Improving the Tracker IV

[Silver Dollar]

I drew up part of the Tracker IV schematic a while ago (the rest was difficult unless you remove
all the components). The resistors were verified by reading the values but caps were measured
in circuit so may be wrong values on schematic.

The circuit looks a bit different than most here and some tricky things are done when switching
modes. (Removing grounds from chips and such).

Mine only goes 2 - 2.5" (50 - 65mm) deep so I was looking for some improvements to gain a bit
more depth. The RX seems to be tuned to 32.5 khz where the TX is 6.5 khz seems a bit funny
maybe tuned to a harmonic?

I thought I might try winding a new coil and maybe messing with the TX (though that's tough
without the complete schematic. Maybe just tuning the RX cap would gain something? I tweaked
the VR1 pot and gained 1/2" or so but it probably messed up the disc.

I'm going to test some baseline targets today then try winding some coils to see what I can do.

[kt315]

thank you for info Silver Dollar. did you draw schematic from SMD board? Tracker IV is nice two tone detector with smooth clear discrimination.

[Silver Dollar]

Yes SMT pcb that's why it is hard to find all traces.

If I had a bad one I could remove all parts and do a better job tracing.

I tested my gold wedding band (21 ct gold) at 5 - 6" in air.
It's hard to keep the target straight until the measuring stick.

I tried adding and subtracting 0.5, 1, 2, 3, 10 nf to C1 but didn't see any amazing increases in range.
A better testing setup will be needed to tell small gains. Adding 2nf seemed to make a stronger
signal but was still around 6".

The TX measures 6.67 khz and around 20V P to P so I don't think you can gain much there.

RX measures 2.46 mh
TX measures 2.68 mh

I still wonder why the RX is tuned with a 10nf cap as that gives resonance at 32.5khz
Maybe I'll try 200nf just for fun!

[Silver Dollar]

Well adding a 100nf didn't really do anything to tune the RX so I'll try winding a new coil next.
I have some 10" housings so I layed out a pattern and wound 43 turns around it. The length
of one turn is 25" and so I calculated the diameter as 7.96" and used Qiaozhi's coil calculator
to find 41 turns should be 2.45mh and 43 turns 2.65. The stock one reads 2.68 so I went for 43 turns
for the TX.

The RX is a bit different. It didn't seem to matter how it is tuned so maybe more turns will give more
signal? I might go for 105 turns at 14.5mh if I don't run out of wire...

That's triple the wire so should bring the signal level up a bit!

[Skippy]

For what it's worth, Designer Dave Johnson (First Texas Products) thinks the Tr4 is a great design, and he's not seen any need to try and 'improve' it. So your attempts may ruin, rather than enhance it.
If you want a larger coil, just buy the correct one. The 10 inch Magnum should be available at a sensible price. Not sure it will work as well as it might on a Fisher F4, for example, but worth considering.

Looks like you've been reading my posts, about the F2 frequency, judging by your notes on the circuit diagram. Ther's an error in your maths. 2.4mH & 240nF give a frequency of 6630Hz (should actually be 6.59KHz, I think)

[kt315]

6,6 khz is whites frequency and any whites xlt, classics coil will work. 9'5 xlt rainbow coil is preferable.

[Silver Dollar]

Well the calculator was off a bit so I had to add some windings.

I ended up using 50 turns for the TX for 2.45mh and 150 turns for the RX for 14.7mh.
I'll try to null them today and see how they do.

No time, money, nor inclination to buy a coil. This is a $60 detector it doesn't make sense to
pay that much or more for a coil that "might" help.

I believe these detectors are generally good performers, I just got one of the 10% that don't
work as well. I figure 10% work better than the others, 80% are average and 10% are poor
performers. Ivonic built many coils and 1 out of 4 just didn't work as well. I'm guessing that is
the problem here. All I'm out is some wire if it doesn't make any difference.

The thing has TL064's and LM324's so I'd expect you could use different opamps and see better performance...

[Silver Dollar]

Hmm not very promising. I had to use 0.250 uf to tune the TX to 7.0 khz and then found the RX chattering
at 1/2 sensitivity. The depth for a gold ring was only 2"! It seems I'm going backwards here...

[Silver Dollar]

Well I nulled the coil and found it to airtest at 5", that's a bit less than stock but the
TX is tuned to 7.0Khz and is 17V instead of 20V. Hnn I guess the coil is not the problem.
Maybe try some other frequencies? There's a lot of filtering on the input opamp I wonder
if it might like to be opened up a bit?

[Silver Dollar]

I tuned it to 17 KHZ today and got 20V peak to peak and 7.8 ma into the coil!
This rig has a nice transmitter! I ended up getting a lot of falsing after a bit
though so will have to sort that out.

The stock coil has TX tuned to 6.67 KHZ and RX appears to be tuned to 33 KHZ
(2.4 mh and 10 nf). Some call it untuned.

I tried to get 13 khz for the RX (14.7 mh and 10 nf) and 15,644 khz (2.4mh and
50 nf) but the TX ended up at 17 khz....

Maybe possible to make a nice gold detector using higher frequencies?

[Silver Dollar]

Hmm retuned it to 7.2 Khz and got 24 ma into the coil (at 17.8V peak to peak) ...

I wonder what frequency peak current would be? I'd expect the batteries to not last long
with that kind of flow. I'll have to measure the stock coil to see what it runs.

It may have something to do with it running in the capacitive side or inductive side as Dave comments about;

" Therefore at 10 kHz the transmitter reactance is capacitive whereas the receiver is still inductive. "

[Silver Dollar]

So I thought maybe the parts on my board are on the outside edge of tolerance.
I have a bunch of PCB's with very precise parts on them. I'll swap the good parts
in for the stock ones on my board and see if it makes any difference.

The other possible enhancement is to lower noise on the input opamp. The
current one is a TL064 which is low power but has 85nv/hz voltage noise.
Maybe put an op27 (4.5 nv/hz) there to help pick up faint signals better?

The stock coil doesn't get noisy at high sensitivity but the 14.7 mh RX coil
I made does so with a quiet opamp in the front end maybe we can gain some
depth?

[Silver Dollar]

Here's a layout of sorts;
TrackerIV.TIF

[Dave J.]

The TK4 should air test in the 5 to 6 1/2 inch range on typical coin size targets. Almost anything you try to do to improve it will actually screw it up although casual air testing may not reveal the screwup.

The BH Junior and its various clones marketed in various venues, is a lot easier to screw around with if you're so inclined. But this is electronics and you still have to understand what the heck you're doing.

The basic problem with trying to modify a commercial metal detector to improve its performance, is that a real beeper engineer already did that before the product went into production. In some cases some sandbagging was done to tame the thing down, and if you know what you're doing you can take out the sandbagging and make the thing more radical. Of course if you succeed, then you get the problems that come with doing that.

--Dave J.

[Silver Dollar]

Thanks for your input Dave. I think I came to the same conclusion that there's not much you can do to improve it.

I actually bought it thinking I got a case, stem and coil for a good price and planned on making my own detector
to go inside but I learned a few things and thought why not try to improve it...

Now about the sandbagging...