Nexus Coil Design

[Mike]

Coil Phase Linearity

I was told this caused phase linearity problems and thus made accurate discrimination almost impossible.

To clarify, do you mean you were told that the phase shift as a function of target resistivity becomes non-linear?

Does anyone know if it makes a difference whether the Rx is tuned above or below the Tx frequency to overcome phase/tuning problems?

Is there a book/patent/other source where all this sort of thing is discussed in this kind of detail? (Perhaps not, since it's all a bit specialised)

Thanks. It's great that people are willing to share their knowledge like this.

[Siliquae_Sid]

To show my ignorance of the matter :o I think "Something like that" would be an appropriate phrase.

I think it was Dave Johnson who told me that there was a phase inversion as the target got closer to the head, something like a filter in nature (look at the group and phase delays for a band pass filter) or was it Robert Hoolko (check the OLD version of this forum) who demonstrated it when he was experimenting with coil design. Either way, I know it happens, just not sure why.:confused:

[Qiaozhi]

Nexus Coil Tuning

The Nexus coils are tunned in full resonance within +/-1% tolerance.
The well known manufacturers arround the world do not manufacture resonance tunned coils for various resons, but the most important of all is costs.

Since the RX coil is a high Q tuned circuit, it is very difficult to keep the receiver circuit tuned to the same frequency as the transmitter. In other words, to keep it in resonance. This is why the major detector manufacturers sell machines with search coils that are sub-optimal. Variations in manufacturing tolerances produce such a large statistical spread that it's like trying to balance something on a knife edge. Tuning the RX coil off resonance effectively increases the manufacturing yield.

If you examine some Tesoro circuit diagrams you can calculate the following:
Eldorado: TX=10.024KHz, RX=10.867KHz
Golden Sabre: TX=14.651KHz, RX=14.714KHz
Silver Sabre: TX=11.963KHz, 13.309KHz

You will note that in all the above cases the RX coil is tuned to a higher frequency than the TX signal. So how close are these detectors tuned to resonance?
Eldorado = +8.4%
Golden Sabre = +0.04% (if the schematic details are correct)
Silver Sabre = +11.25%

Of course, it should be obvious that the closer the RX coil is tuned to the TX frequency, the greater will be the sensitivity. Georgi more or less confirms this in his statement that the abilities of the Nexus are mainly due to the care with which the O-coils are manufactured. This is most probably a system that the major detector manufacturers cannot afford to adopt. Basically, it's down to cost at the end of the day. If this wasn't the case, then people like Charles would not be able to produce coils that can outperform the commercially available alternative.

The Golden Sabre may be an exception, as it's not a coinshooter, and requires the additional sensitivity for detecting small gold targets.

I was told this caused phase linearity problems and thus made accurate discrimination almost impossible. If you have found a way of getting around this then can you explain how you have done this, and the principles and reasoning involved.

If you examine a Bode plot for an LC circuit you will see that the phase versus frequency relationship is nowhere near linear, except perhaps close to the resonant frequency.

[Charles (Upstate NY)]

The problem is once tuned, how to keep them tuned as even temp changes will knock them out. I heard about a VLF (I cannot say where) that can do that. Maybe we'll see it produced one day.

[Qiaozhi]

Target Phase Shift

If you examine a Bode plot for an LC circuit you will see that the phase versus frequency relationship is nowhere near linear, except perhaps close to the resonant frequency.

OK - I've thought about this statement a bit more, and realised that it's total nonsense. The phase shift associated with the RX coil has little to do with discrimination. The phase shift that is seen at the RX coil when compared with the TX signal is (of course) produced only by the target. However, when the RX coil is tuned to a slightly higher frequency than the TX signal, this provides a built-in phase offset that helps to mask iron targets. You can easily test this by building a simple test setup with a Colpitts oscillator driving a test coil, and monitoring the TX and RX signals on a dual beam scope. The phase relationship can be adjusted by changing the value of the RX capacitor. This is similar to what happens when you adjust the ground balance control. If you then place various metal targets near the search head, you will see the phase of the RX signal change.

Two things I've noticed are:
1. The discrimination sensitivity is heavily influenced by how well the coil is nulled.
2. The closer the RX coil is to resonance, the harder it is to null the coil.

As Charles says - drift could be a big problem.

[digger barns]

So how is it done on Saxon X-1 then. I know of someone who has just bought X-1 that is 7 years old and bashed up to hell. But discrimination and tuning are still spot on. And going by depth tests that are published the X-1 is deeper than Nexus. What is the secret here is it all to do with weight.

Digger

[Siliquae_Sid]

Didn't the AH Pro by Alan Hametta use this sort of arrangement, but he used some novel design to get around the problem.

What about putting PIN diodes in the head (in parallel) and real time tuning the head?



There are a few commercial designs, I believe, which use tuned RX including some Fisher and Compass models.

Nice Explanation of the theory Qiaozhi, but ARGH BODE PLOTS!!1 I hoped I had seen thelast of these at university, but apparently not .

Check out Google for a design called the "Radiometric Ferret" This is a VERY simple design, but I built it with centre nulling meters. One gives amplitude, the other phase, and it makes aligning coils a REAL DODDLE as when both needles are centred, you have a perfect null. I use addictional amplifiers to get even more accuracy, but generally I can get within 0.05%

[digger barns]

The trouble with real time tuning or tracking is how do you stop it real time tuning the target

[Siliquae_Sid]

Probalby the same as you would stop auto ground tuning knocking out a target, by applying thresholds which, when exceeded, would stop the system tracking beyond sensible limits I guess, pretty much the same as Whites auto ground track does.

A slew rate limited system might prove more practical, as invariably, ground and tuning will exhibit a slow change, whereas a target signal will have a higher dv/dt.

I'm not sure, no one has ever tried this before, it may be worth someone with some time and a little money having a play.

[Mike]

Nice Explanation of the theory Qiaozhi

I'll second that - thanks Qiaozhi.

Check out Google for a design called the "Radiometric Ferret" This is a VERY simple design, but I built it with centre nulling meters. One gives amplitude, the other phase, and it makes aligning coils a REAL DODDLE as when both needles are centred, you have a perfect null. I use addictional amplifiers to get even more accuracy, but generally I can get within 0.05%

Excellent tip - especially if you don't have a 'scope yet like me. But Google for "Radiometer Ferret", the other one don't work so well.

Thanks,

Mike

[Unregistered]

Saxon X-1

So how is it done on Saxon X-1 then. I know of someone who has just bought X-1 that is 7 years old and bashed up to hell. But discrimination and tuning are still spot on. And going by depth tests that are published the X-1 is deeper than Nexus. What is the secret here is it all to do with weight.

Digger

Can you please tell where to find those test results for the Saxon X-1. I have seen the test results in air for the Nexus in the Treasure Hunting magasine. It will be interesting to put those together.

[Unregistered]

I'll second that - thanks Qiaozhi.

Excellent tip - especially if you don't have a 'scope yet like me. But Google for "Radiometer Ferret", the other one don't work so well.

Thanks,

Mike

You can also null coils with just a voltmeter.

[marin]

You can also null coils with just a voltmeter.

would you tell us please what`s the way of doing this?

[Delbert grady]

Can you please tell where to find those test results for the Saxon X-1. I have seen the test results in air for the Nexus in the Treasure Hunting magasine. It will be interesting to put those together.

I would not take any notice of tests done in any magazines. Magazines will print anything that they think will increase sales of detectors and thus increase advertising revenue. Also always bear in mind that in air tests are the realm of the scoundrel they have no relationship to in the ground performance. There have always been lots of detectors just designed to give good in air ( showroom ) results. Magazine test are next to useless as a guide to how good a detector is. People who do tests for magazines are usually working for some importer or manufacturer and would sell their grandmothers for a cheap detector.

[Unregistered]

would you tell us please what`s the way of doing this?

By measuring the offset voltage in the receive coil. Should measure as minimum as possible.