Tesoro Toltec 100 schematic and adjustment?

[remus_im]

anyone built one after the Toltec 100 scheme?

[remus_im]

anyone construit cineva dupa aceasta Toltec 100 schema ?

[Merc]

Resistor value incorrect on pin2 of U4. Resistor should be 200K not 100K.

Is anyone else attempting to build one of these? Am I the only one foolish enough to try?

Merc

[Merc]

This is revision 5 of the Toltec 100 schematic. Porkluvr pointed out a missing 68pF capacitor in the notch circuit. Thanks Porkluvr. If anyone else knows of any errors, please let me know.

Merc

[porkluvr]

Phase simulation file (LTspice)

Comparing the Golden Sabre and Toltek, I see very little difference in the notch, discrimination and ground balance circuits but I already said that.
The filter circuits are different and of course the Toltek 100 adds target ID. I'm not touching that today.

This is the basic tx/rx and phase reference circuits as best as I can duplicate them. There are two different pot symbols in my simulation file and the only difference
is the way I drew them. Either an arrow points to clockwise, or else a little o indicates counter-clockwise. I can not be certain they are installed with
the proper orientation because I don't have a Toltek or Golden Sabre to look at for verification... so I connected them the way I think they are supposed
to work. If somebody knows otherwise I surely would appreciate a heads-up.

There are two things I need to point out.
1. The Toltek drawing does not show a connection to ground in the transmit coil although I believe that there is a ground connection in the coil connector. The
receive signal lags by about 20 degrees without the ground, which would probably be a malfunction.
2. I added a 68pF capacitor in the notch width circuit. Without a capacitor the width control has no effect. I got the 68pF value from the
Golden Sabre schematic.

5.7kΩ parallel resistance is added to the transmit coil - which would equate to loss in the circuit. I arrived at that value through trial and error.
Everybody who runs this simulation should remove the parallel resistance and watch the transmit signal for about 10ms and see what happens.

I think that all the necessary symbols and models are included. If something missing - let me know. Some of the extra
opamp models in opamp.sub and comparators.sub might not work. Nothing I can do about that.

One more thing: I modified my JFET symbol and maybe some of the others. When your copy of LTspice reassembles the simulation
file there may be some skewed components and missed connections that you will need to correct.

[porkluvr]

mistake in formula

Now, that was stupid. I arrived at a more or less correct answer because even though I wrote parallel, I used the series combination of the 220nF and 22nF capacitors. The coil sees the 220nF and 22nF capacitors in series and the output frequency is approximately as shown below.

LTspice outputs about 14.4kHz with a 5.625 mH coil and 21Ω series resistance, 1234pF parallel capacitance. I add parallel capicatance in the coil to try and duplicate interwinding capacitance - but I have no idea of the real value. 1234 might be the wrong number but zero would also be wrong.

Besides interwinding capacitance we would also need to include cable capacitance in the calculation. So I just make a guess and call it 1234pF.
While I am on the subject, the receive coil and tank circuit would also affect output frequency to a small degree but I'm not going there!

Two parallel bars ( '||') depict a parallel combination. Does anybody know shorthand for a series combination?

[porkluvr]

I'll see what I can do to get some spice files together. They need some dusting off.

It's relatively easy to get portions of the detectors running in LTspice but it gets hairy as you add demodulation and filters. One other thing - it's hard to see what is happening unless you inject some varying target phase into the system. Even then, I doubt that simulation can accurately show what happens in the field. Not with any of my current methods, anyway.

I believe I wrote the correct formula last night BUT when I calculated the parallel combination of 220nF and 22nF I used the series combination formula (for whatever reason) and got an incorrect answer. (doh)
I think the correct answer should be closer to 14.5kHz. That is about what the oscillator does in LTspice.

Unfortunately I can't seem to punch the right numbers into my calculator this morning and keep getting a very different result from what I got last night.
I hope that's not an presage of how the rest of my day might go. I don't want to cancel a Saturday!

[Merc]

LTSpice

I would love to have a copy of those LT Spice files and the spice models to support it. For the Golden or whatever.

thanks,
Merc

[porkluvr]

So be it!

I think that saying the Toltek 100 operates at 12kHz is probably about as accurate as saying 5x7=42. Close, but no cigar!
I just wanted to clear that up. I'll set up the circuits in LTspice and see if I can figure out what they're doing.

I expect that your freq meter will indicate closer to 14.5k because it looks like Toltek 100 and Golden Sabre (which has been pretty well established as operating at around 14.5kHz despite what the GS manual says) use the same frequency determining components and power supplies. There are also a whole lot of other similarities, so it should not take very long. "I've been there, done that." Made lots of mistakes along the way.

I have something else that will keep me busy for a day or two before I can get started.

edit: I think the cable and coil both contribute some added capacitance that is not accounted for in the basic resonance formula.

[Merc]

Coil

After letting scope warm-up a bit, measures 15384. I have a meter a work with a built in freq. counter. I'll try to use it to get a more accurate count.

Merc

[Merc]

Coil

The Toltec 80 and 100 used the standard white solid 8" concentric coil. According to my notes when I checked it years ago the the TX coil is 5.62mH, the RX coil 6.25mH. I just checked my Toltec 100 oscillator with a scope and calcualted the freq. to be 15625. My scope has been sitting cold for a while so not sure that I trust it. I have the original manual and it says 12KHz.

Merc

[porkluvr]

Merc,

I'm wanting to simulate parts of the Toltek 100 circuit to try and figure out why the demodulators are set up like they are but I have a basic problem.
I read in the operators manual tht the operating frequency is 12kHz but when I use the schematic values and the best data I have on the Tesoro coils, LTspice gives me a 14.3kHz output.
I'm assuming that the transmit coil is about 5.6mH (21 ohms) but that is only because most of Tesoro's other IB detectors use that value coil.
Is the Toltek coil any different? And what about the receive coil, is it approximately 6.2uH (23 ohms)? These numbers don't have to be very precise but ballpark figures would sure help. I can't find any information on the search coil in the schematic.

[Merc]

Is it mathematically determining the ratio of the absolute values of two samples?

[Merc]

I would think the same thing could be accomplished with 2 or 3 opamps. There may also be a newer chip available that could be used but probably not a drop-in replacement. Still, I'm trying to understand how the ID is working. I thought ID was accomplished by comparing the absolute value of 2 samples. I've been reading through the Tesoro patents on this site trying to understand. I think I'm just missing something. I'll read through again.

Merc

[porkluvr]

Hello Merc.
I thought that RC4200 was unavailable but I found a dealer on ebay selling them. $72.00 apiece.
I refuse to help put that guys kids through school, so, Do you have any thoughts on what could be used in place of RC4200?

I think that the idea around using a divider (pin-4 output = pin-1*pin-8 /pin-5) is that if you divide 'X' by 'Y' you arrive at the tangent of the phase angle. You get high resolution for small amounts of phase shift as you approach tan90 degrees. I have not really analysed the circuit to see if that is what it is doing but it seems plausible.

If on the other hand Y is being divided by X, then the cotangent would give you high resolution as zero degrees is approached.

My math is really rusty and was never that good to begin with so it's a big maybe.