Whites 6 1/2" VLF concentric coil

[Skippy]

While you were away, I did briefly attempt to work out the full oscillator circuit. I got bored, but suffice to say the 4069 chip is very important in its operation. The inverter on pin1/pin2 is biased linearly with the extra resistor tacked on diagonally, and this appears to be fed with an AC coupled signal through the cap nearby. The output of this inverter is fed pin2 to pin3, pin4 to pin5, pin6 to the coil via 7k5. This chain of inverters has a lot of gain, and results in a square wave drive to the tuned coil, but the 7K5 resistor allows it to resonate sinusoidally. My guess is the oscillator is based around these 3 inverters, and the feedback coil. The high loop gain probably means that you might indeed need only a small number of turns on that feedback winding,1 to 10T, I did say 5 to 50 before, because I knew nothing about the gains elsewhere, clearly one turn might be sufficient, in principle. Slightly too many turns shouldn't be so important as too few, as a guide. As an experiment, try forcing the chain of inverters, eg. stick a length of loose wire, or your finger, on pin 1, so it picks up mains crap, and see if it honks at 50Hz on pin6.

[Skippy]

While the coil drive inverter, pin 6 of the 4069 will have a square wave on it, the light coupling via the 7k5 will result in sine wave oscillations on the coil. The Q factor will be roughly Q = 12, you've got 7k5 across the LC cct, and a series resistance of 10-ish Ohms (the coil resistance). This is a decent figure, typical for a detector, I think. The diodes (regular p-n, and a Zener) that are part of the drive circuit will be to clamp the oscillations, it's anyones guess, but it's not implausible that there could be 20 Volts pk to pk on the coil, so even a single turn feedback winding could give 50mV RMS, enough to drive the gain stages. I'd try 5 turns first, then if and when you get it going, maybe reduce the turns, to see if it improves any aspect of operation.
If yo have a signal generator, you could feed 100mV pk-pk sinewave at 6600Hz into pin 1 of the 4069 via a capacitor, 10nF? and see if it makes the coil sing, adjust the freq and see if there's any resonance peak detectable.

[parun14]

Need some help
Hi to all
I like to build 6" DD coil 50 kHz for white's GM3 .Some one know how to go?
thank you

[nalag46]

I have connected the coils in every possible configuration and I just cannot get anything out of them. There is something I am missing here because with the adding and taking away of turns etc. one should get some kind of reactance, anything to work on and just take it from there, but there is nothing. The oscillator is running and I can control the threshold level to a low tone so that I can hear even the faintest whisper, but bringing any metal near the coils does not produce any signal/tonal variation to the threshold whatsoever.
It looks like I am going to have to wind up one or two extra coils of different lengths but it will pay for me to wait until I receive my used 4B coil in the next few days and examine how that one is set up before going any further.
By the way, after digging in deep into my archive of old detector data, I found the attached diagram of an old Coinmaster 2D. A detector of the same period that is the closest yet I have seen to this one. (I don't think Whites ever designed two models alike).
Looking at the TX coil circuit, it can be clearly seen that our "elusive Screen" is just a centred tap earth end on this one??? while the White and Green cables go the exact transistor arrangement I have here.

[Skippy]

So you've got the oscillator running, that's a start. What frequency does it run at? Is it even close to 6600Hz? This is important, because some of the other circuitry will likely be 'optimised' for 6600Hz, too. Does it look like a decent sine-wave on the TX? What voltage level is it? Did you wind the coil with extra windings like i suggested (read :told) ? Can you look at the receive coil output on a scope? Does it look remotely small, like less than 100mV pk-pk?
Regarding the rest of the circuitry, that sample-hold circuit based around the 3140 opamp is critical to correct operation. Assuming the IC works, I would recommend cleaning all the flux crap off the board in that area, it isn't going to help leakage-wise.

[nalag46]

At this stage is where I let myself and you in this case down I am afraid. You see, I have been dealing with metal detectors (both using and repairing them) for well over 30 years. I had a shop premises with a good and well equipped workshop area but when I sold the business, some years back, my other half was quite adamant that I should not bring all that gear into our new house unless I wanted to carry on doing it in the garden shed, which I was not, and therefore it was just mainly going out detecting than anything else. That said, I continued to get involved helping others with choosing and using equipment, sorting out all their inevitable problems etc. and so all the repair work I do now is based on my past extensive experience of circuitry fault finding but without "serious" testing equipment, which translate to lots of guesswork and more often than not, diagnosing by "cold" resistance tests of passive components and straight replacement of others.
So, apart from a frequency counter, multimeter and a pro digital power supply, I am sorry, I have nothing else to ascertain the readings you are looking for here. The running frequency I am getting is
just 50khz.
Regarding the opamp CA3140, I am certain to be working. I know how this simple sample-hold circuit functions (it dates back to this VLF time) and it is working fine here.
The 4069 hex inverter CMOS is another matter. I have had to replace these many times as they fail easily

[Skippy]

I don't know what to say, really, without a 'scope, you're stuffed. You're not fault-finding, you're rebuilding, reverse-engineering, designing, prototyping, a scope is pretty much essential.
And as for the oscillator running at 50KHz, either your meter is picking up harmonics or such, or those caps in the coil are not soldered right, becausethat thing has to run at 4 to 8KHz absolute worst case.

[nalag46]

I agree but that is the reality. I always get there though mainly through sheer perseverance and I am sorry, I meant the oscillator freq. to be 5khx not 50, lol.

[Gwil]

I'd like to suggest that the four diodes shown in posts 50 and 64 are not part of the Tx circuit, but form a switching gate, a type of synchronous demodulator. Note that the string of inverters is driven from the output of the disc. control.

Gwil

[Skippy]

I admit, I only gave that circuit diagram a quick glance, with it being sideways on, and half the width it should be, but now I've screengrabbed it, rotated it, stretched it out and compared it to the PCB photos. And Gwil is correct, it's a synchronous detector, the following stages are all near-DC operating, not 6KHz, and there is a second 7K5 resistor from the 4069 pin4 driving the diodes. I think all 4 diodes are the same, one is just a different batch/manufactuer, probably... The only 'error' obvious from that circuit is the bucking coil is not shown. The oscillator is just based around the single transistor. The good news is there is a DC voltage out of the first post-demodulator op-amp, so you can examine that with a multimeter to see the null. The bad news is possibly the two tuned circuits, one on TX one on RX. Are they tuned together(might explain the extra caps on the TX) or is the RX tuning cap deliberately off, + or - 10% to 20%. It's pretty tricky to determine operating frequency from other aspects of the circuit, like the discrimination phase shifter, for example. Bearing in mind how an LC circuit has a resonant freq varying with Square-root of L, running at 5KHz instead of 6.6kHz means the L is miles off, about 30% more turns than neccessary, which is a lot..... I'm still puzzled by the low resistance that nalag quoted for his TX coil, are you sure on these details, nalag?

Here's a good experiment, in the absence of 'scopes, inductance meters. Accurately measure the osc. frequency. Now add extra capacitance to the three existing ones, say another 10n, then another 22n,33n etc, measure freq. If the existing 72n is all there is affecting frequency, then the extra caps will change the frequency in a mathematically predictable way, and an L value can be calculated. A pretty darn good approximation is that L varies with turns squared, so you could work out how many turns are needed on the TX to get 6.6KHz. And wind a new TX coil.
If I get any other ideas I'll post tomorrow.

[nalag46]

You are right. I made a mistake with the colour of the wires. It is the black and red cables, or RX side of the coil that goes to that part of the circuit.

[nalag46]

Thanks Skippy. I will work on that. Another bit of good news, well may be, is my "used" 4B coil has arrived and it looks in pretty good clean condition with no signs of any work being done to it. I am not going to open it just yet, rather I will connect it to the circuit and see what happens. I will check the resistance of the two coils and take it from there.

[nalag46]

Hi again Skippy,
I hope you still remember me? I put this project to one side while I attended to some other businesses that required my attention and now that I find myself with more free time on my hands,
I am coming back to try to finish this job if I can. Recently, I managed to get a really good Coinmaster TR 2 DB metal detector in most unbelievable condition as it dates back to 1989, but it looks
like it has never been used before, practically brand spanking new and in full working order of course. Anyway, after examining the pcb components etc, it all seems to tally up nicely with the circuit diagram of this mode I posted here before. It runs at 6.5 kHz as stated in the manual that comes with it. It uses the 8" concentric type of coil, same as in the Treasuremaster 6 1/2". The coil resistance are (accurate readings) 32.3 ohms for the RX and just 6.5 ohms for the TX, which are not a million miles away from the other readings I have on the smaller coil, so in theory, this bigger coil should work here but unfortunately, it does not appear to do so. So far, all I have had time to do is to connect the working coil to the T/Master circuit board. I noticed that the 2 DB coil lead has five wires but the braiding/earth cable is not used in that circuit whereas it is needed in the T/M?, so connecting the four wires here results in some kind of reactance from the pcb (pcb seems to be acting as the coil itself reacting to any metal that comes close to it) but nothing from the coil itself. This symptom can be eliminated by simply soldering the braiding cable to the pcb earth, but, as mentioned no reaction from the coil. I have replaced all 3 ics and double checked all transistors and diodes. After turning on, the tuning level can be brought down to the preset threshold level needed by pushing the reset button, so indicating everything is working ok on the pcb. I have also coming a new multimeter that can measure inductance and I will be able to check that as well when I get it. So do you think there could be much difference between the two coil sizes here?

[Skippy]

The fact that one uses 4 wires, the other 5 tells you they're not the same. I thought you got the oscillator working on the 6.5 incher? I know it's not at the right freq, but it's a starting point.

[nalag46]

Thanks for getting back. I was wondering about the wires and I am tempted to open the 8" coil to find out how it is set up? The only hold back I have is a new coil from a new detector which I do not want to keep and will be reselling it when I have found out as much as possible from it (ony reason I got it). I did also get a Whites 8" 4B and that is the multi tapped coil I was thinking about when I started this thread, needless to say that one most definitely did not work either. I forgot to mention a while ago I received a private message from someone who had been following the thread to tell me that in 1991 or thereabouts, he had one of these 6 1/2" coil opened to replace a damage cable. He says that in the process he ended up winding a new TX coil as all the connections and parts of the windings looked rather poorly (whatever that meant) and while he could not be 100% sure, he reckons the length of the damaged coil was a tad short of 40 meters which worked ok when he wound and fitted a new one. Now, 40 meters would be less than 90 turns I believe so can that possibly work? He said he kept some notes and would let me have more details if found but have not heard from him any more (over two months now). Anyway, I still believe this C/Master 2 coil should work with my circuit board which is almost identical to the one inside the bigger detector.