TESORO GOLDEN SABRE

[dbanner]

On separate note, I discovered easy way to distinguish real NE5534 chip from fake.
Their are anti parallel diodes across the inputs.
Unique to this chip.
A multi meter on diode check range will confirm this if you have the genuine chip.
If not, you are sure to put in trash bin.

[dbanner]

Perhaps you can test.
Looking at internal schematic for lm393.
If you put multimeter on diode check, you should see voltage drop of base/ collector junction between either input pin and pin 4 of comparator chip lm393.( Remove chip from socket)
A regular op amp will not show up like this.

[dbanner]

You might be suspicious that your comparator lm393 chip is not comparator chip but is really a normal op amp masquerading.

[dbanner]

Check the legs of jfets for correct pinout.
Which jfets are you using for tr4 and tr5?

Oscillogram for gate of tr4 does not look right.

You could remove centre wire from switch and take oscillogram of output of pin 1 of comparator 102a.
Check your comparator circuits.
You must have proper switching waveforms for the jfets.

[Majster]

Check all metal/discrimination switch is wired correctly.
Check for waveform at gate of tr4 when switch is set for discrimination.

On a different note, many of opamp chip on the market are fake, especially from eBay.
Make sure you have genuine opamps.

Welcome back,
I checked all metal/discrimination switch: they are wired correctly.
I did this: I finished the second PCB (PCB2). I applied changes to it:
both 2N2907 → BC557
both 2N2222 → BC547
both LM308 → TL071
I chose all capacitors and resistors so that they were as close to the measurement value as possible to the declared value. So that they should be equal to the description in the diagram, not just close. The differences from the declared one do not exceed 1%.
I got the values of the supply voltages + 8.0V and -5.5V. On PCB1 it was +8.0 and -7.0V.
I tuned both coils to this board according to your directions, which is exactly as before, but I left 14.2kHz on TX (i.e. C1 = 220nF, C2 = 22nF, no correction capacitance) and I adjusted RX to 15.8kHz for a shift of 1.6 kHz (it required 220pF soldering parallel to C6, so now C6 '= 15nF + 0.22nF = 15.22nF).
I connected both coils and tried to get the minimum AC voltage between IC101 / 7 and ground. About 0.38V AC was obtained. It is not possible to receive 5mV AC. I assume that with this minimum value, the voltage at the RX connectors is also minimal, I did not check it. I checked how detection works in the All mode. It works, maybe not great, rather weak, but it works.
I switched to discrimination mode. PCB2 behaves almost exactly like PCB1. There is no possibility of obtaining discrimination.
I went back to PCB1.
I tried to set the required AC values (as described above). With the minimum AC voltage value between IC101 / 7 and ground (about 0.45V AC) I recorded the waveform / oscillogram on the TR4 gate. File in attachment.
I know that many of the elements on the market are non-original, but I cannot tell them apart. I try to buy from reliable suppliers, but it doesn't always work. You can check the transistors by yourself, but I can't check the integrated circuits myself.
Greetings.

[dbanner]

Check all metal/discrimination switch is wired correctly.
Check for waveform at gate of tr4 when switch is set for discrimination.

On a different note, many of opamp chip on the market are fake, especially from eBay.
Make sure you have genuine opamps.

[Majster]

Welcome back,
I have just tried to check how GB and discrimination work. And unfortunately there are problems again. At the beginning I switched to All mode. I set all the potentiometers to the middle of the scale. I found a position for the coils that gives the minimum AC voltage between ground and IC101 / 7. Today I was able to get 0.4V AC. Then I started turning the GB potentiometer until it rejects ferrite. The documentation says so. The position of the potentiometer at which the ferrite is rejected is basically at the beginning of the scale (left). When I turn the knob clockwise, the ferrite is still visible. Other metals too (copper, aluminum, steel) are visible. So I left the GB potentiometer at the beginning of the scale on the left. Detection may not be the best (15-20cm), but acceptable at this stage of construction. I switched to Diss mode. Detection has ended. Regardless of the position of the Diss and GB potentiometer and the sensitivity on the PCB, there is no detection. Everything is rejected, every metal. When I go back to All mode, every metal gives a signal again. When I go back to Diss mode, all metal disappears. There is no combination of potentiometers where some metal gives a signal. I'm worried. What could have happened? What should I check now?
Greetings.

[Majster]

Welcome back,
The sound is present from the beginning of the construction. Reacts to any metal in All mode. I tried to test the discrimination when the coils were not yet well corrected with capacitors. It worked so that it was rejecting ferrite by then, and sometimes I thought it was rejecting other metals as well, or all of them. You know how it is: when you are not sure about the reaction, you imagine it, and the more you want to see it, the more you imagine it. But in such cases, I prefer to remain skeptical and not rely on my imagination.
Looking at my waveform, the RX seems to be ahead of the TX, but I don't know if it's good or bad. Should it be ahead or delay? How is this done? How to cary about it?
I plan to attempt GB tuning and discrimination this afternoon. I just preferred to make sure it made sense beforehand, because my previous attempts did not satisfy me.
Yesterday I chose elements so that their values were relatively constant and equal, by the way I calibrated the second capacitance meter, now I have 2.
And what about these elements? Can any more be changed?
Greetings.

[dbanner]

Welcome back,
Thanks for the clarification, you're really good at this, really. I'm just learning and I can't judge my oscillogram like that, but I guess it's not bad. Besides, it is not a real oscilloscope, but a computer adapter. I can't expect her to work better.
But let me tell you something: with my PCB, I visited a friend who has an old Russian oscilloscope, still analog (I can't find what model it was on the network) and not fully functional (no electrodes and problems with switches). And when I connected the PCB to it, for the TX waveform it showed exactly the same sine wave distortion as in the Don Bowers documentation.
What do you think - can I already think about GB and discrimination at this stage?
Greetings.

Yes.

What about the audio? Does it respond to metal object.

Also notice as you move Rx coil, the signal amplitude dips to minimum and then rises again as you emerge on the other side of the null point.
So you want to be 20 degrees, but how do you know which side?
20 degrees leading or lagging? This is where you look at pictures of dbowers.

Also as I noted previously, so long as you can reject ferrite within the range of the gb trimmer and get good distance and discrimination, you can be assured that it's going to be pretty close to the 20 degrees required for this circuit. If you have 19 or 21 doesn't matter.

[Majster]

Welcome back,
Thanks for the clarification, you're really good at this, really. I'm just learning and I can't judge my oscillogram like that, but I guess it's not bad. Besides, it is not a real oscilloscope, but a computer adapter. I can't expect her to work better.
But let me tell you something: with my PCB, I visited a friend who has an old Russian oscilloscope, still analog (I can't find what model it was on the network) and not fully functional (no electrodes and problems with switches). And when I connected the PCB to it, for the TX waveform it showed exactly the same sine wave distortion as in the Don Bowers documentation.
What do you think - can I already think about GB and discrimination at this stage?
Greetings.

[waltr]

The Phase is the time difference of when the peak (or Zero crossing) is of the RX relative to the TX wave forms.
14.5kHz has a period of 68.9micro-seconds.
20 degrees divided by 360 degrees = 0.055 of a rotation. Multiply by period equals 3.8 microsecond time difference between TX and RX.

Look at the TGSL 101.pdf again, Scope picture #4 on page 6. Top trace is TX, bottom trace is RX. Note that the RX peaks are slightly (20degree) after the TX peaks. Also note the scope vertical scale and then read the peak-peak voltage in the picture.

[Majster]

Welcome back,
I did as you advised: I turned on the power, set the coils so that the voltage value between ground and IC101 / 7 was minimal. I was able to get 0.48V AC. I turned it off and on again. Without moving the coils, I re-measured the voltage between ground and IC101 / 7, I got a repeatable result, i.e. 0.48V AC. So the position of the coils is rather stable. Then I turned off, unhooked the multimeter electrodes and turned it on again. Then I measured the AC on the TX connector and got 6.16V AC, I measured the AC on the RX connector and got 5-11mV AC, not very stable but not higher than 11mV. My multimeter is APPA 305, so it shows frequency during these measurements. It is 14.14kHz on the TX connector and 0 on the RX connector. I assume that the voltage on RX is too small for the meter to pick up the frequency correctly, besides, I saw an oscillogram showing two sinus after subtracting the dc, so that doesn't bother me.
Are the AC voltage values that I have read promising? What should I do now?
When you write about angles, I don't really understand what's going on, but I know it has to do with the TX and RX sinus waveforms. When they are not completely in phase, it means that they are moving away from each other by some angle (ahead or lagging behind) and that it depends on the independent frequencies TX and RX which differ by 1.6kHz. But I do not know exactly where or how to measure this angle, or how to regulate it. I think, that only by changing the frequency of the coils, which in turn is regulated roughly with capacitors, and more precisely with the final positioning and bending of the coils. Do I think right?


BDW I have already started the second TGSL PCB. I assumed that if I did something wrong on the first PCB, the second PCB would be comparative. I have already soldered the IC sockets and bridges. I read in the documentation that a few elements can be replaced, for example LM308 to TL071, and pnp transistors 2N2907 to BC557, npn 2N2222 to BC547. Are any other substitutions possible?
Greetings.

[dbanner]

Measure the RX residual voltage across the RX coil, before the preamp. This is true residual value of voltage.

It is not possible, I don't think, to get absolutely zero volts. There is always some small voltage even at deepest null.

But for tgsl, you are aiming for slightly offset from deepest null.

You want approximately 20 degrees phase difference between the tx waveform and RX waveform as your balance point. This is the point at which everything works correct in the circuit for ground balance trimmer and discrimination.
At 20 degrees you will have slight voltage on Rx coil greater than deepest null.

[Majster]

Welcome back,
I did this: I desoldered C6 (it was 15nF) and put 2 capacitors 10nF + 4.7nF = 14.7nF in parallel there. I connected the TX and RX coils to dedicated ports. I connected the mV AC multimeter to ground and the IC101 / 7 pin. I turned on the power. I moved the coils to get the minimum value on the multimeter. Unfortunately, the lowest value that can be obtained is about 0.5V AC, which is still 100 times too much. I connected the osciloscope electrodes to Visual Analyzer, and to the TX and RX ports. I recorded the waveforms from both coils (attachment file). Unable to get flat RX wave. Do you have any idea what I could do now?
Greetings.

[dbanner]

Welcome back,



I think I understand it now, but it took the Colpitts generator formula.
F = 1 / (2Pi * sqrt (L ((C1 * C2) / C1 + C2)))
I understand that from it you calculated the C2 correction and it is 470nF in series with C2 which gives a combined capacitance C2' = 21nF. It is true. The capacity corrected in this way corresponds to the frequency of 14.5kHz. I verified this value in online calculator.
I soldered 470nF as you said, and measured the TX frequency. It is f = 14.45kHz, but it is easy to change it with a very slight curve/bendof the TX coil and then f (TX) = 14.50kHz. This is really a very slight bend of the coil, not more than 0.5 - 1 cm. It's probably acceptable.
I didn't take anything out of the PCB, 470nF is still there. Then I connected the RX coil to the TX port (J1.1 and J1.2) and measured the frequency. It is f (RX) = 14.06kHz. Based on this data, I calculated the inductance of the RX coil. It is L (RX) = 6.68mH.
I guess it's still easy. I need to calculate C6 because RX coil will not work in TX port, but in RX port, so it will be a normal oscillator with frequency f = 1 / (2Pi * sqrt (L*C6)), and this frequency should be 16.1kHz. From here I calculated that C6 should be 14.63nF. If so, I will have to desolder C6 (currently 15nF) and insert 2 capacitors in parallel there, which will give a total of 14.63nF. The closest sum will be 10nF + 4.7nF.
Did I calculate it correctly? Do you think I understand it correctly?
Greetings.

Yes.