new MineLab Machine

[Mark-VA]

It all comes down to one thing, will it perform better than other multi-frequency machines, DFX, V3i etc.
Time will tell.

[Davor]

Concurrent (simultaneous) multifrequency generally requires frequency domain, so I assume that is what the Equinox will use. As such, all those high-order harmonics get filtered out.

I don't think it is at all necessary.
Minelab mentioned they use only quadrature (R) signals for multifrequency discrimination. I tested this in simulation, and to my surprise it works better than I expected. It has some saturation effects at ferrous-silver far ends, but for targets in between it is OK. I used some simple formulas to normalise and stretch the VDI and got a nice S-shaped curve. With some poking with a few more formulas I may get a more straight line.
Fun part is that obtaining a nearly perfect quadrature from such a Tx stimulus is easy by mere sampling and adding/subtracting samples - effectively crude digital filtering. I have no idea what they did, of course, but I know how it is possible to achieve. It is all in a Tx signal. It can also be done with BBS, and considering it is a fair game now - I'd go in that direction. It is also less complicated if you want it implemented in hardware. For software it is the same. Therefore the whole idea of complex Tx scheme is simply in function of bridging the patent gap after BBS stepped into public domain.
Because only the quadrature signals are used per frequency, and a basic VDI information is in something like (R1-R2)/(|R1|+|R2|), all imperfections due to poor filtering are equally mapped into slightly more skewed VDI curve, and hence of no importance. Same goes for small quadrature imperfections.

If anything, the whole idea of quadrature-only discrimination seem a good route for making a scarrily simple multifrequency rig.

[sled]

I don't think it is at all necessary.
Minelab mentioned they use only quadrature (R) signals for multifrequency discrimination. I tested this in simulation, and to my surprise it works better than I expected. It has some saturation effects at ferrous-silver far ends, but for targets in between it is OK. I used some simple formulas to normalise and stretch the VDI and got a nice S-shaped curve. With some poking with a few more formulas I may get a more straight line.
Fun part is that obtaining a nearly perfect quadrature from such a Tx stimulus is easy by mere sampling and adding/subtracting samples - effectively crude digital filtering. I have no idea what they did, of course, but I know how it is possible to achieve. It is all in a Tx signal. It can also be done with BBS, and considering it is a fair game now - I'd go in that direction. It is also less complicated if you want it implemented in hardware. For software it is the same. Therefore the whole idea of complex Tx scheme is simply in function of bridging the patent gap after BBS stepped into public domain.
Because only the quadrature signals are used per frequency, and a basic VDI information is in something like (R1-R2)/(|R1|+|R2|), all imperfections due to poor filtering are equally mapped into slightly more skewed VDI curve, and hence of no importance. Same goes for small quadrature imperfections.

If anything, the whole idea of quadrature-only discrimination seem a good route for making a scarrily simple multifrequency rig.

This sounds really interesting, would you mind to share some details like your sinulation setup/results or formulas you derived?

[Carl-NC]

A square-wave drive TX produces a triangle-wave current and magnetic field, which produces exponential eddy responses in targets. You can sample the RX response with early/mid/late sample windows and effectively get X and R from this.

In a sequential multifrequency (SMF) where short & long duration TX drives are temporally separated, you can easily use parallel demod channels with proper timing to extract the X & R for the different frequencies. No pre-filtering needed since the responses are not intermixed.

In a concurrent multifrequency (CMF) where short & long duration TX drives are intermixed, the RX signal is bandpass-filtered into each frequency channel for X & R demodulation. Then the signal is closer to a sine wave and can be demodulated using continuous-time methods, exactly the same as a single-frequency design. The White's V3 demonstrated that when the channel filtering isn't good enough, frequency cross-talk makes ground balancing exceptionally difficult. That's why the V3 GB doesn't work as well as it should.

It may be possible to do CMF using a similar discrete-time sampling scheme as SMF but I've never explored doing so. SMF has other benefits and I will probably never do a CMF design. In looking at the photo of the Equinox PCB, I don't see any analog demod channels so I suspect it has a direct-sampling scheme like the X-Terras, and that ML is sampling the waveform and doing the channel filtering and demodulation in software.

[Davor]

That would be my take as well. Digital filters are easy, and in case of synchronous sampling it becomes ridiculously simple. So much so it can be done in switched analogue hardware.

@sled, I'm not ready to share much more than I already said, mostly because I still haven't got a good model of a ferrous target, which in nature has some odd frequency dependencies. Basing conclusions on completely linear model based on increased permeability is OK for testing a hypothesis, but not good enough for mapping a VDI.

I'll first try to make a good model of a ferrous target, and then we'll speak business.

[Koala]

No more guessing

Frequencies used in this video https://www.youtube.com/watch?v=qUz1xl-d-nE

[Skippy]

That video just states the same info as the "Equinox Technologies 1/2/3" articles. It doesn't state what they actually transmit/receive/process.

The 600 & 800 may transmit different waveforms, the transmitted waveform may vary depending on what mode is selected. Or they may transmit the exact same signal all the time, but only make use of some of the frequencies, operating-mode dependant.

This is where reverse-engineering stalls, of course. If they're always transmitting, say, 5 frequencies, but only processing 2 in a particular mode, it won't be obvious.

Presumably this Park1 / Park2 bias towards higher or lower freq targets is done by the processing, rather than changing the TX waveform.
Example: if they're using 5K and 20K, 'Park1' is basically a 5K mode, and the 20K is used to lessen the ground response, whereas 'Park2' is a 20K mode, the 5K used for reducing ground response.

I'm intrigued what the gold programs use, as they are both multi-freq, it seems. Would 20K and 40K simultaneously give much advantage over just one freq? They are a bit close together.
I'd often wondered why ML didn't use BBS in a nugget-hunter, eg 30K/120K, maybe it's benefits are minimal?

[Koala]

I viewed it at as

The 600 transmitted @ 5 10 15khz all the time
And the 800 transmitted @ 5 10 15 20 40khz all the time

Whether it receives all in every program isn't mentioned

[Carl-NC]

No more guessing

Frequencies used in this video https://www.youtube.com/watch?v=qUz1xl-d-nE

Even Minelab explicitly telling you something doesn't stop the guessing. 28 frequencies?

[Chiv]

I could be wrong but I thought I read that gold mode was transmitted in a single frequency Eg either 20 or 40khz(800), or use single frequency for pin pointing. I have gotten caught up in the hype last year and have ordered one, on when I turn up is any ones guess, but was told last December it was still been tested in the Australian gold fields haven't been back to shop to see how they went.

[Skippy]

I think "WAS" is correct, they appear to have done some real-world development/testing, and made multi-freq the 'standard' gold-prospecting setting, presumably with the single-freq choices still available.
From "EQ Technologies 3"
"Previously we have stated that Park, Field and Beach would run in multi-frequency and that Gold would only use the single frequencies of 20kHz and 40kHz, giving better results for gold nugget hunting. Our ongoing collaborative field testing feedback from around the world has resulted in further improvements to Multi-IQ to the point where multi-frequency is now the best option for Gold Mode as well, and will be the default setting."
Presumably it's possible to run at 40KHz, and use, say the 10K freq to lessen the ground signal? Ditto, 20KHz could use 5K for the same result?

[daverave]

give me a good GS5 or tdi pro any day verses this new detector

[Taktyk]

In the latest marketing materials of the MineLab company, we must pay attention to the content:

Simultaneous Multi-Frequency

True intelligent Simultaneous Multi-Frequency for maximum performance, plus a wide range of single frequencies.

So what, the previous definitions of BBS and FBS solutions did not rely on the transmission and analysis of the signal simultaneously? No, they certainly were. I am very skeptical about the marketing slurry that various manufacturers offer us. In the case of MineLab, we have already dealt with the introduction of the term VFLEX with respect to Xterra detectors.

I allowed myself to look at the patents that the producer placed on the occasion of the premiere of the new detector model, patents referring directly to Equinox 600 and Equinox 800.

These are two US patents: US7432715, US7579839 and the Australian equivalent of the first AU2005276953.

The first one seems to be fundamental: Method and apparatus for metal detection employing digital signal processing. The patent belongs to Laurent Stamatescu and is from 2008, in 10 years, the protection of the invention will be over. What is the patent about? Presents a method and apparatus for a metal detector that uses digital signal processing, a solution used in most modern metal detectors. However, when reading in detail about a patent, one should pay attention to the transmission method of TX transmission signal. Well, the DDS (ang.direct) method is used here. To my knowledge, not used in other commercial solutions related to MD. However, what does it give? According to the author of the patent, for accurate and simple emission of signals by means of two DAC. What enables the emission of a single frequency signal, i.e. a sinusoidal signal and a rectangular multifrequency signal. So much theory. And what is in the claims. In patent claims, we have:

1. Digital metal detector that generates two sine signals in a square.
2. Digital metal detector that generates a rectangular signal.

However, as it exactly looks in the practical solution used in Equinox, it requires measurements of the transmitter system in the laboratory.

Source:

http://i-metaldetector.com/2018/03/0...tenty-minelab/

Regards
TAKTYK

[Orbit]

And what they accomplished would they penetrate the ground better? Personally, I do not think so!

[Davor]

So what, the previous definitions of BBS and FBS solutions did not rely on the transmission and analysis of the signal simultaneously?

Actually that's what precisely happens with BBS, and because you process it sequentially it is/was a brilliant idea. The only reason this new idea is marketed this way is because it is new, shiny, and patented, while BBS/FBS is now in public domain.