What Features And Performance Improvements Would You Like To See In The Next High End Minelab Coin Detector?

[TripleT]

I'd like it to be able to find Can$ clad as well as the AT Pro.  ?

[stevejhgla]

I would like to see a "smart detector".  One that becomes more intelligent the more you use it.  Collecting data about each target and using an algorithm for a more accurate ID.  

Soil condition, reactive signal,  resistive or loss component, shape and strength of the  eddy current are just some of the factors that determine the ID of an object. 

These could be captured and after digging up the target, the user would input what the target was.  Perhaps a secondary number like the ctx 3030 would then display what the target would be based on past targets.

The machine would become "smarter" the more it is used. 

 

I'm not asking for too much am I? ?

[GB_Amateur]

1 hour ago, stevejhgla said:

The machine would become "smarter" the more it is used.

I'm not asking for too much am I? ?

That kind of artificial intelligence is doable today, although the R&D cost might be an issue given the limited market.  (Metal detectors don't have the distribution volume of smartphones....)

Maybe doing this in steps over the next few years is more reasonable.  White's certainly was headed in the right direction for many years providing visual information to aid our brains.  ML has done some, too.  Much of that direction seems to have gone stale, though, from my viewing angle, anyway.

[Dug D]

I think a real breakthrough would be a setting/button that could help discriminate between aluminum and everything else.

There must be something "special" about aluminum that could help identify it specifically, and save digging all of it from the tiny pieces to cans 15 inches down. (when you don't want to miss gold)

 

[PimentoUK]

I posted this length reply on Tom D's forum, but it's probably not the niche reply he's wanting, so maybe it's more relevant here:

I've had my Eqx 600 over a year now, so I've had time to form some opinions on what could be changed and improved, plus some thoughts on what could be incorporated into the new 'high end' machine.
Starting with the hardware/mechanicals:

* The elbow cup/rear stand has shown itself to be weak. There's several reports of a broken upper section, it just needs to be made chunkier. The foam padding is too thin, and surely it could be designed to actually fit the plastic part. The lower 'stand' part at first sight appears to be heavily ribbed and strong ... except the ribs all stop on the centre-line, leaving it with no reinforcement at all. It will bend/break very easily along that line, as I discovered. The ribs should go all the way along, and ribs added in the 'hidden' part. I think the 'stand' part is a bit small, too. Make it wider on the new machine, similar width to the elbow cup section. See attached photo.

* Having the machine break down into 3 parts is a good idea BUT the 3 parts should be comparable in length, so it packs away neatly. Two short sections and one long one is poor design. Allow for the fact that a plain shaft will pack 'diagonally' relative to the bulky upper section. As I've done this re-design myself, I think the upper section needs to be 25mm longer ( between the handgrip and the camlock) , the middle section needs to be 25 - 30mm longer, and the lower rod 65/70 mm shorter. Have 10-15mm less insertion overlap of the lower rod, maybe 10mm less insertion of the middle rod.

* The hole in the lower carbon rod for the locating pip has been seen to enlarge over time, with resulting minor misalignment. Perhaps if there was at least one 'spare' pip hole, the service life of the shaft could be extended.

* The coil-fixing nut/bolt are difficult to adjust. The combination of just two flats, and the closeness of the clevis to the coil body is awkward. Make the nut heavily knurled, so it can be gripped firmly regardless of orientation.

*Camlocks: make the hole for the 'wedge' part on the opposite side of the shaft to the spring-pip holes. This avoids the problem of the sprung pip scraping across the 'wedge' every time the shaft is assembled/disassembled. ( First Texas please take note of this, too )

* Please make 'solid' coil covers available for the 6" (and 11" too). An open 'spider' design is not good on a small coil that's going to be used in difficult terrain, bushes/scrub, tight spots with debris.

* The 6" coil is too big for a small coil. If there's going to be new coils made, make a solid 4.5" or 5", basically copy the Fisher/Tek one, and put ML logos on it.

* Make the new machine compatible with existing Equinox coils, and if possible, make coils for the new machine compatible with the 600/800 machines.

* I have big hands, so don't find the fat handgrip a problem, but I can see why some find it excessive. As the size is almost certainly related to the large (26mm diameter) 26650 Li cell inside it, I recommend changing the cell on the new machine to a 21700 size (21mm diameter). These are becoming increasingly popular in electric vehicles, so are well-developed. Capacities of 5000 mAh are available, almost the same as the better 26650's. Whichever size they use, ML should use their business clout to get high capacity cells sourced. Intermediaries like Shock-Li seem to do this for the vape industry. Allowing for the low drain of detectors, 26650's with a capacity of 6000 mAh are available.

* The icons/labelling of the buttons/functions could do with some thought. There's too many pictures of search-coils with random arrows, lines etc around them.
--------------------

And my electronics/software thoughts:

* The On/Off power button has shown to be a weak point. A more robust switch is needed, maybe a custom version of the standard one with solid fixings, like a metal frame with through-hole mounting. Perhaps a custom-designed switch?

* The USB/charger cable should be 30cm / 12" longer, it would be more convenient for charging.

* On the current machines, have the search-mode selector give a distinctive double-beep when it returns to 'Park 1'. That way, it's easier to work out where you are with audio only. It's not always easy/convenient to see the screen, but beep-beep-biddip-beep would tell me I'm in 'Park2' mode. This audio cue should also be applied to other functions. Engaging 'all-metal/horseshoe'=> double-beep; disengaging => single-beep.

* Assuming the new machine will have more modes, 'Scroll-left' and 'Scroll-right' would be preferable to single-direction action.

* Have the detector give an audio warning of very-low battery, perhaps 30 minutes before the 'musical finale' & switch-off.

* Make the green LED indicate that battery charging is near complete. For example having a rapid-rate flash once the battery goes from constant-current to constant-voltage charging.

* Pinpoint button: it's the most important button to locate by tactile means. I've attached a small sticky-backed rubber button over mine, and it's a great help. Could ML think about making something similar as a standard feature?

* Pinpoint Mode: The unusual 'auto-cal' operation takes a bit of learning, and has its quirks, but it does work OK. But one thing I don't care for is the resulting lack of 'depth' information it gives the user. A conventional pinpoint mode will immediately inform you if a target is weak/deep, or shallow, etc. A good example is the Fisher F75's pinpoint mode. The Eqx really needs the option of regular VCO pinpoint mode. It would seem pretty straightforwards to achieve - use a short press for 'auto-cal' pp mode, or press and hold for a second ( or a bit less ) and it gives conventional VCO pinpoint. Use the 'single-beep' / 'double-beep' to alert the user which mode is engaged. A single short button push exits pp mode, as normal.
In regular VCO mode, have the '88' display indicate depth, calibrated for a US 5 cent 'nickel'. I'd prefer centimetres displayed, as the 1cm increment size helps with pinpointing. I find the 1" step size on the Inches scale ( eg. on the F75 ) a little blocky.
The standard Eqx jumpy flickering double-bargraph pinpoint strength feature is awful and unusable. Perhaps replace it with a non auto-calibrating indication, there's 25 bargraph points available, so it should be possible to make it have decent resolution, 0.5" / 15mm step size roughly.

* Non-motion / pinpoint hunting, for deep targets, hoards/caches: This seems to be something the Eqx is quite capable of, and only a few changes to the standard pinpoint mode are needed. A "Re-cal" function would be useful, and it seems possible to use the 'horseshoe' button to do this, as it serves no purpose in pinpoint mode. The '88' readout could display a sensitive 00 - 99 signal strength, plus if possible use -1 to -9 to indicate weakening signal, eg. drop in ground signal, either natural, or by the operator raising the search-coil too much.

* It's known the Multi modes use 7.8kHz / 18.2kHz / 39kHz ( Beach has 13 kHz too ) , and the machine is evidently capable of operating at 5Khz, 4KHz with conditions. So is it possible to gain anything by running a 'Low Multi' mode, such as 5.4k / 12.6k / 27k (same ratios), which could be better for the US-style milled silver coin hunting? Maybe ML have tried it and decided it gave no advantage, or worked less effectively, but I think it's good to query it.

* Unlike many people, I don't dislike having 40 points of non-ferrous disc resolution. However, I don't like the calibration of the ID scale. A mid-range '20' target is roughly a 5kHz target, and to get the '5th tone' 30+ ID, the target needs to be quite a decent size copper/silver item. It's quite a novelty for me to find 5th-tone targets.
It's pretty certain that Park1/Field1 favour the 7.8k frequency, so perhaps in these modes a 5kHz mid-scale target is OK, but the scaling needs stretching out so that real-world targets read in the 30's.
But Park2/Field2 seem weighted more to the 18.2kHz freq, and the ID scale should be calibrated accordingly. A US 5c 'nickel'( a 17kHz coin) should be a '20' target, with correspondingly different calibration over the full range.
Perhaps there could be some way of choosing between 'Legacy' ID scaling, and 'dual-scaling' tailored to the machine's mode?
( I haven't commented on Beach & Gold modes .. I don't have gold modes, for a start )

* I see some are wanting a First Texas style mineralisation meter, and I see the value of it. But the '88' display is capable of providing better info. I proposed how this could be done in an old thread on Tom Dankowski's forum ( see my later post in this thread ), but the 1/3/10/30/100/ scale is a bit blocky, and if you could resolve it to: 10/16/25/40/64/100/ etc rather than just 10/30/100/, you could likely discover more about the land. I was originally thinking about it use on the Fisher ProArc-F75 ( the blue one for archaeologists), but it could be relevant to us hobbyists.

* True all-metal mode is conspicuous by it's absence from the Eqx. Is it possible, in single-freq mode, at least? Multi mode all-metal is perhaps more challenging, but it should be considered.

* Stereo Operation: I hope the Eqx is stereo-capable, it seems likely. Its successor certainly should be. The standard 'Nautilus' mixed mode ( disc in one ear / all-metal other ear) would be an obvious implementation, but there are other potential uses.

* When you scroll through the options and reach the discrimination/notch function, it defaults to '-9', which is inconvenient, as it's usually the non-ferrous zone that users want to adjust. Can it be modified to default to '00', to save some button-pushing?
 

[GB_Amateur]

4 hours ago, PimentoUK said:

I posted this length reply on Tom D's forum,...

You sure went to a lot of trouble.  I hope this at least gets serious consideration at ML.

One thing you brought up is something I've thought about but never really vented about:

4 hours ago, PimentoUK said:

* Pinpoint Mode: The unusual 'auto-cal' operation takes a bit of learning, and has its quirks, but it does work OK.

I don't think it's OK, at least for my style.  I'd much rather it give me the option of reducing the pinpoint signal than thinking it knows better than I do what I want.  From my experience, not knowing the gain of the pinpoint response makes it more difficult to discern depth, size, shape, etc.  And I wonder if that quirky loudness problem that occurs occasionally at initialization of pinpoint is related to this supposed feature.  I also wonder if it leads to less usage of the pinpoint function by many users.  (Several here say they never use it.)

It's so easy to place the coil nearer to the intended object and toggle pinpoint off-on to lower the strength of the signal.  Isn't that how most detectors with pinpoint work?  This auto-recal feature seems like an engineer's idea of something cool to brag about with little concern for the in-field user.  But you've recommended two modes -- one with and one without auto-recal -- which would suit me just fine as I could then leave auto-recal off (forever!).

[nickelNdimed]

Surprised this has not been brought up more. Seems to address some of the complaints at least:

https://m.facebook.com/Sea-Ghost-101743838386442/

[PimentoUK]

I found my post on Tom D's forum about using the '88' display as a ground mineralisation meter. Rather than link to it, I'll revise my thoughts here:
Ground signal strength, or Fe3O4 strength, varies a huge amount depending on location. It can be 1000 times stronger in the red Tennessee dirt than it is in Florida's sandy ground, or on freshwater beaches. This range makes it harder to describe in an easily presented way.
Fisher's F75 uses a 6-bar bargraph, scaled as:

Bar's labelling : Fe3O4 strength
0.01 = 25 x 10-6
0.03 = 75 x 10-6
0.1 = 250 x 10-6
0.3 = 750 x 10-6
1 = 2500 x 10-6
3 = 7500 x 10-6
(I'm not sure what the units of strength are, I'm curious. Edit: see my later post below)

The wide range of strength means some logarithmic type of compression is needed to present this in an understandable way. Two viable techniques are:
* Create an arbitrary scale, where stronger ground gives a bigger number on the display, but to find the 'true' strength, you would need to look up a conversion table, eg. in the User Manual. Similar to how the Fisher F75 bargraph works, if you have "0.1" on the bargraph, you can use the manual to convert to "250 x 10^-6 units".
* Use some type of 'mantissa and exponent' scientific notation ( eg. 3 x 10³ ), to take advantage of the logarithmic nature of the exponent function. This is essentially how electrical resistors are labelled, for example '473' = 47 x 10³ = 47000 Ohms.

If you're limited to just an '88' display, the first of these options is by far the best, you could achieve all the range you'd want, at quite a fine resolution, good enough for amateur archaeological surveys.
The second option is very restrictive, but you could achieve a scale slightly finer than the F75 bargraph, with about 10 values covering the range of the Fisher's 6 bars.

A proposal for an arbitrary scale:
Use an increment of 10 on the scale to represent a 10-fold increase in ground strength, giving the basic values:
88-readout : Fe3O4 reading
10 = 10 x 10-6
20 = 100 x 10-6
30 = 1000 x 10-6
40 = 10000 x 10-6

Then for a hobbyist machine, fill in the gaps with 4 intermediate values, to give 5 readings per decade. These need to be ideally 1.59 : 1 ratio ( 5th root of 10 ), so giving readings of:
88-readout : Fe3O4 reading
10 = 10 x 10-6
12 = 15 x 10-6
14 = 25 x 10-6
16 = 40 x 10-6
18 = 65 x 10-6
20 = 100 x 10-6
etc
This gives about 15 different values covering the full range, the low strength end could go lower; '8' = 6.5 x 10-6 etc.
If you want a pro job, fill in the gaps with 1.26 : 1 ratios, giving: 10, 13, 16, 20, 25, 32, 40, 50, 64, 80, 100, for 88-displays: 10, 11, 12, 13, 14, 15 etc.

[GB_Amateur]

5 hours ago, PimentoUK said:

(I'm not sure what the units of strength are, I'm curious.)

From the Fisher F75 user manual:

I don't know what 'micro-cgs' means.  'cgs' is commonly meant to refer to a particular form of metric system units, where everything can be simplifed to combinations of centermeters (c), grams (g) and seconds (s).  The unit of magnetic field strength in cgs units is 'Gauss'.  So maybe they mean micro-Gauss (one millionth of a Gauss)?  I like the left hand scale myself which (as can be seen in the cut & pasted snippet from the manual) is % magetite equivalent by volume.