Which Frequency Is Running In Each Mode?

[☠ Cipher]

Not at all lol. I think your position is the most prevalent. Most people just want to use any given product whether it be a metal detector or a phone etc. I'm the odd man out in this scenario. Jumping into the engineering behind electronics requires a few elements. One is the desire to understand it. Two is the ability. Three is a sustained effort. You could have element 2 but if you have no desire to get into it, it's no reflection on intelligence. I'd never read into that at all. Where we probably start making judgments about that is where people have element 1 without 2 or 3 and use "a little bit of knowledge" to base or start an argument on without qualifiers in areas of uncertainty. 

[Steve Herschbach]

On 4/8/2018 at 11:01 PM, ☠ Cipher said:

I believe what Minelab has always done with multifrequency is run it in a sequential pulse like manner (utilizing time and frequency domain). This has an advantage over concurrent multifrequency, like the V3i, where there is very little depth loss vs any single frequency and it doesn't require filters. The dead giveaway that multi-IQ is using a sequential approach is the reports of no loss of depth between single frequencies and multi-IQ, and the way Multi-IQ seems to perform as well as single frequencies in nearly any scenario. This probably wouldn't be the case if the frequencies were run concurrently.

The transmit waveform is no less simultaneous for Equinox than with V3i. The received combined information is separated into the different frequencies. Some combination of those frequencies is processed sequentially because that is how most programs run - one step at a time. I believe Multi-IQ outperforms any one single frequency component because you are getting a cumulative response from multiple frequencies that combined give a stronger target response. All targets have a frequency they respond best at, but they do respond at the other frequencies also. I believe Equinox gets a boost in Multi-IQ in the way the information is processed differently than in V3i more so that responses over multiple frequencies literally add up. V3i does not appear to have the same additive effect from employing multiple frequencies at the same time.

From http://www.detectorprospector.com/metal-detecting/minelab-multi-iq-technology-details-explained.htm

The Multi-IQ transmit signal used in EQUINOX is a complex waveform where multiple frequencies are combined in a very dissimilar way than our proven BBS/FBS technology in Excalibur II / Safari / E-TRAC / CTX 3030 detectors.

If you view the BBS signal amplitude on an oscilloscope, it looks something like this:

In comparison, Multi-IQ looks something like this:

Hence – Multi-IQ is not a derivative or evolution of BBS/FBS. Multi-IQ is a DIFFERENT method of simultaneous multi-frequency metal detection. We could also debate “simultaneous” versus “sequential” semantics; however the real detection ‘magic’ doesn’t happen with what is transmitted to and received from the coil alone. Remember, in Part 2, we discussed how frequencies are “combined AND processed” as being important for achieving better results?

 

Here is how V3i does it. This from page 7 of the White's V3i Owner's Guide can provide some insight into both machines...

 

I think that both machines are transmitting and receiving in a similar fashion (though with different waveforms) and that the main differences are in the processing.

[☠ Cipher]

That's an equally good hypothesis. I just don't know that I can see how the V3i and Equinox machines transmit in a similar way. I'm still very early in my study of these concepts and it's very possible I don't quite have this down so, I have to defer to your knowledge and I think anyone else here reading this thread should too, but let me explain why I think a bit different on it.

 With the V3i, we know for sure that it transmits concurrently because Carl has told us so. We also see evidence of that in multi, with the transmit power behind 3 signals, there's a noticeable depth loss. When it is focused on one signal we see a gain. This is described in the V3i manual. This has been, up to the Equinox, generally accepted as how concurrent multifrequency stacks up against a single frequency when power is focused rather than divided. In my mind, there should only be two ways around such a power difference. One would be amplifying the multifrequency tx  signal in some way. The other would be to fire each frequency in a sequence so that there's not a constant division, but rather an ultra rapid firing of the signals. Sequential Multifrequency. They are then demodded in a sort of time domain analysis without filters. The use of time domain is where I believe Multi-IQ builds on FBS/BBS. The way in which they almost act like a pulse machine, while preserving the best of VLF discrimination. Below are the basis for how I'm currently thinking on it. The first is a quote from Carl Moreland. The second is a screenshot from Tom's forum. 

“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.  - Carl- Geotech forum..”.     

[☠ Cipher]

On the receive/processing side of things I have no doubt that Multi-IQ is doing something fundamentally different than Spectra and even BBS/FBS as you're describing and I've been wondering if the Equinox having its fundamental frequencies so closely harmonically related gives a boost, particularly on the high end, which was the weak spot of BBS/FBS. 

[Jackpine]

So where does this statement fit in?  since it is a "key difference in technology"

Quote

 

We can go to a statement from Dr Philip Wahrlich, our principal technology physicist, about a key difference of Multi-IQ compared to the demodulation taking place in conventional single frequency VLF detectors:

“Within the Multi-IQ engine, the receiver is both phase-locked and amplitude-normalised to the transmitted magnetic field – rather than the electrical voltage driving the transmitted field. This field can be altered by the mineralisation in the soil (in both phase and amplitude), so if the receiver was only phased-locked to the driving voltage, this would result in inaccurate target IDs and a higher audible noise level. Locking the receiver to the actual transmitted field, across all frequencies simultaneously (by measuring the current through the coil) solves these issues, creating a very sensitive AND stable detector”

 

Quote

Precisely measuring these extremely small current variations is quite remarkable if you consider the levels involved. It’s actually parts per billion, or nanoamp signals, we are talking about here!

The claim is better processing of the received signal but what if ML is keeping other details close to the vest on this. Using this information to vary the frequencies used in the muti-freq sequence on the fly as the coil is swept over varying ground is one thought that comes to mind.

Tom

[dewcon4414]

How does having that chip IN the coil bring this all together better......based on the batteries, it doesnt seem to require more power to the coil.

[Steve Herschbach]

Or this...

“For each frequency the detector transmits and receives there are two signals which can be extracted which we refer to as I and Q. The Q signal is most sensitive to targets, while the I signal is most sensitive to iron content. Traditional single-frequency metal detectors use the Q signal to detect targets, and then use the ratio of the I and Q signals to assess the characteristics of the target and assign a target ID. The problem with this approach is that the I signal is sensitive to the iron content of the soil. The target ID is always perturbed by the response from the soil, and as the signal from the target gets weaker, this perturbation becomes substantial. With some simplification here for brevity, if a detector transmits and receives on more than one frequency, it can ignore the soil sensitive I signals, and instead look at the multiple Q signals it receives in order to determine a target ID. That way, even for weak targets or highly mineralized soils, the target ID is far less perturbed by the response from the soil. This leads to very precise target IDs, both in mineralized soils and for targets at depth.”

There is more to it of course than just frequencies. Or just processing. Minelab is doing multiple new things with Equinox, that all add up to an advance in the technology.

At the end of the day in my opinion when dealing with microprocessors a debate about simultaneous versus sequential is mostly just semantics. We are never going to get more than vague ideas about what’s really going on under the hood with plenty of room left for debate. It’s all kind of moot. The reality is White’s was way ahead of their time in the basics of what V3i is doing. The V3 in 2009 does pretty much what the Equinox does. Multifrequency plus ability to run single frequencies and built in wireless capability. If White’s was not so married to large metal boxes, the next obvious step would be to put the same basics into a modern housing with a simplified control interface. If the MX Sport a couple years ago had been a V4 instead of a repackaged MX5 the entire industry dynamics would be different right now.

But they did not. The industry sped up and the old U.S. manufacturer timeline of one truly new machine every ten years or so just can’t keep up any more.

[☠ Cipher]

I should probably explain that my curiosity about what any machine is doing, or doing differently extends beyond a brain teaser. There's that element too, and I'm sure I'd wonder regardless, but I have a background in building and repairing electronics, mostly computers and drones. This had been up to recently just dealing with preprogrammed parts. 

Some time ago when Radio Shack was going out of business I bought out the electronic components of a couple local stores when it got to be pennies on the dollar, and so I've been studying electronic engineering and experimenting with prototype boards, breadboards, arduino, Raspberry Pi and learning kits. I've ordered Carl's book Inside the Metal Detector. I've built some simple, very basic metal detectors. So it's something I have a growing interest in, and the resources to experiment more and more as time goes on.

One thing I've not seen on the DIY side, are any multifrequency kits. We have single frequency and pulse, but nobody has ventured into that area. Where it will eventually lead, I have no clue. I've taken a similar interest in auto mechanics, home renovation, paralegal/PI studies, cosmology, social sciences, political science, earned varying credentials, and then I'm happy to know as much as I can.

I always have to be learning something, and the magic that goes on inside these boxes and behind these buttons has always captivated me. I was like a kid in a candy store when I programmed my first micro controller. Over the next few years I want to soak up every bit of information as I can and by the end of it, have a deep understanding of how these machines function. I have as much fun with this as I do actually using them, maybe more. 

[Dubious]

Yeah, well, here it's mostly the processing on the receiving end that matters. You would need to understand the hardware then yank the code from the firmware and run a dissembler, and then figure out what it all means.  But even if you do all that, you're still not going to know any more about how to use the machine effectively. The devil in implementation is always in the details, and there's a big divide between theory and practice.  What is going to help someone use the Nox effectively is simply using the machine (a lot of hours) and experimenting with different targets and conditions. The best guide to that is what's been posted by actual users, on this forum and others.

[EL NINO77]

.I thought you were responding to Equinox programs so I ... I compared Equinox 800 to 3 buried items: -N2,N5,N6 ....1-AU segment  O  -0.05 gram-in depth 10cm, .... 2.- 14mm denar in depth 23cm ... and 3-13.5mm silver earrings at a depth of 20cm .... Equinox on Park 2, Field 2, Gold1, Gold2 has a signal for all 3 items at ease from sensitivity-19 -23 even with good ID -9,10 -Denar, 9ID- earrings., 1-2 ID Au segment O -0.05 gram.. but surprised me Park1, and Field 1, at sensitivity 20-23,24, -nice signal of Denar and silver earrings with a bit higher ID 11,12-Denar, even at such a depth .... but again on small  AU  segment O-0.05 grams did not respond to the maximum sensitivity 25 ... before I tried to dig up and AU segment O- just for 8cm .. and program park1, field 1 it also did not take away ..... Recovery speed I skipped from 1 to 8 ... good response had recovery 3,4,5,6 - even when the best was -4 ... Wink...