Balancing "pancake" stacked coils for a VLF pinpointer project

[Glasswalker]

Hey everyone, so I've been lurking around, and doing much reading, and decided a good starter tinkering project would be to try and design a pinpointer from scratch...

I thought, all the other pinpointer designs I've seen are either PI or simple BFO designs, I'd love to make a discriminating pinpointer, thought that would be a fun project.

So I thought about coil design that might fit into a pinpointer formfactor, but allow for induction balance, and then I thought "Ok, a pancake might actually be best here" and so I did some quick calculations and made a template for a coil bobbin that can either be air core, or slides neatly onto a 10mm diameter round ferrite rod. Thinking I could experiment with using the ferrite or not as a core instead of air core to make the field more "linear".

Ultimately it worked out ok as a quick bench experiment, winding the coils to 10uh each with the 0.25mm enamel wire I had on-hand resulted in about an 8mm thick, 22mm OD coil, which could either be air core or fit onto a ferrite core.

I slid that onto a 100mm long ferrite rod, and kept the coils near the end, and tried to balance it up, with TX in the middle, and the RX coil in front, and nulling coil in the back. All 3 coils were wound approximately the same (400 turns). I found of course the ferrite significantly altered the inductance of the coils as measured individually, and depending on where they sat on the rod. So I ended up with the RX coil first, TX immediately behind (and right against it), and the null coil behind that, spaced about 1cm away...

I fiddled about with some caps to match it to 10Khz roughly (just an arbitrary frequency for now, haven't taken the time to figure out perfect resonance of the coils as they were one-off experiments, but this seems to work), and sending in a 10Khz PWM frequency at 50% duty cycle from a pulse generator, I was able to monitor the rx+null on a scope and tweak the location of the nulling coil to balance it out...

Unfortunately the balancing position was VERY sensitive, so only 1mm either direction made a huge difference in balance. I was able to get it down to a reasonably small ripple signal (directly coupled from the tx coil) which with a roughly 5V pulse amplitude, came in at about 20uV ripple, and when I brought a coin near the end of the assembly, I could see the signal distort nicely (it was barely noticeable at about 2cm distance, and got exponentially larger as it got closer). So at first I thought "Ok great! that's going well!" so I moved onto designing the TX circuit to run the PWM from a microcontroller, give it some power amplification, etc... And then thought "Ok now onto the RX side circuit"... At which point the balancing I achieved started to bother me a bit...

The thing I'm struggling with now, is designing the pre-amp circuit for the RX side, I've designed the TX PA circuit, and that's now driving my PWM from a microcontroller, but the RX side, I'm planning on going through a pre-amp first, then a pair of circuits to extract the DC amplitude, and DC phase difference between TX/RX signal, and feed those two voltages into the ADC inputs on the microcontroller... But in trying to piece together the pre-amp, I'm concerned that I can't get that last ripple out of the signal, and it seems fairly significant in amplitude... I tried removing the ferrite, and while that made it much easier to balance, it also made the return signal from a coin completely undetectable on the scope unless I touched the coil with the coin basically (and even then, barely detectable).

So my concern is I won't be able to get any receive sensitivity.

My suspicion is that I need to not wind the coils to equal windings, probably have all 3 coils at slightly different specification, with the TX, RX, and Null coils all having their own winding counts, in order to optimize balancing, and sensitivity... But I'm not sure if this is on the right track or not...

Just thought I should pop on and ask on here, since others have made their own coils, and gone through this exercise...

So if anyone has any tips/advice I'd greatly appreciate it.

(I'm happy to share more details on the project as well, plan on open-sourcing it if it works out ok anyway, just didn't have the material and prototype bits on-hand when I decided to write this, but can do that pretty easily upon request).

Thanks!

[Skippy]

Come on, show us some pictures/drawings/diagrams, a wall of words is hard work.

For what it's worth, I'm slowly working on a discriminating pinpointer project, with a search-coil that should be compatible with my Fisher F75, so there's the possibility of using the design as a 'Sunray style' pinpointer, or a stand-alone very-small coil. No-one has ever reverse-engineered a Sunray ( there's loads of threads on here asking 'how are they made?' all with no answer. ) So I'm just tackling the design from scratch. The pinpointer electronics will be inspired by this Polish project, which is worth a look:
https://www.geotech1.com/forums/show...387-PINPOINTER

I should tell you - the raw output of a search-coil is tiny, so you shouldn't be surprised if it's hard to see. On a full-size coil ( eg. 8 inch / 20cm ) you'll typically get just a couple of millivolts signal from a coin like a USA 25 cent at 5cm / 2 inch distance. At the limit of detection, the detector will be triggering off signals under 5 microvolts in size.

[Glasswalker]

Come on, show us some pictures/drawings/diagrams, a wall of words is hard work.

For what it's worth, I'm slowly working on a discriminating pinpointer project, with a search-coil that should be compatible with my Fisher F75, so there's the possibility of using the design as a 'Sunray style' pinpointer, or a stand-alone very-small coil. No-one has ever reverse-engineered a Sunray ( there's loads of threads on here asking 'how are they made?' all with no answer. ) So I'm just tackling the design from scratch. The pinpointer electronics will be inspired by this Polish project, which is worth a look:
https://www.geotech1.com/forums/show...387-PINPOINTER

I should tell you - the raw output of a search-coil is tiny, so you shouldn't be surprised if it's hard to see. On a full-size coil ( eg. 8 inch / 20cm ) you'll typically get just a couple of millivolts signal from a coin like a USA 25 cent at 5cm / 2 inch distance. At the limit of detection, the detector will be triggering off signals under 5 microvolts in size.

Hey, thanks for the reply, yeah sorry I should post some pics. Right now no real diagrams, (hand experimenting on breadboard in pure prototype mode) but could whip some up easy enough.

Will take some pics and post them up shortly, apologies for the wall of text

Yeah I hear tons of references to the Sunray, but have no idea what it is.

Good to know on your last comment, I had no idea we would be talking that kind of a small signal. That's fine then, I just need tons of gain (in both pre-amp and processing stages). But in that case, I suspect balancing the coils to be that much more important (to get it exactly precise?)

In my original post I mentioned I'm seeing ripple of 20uV, in fact I meant 20mV (millivolts, not microvolts). What I mean by that is at the most precisely balanced I can get it, I'm still seeing 20mv of constant signal amplitude (with no metal targets) coming from the coupling of the TX/RX coils. I can shift it either way and that becomes significantly larger (with the phase of course inverting if I'm either too close or too far away). So with 20mV of signal, I assume I'm going to swamp the pre-amp trying to get enough gain to reliably detect a signal shift from a target of only microvolts...

Anyway, I'll try and get some pics up shortly to make this easier to look at

Thanks!

[Carl-NC]

If you place the TX & RX coils on a common ferrite rod you can make it work on the bench but when you pick it up and rotate it around in free air you'll find you have created a dandy magnetometer that responds strongly to Earth field.

I designed the coil & circuitry for the White's TRX. I had started with the premise of a RX- TX RX+ coil stack but it turned out the Sunray approach was simpler and performed better. It (and the TRX) has a TX coil wound on a ferrite rod and the RX coil is air core. The wires for the RX coil must pass over the TX coil to get back the the circuit board, so for induction balance one of the wires is reverse-wound a few turns over the TX coil. It's difficult to get it "just right" in production so an additional piece of shield tape is moved around to tweak in the IB and glued down.

It worked very well and to date the TRX is still one of the most sensitive pinpointers ever made. And probably the best with tiny gold.

[Skippy]

Wow ! Someone actually has seen the guts of a Sunray. There's a few curious people on here who would like more info, if you could spare the time to describe the physical aspects of it.

The ferrite rod, is it a common 9mm rod, or something fatter etc ? How long is it ? I presume not the length of the plastic pipe the complete probe is encased in.

How close is the end of the rod to the air-cored RX coil ? Presumably if it was close, it would slightly affect the inductance of the RX and affect the coupling from TX to RX. but if it was too far away, the TX magnetic field seen by the target would be less, and sensitivity would be reduced.

How would fine nulling be achieved ? Moving the rod+TX closer/further from the RX ? Fiddling about with small loops of wire to adjust the bucking winding?

What about shielding. Is the RX screened seperately, or is the whole assembly shielded? Graphite paint ? Copper tape ?

It seems the entire Sunray probe is filled with expanding foam. Do you think that's for mechanical stability, eg. stopping wires moving about; or water-resistance, or just to stop people tinkering with it?

My homebrew 'Sunray' IB coil was totally air-cored, I felt that introducing ferrite would be asking for trouble, such as temperature instability, drifting null. My opinion was influenced by observing how much drift the Garrett Propointer shows on sunny days. ( though how much is ferrite-related and how much is the other bits I never investigated )

[Glasswalker]

Wow ! Someone actually has seen the guts of a Sunray. There's a few curious people on here who would like more info, if you could spare the time to describe the physical aspects of it...

Agreed, My other response directly to Carl, somehow ended up in moderator land... Not sure why that is. But yeah any tips you could give seems like both Skippy and I are working on pinpointer projects, and are eager to experiment a bit, and real concrete info on coils for IB type pinpointers seems extremely scarce right now, making it a bit of a dark art it seems... Would love the privilege of benefiting from some of your experience Carl!

...My homebrew 'Sunray' IB coil was totally air-cored, I felt that introducing ferrite would be asking for trouble, such as temperature instability, drifting null. My opinion was influenced by observing how much drift the Garrett Propointer shows on sunny days. ( though how much is ferrite-related and how much is the other bits I never investigated )...

Yeah and as it seems I'm having major issues with my design, which is using a common ferrite core for 3 stacked coils. When I remove the ferrite and use air core, of course they lose a ton of their inductance, so I'm having trouble measuring the outputs on my scope, but I'll be making a few revisions to my driver circuit (schematic incoming) which should help bring the output up. (right now, just to experiment with balance and coupling I was driving it from 3.3v with a simple transistor from the micro's pwm output, but will switch to a mosfet driver circuit running from 9v instead, which should significantly improve output from the coil)

Anyway, eager to continnue this discussion to see if we can't help one another out

[Glasswalker]

Here are some images of my prototype so far (though as per previous posts I may be heading in the wrong direction, would love to try and re-create the coil concept described by Carl for the SunRay or the Whites TRX pointers).

This is my prototype:

The coils are on individual bobbins with an opening in the core for a standard 10mm ferrite rod with flats. I tried on that originally, this pic shows it with a 3D printed plastic rod that is the same shape/size (to test in more of an air core scenario)

And here is the schematic I'm currently experimenting with to drive the TX coils (right now I'm just looking at the RX coils with my scope directly):

[Skippy]

You need very close coupling between the bucking coil and the RX, so that you don't have to put much TX power into it to cancel the main TX field. Either wind bucking on top of RX, or if they are alongside, have very little seperation ( 0.5mm ? ) between them. The TX and RX need to be reasonably far apart so they don't couple too strongly, though you can't have them many cm apart, because that would mean one of them was miles away from the target, not good for target detection.

( love the tiny dot matrix OLED display. I found a lost high-end vape device in a public park recently. Water had got in it a bit, but it still powered up, and the little screen, about 6mm x 18mm came to life, offering me Wattage settings and telling how many Ohms my vape coil was, clever little gadget, an interesting mix of microcontroller bits and 30 Amp SMPSU circuitry, plus battery charger stuff.)

[Carl-NC]

I'll try to answer questions, but this was 8-9 years ago. Somewhere I still have the Sunray probe I took apart but it's part of stuff in storage, likely another year before I can look for it.

A picture is worth 1000 words:



I don't recall exact parameters but I think the TRX had 100 or so turns on the TX and several hundred on the RX. The RX wire was pretty thin, maybe 35-36 awg. Coil spacing was probably 1/2" or so. Don't recall the number of reverse windings, maybe 5-8. The coil assembly was slid inside a short piece if 1/2" fiberglass tubing (made by the same company that made the White's lower rods) and potted with epoxy. After complete assembly of the circuit+coil, a piece of shield tape was maneuvered on the coil tube to tweak in the IB and affixed. Again, I don't recall the kind of tape.

The Sunray may have been shielded, I don't think the TRX was shielded. I don't remember. I think the TRX uses a 6mm ferrite, the Sunray was probably a lot bigger. The TRX was epoxy-potted for stability and I'm sure the Sunray was foam-potted for the same reason.

[6666]

I'll try to answer questions, but this was 8-9 years ago. Somewhere I still have the Sunray probe I took apart but it's part of stuff in storage, likely another year before I can look for it.

A picture is worth 1000 words:

[ATTACH]54226[/ATTACH]

I don't recall exact parameters but I think the TRX had 100 or so turns on the TX and several hundred on the RX. The RX wire was pretty thin, maybe 35-36 awg. Coil spacing was probably 1/2" or so. Don't recall the number of reverse windings, maybe 5-8. The coil assembly was slid inside a short piece if 1/2" fiberglass tubing (made by the same company that made the White's lower rods) and potted with epoxy. After complete assembly of the circuit+coil, a piece of shield tape was maneuvered on the coil tube to tweak in the IB and affixed. Again, I don't recall the kind of tape.

The Sunray may have been shielded, I don't think the TRX was shielded. I don't remember. I think the TRX uses a 6mm ferrite, the Sunray was probably a lot bigger. The TRX was epoxy-potted for stability and I'm sure the Sunray was foam-potted for the same reason.

Invalid Attachment

[Glasswalker]

You need very close coupling between the bucking coil and the RX, so that you don't have to put much TX power into it to cancel the main TX field. Either wind bucking on top of RX, or if they are alongside, have very little seperation ( 0.5mm ? ) between them. The TX and RX need to be reasonably far apart so they don't couple too strongly, though you can't have them many cm apart, because that would mean one of them was miles away from the target, not good for target detection.

( love the tiny dot matrix OLED display. I found a lost high-end vape device in a public park recently. Water had got in it a bit, but it still powered up, and the little screen, about 6mm x 18mm came to life, offering me Wattage settings and telling how many Ohms my vape coil was, clever little gadget, an interesting mix of microcontroller bits and 30 Amp SMPSU circuitry, plus battery charger stuff.)

Ok, so in my setup I had [END OF ROD] | [RX Coil] | [TX Coil] | [Bucking Coil] | [Rod Tail]
In my setup I was going with RX+/TX/RX- setup (so Bucking = RX inverted polarity)

When using Ferrite rod, I found I actually had to keep the TX/RX coil "touching" (no gap), but had to move the bucking coil further away in order to achieve balance (had to move it like 1-2cm away from the others). This is likely due to the effect the ferrite has on the inductance (closer to the middle of the rod increases inductance), and the wierd things it does to coupling...

But with the "plastic rod" (air core) I had them all smashed together as in the photo to get the coupling so that RX ripple is below the threshold of the probe on my scope... But unfortunately that still means that signals below about 5mV are in the noise floor, and so sensitivity wise, I couldn't detect a coin unless it was practically touching the RX coil...

I think looking at something like Carl described for the SunRay, where it is RX Air, TX on Ferrite, and bucking coil wound outside the RX.

Would love more details from Carl though if he has the time to help out...

And yeah the OLED I was playing around with as well, they are neat little devices, and super tiny, so able to fit into a small project. Have a few little ones like that, and thought this might be a neat project to incorporate it to for giving some user feedback for menus or whatnot...

EDIT: Apologies Carl, I started writing this last night, before you had responded, now I see you have sent more info. I'll make another post shortly with a "Visual Aid" hopefully that will assist our discussion (because you're right, a picture is worth a thousand words, but a picture with dimensions is even better from an engineering standpoint lol)

Thanks!

[Glasswalker]

...

A picture is worth 1000 words:

...

I don't recall exact parameters but I think the TRX had 100 or so turns on the TX and several hundred on the RX. The RX wire was pretty thin, maybe 35-36 awg. Coil spacing was probably 1/2" or so. Don't recall the number of reverse windings, maybe 5-8. The coil assembly was slid inside a short piece if 1/2" fiberglass tubing (made by the same company that made the White's lower rods) and potted with epoxy. After complete assembly of the circuit+coil, a piece of shield tape was maneuvered on the coil tube to tweak in the IB and affixed. Again, I don't recall the kind of tape.

Awesome, that's some useful detail. Thanks Carl! Ok, so based on this, I've done a mock-up to use for discussing dimensions. This is not meant to be scale, just to identify the parts, make sure I'm interpreting your notes correctly, and possibly attach some real numbers to the diagram, from there I can do a more accurate mock-up:

So I'm interpreting a device which has the RX coil (blue) out in front (at the tip of the probe), and TX coil (red) in-behind. TX coil is on a ferrite core, RX is Air Core. And the RX then has some reverse windings coming back over the TX coil for balance...
See this crude mock-up:


And a Cross-Section of the mock-up:

And finally, a Flat Cross-Section Diagram with dimensions added:

Ok from that diagram, the following dimensions are described:

• A - Diameter of the TX Core (you've already stated this was likely around 6mm for TRX)
• B - Depth of the TX Coil windings (from the fact you said 1/2 inch tubing held it, probably 3mm approx)
• C - Length of the TX Coil
• D - Length of the TX Core
• E - Winding Count for TX Coil (you've stated this was likely 100 or so turns on the TRX)
• F - Diameter of the RX Air Core (ID of the coil... I'm assuming this is the same as the TX coil? so 6mm approx? or is it different?)
• G - Depth of the RX Coil windings (this is hard to extrapolate without other dimensions to work with such as length and core ID/Core Diameter)
• H - Length of the RX Coil
• I - Winding Count for the RX Coil (you've stated this was likely several hundred)
• J (not on diagram) - Number of reverse windings for Induction Balance (you've indicated this was only a few windings, maybe 5-

From this, can you confirm if I'm even remotely on the right track for the concept? And could you help maybe nail down any more of the above dimensions (even with rough/approximate numbers).

With this we could then finish a more accurate mock-up, and I could prototype a coil

Cheers!

EDIT: I guess I should have said "TRX MockUp" instead of SunRay, in that we seem to have more data on the TRX, and the SunRay (while similar in concept) is different dimensions. The idea here is to get enough info around one pinpointer probe design, to be able to figure out ratios/aspects etc of coils, which is all important to get it "just right". So for the purposes of this exercise, let's focus on the TRX dimensions above until we get a design figured out.

Thanks!

[Skippy]

Thanks, Carl, though despite my best hacking attempts, I couldn't get the image to load. I'm surprised by how small the probe coil is, I assumed the diameter would be larger.

Glass: I'm not sure putting all the coils on a ferrite rod is the way to go, the rod is just going to couple them all together too well. And you don't really want much direct coupling between the TX and RX.
The air-core assembly at least has a 100% chance of working ( hence my choice for my coil construction ).
Aim for around 10% coupling coefficient between TX and RX. RX to bucking coil coupling should be high, 70%+ , so you don't need loads of turns on the buck to cancel out the TX direct coupling.
( I don't have ny notes handy, and I haven't worked out the details yet, anyway, and it's a year since I did any work on it, so I can't remember much, TBH )

Cute though the dot matrix display is, it unfortunately has no true place on a pinpointer, which will be absolutely cacked in mud in use. A 5mm LED is going to be hard to see in real life.

[Carl-NC]

I've fixed the image attachment above. A longer TX (or at least a longer ferrite rod) will help with depth as it keeps the magnetic field better focused, up to a point.

[Skippy]

Thanks for fixing the image link, Carl. It's not too far off what I was imagining, but I thought a larger diameter larger-bore shorter-length air-cored section may have been used. And I agree that a longer ferrite rod would be better, something 60mm - 100mm, as used in most handheld pinpointers.

It would be possible to fine-tune the induction balance of this setup by moving the air-cored coil closer/further from the ferrite rod end, as well as all the other possible methods.