Minelab Multi-frequency

[Lanny]

On 12/24/2018 at 8:49 AM, Steve Herschbach said:

The why of it is simple - you are reading things into the careful gaps Minelab leaves in their advertising. I am not reading it the way you are, and so I am not offended because I do not feel I am being mislead. You are reading something else into it and feel unhappy because you feel you have been mislead somehow. It all boils down to debating perceptions of what you think the advertising means as compared to what is really going on under the hood. Since you don't 100% know what is going on under the hood, you just don't know.

All detectors are transmitting on at least one primary frequency and many harmonic frequencies. This can be determined with a scope and if you count harmonics it is just a matter of picking a number you want to advertise when it comes to transmitted frequencies. You have no idea how many are being processed unless you know the processing algorithms.

This back and forth with Minelab and the others has been going on for years. Garrett picked 96 frequencies for a Garrett Infinium ad as a direct poke at Minelab. How many people thought the machine was really transmitting and processing 96 frequencies based on the ad? It is all about transmitted because a PI detector does not receive and compare frequency information. Meaningless fluff, an industry insider joke.

Get your 96 frequency detector right here....

Garrett Infinium ad - 96 frequencies!

Carl Moreland weighed in with his frustration on another thread. White's even gave room over in the V3i manual, probably with Carl's input, to comment on the situation. From the White's V3i Owner's Manual, page 6:

"There is much confusion - some of it deliberate - over how many frequencies a detector actually uses, and whether multiple frequencies are truly better than a single frequency. What defines a multi-frequency detector? What do multiple frequencies really do for depth and discrimination?

A multi-frequency detector is defined as one that simultaneously—or, in automated sequence— transmits, receives, and processes more than one frequency. Some detectors have the ability to operate at one of several selectable frequencies, but they still are single frequency detectors because during operation they can only transmit and process a single frequency. The same is true of detectors that have a control to slightly vary their operating frequency to minimize interference; even though they have the ability to operate at many (slightly) different frequencies, they are fundamentally single frequency designs. Currently, all multi-frequency hobby detectors run their multiple frequencies simultaneously as opposed to sequentially; they are all characterized by having multiple processing channels in the receive circuitry. Therefore, a 2-frequency detector will have two processing channels. Spectra V3i has three independent processing channels; it is a true 3-frequency detector.

This all sounds easy, so where is the confusion? It turns out that (currently) all multi-frequency detectors create a transmit signal that is composed of digital waveforms which are designed to produce peak energies at the desired frequencies. As a side-effect, these digital waveforms also produce undesired harmonic frequencies. Lots and lots of harmonic frequencies, 10’s or even 100’s of them. These harmonics have no useful energy and are not part of the signal processing. So while we can claim to transmit many, many frequencies, we cannot claim to process or use them. Therefore, we could easily claim the Spectra V3i transmits 17 frequencies, or 28, or 39, or 55—we could get plain silly with this. And such a claim would be true, technically speaking, but since all those extra frequencies are not actually used, it would be misleading to make such a claim. White’s chooses, instead, to claim the number of frequencies we are actually using and processing. It may not sound as impressive as a 55-frequency detector, but it’s honest and accurate.

Pulse Induction (PI) detectors also utilize a digital (pulse) transmit waveform, so they transmit a tremendous number of harmonic frequencies as well. So is it fair to include pulse induction as a multifrequency technique? Not really, because PI detectors process in the time domain, not the frequency domain. So even though they use broad-band signals, they are not frequency-based detectors at all. Calling them “multi-frequency” is simply another attempt to confuse the consumer.

Once we get beyond the marketing hype, the real question is: What does multi-frequency do for depth and discrimination? The truth is, any time a detector is simultaneously transmitting more than one frequency, the transmit energy must be divided amongst the frequencies. Therefore, a single frequency detector can usually squeeze out slightly more depth than a multi-frequency design at that certain frequency. But this is an advantage only at one frequency, which tends to favor only a narrow range of targets."

 

These are old debates going back to Minelab BBS and it’s 17 transmitted frequency claim. I reference it all in my discussion at Selectable Frequency And Multiple Frequency

Anyway, interesting subject for armchair engineers and armchair attorneys. Hopefully you all get it sorted out while I am out metal detecting! 

Loved this post Steve for how it explained the Multi-Frequency confusion, many, many thanks.

All the best,

Lanny

[Lanny]

On 12/24/2018 at 11:46 AM, Steve Herschbach said:

This is all based on an assumption people make - that more frequencies are better. Are they? Why? Other than advertising, what makes anyone think that?

There is no reason to compare items at close frequencies. Comparing 10 khz and 12 khz is a waste of time and processing power. What you want to do is compare the targets and ground at a few widely differing frequencies. That way you see what happens that is different at say 4 khz and 32 kHz. That difference or lack of difference gives you extra data to work with. And as Minelab points out in the article about Multi-IQ the more interesting question is how few frequencies are needed to get useful information. And that seems to boil down to two or three.

"'How many simultaneous frequencies?' you may ask, wondering if this is a critical parameter. Minelab has been carrying out detailed investigations into this in recent years. Just as you can color in a map with many colors, the minimum number to differentiate between adjacent countries is only 4 – a tough problem for mathematicians to prove, over many years. Similar to the map problem, it’s perhaps not the maximum number of frequencies needed to achieve an optimum result, but the minimum number that is more interesting. When it comes to frequencies in a detector, to cover all target types, how the frequencies are combined AND processed is now more important, with the latest detectors, than how many frequencies, for achieving even better results."

 

My assumption always when dealing with frequencies and what is processed versus what is transmitted goes like this.....

Manufacturer claims multiple frequency operation, simultaneous or near instant sequential, I don't care.

They state a number of frequencies.

I then assume that out of the frequency range mentioned, that at least two or more frequencies are being received and compared at the processing level to justify the claim of multifrequency. Just exactly what frequencies are being processed and exactly how they are being compared.... I do not expect any company to tell me other than in vague terms.

That's it. That way I stay calm while others try and divine the magic. And with that I leave the subject to better minds than mine.

Truly great summation Steve, and thanks for your work and research to clarify my understanding of what's going on with Multi-Frequency detectors. Outstanding post!

All the best,

Lanny