EQUINOX Technologies (Part 4)

This is the fourth installment in a blog series introducing and explaining the technologies inside our new EQUINOX detectors… (Read Part 1 here. Read Part 2 here. Read Part 3 here.)

In Part 3 we ended with mentioning the different Multi-IQ “frequency weightings” for the different search profiles. Part 4 explains further why it is not a simple matter of just referring to specific individual frequencies for learning more about Multi-IQ technology. Let’s now consider one of the key practical detecting outcomes and then discuss how this was achieved…  ​

“A lot of people are going to be surprised at how well the machine works in saltwater. At the outset we weren’t sure whether reliably detecting micro-jewelry in a conductive medium was even possible, but – with the help of our field testers and the subsequent fine-tuning of the Multi-IQ algorithms – we’ve found the EQUINOX to be more than capable.” Dr Philip Wahrlich

Background and considerations

While Multi-IQ may appear as ‘magic’ to some, to our team of signal processing experts, it’s the result of a significant number of man-years of development.  So where did they start?  By assessing the metal detectors and technologies available in the market at that time, along with typical customer perceptions about their practical applications; and actual detecting results achieved:​

So, an important goal with developing Multi-IQ technology was to retain the above simultaneous multi-frequency advantages AND greatly improve performance in the two key areas where many single-frequency detectors typically excel – fast recovery in iron trash and finding low conductors in all conditions.

Speeding up the process

Most comparable low-power Continuous Wave transmit-receive detectors (for the same coil size) will have a similar raw detection depth at which the transmit signal penetrates the ground and has the potential to energize a target. To increase detection depth significantly typically requires higher power and Pulse Induction technology. This has advantages for gold prospecting, but discrimination is poor for identifying non-ferrous targets. While we continue to push for depth improvements, Multi-IQ also aims to provide substantial speed improvements, resulting in being able to better find ALL non-ferrous targets among trash in ALL locations. You could therefore say “fast is the new deep, when it comes to EQUINOX!”

Let’s start with considering signal processing not as a ‘black box’  where ‘magic’ happens, but more as a complex chain of applied algorithms, where the goal is to more accurately distinguish very small good target signals from ground noise, EMI and iron trash. Now, ‘fast’ by itself is not enough – you can have fast with poor noise rejection and poor target identification, giving no great advantage.  Fast is also not just a result of microprocessor speed. Processors operate at much higher speed than is needed to ‘do the signal processing math’.

You can think of the signal processing chain broadly as a set of filters and other processes which are applied to the metal detector signals to convert these signals into useable, informative indicators, such as an audio alert or a target ID.  For Multi-IQ, keeping the ‘good’ properties of these filters, while keeping them lean and removing unnecessary processing, was an important step towards achieving ‘fast’ for EQUINOX.

It’s also important to recognize that these filters are not the coarse filters of the analogue electronics hardware of last century – it all happens in software these days. Perhaps think of the older analogue TV standards versus current digital TV. (Standard digital HDTV has approx. 10 times the resolution of analogue NTSC.) With metal detectors, a fast higher resolution filter set will result in improved target recognition. 

Factoring in the ground conditions

However, speed without accuracy is not enough to produce a “game changer” detector – and improved accuracy cannot be achieved with a single frequency alone. Why? – “multi-frequency has more data-points” Philip Beck, Engineering Manager. This is worth explaining in more detail…

All transmit-receive detectors produce in-phase (I) and quadrature (Q) signals that can be processed in various ways depending upon the response received from targets, ground and salt. This processing happens through ‘channels’ that have different sensitivities to the different signals received. It is important to recognize that channels are not exactly frequencies. This is why it is more complex to explain than just correlating optimum frequencies to specific target types.

With a single frequency detector there are two basic channels for information (i.e. I and Q) that respond differently to good and bad signals, depending upon the frequency of operation and whether you are looking the the I or Q signal. It is also possible to scale and subtract these signals, while taking ground balance into account, to best maximize good signals and minimize bad signals. You could thus think of single-frequency being Single-IQ, with a limited set of data (e.g. I, Q, I-Q, Q-I) that works well for a particular set of conditions. To further enhance performance for a different set of conditions, you need to change frequency and detect over the same ground again. Therefore a selectable single frequency detector has an advantage with more data available, but not all at once (e.g. I1, Q1, I1-Q1, Q1-I1 OR I2, Q2, I2-Q2, Q2-I2 for as many frequencies that you can select from).

Now, getting back to Philip Beck’s “more data-points”, and just looking at two frequencies, a simultaneous multi-frequency detector would be able to process (for example) I1, Q1, I1-Q1, Q1-I1 AND I2, Q2, I2-Q2, Q2-I2 AND I1-Q2, Q2-I1, I2-Q1, Q1-I2 to give better detection results. Increase the number of frequencies further and the number of extra data-points also increases accordingly.  What Multi-IQ does is process different optimized channels of information (not just individual frequencies) for the different modes. We have previously explained this as “frequency weighting” (in Part 3), where the various EQUINOX Search Profiles are matched to the respective ground conditions and target types.

Here is a very simplified example where you can see the result of processing more than a single channel of information (remember, a channel is not a frequency):

Channel 1 has a strong target signal, but the salt signal is stronger still. Channel 2 has weaker signals for soil, salt and the target.  If the detector just responded to either Channel 1 or Channel 2, the target would not be heard through the ground noise. If the detector processes a subtraction of the channels (e.g. ch.1-ch.2), then it is possible to ignore the ground noise and extract a strong target signal.  Now, think back to the high number of possible combinations of I and Q for simultaneous multi-frequency compared to single-frequency and the frequency weightings for the modes.  All of the EQUINOX Park, Field, Beach and Gold Search Profiles have dedicated signal processing to best suit the conditions and types of targets being searched for.

Conclusion

Multi-IQ = more data-points = sophisticated processing = better ground noise rejection = more finds

Just as targets are more sensitive to certain frequencies, so is the ground – an important reason why air testing has inherent limitations when comparing detector performance. As soon as you have ground to consider in the signal processing equation, it can greatly impact on the ability of a single-frequency detector to accurately identify a target. Also, the deeper a target is buried, the weaker the target signal is, relative to the ground signal. The most difficult ground response to eliminate is the salt response, which varies greatly between soil, dry sand, wet sand and seawater. It is not possible to eliminate the salt response and the soil mineralization response (e.g. black sand) with just one frequency. However, within the carefully calibrated Multi-IQ channels, EQUINOX is able to identify both signals and therefore mostly ‘reject’ them (just as you would notch discriminate an unwanted target) BUT still detect gold micro-jewelry.

If you haven’t tried EQUINOX yet – why not give it go – with Multi-IQ being fast AND accurate, those diminishing and elusive targets are running out of places to hide!

(Part 5 to follow…)