TGSL 2 tones ID

[hugo]

Hello,
This weekend, I've worked on my TGSL.
My goal was to add a dual tone ID on it when the MD is in All metal mode.
On the bench, it works quite fine.
To do this, I use an additionnal LM393 which drive a FET transistor to switch another output of the 4024 divider to the audio stage.
In Full metal mode, the Output 6 of U104 is disconnected from U106a. And the input 3 of U106a is shorted with the input 5. (Both inputs see the GEB signal).
The output 6 of U104 is rerouted to the input 3 of the new LM393.
The output 6 of U105 is routed to the input 5 of the new LM393.

At this time, I still have some points to adjust :
- The initial LM393 is more sensitive than my second stage. There is a distance range where the audio output is active while the FET has not switched yet.
- The outputs of the new LM393 can oscillate under some circumstances.
- And at least, I don't really understand why Tesoro has added a small delay on the GEB channel (R23 and C11).

Any help is welcome,
Hugues

[Qiaozhi]

And at least, I don't really understand why Tesoro has added a small delay on the GEB channel (R23 and C11).

Any help is welcome,
Hugues

When the detector is switched into All-Metal mode, the sample gate for the DISC channel is connected to a slightly delayed version of the in-phase sample for the GEB channel. In this case, both channels will now respond to all targets (i.e. All-Metal). An alternative method would have been to simply put a switch in series with the output of the DISC channel comparator to disable it.

[hugo]

TGSL 2 tones ID

Thank you Qiaozhi for your answer.

But my question is not why (in all metal mode) both channels see the GEB signal but why there is a small delay between them ?
Is it really useful ?

During the night (It's amazing how the brain still working even if you are sleeping !), I think I've found a better way to do the mod. I have to test it ASAP :
- Re-route the output of the LM393 (U106) to drive the FET.
- Send the GEB channel to U107b (unused) to control the audio stage (Together with a transistor because of drained output command).
- Plus a switch to select between DISC and all metal mode.

I'll let you know ...

[simonbaker]

Hello,
This weekend, I've worked on my TGSL.
My goal was to add a dual tone ID on it when the MD is in All metal mode.
On the bench, it works quite fine.
To do this, I use an additionnal LM393 which drive a FET transistor to switch another output of the 4024 divider to the audio stage.
In Full metal mode, the Output 6 of U104 is disconnected from U106a. And the input 3 of U106a is shorted with the input 5. (Both inputs see the GEB signal).
The output 6 of U104 is rerouted to the input 3 of the new LM393.
The output 6 of U105 is routed to the input 5 of the new LM393.

At this time, I still have some points to adjust :
- The initial LM393 is more sensitive than my second stage. There is a distance range where the audio output is active while the FET has not switched yet.
- The outputs of the new LM393 can oscillate under some circumstances.
- And at least, I don't really understand why Tesoro has added a small delay on the GEB channel (R23 and C11).

Any help is welcome,
Hugues

This is a nice idea, I was thinking of similar thing. Not quite as good as IGSL because you won't be able to get overlapping ranges (correct?), but worth doing. I will follow your experiments.

-SB

[dfbowers]

When the detector is switched into All-Metal mode, the sample gate for the DISC channel is connected to a slightly delayed version of the in-phase sample for the GEB channel. In this case, both channels will now respond to all targets (i.e. All-Metal). An alternative method would have been to simply put a switch in series with the output of the DISC channel comparator to disable it.

I always wondered why one is phase-delayed as well. I guess I could clip out C11 and see what happens. Maybe it's to quiet things down a bit in the AM mode?

[Qiaozhi]

But my question is not why (in all metal mode) both channels see the GEB signal but why there is a small delay between them ?
Is it really useful ?

I always wondered why one is phase-delayed as well. I guess I could clip out C11 and see what happens. Maybe it's to quiet things down a bit in the AM mode?

Although you could just as easily connect the GEB sample pulse to the DISC channel without the delay, there may be a problem (as Ivconic discovered with the IGSL) of the two channels sampling at slightly different times. If the DISC channel samples after the GEB channel, then everything would work as expected. But if it sampled earlier, then you might get some strange behaviour. To allow the design to work with a wider component tolerance, the extra delay guarantees that all targets will be accepted. Even removing the delay may not prove anything; only if that particular circuit works or not. You might want to try swapping over the connections so that the GEB channel is delayed relative to the DISC channel. This would emulate the condition I mentioned earlier.

[simonbaker]

I always wondered why one is phase-delayed as well. I guess I could clip out C11 and see what happens. Maybe it's to quiet things down a bit in the AM mode?

The last time I looked at the all-metal sync pulse, it was not only delayed, it was a completely different shaped pulse - virtually triangular.

The triangular pulse would appear to selectively weight a small phase region rather than integrating over a full half-cycle. The purpose of that eludes me, and would even seem to reduce the gain of the GB channel.

Even if the pulse were square and delayed, the only purpose I can see would be to more positively reject the "ground" signal (unless I have it backward, in which case it would include ferrite/ground as a target). Delaying it might also make silver come in stronger at the expense of iron (which is usually not a problem!); but then you would think you would want to delay both sync pulses.

A good question.

-SB

[hugo]

TGSL 2 tones

Hi,
I've done what I've said in my previous post ... and it works !
Re-use the second half of U107 to drive the audio and re-route the outputs of U106 to select the audio tone.
Voila !

[Jerry]

Now that is a very elegant solution!

Thanks for posting it.

Jerry

[simonbaker]

Hi,
I've done what I've said in my previous post ... and it works !
Re-use the second half of U107 to drive the audio and re-route the outputs of U106 to select the audio tone.
Voila !

Looks like a very nice idea! Great way to make use of that extra LM358.

-SB

[hugo]

Looks like a very nice idea! Great way to make use of that extra LM358.

-SB

Now that is a very elegant solution!

Thanks for posting it.

Jerry

Thanks. I hope it will be useful for many.

[satdaveuk]

Thanks. I hope it will be useful for many.

Nice one hugo, over the last few days i been looking at just that topic comparing itgsl and tgsl to be able to do that very same mod, but now vola you already cracked it
Matter of interest how good did your pcb look after the mod, or is it worth making another one.
Pics of your pcb after mod would be nice

Many thanks

Regards

[hugo]

@satdaveuk :

The visual impact of the mod is very small, you don't need to redraw the PCB.
In fact, every holes except three ((Pins 5 & 6 of U106b and the gate of the FET for audio) are already present on the board.
Remember the second half of the LM358 was populated !
You just have to remove the resistors around it and add the good ones.

I will post some pics asap.

Have a nice day,
Hugues

[ivconic]

Is there any chance you to record video and upload on Youtube?
I would like to see and hear behavior.
If it turns so interesting - i will make it for sure!

[hugo]

Hi,
Here are some pics of the mod.
On the top side of the PCB, the two feedback cells (100K+1N4148 ) for the comparators.
On the bottom side, the links (green wires) between the LM393 and the LM358.
And near the CD4024, the FET (2N3819) and the diode.
Pin 7 is always connected to the audio.
Pin 11 is always connected to the audio but in DISC mode together with 2 TONES mode, when the target is rejected.
The setting may be different is a Tesoro coil is used (Different frequency).
Mine is for a Minelab TS1000 coil.

Note : Take care to connect the Drain of the FET to the audio side and the Source to the CD4024.
Finally, I've connected the Gate of the FET to the ground when the switch is not in 2 tones mode (Not to the -5V).

@Ivconic :
I'll try to find a cam and post a video.