ~equal time for the equal output. So in effect you may simply make a small variation in a, say, middle frequency, and thus fine tune GB without any other adjustment. this fact makes me believe I can do it completely analogue and scaringly simple. The simplest instalment would be a zero-brain rig akin to a TR that would just indicate targets and ignore soil and sea. A more complex one would have an audio VDI, etc. If there is only more time to play with it...

[Skippy]

It sounds like you have some experience and understanding of ML's system. So my next obvious question is - how do you know this? Do you have a partial or full circuit diagram? Is there still a problem with publishing/distributing such circuits, due to patents, IP, etc? I have searched this forum, and found threads that suggest that Sovereign circuit diagrams have been posted in the past, but have been withdrawn.

On to practical matters:
I understand why there is a pre-amp inside the search-coil. But I'm sure an equally capable machine can be made with a 'simple' coil, and the pre-amp in the main electronics box. So my question is, what commercial search-coil would be most suitable for a ML-style machine? Obviously any existing BBS-compatible coil would be one solution, but how well it would work at other frequencies is unknown. Other 'simple' coils that would be worth considering are the Fisher CZ-series coils (5 / 15KHz, force-driven TX coil) , and the Whites DFX . What about First-Texas broadband coils made for the Teknetics Greek series (8KHz), G2 / GB-Pro / F19 etc (19KHz) ?

As there is a significant difference between wet-salt beach operation and dry land operation, would it be sensible to make one or the other, rather than possibly make things too complicated by attempting a dual-mode build.? I'm a land detectorist, and don't have any interest in beach machines, other than for educational purposes.

[Davor]

It sounds like you have some experience and understanding of ML's system. So my next obvious question is - how do you know this?...

No, I have not dissect any marketing leader schematic, apart from a SD2000 that was posted on this forum, but irrelevant for this matter. Accidentally I was playing with a way to tailor a Tx that would be capable of running multiple frequencies sequentially, as a mind game. I came up with a scheme that looked familiar, and it turned up to be used by ML machines. Next I searched through the patents covering the technology, only to learn that they are kaputt, and a fair game for everyone.
Knowing what I can do with such a signal I learned what they possibly could do with it, and surely they did many of those things. Also their marketing hogwash suddenly made half way sense, as the real benefits creeped through the huge pile of nonsense. I think they'd sell more machines if they just stop treating their buyers as complete idiots.
Also I like to keep my projects simple, and reproducible in 10 years, which makes copying ML detectors impractical as they tend to shove plenty of redundant stuff into their boxes to justify their price tags. The mojo of their machines fits into just a small fraction of chips they put there, and the rest are just gimmicks.

Long story short - no, I have no interest copying ML detectors. I'm just very happy the Tx scheme that is capable of lots of magic is now free for all.

[Carl-NC]

Sovereign schematics haven't been posted here, but if anyone has non-copyrighted schematics (reverse-engineered) they are free to post them, or send them to me and I'll post them.

I have done some probing on a Sov and sketched out a little of the front end, but not much else. Doing that is a good way to see what's going on. TX timing would be super-simple to pound out on a micro, and because the RX doesn't have all the channelization filters it's simple as well. Davor is right, you could do an all-analog easy-peasy detector, though I'd still use a micro for timing.

There is nothing magical about putting a preamp in the coil. Easier not to, just use a good wideband coil. Salt operation notches the salt response, which also notches fine jewelry like thin necklaces. If you don't need salt notch, then it's best to leave it off.

[Skippy]

I understand there's no magic to the pre-amp, I'm familiar with the circuit diagrams of the Sov coil, and the Xterra ones, too. It's just that I don't like the idea of making 'control box' electronics, AND a wideband search-coil with frills. It sounds like a recipe for frustration and failure. If there's a coil out there that is suitable, any new project should be designed around it. If anyone wants to homebrew something different - extra-large, bigfoot etc, then they can still do it.
And it's worth pointing out that well-regarded machines like the Tek T2 don't have pre-amps in the coil. The front-end op-amp they use isn't even the lowest-noise device about, and the components around it aren't minimal noise choices. Also, the RX coil could easily be half the resistance it currently is. So there's room for improvement.

I do own a Fisher CZ-series 8" concentric search-coil. When I get time [not soon] I could have a try at a frequency 'sweep' and see how wideband it is.

Another thought I have had about the BBS machines in the past is: why did they not make them capable of running single-freq, at either 3 or 25kHz? It seems like it ought to be possible, yet despite improving the system (FBS/FBS2) they a) never used different frequencies and b) never offered single-freq operation. It would be one option worth considering if we built our own.

[Davor]

Single frequency processing is entirely different from multifrequency. In single frequency Rx you need to eliminate ground by multiplying a GB channel with a local oscillator at quadrature to a ground component, while in multifrequency you deal with simple summation. Plus you need to provide some sort of GB tuning mechanism for a single freq part. It may surprise you but programming such entirely different jobs takes a lot of extra room, and micros from the era these devices were designed had problems squeezing in lots of code.

[Skippy]

I was assuming that a ground-balance control would have to be added, possibly two of them if you wanted the machine to run at either of the two freqs. And I suppose you would have to set them at the "zero" position when using multi-freq operation.
Obviously I'm generating these ideas with almost no knowledge of BBS circuitry.

[Davor]

Just consider BBS is transmitting frequencies one after the other. Whatever you can do with that is up to you, but if you are thinking multifrequency, and minimum buttons, you can have a turn-on-and-go solution that even trained monkeys can use.

[Skippy]

I understand the concept of 'two frequencies one after the other' etc etc etc. But I'm an engineer, it's the detail that matters, and I'd need a circuit diagram to see that, or some of it. The micro and it's function would still be an unknown, I realise.

[Davor]

That's not because I'm holding any critical information from you. I can not withhold what I don't have. The concept with a micro is possible, but I consider it exercise in futility because there are no micros capable of doing proper ADC at this time, and also before this time. Perhaps in a near future we'll have a decent micro with a proper ADC and that will be another thing.
You may use a dedicated proper ADC and pair it with a micro, or you can do some oversampling with an existing micro. Just go ahead. After you fix that, the very summation/subtraction algorithm is nearly trivial.
Also when you have target information extracted, you must decide how to do a human interface. Most rigs out there suck at that part.
So whoever did multifrequency by now had to do it in analogue way, or had to shove quite a few chips to crunch it into digital. BBS favours digital, but I think it'd be more fun to do it in analogue instead. At least you'd be able to reproduce such a rig any time in the future, while with micros ... not so much.
You may also do it 50:50 but in my book it is stupid. When you do the effort and make a proper analogue front end, why give up half way through?
When I make time, I'll start it as an analogue project. I have a few ideas how to tame an IB coil to perform well over wide frequency span. I'm still at loss how far I'll go complicating it, but a turn-on-and-go solution seem a right way to do it.

[Skippy]

I'm pretty sure that the diagrams posted in this thread must have been copied and distributed:
http://www.geotech1.com/forums/showt...eign-schematic
I wonder if Sean Goddard has anything, he seems to know something about these machines...

Slight change of topic: Does anyone think that the BBS approach has any merit when applied to 'hoard hunting', either with a 2-box set-up, or with a large 18-24 inch DD coil used like a 2-box? You may have seen Funfinders recent thread on large DD's.
Experiments I have done related to the ground pickup when a DD is used clear of the ground suggest that the ground-phase lag moves quite a way towards 'salt', 25-50 degrees relative to ferrite. As BBS machines are known to cope well with salt beach work, might they also be good for this kind of 'deep-searching'?

[Davor]

All physics apply equally.

[Skippy]

I realise that. Perhaps I should have been more specific.
One main advantage of multi-frequency is its ability to allow discrimination to work 'deeper'. Now many people would use a 'hoard-hunter' in non-discrimination mode, they are just looking for anything large and deep. So my question is : would multifrequency offer any real advantage over a correctly ground-balanced single-freq machine in this kind of searching?
My hunch is 'no advantage' unless you then want to try and use a motion-discrimination mode after locating the target, to try and reject deep horseshoes etc.
I have tried single-freq non-motion all-metal hoard hunting, and it seemed sensitive enough. But setting the ground balance was a bit troublesome, and I wondered if the 'auto-ground-balancing' aspect of BBS machines would help or hinder this kind of searching.