new tgsl sensitivity problem

[windaube]

hey

the beginning of the discussion is here, this topic is dedicated only to the construction of the tgsl
https://www.geotech1.com/forums/show...-step-building
I therefore open this new topic for problem research


i built a tgsl v3.0, but it is really not responsive
I made the coil with the tgsl101 manual, the TX coil is at 14.5 khz and the rx coil is at 13.55 when it is connected to TX
the nulling was done with pin 7 of the U101

I tried the tgsl mod with -5v on the sensi pot, but the tgsl refuses to speak when I turn the sensi fully up

to better understand here is a video of the problem
https://www.youtube.com/watch?v=iwme...ature=youtu.be

I think my problem is towards U106, when I pass a 1euro coin at 20cm I have a response on pin6 U105 and Pin6 U104 but the output of the u107 does not switch.
if I spend the 1euro coins at 5 cm, I have a "longer" reponce and the output of U107 switches

I took care to test all the points of the tgsl101 manual

[windaube]

next...

[waltr]

A few of the scope traces do not look right.
Biggest is TR4 gate. This should be more of a square like in the TGSL 101 doc.

The other traces in the
GB and DISC also have bad shapes.

Going back over your scope pictures I see the TX output is not a clean Sinusoidal but has a strange 'tail' on the lower part of the wave. So something does not seem right in the TX oscillator.


Are you sure the PCB is correct and matches the EDU schematic?
Sure you have all the correct part values?
No Solder shorts between pads?
All soldering good with no cold joints?
Did you clean all the soldering flux off the board?

[windaube]

A few of the scope traces do not look right.
Biggest is TR4 gate. This should be more of a square like in the TGSL 101 doc.

The other traces in the
GB and DISC also have bad shapes.

Going back over your scope pictures I see the TX output is not a clean Sinusoidal but has a strange 'tail' on the lower part of the wave. So something does not seem right in the TX oscillator.

I think my oscilloscope is joking ..... I added a resistance at the input of my osciloscope because at 1V div it bugs
the test points seem much better, I have a doubt about the output of the U101, I don't know if it is a measurement problem but when I nulling the coils the wave is not beautiful

Are you sure the PCB is correct and matches the EDU schematic?

yes I checked twice

Sure you have all the correct part values?

correct part values YES, are an original or a copy, I don't know

No Solder shorts between pads?
All soldering good with no cold joints?

I do not check the pcb but normally it works

Did you clean all the soldering flux off the board?

I just cleaned it but without improvement


I replaced the LM308 with 2 TP072 (adaptation of wiring) the sound seems better, it "cuts" less

[waltr]

Those new scope pictures do look much better.

The TL072 op-amps are very nearly as good as the TL082 and are 'drop in' replacement for the LM308.

[windaube]

I would like to replace all the IC with a "serious" supplier (farnell)

for TL081 I find several version : TL081BCP, TL081CP, TL081IP which one to take?
For the lm358 I can't find an LM358N but an LM358P, it works the same?
I found some LF353N and LM393N

does that sound correct?

I am also looking for an equivalent to j107 and bf245c, an idea?

[waltr]

I use the TL081CP in my TGSL.
https://www.ti.com/lit/ds/symlink/tl...=1596331629893

P = plastic package
C = 0 to 70C Commercial Temperature range
I = -40 to 85C Industrial Temperature range

B has 2mV instead of 3mV Input Voltage offset.
CP ACP & IP have same specs as far as I see in the DS.


I used BF256B for the FET switches (TR4, TR5).

Either LM358N or P work.

I also used the LF353N and LM393N.

I buy from either DigiKey or Mouser so I am ensured of getting 'real' parts.

Mouser does ship to Europe so if Farnel doesn't have what you need check them.

[windaube]

thank you for this information

the problem with Mouser, shipping costs, 20 ? minimum for europe...

I took everything in texas instrument (farnell)
the lm353n and the lm393n have "nopb" in the reference I don't know if it matters ?

if my oscilloscope is correct, it proves that I have a problem directly at the output of pre amp (pin 7 U101)
I hope the change of IC will solve my problem

[waltr]

Yes, shipping from US would cost.

No Pb (lead) is not a problem. Almost all parts now are no Pb to meet RoHS standards.

The U101 output does look wrong.
Did you ground one side of the RX coil or leave the RX coil un-grounded. I left mine un-grounded.

Did you shield the coils yet? If so are the shields insulated from each other and each have a gap (no shorted turn).

[windaube]

I realized an error, for C18 and C21 I think put 47nF instead of 470nf ... now it sounds better, I almost have a stutter as to the maximum sensitivity.

the farnell command is also realized
for the moment the coils are just protected with tape, without shielding (I thought my problem came from here)
the cable is connected to the gnd only pcb side

for the shielding I don't know what to understand

-I saw some use of graphide in the hull and on the coils
-I saw some use of graphide only on the coils
-I saw some use a shield only on TX
what to choose ? I am thinking of making the two graphite coils

I remember that the aluminum foil was bad, I am thinking of using graphite powder and spray varnish (or glue)


do you have a technical to adjust C6? I read that this was important to create a 20 degree phase shift
I went from 15nf to 19nf only by looking at the maximum wear, I would prefer a visual method

[waltr]

47n instead of 470 would hurt.

The Phase adjustment , coil over lap, C6 or C2 adjustment requires a Two channel O'scope. No way around it.

I'm sure I posted in this TGSL category how I did Graphite shielding.
Really did the same as the Russian guy's Video.
https://www.youtube.com/watch?v=XnrStRklcUQ

I mixed very fine graphite power with polyurethane (clear for wood finishing) to coat the DD coils. With electrical tape under and over the graphite. What ever you use mix some small batches and paint on a piece of card board for resistance testing.
Same graphite mix is used to coat the inside of the coil shell for the 8 inch Concentric. Just like DBowers shows.

[windaube]

The Phase adjustment , coil over lap, C6 or C2 adjustment requires a Two channel O'scope. No way around it.

perfect ! I found a used real oscop

I'm sure I posted in this TGSL category how I did Graphite shielding.
Really did the same as the Russian guy's Video.
https://www.youtube.com/watch?v=XnrStRklcUQ

the video is very well explained, thank you

Just like DBowers shows.

What is the DBowers shows?


I replaced all the ICs and j107 by bf256B ,now I have a very very high sensitivity, especially the cable! just move next to the cable and the tgsl rings
it is now impossible to exceed the 60% sensitivity ( sensi pot )
without shielding, the tgsl rings with 1euro at 20cm,I hope to reach 30cm with the shielding

at the moment i'm trying to modify C6,I added 2X2.2nf parallel to the 15nf
I have a phase shift of 8-9us, I read that to have 20degres of phase shift it is necessary to have a phase shift of 11us
I will continue to look for the perfect value of c6

now i have another problem,now the discri does not work anymore
same metal the discri "cut" always at the same point

I looked at tr4 and tr5, now the discri pot and the GB pot cause a phase shift, the "pulse time" does not change, just signal shift

[waltr]

What is the DBowers shows?

In his building a concentric coil document:
https://sites.google.com/site/dbcoil...home/resources

Well the scope traces of signals look much better with your new scope.
Yes, 50-60% sensitivity is max with the -5V mod. I typically run about 40% in the field else it continuously chatters.

20 degrees is about 4us = (1/14500) * (20/360)
Looks like you are near enough to that with 8us.

There is timing changes with GB & Disc at min & max. Compare with scope pics in the TGSL 101.
The phase shift of the TR4 &TR5 pulses is what is important not the width.
For Disc make sure the switch is in Disc mode not All Metal mode.

Is your coil still unshielded and laying on a table?
Also be sure there is NO METAL within a meter or two of the coil and you remove rings, watch, etc.

Sounds like you are getting nearer to having this work.

[windaube]

the video says 11us
the tgsl 101 says: 14.5khz tx and 16.12khz off resonance
14.5khz = 69us period
16.12khz = 62us period
69-62= 7us phase shift

the coils are still not shielded, I don't know if the discri is influencing ?
which is weird, the cable seems to be detected and very sensitive

[waltr]

the video says 11us
the tgsl 101 says: 14.5khz tx and 16.12khz off resonance
14.5khz = 69us period
16.12khz = 62us period
69-62= 7us phase shift

the coils are still not shielded, I don't know if the discri is influencing ?
which is weird, the cable seems to be detected and very sensitive

Yes, RX is resonant at 16.12kHz but you will never see this since it is the TX frequency the RX coil picks up.
So you see 14.5kHz on RX circuit. 20 degrees of 14.5kHz = (1/14500) *(20/360) = 4us

Carefully watch the phase difference between TX and TX when adjusting coil Null.
The phase shifts as the coils go through the null.
What you want is the RX signal lagging the TX by about 20 degrees when the RX signal is near minimum (null).

This is a VERY sensitive adjustment, a faction of a mm movement makes a difference.
I never got this quite right on my first DD coil. On the second one I built I did put in an adjustment screw like in the Vedio I posted a link to.
Even then, after the poured epoxy set, the slight shrinking pulled the coils out of Null and changed the phase shift.
I ground out the epoxy with a thin grinder wheel and readjusted null & phase then re-epoxied.