ML EQUINOX coil discussion thread

[nordic]

So, I have made this ugly little thingy running off the 9V battery through a 3.3v regulator, using an etched pcb from another project. The ripple in the Vdc line is now completely gone. There is HF noise present, but it is present even when the MD is turned off.
I can't tell if this made any different in air tests, so will have to take it out for field testing, it didn't make it worse sure.
In air tests, generally, filtering the Vdc seems to improve TID recognition - it comes a little sooner and is more consistent from extreme distance. I don't think it improved any air test depth, nothing obvious anyway. But if TID comes stronger, it can't be bad.
Was it worth the effort, maybe, I enjoy this stuff. Gave me a good deal of information on how parts of this stuff work.

[EL NINO]

Nordick ... my view is to improve signal parameters - and signal / noise .. will still contribute to better detection .. Interestingly, it will be tested on ground tests ... - which can confirm the virtually the benefit of this improvement.
I would be especially interested in the result of this improvement on better work ID in Allmetal mode..on horseshoe..where in tests for 1 gram gold 24k brick at a depth of 16cm -in heavily mineralized "Black sand" I get mostly signal from negative -9, -8ID ... with occasional corect -management+ ID ... In discrimination mode by using discrimination to adjust detection to correct positive and accurate response of Tone and ID ..

I tested all 3 coil dimensions 6 ", 11", 12x15 " coils into Equinox .., big 12x15" coil gave the best response in Allmetal..as much less positive ID- I think it is somewhat better reach this coil - stronger received signal.

The test results confirmed that the appropriate coil size is superior and contributes to the depth of detection.
Even in heavily mineralized soil ... standard 11 "and large 12x15" coil proved well to detect 1gram of gold brick at 16cm depth in "black-sand" = 4% magnetite ..
The small 6 "coil was beyond the depth of detection ... and could not detect 1g of gold brick ... at a depth of 16cm ..

[nordic]

Hi EL NINO. I definitely see some improvement in target recognition after regulating the coil supply. I used soic ADM7150 for 3.3 Vdc. It doesn't feel going deeper, but TID does come more precise and stronger. Not a day and night, but noticeable improvement. I think it gives an extra sensitivity point+, in other words, for example, if I set sensitivity to 20 and can hear audibly a coin under the coil from a measured distance, but no ID, with the dongle, it does get the ID.

It happened I have bucket of sand, that I use for some cement jobbies around home, it is heavily mineralized as well, GB comes at around 60-70 over it. I did bury children's gold/silver ring about 10cm into it + it has another 10cm air gap, it was picking it rather nicely. I did similar tests before and that sand would just hide things and only give negative grunt, unless they are right on the surface.

I've read some minelab white papers, they claim, that beside the user sensitivity setting, they also do one in software in background, which, kinda, makes my tests, or anyone's tests in fact, a little invalid. They don't say however if they do so at startup only, or all the time.

Unfortunately I don't have any gold pieces to test, only few silver coins and gold rings. My floor at home is nails infested and pretty much gives iron grunt everywhere I move the coil. I find that with the dongle it seems to discriminate good items on it a bit better, or from further distance. I have a pistol ball lead bullet, about 5mm in diameter, with which the 11" coil always struggles at any distance - could never see it on the floor, but with the dongle I does come through in one direction. So something is working a tad better.

I do have all 3 coils as well, but for the sake of consistency I'm taking the stock 11" coil this weekend, and will report back if of any interest. The difficulty here is that these hunts are very subjective, but I'll be keeping the dongle permanently, I like it. Nox is less erratic with it. Have modified it heavily yesterday for more permanent setup, using proper gauge wires and shielding for different types of signals.

[EL NINO]

Thank you Nordic


One possibility may be a test on buried objects on the boundary of the Equinox range. This allows you to quickly determine if a given power adjustment allows you to improve the depth of detection or better-accurate ID ...
Another good test could be the reach in Emi ...- where it would be found to work as such in an environment of electronic interference.

[nordic]

Most of my tests are around the detection threshold - the goal is to be able to sniff out smallest and deepest of things, to me anyway - if it can do that, it can do larger stuff too. Having such RX gain in the coil circuit and also gain in the digital domain (I think it is PCM1861 ADC), one of keys is clean power lines to lower the noise floor as much as possible to both the analog and digital sections. I think that ML could've used a better cable for the coil. The ADC setup in the control box looks minimal - looks like they run both analog and digital sections of the ADC from the same rail. I almost wish I could open the control box, but then knowing myself I would never end tinkering with it.

For the EMI, I think most of it is introduced through the coil... In all fairness, ML made Nox a very sensitive machine, because it reacts to any rapid changes of even a smallest magnitude, so any conditioning of the signal I tried so far only degraded the discrimination ability. I consider myself lucky to be able to hunt places where I can run the Nox at full or near full sensitivity, while some fellow users can't even go over 2/3.

[nordic]

Pulled good few smaller and bigger items today from one of fields. Did really well compared to my friends Deus, pretty much 1:2 in finds, perhaps luck or maybe not. Run full throttle sens 25, no bother at all - totally quiet off the ground. All targets were coming really clean. There was not one situation of an iffy signal without an ID, where I had to go over and over the target pressing coil down. Anything audible had an ID, which coincides with my findings in air tests. Field, however, had a lot of coke and iron too, but no problem with discrimination, not sure about unmasking, but pulled few small item from considerable depth, some took a while to pinpoint with external PP.

One thing noticed, that MD became very sensitive to larger metal objects, passing near fences or concrete armature bridges was giving long flutey sound, that I only ever heard when running recovery 1-2. But I was on 4. Also having a coil on the ground for 10 seconds or so, and then taking it away sounded this flutey as well, but not as prolonged as in mentioned earlier. Very interesting sensation of this coil "stickiness" to the ground.

Anyway, more testing some time soon. The dongle stays.

[EL NINO]

It seems to me that Equinox is now more sensitive (less filters) to low-frequency signals - this is the ground signal area and the area of large metal objects.
Theoretically, this would also have an impact on better detection of high-conductivity large silver objects and the like.
Even a simple Airtest with a large silver with a dongle and no dongle on various recovery speeds could confirm that ..

[EL NINO]

I got another good knowledge about Equinox coils .... in the new Black Sand box I mixed 3x more magnetite = 12% magnetite than in the first test box ... Tests are done at 10 cm depth of Black sand ..
The first tests showed:


1. Test result ... small 6 "coil at multifrequency ... / except Beach programs / no longer manages such mineralization and ... overloading ... secret weapons are here working at 1F at 5kzh, 10khz, 15khz, 20khz. ..by me there is a lower Tx performance so the received signal for the coil is still at an acceptable level ..
But don't expect much deep reach ... Detection of tested coin size targets is impossible through 10cm black sand ..


2. Surprise ... is of 11 "standard Equinox coil works well ,, and detects even at multi-frequency without problems coin size targets ...- but 12% magnetite mineralization no longer breaks strength and accuracy of multi-IQ ID .. showing ID is more like a classic Vlf detector ... shows higher and high ID numbers ...
In previous tests on 4.4% Magnetite mineralization, the strength of Multi-IQ ID is still maintained.


I think after these tests I get closer to the definition of how big should be the optimal coil for VLF or Multifrequency detector ... according to me it can be coil size from 8 -9""to 11" ..

[WM6]

I got another good knowledge about Equinox coils .... in the new Black Sand box I mixed 3x more magnetite = 12% magnetite than in the first test box ... Tests are done at 10 cm depth of Black sand ..
The first tests showed:


1. Test result ... small 6 "coil at multifrequency ... / except Beach programs / no longer manages such mineralization and ... overloading ... secret weapons are here working at 1F at 5kzh, 10khz, 15khz, 20khz. ..by me there is a lower Tx performance so the received signal for the coil is still at an acceptable level ..
But don't expect much deep reach ... Detection of tested coin size targets is impossible through 10cm black sand ..


2. Surprise ... is of 11 "standard Equinox coil works well ,, and detects even at multi-frequency without problems coin size targets ...- but 12% magnetite mineralization no longer breaks strength and accuracy of multi-IQ ID .. showing ID is more like a classic Vlf detector ... shows higher and high ID numbers ...
In previous tests on 4.4% Magnetite mineralization, the strength of Multi-IQ ID is still maintained.


I think after these tests I get closer to the definition of how big should be the optimal coil for VLF or Multifrequency detector ... according to me it can be coil size from 8 -9""to 11" ..

Interesting outcome. Thanks for share results.

[EL NINO]

I did an Equinox test and an 11 "coil, I used a compact Black-sand box which contains 12% magnetite.

I did a test on Park1 for multifrequency, recovery 6-7,iron bias F2=0, discriminacion =0, and sensitivity from 10 to 24.
The Ground balance on the Black Sand was the same for the multi-frequency as well as for the individual 1f frequencies and had a GB value of 2.
In another test, I have also tested Equinox on individual frequencies 5,10,15,20, a40Khz - where I have investigated how high the sensitivity is needed for a given frequency to correctly detect with IDs on various size.. and various conductive objects ..
The test results are a bit of a surprise for me ... Equinox detects all objects over a 10 cm thick layer of Black Sand. on the multifrequency already at c.uz on sensitivity 12-15 .... the only exception is an open gold earring which needed a sensitivity of at least 22-23 ... and more to correctly detect...
----------------------------------------------------------------------------------------------
The biggest surprise was the direct comparison of the different frequencies on the low conductive-gold target, where it showed the advantage of having 20 and 40 kzh frequencies as opposed to 5Khz or 10 khz.
2cm -0.9gram gold ring was on 1f- 40khz detectable with ID already on sensitivity 11 ...
But when testing the ring at 1f-5khz ... I had to increase the sensitivity to at least 18 to 19 to get a similar good signal from .. ID..but from smaller detection window as it is at detection to 40khz ..
A similar but opposite situation occurred when testing Us 25cent / quarter / ...
When using 1f 5khz I got signal with ID already on sensitivity 11 ..
After changing the frequency to 40 khz, I needed to increase the sensitivity to at least 17-18 to get a similar good ID signal ... but still with a smaller detection window as it is at detection to 5Khz...
For a small open earrings you will need a sensitivity of 22-24 .. to detect it at 1 frequency of 40KHZ .. Other frequencies 5-20khz here do not have sufficient range ...
Multifrequency detects it at sensitivity 22-23 ...

At the end of the Test,... I tried to find a frequency that would have a good ID range on the high conductive Quarter .. as well as the 2cm low conductive 0.9 grams gold ring ..
As optimal frequency in this test showed 1f -15khz frequency..t but needed good detection of both objects to increase sensitivity to at least 14-15 ... Yes it is a compromise for universality ..

[EL NINO]

WM6 ... thank you ... Just your standard testing process on your 4-Bar Fe3O4 mineralized ceramic grit "Liapor" ... inspired me to further standardized tests for extremely strong Magnetite mineralization - type Black Sand ...
Such testing reveals many physical facts in terms of detector and coil techniques, as well as the impact of various detector settings and frequencies on the real range and also the stability of detection under conditions of heavy mineralization.


I use 3 "Black Sand" boxes with 4.4%, 12% and 33% Magnetite content to test the detectors in extreme mineralization.

[nordic]

Hey guys,

I just wanted to update on the latest that I was field testing and working on, as a result I have a very stable operation on max sensitivity and noticeably clearer feedback from the MD on real targets. To what it looks to me, the main improvement lies with the rejection of noise in the power lines, both Gnd and Vdc, which is primarily generated by the TX line cross talk to adjacent lines. The magnitude of the noise is quite severe, around 180mVpp on every wire that is in the coil cable, because they are wrapped around the shield of TXN (TXP is the core) - they don't cancel each other that well really. I think it is just a poor choice of otherwise unshielded unscreened cable, that is susceptible to EMI and noise. I think, ideally, ML should've shielded the twisted TXN and TXP together, similar to CAT6A network cable.

While the preamp circuit in the coil appears to be quite immune to PS noise (nor anything can be done to suppress it really due to enclosed design), the very same contaminated Vdc and Gnd lines spread back to control box and seem to have negative impact on the operation. I think ADC in use has a fairly low PSRR if I recall too. I went through a number of circuits, that included even lifting the ground and letting only the differential RX signal feed directly to the control box without ground reference, for some odd reason it elevated EMI sensitivity. The best result I could achieve is through:
- having a ground organized in a star arrangement to avoid any loops - addresses the Ground noise going into the control box.
- battery powering the Vdc line through a low noise regulation. The Vdc line from the control box is lifted and decoupled to the Ground at the star point. This eliminates the noise feeding back to the control box via Vdc.

It looks like this in the end and seems to work very well. Fits right in between the Coil and MD connection. The copper pour on the board is NC.

[Tim Mason]

I need some help.

Working on replacing an equinox 800 coil plug that was damaged. I ordered a plug from Digi Key part # NOR1413-ND.

Once soldering was all completed based on the original plug pin assignment I tested it and the unit started with the two dashes on the screen on typical start up but it would cycle just doing the same two dashes flashing over and over. unscrewed it and connected it again this time it went CD error 41.

Your Pinout update suggests I have P1 and P2 reversed but that is actually what I see on the old damaged plug. I cut the original plug so that I could see the cross section that shows the red outer wire in the P1 position

Also I used shrink tube to isolate the completed soldered pins from each other instead of the hot glue fill that is in the original plug.



Pin 1 : TX cold (Shielded wire outer) Red coated copper wire
Pin 2 : TX hot (Shielded wire inner conductor)
Pin 3 : 3.3 V DC power (Brown)
Pin 4 : RX + or - (Red)
Pin 5 : RX - or + (Blue)
Pin 6 : NC ?
Pin 7 : 0 V ground (Green)
Pin 8 : Data (Yellow)

Any help would be greatly appreciated.

[Skippy]

I'm pretty certain my updated pinout (Post number 10 ) is correct. I have also supplied this info to another detectorist in the UK who was replacing a broken connector, and he made a successful repair.
I suggest you double-triple check the pin numbering sequence, the only pin you can't get wrong is Pin 8, in the middle. But if you're erroneously going round the outer pins in 'mirror-sequence' , you will find Pin1 and Pin 2 become swapped over.
( and to be honest, I think that is why Post 1 has the wrong info, I got the sequence mirrored. It bothers me the first post has the error ... I will contact George and see if he can edit it for me. )

Just as a confidence check, here's something simple you can test, if you have a multimeter:
Coil connector Pin 2 is the main circuit ground, and it's also the 0V ground on the USB port. If you look in the user manual, you'll see the USB port pinout, and 0V is the bottom right pin, as you look at the rear of the control pod. So a multimeter shows zero Ohms ( pretty much ) between Pin 2 and USB 0V. If you're not seeing this connection, you're perhaps probing on Pin 1 in error ?

[Qiaozhi]

I'm pretty certain my updated pinout (Post number 10 ) is correct. I have also supplied this info to another detectorist in the UK who was replacing a broken connector, and he made a successful repair.
I suggest you double-triple check the pin numbering sequence, the only pin you can't get wrong is Pin 8, in the middle. But if you're erroneously going round the outer pins in 'mirror-sequence' , you will find Pin1 and Pin 2 become swapped over.
( and to be honest, I think that is why Post 1 has the wrong info, I got the sequence mirrored. It bothers me the first post has the error ... I will contact George and see if he can edit it for me. )

Just as a confidence check, here's something simple you can test, if you have a multimeter:
Coil connector Pin 2 is the main circuit ground, and it's also the 0V ground on the USB port. If you look in the user manual, you'll see the USB port pinout, and 0V is the bottom right pin, as you look at the rear of the control pod. So a multimeter shows zero Ohms ( pretty much ) between Pin 2 and USB 0V. If you're not seeing this connection, you're perhaps probing on Pin 1 in error ?

We don't normally modify old posts without a good reason, but it makes sense to do so in this case, especially as it seems to be causing some confusion.
I have updated post #1 in Skippy's thread with the correct pin connections to match post #10.