TESORO GOLDEN SABRE

[Majster]

Little can go wrong with the circuit from tx to gates of switchers.

How about dropping the frequency to around 5khz. At this lower frequency, VA will work fine. You will be able to render the waveforms at the switches gates through the soundcard?

If a 100n cap is put across the tx coil.

All you want is to confirm that the phase shifters are working.

Welcome back,
I look at the diagram and see that you are right. All mode signal bypasses U101b, only goes through U102b. Diss mode signal goes through U101b and U102a, could that be the cause of my trouble? I think so. These are obvious things to you, but I don't know how to read diagrams yet. Thank you for your attention, it prompted me to re-examine the pattern.
I think I will have to buy some better IC somewhere (LM393 and LF353).


I am not going to lower the frequency for now, but I understand what this could do and how 100n could do that. I could probably replace the C1 and C2 pair with a larger capacity and that would lower the TX frequency. That would not do the trick, as I would probably have to lower the RX frequency similarly to read it reasonably. And I know this would only be done temporarily, just for checking U101 and U102. But for now I'll hold on to it, we'll see what happens when this Hantek arrives.


I understand your initiative and I am grateful for it, because it helps me, but I am not able to do everything without thinking, I have to be patient, this is what the lesson requires. I am not as fast or as proficient in electronics as the specialists here. I'm just learning. I've already learned a lot from you guys.


Do you have any way to get these safe IC measurements in the socket? Or maybe you know some tricks in choosing real / genuine ICs? 'Cause I'm not good at it.
Greetings.

[dbanner]

Little can go wrong with the circuit from tx to gates of switchers.

How about dropping the frequency to around 5khz. At this lower frequency, VA will work fine. You will be able to render the waveforms at the switches gates through the soundcard?

If a 100n cap is put across the tx coil.

All you want is to confirm that the phase shifters are working.

[Majster]

All that is left is U101b. You are right about VA, it is limited. But if you deselect D/A you might be able to discern some kind of waveform at the output of comparator. All you really want to know is that it is spitting out something.

Do you have suspicion on 353 chip?(101b)?

Welcome back,
I thought that when the IC is on fire, it is all right at once, i.e. both channels. Is there a chance that one channel of the amplifier will burn out and the other will remain operational? If so, maybe my LF353 (U101a / b) is damaged too. I didn't suspect it of this before, but on the other hand, if the LM393 was relatively working (maybe not perfect, but still) then the cause of my trouble is with the LF353? Is it possible?


I am not measuring anything yet. I decided to buy this Hantek 6022 oscilloscope. I just stated that I would need such a device someday, so there was no point in waiting for it. Sensible measurements of my PCB only make sense when this oscilloscope reaches me. As soon as I measure something - you'll be the first to know about it.
Do you know any way to safely connect to the IC with measurement electrode during work? Any needle? I do not want to solder a few wires from the bottom of the PCB, and by adding electrodes from above there is a risk that I will short-circuit something and burn it.
Greetings.

[dbanner]

All that is left is U101b. You are right about VA, it is limited. But if you deselect D/A you might be able to discern some kind of waveform at the output of comparator. All you really want to know is that it is spitting out something.

Do you have suspicion on 353 chip?(101b)?

[Majster]

You MUST see rectangular waveforms. There is no escape from this.

Even opamp like a TL072 in this position should give rectangular waveforms. You can try. They have same pinout. So you can just stick one in.

Welcome back,
I did this: I inserted the previously used LM393 into the IC102 socket. It turned out that the work of the plate returned to normal. Unfortunately, its "normal" is to work in All mode, and not in Diss mode. So it works like it did a few days ago, before I didn't put different LM393s from different manufacturers into the PCB. I don't think anything else has burned. As far as I know, the FETs (BF245c) are quite resistant to being burned by the gate, because the FET junction is mainly controlled by voltage, not current (similar to MOSFET). On the other hand, before the IC102, there is probably nothing that could burn, except the LF353. But if the All mode works then the LF353 doesn't seem to burn out. Do I deduce well? I hope so.
Additionally, I just understood something else:
1. My waveforms are incorrect. I use a computer adapter (simple resistance bridges diode protection through the sound card, and Visual Analyser software), so the maximum sampling will be as described in the specification of this external sound card. Most often it is 48kHz or less. And it does not depend on the program, but on the hardware capabilities of the sound card. This means that describing the full period of the frequency wave around 14.5kHz takes only 3 points. This is definitely not enough in my opinion. The program that supports the input of the sound card (VA, or other) creates then not a real waveform, but a joyful, wonderful interpolation of what it sees on these three points, which is usually a sine, because that is the easiest way to interpolate. Even if there is a rectangle at the entrance. Such adapter is suitable for impulse detectors, because they work below 1kHz (but carefully anyway, because impulse detectors generate a sharp signal of several dozen volts or even more). Then I have to buy a real oscilloscope. These are rather expensive devices. I do not want to buy a used one, because there can be problems with that, and I know too little about the subject for me to recognize failures of the used apparatus. Hantek 6022B will hold my pocket. What do you think about such an oscilloscope? If I don't burn it (unfortunately, with my talents, it's possible) then it should be good, right? It is dual-channel up to 20MHz, which means that the entire period of the ~ 14.5kHz waveform will be very densely sampled.
2. I read in the Hantek 6022B specification that it supports a 0-5V measurement range. If so, can I use it to test the TGSL PCB? Will I have to do any additional protection for the oscilloscope input? Resistance bridges?
3. I don't need to desolder T4, because in my PCB its gate goes through a three-position switch (Left, 0, Right). Just set it to 0 and connect the oscilloscope electrodes to L or R, then I will see the T4 input waveform in Diss or All mode when T4 is off. But it only makes sense when I buy this real oscilloscope.
4. The waveform on the T5 gate should be very similar to the waveform of the output from IC102b (7), a bit higher of course, but consistent in phase and amplitude, do I think right? But that also has to wait for the real oscilloscope.
5. I have a TL072 at home, and even a TL082 here, but I don't think I should have connected them without an oscilloscope. I suspect so, am I right?
6. All my LM393 are tested and rejected, but maybe I could find the LM386 here, would that also be suitable?
7. Do you know any safe way to record waveforms on IC pins without removing them from the socket? Any needles? Any sharp nails?


Thank you very much for your help and commitment. Without your help, I would have stood still.


Waltr - you are probably a professional and I am an amateur. I'm just learning. I do not always understand your tips that you give very professionally. I know how it works. In my job / profession (not electronics, but also a very technical profession) I am a professional and it is often difficult for me to understand that certain issues, very easy for me, are difficult for my students. Then I try to build such a simple model and describe everything as simply as I can, build examples so that they are easy to understand and similar to the problem in question. Then I give instructions so detailed, that even a child can understand. I am not afraid to say to them: take this in the left hand, take that in the right hand, turn on the green button, wait 2 minutes, connect the tube here and close this valve, etc. Why? Because that and that, etc. I know it works. But I also know that here I am a student and I do not have teachers from my dreams, only those who know something very well and talk to me with the hope that I will understand it, because something is easy for them. I think you know what I mean. I am so happy when anyone helps me.


KT315 - thank you very much for your suggestions. I looked at your pictures of Your IC tester, it's very interesting, but I won't be able to make it. PCB in one image does not match what I see in the photos (different number of IC sockets). I suspect it's a few different devices, maybe similar but not identical. I will not be able to understand it. If I had the actual PCB image and instruction manual, I would do it tomorrow, or when I get the necessary components.

Greetings.

[kt315]

look my album https://www.geotech1.com/forums/album.php?albumid=109
look my video https://www.youtube.com/watch?v=3S8JT0S0mmc

[kt315]

I think your problem is very simple. There is some error in the circuit of the lm393.
I suggest re-evaluating the PCB layout and components around this area. Check every component.
LM 393 getting hot is very bad sign. Now you describe continuous beep and overheating.

Check every component. the op amp and comparator tester/pcb was designed especialy for geotechers in hardest same cases.
if you dit not see i may give a link.

[dbanner]

You MUST see rectangular waveforms. There is no escape from this.

Even opamp like a TL072 in this position should give rectangular waveforms. You can try. They have same pinout. So you can just stick one in.

[dbanner]

If you cannot find anything wrong with your PCB layout and components are good, then you must check for correct waveform on gate of tr4.

Remove tr4 and tr5 from PCB. Or just desolder gate legs and lift them out of PCB.

Put probe and see what waveforms are present on output legs of LM393.

[dbanner]

I just realized that something else could burn on the PCB besides this unfortunate LM393. If so, then I won't find what it might be at all. I'm worried about the progress of my work, but I can't do enough. I don't think I will finish this project, it is too difficult for me. It's a pity because I made 2 PCBs and I think I have properly wound and tuned coils, it's a shame about it all.
Is there any simpler detector with discrimination? Obviously not BFO, something with better sensitivity. Could you recommend something with good installation and commissioning documentation? I would be very grateful. I am grateful anyway, because you help me a lot, but I must be for a weak student.
Greetings.

I think your problem is very simple. There is some error in the circuit of the lm393.
I suggest re-evaluating the PCB layout and components around this area. Check every component.
LM 393 getting hot is very bad sign. Now you describe continuous beep and overheating.

[waltr]

Sorry to hear all the trouble with this circuit.
The TGSL is about the simplest one with discrimination.
It is the TGS without the notch circuits.

The IDX is about the same complexity wise.

[Majster]

Yes, tr4 scope picture is not right. This is why it does not discriminate.
Refer to the TGSL 101.pdf for what the Gate of TR4 should look like.

I just realized that something else could burn on the PCB besides this unfortunate LM393. If so, then I won't find what it might be at all. I'm worried about the progress of my work, but I can't do enough. I don't think I will finish this project, it is too difficult for me. It's a pity because I made 2 PCBs and I think I have properly wound and tuned coils, it's a shame about it all.
Is there any simpler detector with discrimination? Obviously not BFO, something with better sensitivity. Could you recommend something with good installation and commissioning documentation? I would be very grateful. I am grateful anyway, because you help me a lot, but I must be for a weak student.
Greetings.

[Majster]

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.

Welcome back,
I did this: I searched for all LM393 in my stocks and it turned out that there are 3 populations of different designations, maybe that's what the manufacturers mean. Of course, there is LM393 on all of them, but the extra numbers / letters / symbols are different. I chose the first pair other than the ones I had in PCB. I took out the previously used ones (both) from the PCB and put this pair in positions IC102 and IC106. I turned the chip on and tried to set the coils to the minimum AC voltage between IC101 / 7 and ground. It turned out much worse than before, just slightly below 1V AC. I turned on the diss mode. Nothing has changed, discrimination still does not work. I didn't even record an oscillogram on the T4 gate. I found another other pair of LM393. I put in both instead of the previous ones. I repeated the setting of the coils as above. It turned out worse again. I put Diss Mode on and a long continuous beep came from the speaker. I'm back in All mode. Silence, regardless of the presence of metal / ferrite above the coils. I checked IC102 - very hot, I assume it got burnt but why? The previous ones did not heat up. For sure I inserted them correctly and I did not confuse pin 1 with 5. I did not short-circuit anything, I did not change anything on the PCB, nothing accidental is on the PCB or underneath. I took out the IC102, now I have an empty slot. I don't know what to check.
For the T4 and T5 positions I use the BF245c. I checked the legs, in my opinion they are correct, the gate is on the side, not in the middle. Behind the gate is the source, the last drain.
You proposed some way to test the LM393, but I don't understand your tips, I don't even know how I could connect the multimeter electrodes to this LM393. Even if I can check, and even if all my LM393s are fake, I still won't be able to buy the real ones.
This is an optimistic scenario. Because if I buy another LM393 and luckily find a real one, it may be discriminatory. But I don't know how to buy such a real LM393.
The pessimistic thing is, however, that some pair of my LM393 are correct, and the cause of my failure lies elsewhere, and I cannot locate it.
Greetings.

[SkyeHi]

hi cristiano, i have two of these i bought a kit of parts from ebay.............the only downside is you must use a good quality battery and keep an eye on it, the inbuilt battery volt meter tells lies, it will say 8.4v but really be 6.9v so just use a duracell or energizer and disconnect it when not in use.
low power makes the meter readings wrong.
[ATTACH]56476[/ATTACH]

ahhh ha, thanks for that sinclair.
Late to the party as usual but I also have one of these kits, just finished winding a coil and expecting 415/465uH, instead I got 8mH. I was blaming the weave geometry, the thin ply former I used instead of the acrylic one that snapped. Snow showers on a gale of wind atm, as soon as it eases I'll be out in the workshop to check that bloody battery.
Thank you all, great site this.

[waltr]

Yes, tr4 scope picture is not right. This is why it does not discriminate.
Refer to the TGSL 101.pdf for what the Gate of TR4 should look like.