TGSL Tuning & troubleshooting

[sdesic]

>When making a circuit change, keep checking those voltages that were wrong so we make sure you fix the problem.
OK, removed again C12 220nF and now measuring
on 393 U102 pin 7 -3V
on 358 U103 pin 1 is now -6V
pin 3 is now -6V
pin 5 is now -2.7V
pin 7 is now -2.7V

>I would check C12 for short circuit. Then look for solder bridges. Then suspect bad chip or part.
replaced with another but I have that -6V on pin7 I removed it ...

>I don't think a bad C11 (15pF) would drag down TR5 or the LM393 pin 7 to -6V.
OK

it is not C12 that is short circuit

[sdesic]

replaced 220nF with other one and that -6V comes back

[sdesic]

placed back ceramic capacitor 15pF and it still do not beep
I have tried maximum sens and no beeps or other bad sounds
detects on 2-3 cm
I have to replace that 220nF capacitor and then we will see
nulling will be fine I think

[simonbaker]

removed 15pF capacitor connecting to disc switch
and 220nF C12
I have detection but very low ... no beeping ... I think it is not ceramic 15p

removed 15pF capacitor connecting to disc switch
and 220nF C12
I have detection but very low ... no beeping ... I think it is not ceramic 15p

When making a circuit change, keep checking those voltages that were wrong so we make sure you fix the problem.

>> I was wrong with this so edited: [I would check C12 for short circuit. Then look for solder bridges. Then suspect bad chip or part.]

I don't think a bad C11 (15pF) would drag down TR5 or the LM393 pin 7 to -6V.

Actually, I'm wrong about checking C12 -- that would not make a -6V measurement (no harm in checking anyway -- remove power, short the capacitor with a wire, then use ohm meter for rough test).

1. Problem could be negative power supply getting in somewhere. Possibly solder bridge, or bad LM358.

2. But another possibility is if your coils are not nulled and you have a very large null signal at LF353 pin 7 (16 V p-p for example) and your phase is such that the ground channel is in sync with negative going RX cycle, you might get a large negative voltage on C12 and pin 5 of U103a, as you observed.

3. Can you measure the voltage at LF353 pin 7 and show it on same scale as TX signal so we can see how large it is?

4. Also, you can put your voltmeter on C12 or pin 5 of U103a (LM358 ) and then null your coils and see if you can make the voltage near zero?


-SB

[sdesic]

removed 15pF capacitor connecting to disc switch
and 220nF C12
I have detection but very low ... no beeping ... I think it is not ceramic 15p

[simonbaker]

I have that same -5 to -6V on LM393 pin 7 and all pins of TR5
on LM393 pin5 is 0V

Ok, we have found a problem there. Time to start looking for reasons.

-SB

[sdesic]

I have that same -5 to -6V on LM393 pin 7 and all pins of TR5
on LM393 pin5 is 0V

[sdesic]

all metal mode - while beeping
DC voltage on LM358 (pin 3) 1.075V
DC voltage on LM358 (pin 1) 1.076V

same IC lower side
DC voltage on LM358 (pin 5) -5.90V
DC voltage on LM358 (pin 7) -5.99V

I will check for solder bridge ... this kind of PCB prevents almost all bridges... except inside of IC socket and soldering on IC sockets

[simonbaker]

Simon,

I will gave scope in this week so we can check it again...
one channel only

One channel scope is almost useless for most metal detector troubleshooting, since phase relationship between signals is so important to many functions. But it will tell us the shape of signals, which can be useful to see distortions.

If the scope has an external sync probe, it can be used for phase, but with great difficulty.

-SB

[simonbaker]

I must do this next ... how about that -6V I measured on LM358 U103a (pin 5 and 7) ? It that OK ?

No, that is not OK at all -- very wrong -- but first make sure you are measuring the correct pins on the correct IC chip. It is easy to get mixed up.

The voltage should be between - 1 V and 1 V. Typically about -.2 V for example.

If you are sure the voltage is -6V, then look for some kind of short to the negative rail, maybe a solder bridge or bad capacitor C12 for example.

Regards,

-SB

[sdesic]

porkluvr,
I have made screenshots of Disc mode TR4 and TR5 separately vs pin1 4024 (green on both)

[porkluvr]

porkluvr,

in my last post with graphs is TR4 vs TR5
www.geotech1.com/forums/showthread.php?16122-TGSL-Tuning-amp-troubleshooting&p=173608#post173608

audiocard scope does this kind of graph...
I am going to have scope for use this next week so I can do another
it is one channel oscilloscope

Those gate signals look like severely anemic. But, they are consistent and I am unsure about the capability of your scope. Maybe what I should have asked more directly is: (in DISC mode) how does either gate signal look in a side-by-side comparison to the signal at CD4024-1?

[sdesic]

I think what may be the problem for sdesic is that the Silver Dog schematic has an extra-high gain amplifier compared to standard TGSL. That can make instabilities more likely, even involving local EMI noise.

Maybe someone has already suggested this, but assuming the circuit is built correctly (in spite of the puzzling TR4 gate signal phase), I would recommend the following:

1. Change feedback resistors R29 and R31 from 1 Meg to 470K. If that is difficult, consider piggy-backing an additional 1 Meg resistor in parallel with each resistor on the circuit board (avoids unsoldering).

2. Replace capacitors C14a,b and C17a,b with single 4.7u bipolar capacitors. If bipolar are difficult to obtain, use back-to-back 10u capacitors, carefully observing the polarity in the schematic.

This will slightly reduce the gain and increase the center frequency of the filter section, same as the standard TGSL. It should be a more stable MD.

3. Take the circuit and coils outside, far away from power lines and computers. Use an AC voltmeter to null the coils by minimizing the AC voltage at LF353 pin 7. Maybe choose point slightly on either side of minimum.

Let us know if it still beeps.

Normal alkaline batteries should be fine -- you shouldn't need big lead-acid batteries.

-SB

I must do this next ... how about that -6V I measured on LM358 U103a (pin 5 and 7) ? It that OK ?

1. is OK
2. don't have bipolar - will try to find two the same 10uF 25V
3. when I finish first two

Alcaline are fine but I am going to have belt with batteries ... so no problem with weight
less weight to swing arround

[sdesic]

When you attach the leads via the potetial dividers, you also need a ground connection - or screen gound - this will take the huge mains hum away that is modulating the traces
S

How can I do that ... two channels are connected to PC line in of audio card ... both channels GND are connected to TX GND
PC is making so much noise in this mangnification.
I was trying connecting capacitor in signal serial to resistor but I don't know if it changes graph ... AD... DC ... something

maybe this kind of probe ...

[sdesic]

Thanks, your current looks perfect. We'll have to ponder what causes the 50s delay, but perhaps not relevant if this beeping problem is very borderline and affected by very small changes.

Actually, I will guess that the 50s delay is related to capacitors C15 and C12 reaching steady state.

With a DC voltmeter, it might be interesting to check the DC voltage on LM358 U103b non-inverting input (pin 3) and also the output (pin 1). I believe they should be about equal after steady state operation is achieved. And do the same for U103a (GB channel).

But first we must understand you scope pictures of TX signal vs. gate of TR4 -- something not right.

Regards,

-SB

Simon,

I will gave scope in this week so we can check it again...
one channel only

all metal mode - while beeping
DC voltage on LM358 U103b non-inverting input (pin 3) 0.834V (voltemeter stops beeping)
DC voltage on LM358 U103b output (pin 1) 0.827V

DC voltage on LM358 U103a (pin 5) -5.91V
DC voltage on LM358 U103a (pin 7) -6.01V