Minelab Go-Find

[kt315]

I've just probed around the coil with a small ferrite-cored sensor,

connect a signal generator on RX coil and will find where is RX situated.
another way you know.
here DFX coil under X-ray, took this pic from RU forum. it is only example.

[crane]

GoFind Patent...

[Davor]

So it is electronically balanced. Interesting.

[bernte_one]

there are some µController vlf diy Projects from russian Forum where the coil is balanced by electronic/Software so nothing new but patented form ML

[satdaveuk]

I Wouldnt be at all suprised, been a budget type detector if it was produced somewhere like china for Minelab.
Regards

[Hannu]

is this same method? https://across.fer.hr/_download/repo..._IOP2013_r.pdf

[Davor]

Not exactly. The FER solution is partly balanced by geometry, while judging by the patent, go find is entirely relying on a bridge. The former is less prone to Tx noise problems, but apparently go find fixed the noise issue by other means.
It all boils down to how low you can go with the noise. Better IB -> lower noise. Low noise Tx generator -> lower noise.

A lesson to hobbyists building their own rigs: it can't hurt to take care of the Tx noise by all available means.

[Qiaozhi]

GoFind Patent...

Been busy, and not had a chance to continue this investigation until now.

Well done on finding the patent application. This is clearly related to the GO-Find.
So, it appears that the GO-Find coil is neither concentric, nor coplanar; but is indeed a mono coil (which surprises me).

Continuous wave (or VLF) detectors require that their TX and RX coils are magnetically decoupled from each other. This is usually achieved by overlapping the two coils [DD type], or using a nulling (or bucking) coil [concentric and omega] to achieve induction balance. The result is that weak signals from a target, that has been excited by the TX, are not obscured by the strong TX signal. This requires that coil nulling must provide attenuation on the order of thousands.

In general, the only detectors that use a mono coil are either pulse induction (PI) and beat frequency oscillator (BFO). Quoting the patent application: "The BFO is very limited in terms of discrimination and ground rejection". While testing the GO-Find, it is obvious that this detector is neither a PI nor a BFO, and therefore must be a VLF type.

So the question is, how have Minelab achieved such a good result in terms of sensitivity and discrimination with a mono coil. Clearly, this detector is something unusual.


The patent appllication cites a number of other patents that attempt to provide electronic nulling in non-balanced coils:
US4030026 - fixed arrangement
US4006407 - adaptive system
US5961640 - adaptive system
US5729143 - adaptive system

All the above require the coils to be coarsely nulled first. The electronic nulling is only a fine adjustment afterwards.

Fig-3 shows what essentially looks like a Wheatstone Bridge, but closer examination indicates this to be something subtly different.

Fig-4 is the same as Fig-3, except that one arm of the bridge is replaced by an RLC series combination. By driving the left-hand node of the bridge with a TX signal, known as the "reference", and the right-hand node with an identical signal (but 180 degrees out of phase), the signal generated at the top node will be zero. Likewise for the bottom node. In the presence of an excited target, the top node will produce the RX signal (plus noise), whereas the bottom node will only produce the noise. Hence, subtracting the bottom node from the top node will result in the RX signal only. In addition, the in-phase and quadrature components of the RX signal can be extracted in the normal way by synchronous demodulation.

The circuit in Fig-4 was tested (by the inventors) using a 203mm diameter coil with a TX frequency of 7.76kHz (same as the GO-Find). C(73) = 27nF, L(75) = 16mH, R(76) = 18R, R(77) = 18R, R(78 ) = R(7 ) = 100R, R(79) = 100R. The gain of amp(50) = 50, and amp(52) = 23.

Fig-5 shows an embodiment for a metal detector with separate TX-RX coils and non-resonant operation. The voltage generated at the top node depends on the mutual inductance between the two coils (93 and 95).

Fig-6 shows a comparison between a detector with a concentric coil and one with a mono coil. The concentric gives slightly better sensitivity when the target is close to the coil, but the mono has better sensitivity further away.

Figures 7 and 8 show the differentiated (high-pass filtered) result of the in-phase and quadrature components respectively, after removal of the correlated noise. Noise reduction on the in-phase component was 2.4x for RMS values, whereas the quadrature component was reduced by 22x.

Figures 9 and 10 show the residual noise of a concentric coil and a mono coil respectively. This demonstrates that, since the noise floor and the sensitivity are quite similar, then the invention is seen to be effective.

Figures 11, 12 and 13 show other embodiments of the invention. Fig-13 is the most interesting, as it uses a twisted bifilar coil arrangement for the top right-hand resistor, which is also thermally coupled to the TX coil. This arrangement leads to enhanced thermal stability.

The exact coil arrangement within the GO-Find is unknown, except that it is more or less certainly a mono type.

A very clever solution that appears to provide some good results.

[Davor]

Most of the VLF designs employ a free running LC oscillator, which is running in a positive feedback loop, and is virtually impossible to control by any kind of external feedback loop, hence there is no way to silence its noise. The farthest you can go is limit the amplitude by letting it saturate at low noise power rail, but it is still quite noisy. Forced oscillation as in Bandido or Verator may be a step forward, but noise requirement as explained in a patent is a different animal altogether. I'm not sure I wish to go that way.
I found a way to eliminate "air signal" at Rx front end, but it is still prone to Tx noise (attenuated by virtue of IB). Perhaps it is high time to reconsider Tx noise in VLF.
The marketing leader is apparently on the rampage to fix VLF with zed and go find. What's next?

[deemon]

So, it appears that the GO-Find coil is neither concentric, nor coplanar; but is indeed a mono coil (which surprises me).

What surprised me is that they use not a digital but quite an analogue solution ...

Figures 11, 12 and 13 show other embodiments of the invention. Fig-13 is the most interesting, as it uses a twisted bifilar coil arrangement for the top right-hand resistor, which is also thermally coupled to the TX coil. This arrangement leads to enhanced thermal stability.

The exact coil arrangement within the GO-Find is unknown, except that it is more or less certainly a mono type.

A very clever solution that appears to provide some good results.

But you see , Qiaozhi , that my solution described here - http://www.geotech1.com/forums/showt...3-mono-coil-IB - does quite the same , being much simpler And I also use a bifilar winding there , in order to exclude a thermal-dependent part of the balance chain ( coil wire with its PTC ) .... but compare a parts number in both circuits ...

[deemon]

Most of the VLF designs employ a free running LC oscillator, which is running in a positive feedback loop, and is virtually impossible to control by any kind of external feedback loop, hence there is no way to silence its noise. The farthest you can go is limit the amplitude by letting it saturate at low noise power rail, but it is still quite noisy. Forced oscillation as in Bandido or Verator may be a step forward, but noise requirement as explained in a patent is a different animal altogether. I'm not sure I wish to go that way.
I found a way to eliminate "air signal" at Rx front end, but it is still prone to Tx noise (attenuated by virtue of IB). Perhaps it is high time to reconsider Tx noise in VLF.
The marketing leader is apparently on the rampage to fix VLF with zed and go find. What's next?

But what is interesting - if you are able to cancel a TX frequency completely , you are nulling all the TX noise located near the carrier at the same time . In another words , our "balance modulator" can suppress not only a carrier , but also a USB and LSB TX noise components , so they cannot interfere with USB and LSB RX signal components , appearing when the target is moving near the coil . It means that TX noise isn't a problem when we can maintain a good coil balance all the time .

[Qiaozhi]

What surprised me is that they use not a digital but quite an analogue solution ...



But you see , Qiaozhi , that my solution described here - http://www.geotech1.com/forums/showt...3-mono-coil-IB - does quite the same , being much simpler And I also use a bifilar winding there , in order to exclude a thermal-dependent part of the balance chain ( coil wire with its PTC ) .... but compare a parts number in both circuits ...

Yes - after reading the patent application I also searched Geotech for any similar ideas. Your neat solution was one I remembered reading before, and it was also on my list to investigate in the future when I have more time.

Read the claims on page 7/8 of the patent application. I expect some of these claims may be rejected by the examiner as being too wide. The application was filed on the 29th August 2013, and is based on an Australian patent application (AU2012903739) which has a status of "lapsed". The Australian application was filed exactly one year previously.

http://www.ipaustralia.com.au/applic.../AU2012903739/

[crane]

Australia introduced Innovation Patents some time ago, supposedly to encourage more innovation. The innovation or invention isn't made public and you have a year to do something with it and if you don't then it is abandoned. If you do wish to continue then the invention/innovation must be made public. This type of patent is only good for eight years but can't be enforced without an examination and a certificate. This costs a bit extra but is much cheaper than a full standard patent.

Minelab's AU2012903739 was an innovation patent, allowed to lapse, and was converted to a full standard patent exactly one year later.

.

[WM6]

... an innovation patent, allowed to lapse ...

.

Thanks for explanation.

What exactly mean "allowed to lapse": by not paying some annual tax or by not provided some info to examiner?

[crane]

Thanks for explanation.

What exactly mean "allowed to lapse": by not paying some annual tax or by not provided some info to examiner?

I meant it was a deliberate act because they intended to file a full standard patent. If an innovation patent is just abandoned then the applicant can't claim priority but he can if he converts it to a standard patent or pays to have it examined as an innovation patent.