TGSL EDU Mod

[Silver Dollar]

For those that missed it here is a mod to TGSL EDU to allow a bit more gain and cleaner output.

Remove D8, D9, C20, C23, R33 and put diodes at C20 and C23 reversed from D10 and D11
and change R29 and R31 to 1 meg.

Thank Davor as he created this mod on the IGSL.

[manny77]

Hello Silver Dollar,
I made a good working TGSL last year. Very good, it can detect 1 euro coin at 30+ cm. I made some little mods to it to suit to my preferences in antichatter section of the circuit. And dfbowers mod to increase sensitivity (connecting sens pot not to ground but to negative voltage).
Only problem is chattering. The MD is sensitive to some sort of RF emi. It is very quiet at home and in several places, but in some case it tend to be noisy and i had to lower sens at a low level...

You think this mod can improve my MD? Where did it is treated? I cant't find it in the forum...

Thanks in advance!

[Silver Dollar]

Check here #1404. This was done to a IGSL which is 2 TGSL's on one board.

http://www.geotech1.com/forums/showthread.php?17861-IGSL/page57&highlight=igsl

[manny77]

Wow so much fast! Thanks! Now I check this link...

[manny77]

So this mod could improve stability of md reducing sensitivity to EMI noise? It could be a good solution for my tgsl....

[Davor]

This mod significantly improves impulse response of the filtering, and provides compression to better cope with strong signals. Target response becomes faster and snappier as the filters never get into saturation, while the difference between the shallow and the deep target becomes softer and more natural for your ears. It works wonders for many targets in a row as you can hear every each one of them.
It does not provide any EMI advantage apart from making such offending pulses last shorter.
As for EMI, TGSL has a poor design of a front end. It can be improved, but not by much. So far I did not do anything about it because I did not need any better.
I played a bit with the input configuration to reduce noise, and I can recommend a simple upgrade: replace the LF353 with NJM2068. It has 1/4 of Lf353 noise voltage and 1/10 of THD at economy class price. With NJM2068 the remaining dominant noise sources are R13 and R14, and you can stop there. To go lower with noise you can replace those with down to 1k, but in that case the phases get affected and it becomes a little more complicated - nothing to worry about if you can reach GB solution with R22. It gives some extra 5cm or so. It works for my IGSL, but I had to compensate for GB.

[manny77]

I think that shorten the bip of emi noise could be better to my ears. Thank you Davor for your other mod. Probably next thing to try on my md.
I'll keep you informed...

[manny77]

Thanks Davor!
I made these mods:

A- All Silver Dollar mods on top of thread
B- lowered frontend resistors: R13, R14 to 1K;
R15, R16 i made a parallel from original 220K and 100K (total about 68K. Yes it is less than your 100K but amplification is lower and stability is even more I think). 10 pF in parallel to R16. Not changed LF353 with NJM for now, i just ordered it.

Anyway great progress! As you said response is very clear and the sound let me understand better target size and deepness. And the beep is more fast let me sweep more fast too. No fear to lose targets anymore!
I also changed back to back electrolytics (22uF in series to 22uF = 11uF) with 10uF smd non pol. Probably another little improvement.

1 euro coin now detected from 28 cm in air. Lose some sensitivity but stability is a must!

Tested in real field for short time...enough to make me happy for the mods, but not to see EMI noise if any... Now waiting for NJM hope for some other benefits...

C- I also increased R43 from 330K to 1M. This is my experiment. It give some more long beeps for little and faint targets. I find it useful to not lost some tiny signal. Pinpointing is not compromised from this mod. Works well anyway.

Removed completely dfbowers mod (sens connected to negative and not to ground + some resistor). I think it tend to be useful only in air, with no real gain in real soil. It is better to have very silent MD that beeps only when needed and not hyper sensitive MD in wich you have to sweep more times on the same target to understand if real or only chatter.

[Davor]

I don't mind chatters if they are short and faint, but yes, you can set Sens the way you like it.
R43 is a little bit critical because it sets the offset required to overcome Vbe of Tr2. If set properly it promotes softer and longer reaction. I'm not sure what value would work best for TGSL. R44 may be connected in series with a potentiometer which provides a loudness control.
Replacing back-to-back capacitors is a good call. I replaced those with the same value non polars, which is effectively double the capacitance, but it promotes slower motion near targets which is a good thing.

I think you'll be positively surprised with NJM2068

[manny77]

I hope so Davor for NJM!

I twisted every connection from pcb to sens and disc pot and disc switch. Gnd balance pot is shielded.
Somewhere i read about cross talk between twisted cables...do you find that parallel wires like from divided flat cables in pieces of three would be an improvement?

[Davor]

It all depends on signal levels and impedances. If you have adjacent twisted pairs with 10V level in one, and 100kohm impedance and uV level in another - you are in trouble. It helps if you can separate those a cm of distance. Otherwise a twisted pair is superior to parallel wires regarding crosstalk. A coaxial microphone cable with two wires inside is a good choice.

[manny77]

Then probably the signal levels in TGSL wires are sufficiently high so twisted pairs should not be a problem... I'll leave them twisted... Thanks!

[ashefr10]

I replaced the LF353 by NJM2068 except that I was forced to hit the GEB I have not found a better power.

[Tibuck19]

For those that missed it here is a mod to TGSL EDU to allow a bit more gain and cleaner output.

Remove D8, D9, C20, C23, R33 and put diodes at C20 and C23 reversed from D10 and D11
and change R29 and R31 to 1 meg.

Thank Davor as he created this mod on the IGSL.

[ATTACH]30047[/ATTACH]

Is this Mod Good for the TGSL-EDU black board? If so, can you use the photo below and circle what needs to go and be changed please ? Thank you

[Nandor]

It increases the gain and makes the last filter stage faster, try it. Basically they are the same circuit.