TGSL Tuning & troubleshooting

[dfbowers]

Went to beach today and took 2 MDs: dfbowers TGSL and my Tesoro Compadre with 5.75 inch coil.

Literally got my 2 cents worth . But I was with people so not focused.

One interesting test. I put a Euro on top of sand and stuck a 30 cm stick in sand. Sand was in dry area, but if dig down somewhat moist.

By eye, it seemed TGSL had range of about 15 cm to 20 cm. Compadre surprisingly was almost the same, probably a little less than TGSL.

I tried to do air test of TGSL but sorry didn't bring headphone extension cord so hard to hear audio. But maybe air test got 24 to 28 cm.

It seemed beach had somewhat more noise than in mountains. When turning sensitivity full, more chatter than in Mountains - just an impression.

I am wondering if conductivity of sand causes smaller difference between Euro and sand than with less conductive ground or air, thus smaller signal blip.

Why Compadre did quite well not sure. Maybe need to study TGSL over sand/ground some more.

-SB

Somon,

I just did a comparison test. For my coil that gives me 31cm in air (the one that I have setup to deal with wet grass) I get a consistent signal at 28cm for a 1 Euro laying on the ground.. Still a tick at around 30cm. The ground is fairly dry today..

Don

[simonbaker]

Somon,

I just did a comparison test. For my coil that gives me 31cm in air (the one that I have setup to deal with wet grass) I get a consistent signal at 28cm for a 1 Euro laying on the ground.. Still a tick at around 30cm. The ground is fairly dry today..

Don

Thanks for that. I'll retest over dry ground and also next time at beach.

-SB

[Dennis the Mennis]

Just an update on what I'm looking into regarding TGSL tuning....

The main issue I'm interested in right now is why some of us have working MDs but don't get more than 15 - 20 cm depth.

My theory is that the reason is a combination of:

1. noise received by the coil is breaking up the target signal and the pieces can't go through the final low-pass filter.

2. The sensitivity threshold depends on the LM308 "output bias voltage" (DC level when no input) and the "input offset threshold" of the LM393 comparator. These two voltage levels may differ between chips and cause some of us to have poor max sensitivity.

---------
To test (1), dfbowers sent me one of his TGSL models that gets 30 cm air depth. I am able to confirm that in my "workshop", I do not get 30 cm depth, more like 20 cm (maybe one day I got near 30). I believe the reason is the noise in my area. However, it is possible something happened to the MD during shipping. To try to confirm which is true, I plan to take the MD as far from noise sources as possible, into the mountains. However, I have no real guarantee that there isn't noise there (when will it end?). So that is my next test -- hard to believe, but I haven't had time to go there yet (actually I went once but by mistake left my Euro target behind).

If I can find a place where dfbowers MD gets 30 cm air test, then I can try (A) attaching my coil to his PCB, and (B) my PCB to his coil.

If case (A) does not get 30 cm, I know my coil is somehow not compatible with his PCB or my coil is just plain bad. I will concentrate on building coils to see what I am doing wrong.

If case(B) does not get 30 cm, then I know my PCB is incompatible with his coil or my PCB is not good. I will first try tuning my oscillator to same frequency as his, then try to make sure my capacitor C6 is same as his to get same RX resonant frequency.

If I still don't get good air depth, then I suspect the level problems at output of LM308 and input to LM393. And proceed to test (2).

--------------
To test (2), I test by connecting dfbowers coil to my PCB, changing LM308 chips and LM393 chip, and also implementing the "dfbowers" mod which makes the sensitivity threshold go negative. If I still can't get 30 cm depth, then there is something more subtle with my PCB that requires careful analysis, such as audio circuit, bad phases at front end, etc. I hope not!

-------------

If I'm able to get around 30 cm depth with my PCB and dfbowers coil, then we have better understanding what to do to fix PCB.

If I can get 30 cm depth with my PCB and dfbowers coil, but not with my own coil, then I need to concentrate on building coil and learn what I'm doing wrong and what is critical. Then can recommend what I found.

---------------

I hope lunamay, Stefano, Dennis the Mennis, Mullihaka, Ultimate Haze, Aiko, Darko, and others can report current status of their TGSL, air depth, any other problems or successes.
-SB

Hi Simon ,

I'm trying to keep up with the posts after my holiday so there is a little delay ....

Current status of the TGS :

A while ago when you were investigating the noise and dc level on the output of the LM308's I'd managed to get around 30 cm detection on a 1 euro coin in AM mode... with the TL081 instead of the LM308.

With the 308's placed I had poor detection , don't know why but it is still under investigation.

The coils at that time were loose and are now a kind of fixed with hotmelt in order to get a more stabile test enviroment .

Meanwhile I'm trying to get the DISC mode working , all the previous tests I did where in AM mode only , I hope I can reproduce the 30cm airdepht again , I think it is pretty noisy overhere on my workbench so that's one other thing to deal with , enough tests to do ....


regards

Dennis the Mennis

[simonbaker]

Hi Simon ,

I'm trying to keep up with the posts after my holiday so there is a little delay ....

Current status of the TGS :

A while ago when you were investigating the noise and dc level on the output of the LM308's I'd managed to get around 30 cm detection on a 1 euro coin in AM mode... with the TL081 instead of the LM308.

With the 308's placed I had poor detection , don't know why but it is still under investigation.

The coils at that time were loose and are now a kind of fixed with hotmelt in order to get a more stabile test enviroment .

Meanwhile I'm trying to get the DISC mode working , all the previous tests I did where in AM mode only , I hope I can reproduce the 30cm airdepht again , I think it is pretty noisy overhere on my workbench so that's one other thing to deal with , enough tests to do ....


regards

Dennis the Mennis

Wow, that's sounding very good. It does make sense to me since those chips can really affect the sensitivity threshold.

Maybe also put in the "dfbowers mod" in the sensitivity circuit, that can't hurt and can help test these ideas. It's pretty easy to wire, there are a couple of messages about it.

Regards,

-SB

[simonbaker]

solder won't stick

Ok, thanks. I bought three different brands of 6 ft. USB cables to compare.

The cheapest ($3.99 IdotConnect) has a very thin foil shield with a copper shield wire (stranded) and four internal wires.

The next most expensive, ($4.97 Green Products) has a good loosely braided shield around a good foil shield, with a copper shield wire (stranded), and four internal wires.

The more expensive ($7.99 Belkin) has a good braided shield with what may be a silver shield wire (stranded) and more spiral wrapped foil shield, and four internal wires. It appears that the green and white pair may be twisted.

I did not buy the very expensive cables (around $25), but some may advertise separate shields on the pairs.

All three cables looked usable, but the cheapest connectors did not fit nearly as snugly as the others, so I favor the middle cheapest.

Theoretically, I would think separately shielded pairs would be desirable, but dfbowers gets good results with single shield, so... there it is.

-SB

Recently working with the Green Products cable, I forgot about the copper wire (thin and hidden in the foil) and so was trying to solder to the silver braid around the foil shield.

I don't know what that stuff is made of, but man it's like Teflon to solder! Cannot solder to it. So look for that little copper wire and be sure to use that for connecting shield to PCB.

Tip #10115078c for mechanically challenged circuit makers.

-SB

[dfbowers]

Poor mans coil tuning.

I thought about this one for a while and just gave it a try. No way to measure the inductance of your coils? No problem! Just tune it like a guitar.

Since the audio frequency is just divided down by 32 from the TGSL oscillator, we can adjust our frequency by just listening to the audio, and removing or adding turns to our coils. It's easy for me because my coils are made with no binding. I make mine sticky with "Spar Varnish" when it goes on the form and then let it dry. It's easy to peel off a few windings.

Ok.. so here is what to do:

Download this slick little audio oscillator to run on your PC.
http://www.softpedia.com/progDownloa...oad-65501.html

DISCLAIMER: I got mine from the external mirror site. I scanned for viruses so to appears to be OK.

Then, set up your TGSL and null your coils. Crank up the audio oscillator and set it to a 453Hz square wave (14.5kHz / 32 = 453Hz). When the audio of your TGSL is the same as the audio oscillator, you should be tuned to 6.0mH for the Tx coil! It's fairly accurate..

Then.. switch your coils.. Tx for Rx and Rx for Tx We won't leave things that way.. just do it for testing.

Then, set up your TGSL and null your coils again. Now, adjust your audio oscillator for a 434Hz square wave (13.9kHz / 32 = 434Hz). Again, add or remove turns to your Rx coil (which is now connected to Tx) and tune your TGSL to the same frequency. Now, your Rx coil is set to 6.5mH!

What we just did is resonate each coil to .020nF, and assuming C1 and C2 are in spec, your Rx and Tx coils should be real close..

Good luck

Don

[simonbaker]

I thought about this one for a while and just gave it a try. No way to measure the inductance of your coils? No problem! Just tune it like a guitar.

Since the audio frequency is just divided down by 32 from the TGSL oscillator, we can adjust our frequency by just listening to the audio, and removing or adding turns to our coils. It's easy for me because my coils are made with no binding. I make mine sticky with "Spar Varnish" when it goes on the form and then let it dry. It's easy to peel off a few windings.

Ok.. so here is what to do:

Download this slick little audio oscillator to run on your PC.
http://www.softpedia.com/progDownloa...oad-65501.html

DISCLAIMER: I got mine from the external mirror site. I scanned for viruses so to appears to be OK.

Then, set up your TGSL and null your coils. Crank up the audio oscillator and set it to a 453Hz square wave (14.5kHz / 32 = 453Hz). When the audio of your TGSL is the same as the audio oscillator, you should be tuned to 6.0mH for the Tx coil! It's fairly accurate..

Then.. switch your coils.. Tx for Rx and Rx for Tx We won't leave things that way.. just do it for testing.

Then, set up your TGSL and null your coils again. Now, adjust your audio oscillator for a 434Hz square wave (13.9kHz / 32 = 434Hz). Again, add or remove turns to your Rx coil (which is now connected to Tx) and tune your TGSL to the same frequency. Now, your Rx coil is set to 6.5mH!

What we just did is resonate each coil to .020nF, and assuming C1 and C2 are in spec, your Rx and Tx coils should be real close..

Good luck

Don

Love it! What if you can't carry a tune...?

Using oscillator to check coils is nice idea though, even if you use a frequency counter -- very convenient. And if your capacitor is off spec, it probably works out even better than using actual inductance (maybe?).

-SB

[simonbaker]

Love it! What if you can't carry a tune...?

Using oscillator to check coils is nice idea though, even if you use a frequency counter -- very convenient. And if your capacitor is off spec, it probably works out even better than using actual inductance (maybe?).

-SB

Correction to what I said -- if your oscillator capacitor is out of spec the RX coil calc would be off unless you know the true value and calculate the desired frequency for the RX coil. And this assumes the RX capacitor C6 is in spec also.

-SB

[dfbowers]

Correction to what I said -- if your oscillator capacitor is out of spec the RX coil calc would be off unless you know the true value and calculate the desired frequency for the RX coil. And this assumes the RX capacitor C6 is in spec also.

-SB

Yes, I agree.. The whole idea makes some assumtions. I thought that maybe this would help a few guys who were struggling with getting their coils right with no scope, inducatance meter or frequency counter.. They would also have to get at least 3 of their caps in spec to begin with... C1, C2 and C6.

One observation I just made.. Use volume sparingly! I think that the square wave function tends to heat things up a bit in a PC. Duty cycle must be at 50% or something..

Don

[simonbaker]

Yes, I agree.. The whole idea makes some assumtions. I thought that maybe this would help a few guys who were struggling with getting their coils right with no scope, inducatance meter or frequency counter.. They would also have to get at least 3 of their caps in spec to begin with... C1, C2 and C6.

One observation I just made.. Use volume sparingly! I think that the square wave function tends to heat things up a bit in a PC. Duty cycle must be at 50% or something..

Don

Yes, I'm a fan of these clever tricks to work with minimal budget/equipment - like it!

Now that we have singing oscillators, we should work on singing discriminators -- "Heart of Gold", "Silver Bells", "Copper Kettle", "Steel Rail Blues", "Scarborough Ferrite"...

[Dennis the Mennis]

Another poor mans coil tuning

I thought about this one for a while and just gave it a try. No way to measure the inductance of your coils? No problem! Just tune it like a guitar.

Since the audio frequency is just divided down by 32 from the TGSL oscillator, we can adjust our frequency by just listening to the audio, and removing or adding turns to our coils. It's easy for me because my coils are made with no binding. I make mine sticky with "Spar Varnish" when it goes on the form and then let it dry. It's easy to peel off a few windings.

Ok.. so here is what to do:

Download this slick little audio oscillator to run on your PC.
http://www.softpedia.com/progDownloa...oad-65501.html

DISCLAIMER: I got mine from the external mirror site. I scanned for viruses so to appears to be OK.

Then, set up your TGSL and null your coils. Crank up the audio oscillator and set it to a 453Hz square wave (14.5kHz / 32 = 453Hz). When the audio of your TGSL is the same as the audio oscillator, you should be tuned to 6.0mH for the Tx coil! It's fairly accurate..

Then.. switch your coils.. Tx for Rx and Rx for Tx We won't leave things that way.. just do it for testing.

Then, set up your TGSL and null your coils again. Now, adjust your audio oscillator for a 434Hz square wave (13.9kHz / 32 = 434Hz). Again, add or remove turns to your Rx coil (which is now connected to Tx) and tune your TGSL to the same frequency. Now, your Rx coil is set to 6.5mH!

What we just did is resonate each coil to .020nF, and assuming C1 and C2 are in spec, your Rx and Tx coils should be real close..

Good luck

Don

Perhaps also another way to determine the coil value ( also for those who haven’t got a scoop , L meter or frequency counter ) is with F osc , the oscillation frequency of the coil itself .

The first set of coils I've made were complete according the specs : 6 and 6,5 mH ..... I thought so .
Only the L meter on my Digital Multi Meter was a little bit very much out of range , I should have bin more suspicious while removing the big layer of dust of the DMM.

Anyway I went to the electronic shop , measured the coils over there and re-calibrated my DMM according the values given by the expensive L meter from the shop.

But to be sure I double check the coils with F osc. :
In the attached examples is a Function Generator connected to the coil ( Lx) in serie with the capacitor ( 22n for Tx or even better remove the 22n from the pcb in case of any tolerance issues ).

Now use the SweepGenerator or a similar Tone Generator Program on your pc or laptop instead of the function generator

As Don stated :
begin with a very low input sinus voltage round 14 KHz / 0,2 – 0,3 Vac and slowly raise the frequency :

In case of the voltage meter over the coil the voltage will peak at the oscillation frequency of the coil

In case of measuring over the coil and the capacitor the voltage will reach a minimum voltage at Fosc.


Take the value for Fosc ( for example 15,6 KHz and 22n for C ) and use the next formula to calculate Lx :

http://www.whatcircuits.com/lc-resonance-frequency-calculator

The calculated Lx values are pretty close to the values of my re-calibrated DMM J .

I don’t know if this can hurt your pc or laptop so be careful …
I haven’t tried the software SweepGenerator or Tone Generator way myself yet
( it was on my list in the “preliminary” ) .


regards

Dennis the Mennis

[Dennis the Mennis]

Sawtooth sample puls on GEB channel FET instead of squarewave ?

Holy cow.. It finally happened! With the coil described above, coil #7 allows my null phase to occur close to where a ferrite slug ground balances. 31cm for a 1e.

Now.. if I can only make another one just like it.

Hi to all ,

My reply on post 523 and before : in these examples the signal on the gate of the GEB channel FET is a kind of a square wave ...

On my TGS this signal - sample puls is more like a sawtooth , I can remember -SB SimonBaker had the same experience in the past?


Is the sawtooth a correct signal here ?


I know this isn't supposed to be the right way but when I adjust my MD in
DISC mode as described above and according to the attached example I have a great detection , but no discrimation at all .


regards

Dennis the Mennis

[simonbaker]

Hi to all ,

My reply on post 523 and before : in these examples the signal on the gate of the GEB channel FET is a kind of a square wave ...

On my TGS this signal - sample puls is more like a sawtooth , I can remember -SB SimonBaker had the same experience in the past?


Is the sawtooth a correct signal here ?


I know this isn't supposed to be the right way but when I adjust my MD in
DISC mode as described above and according to the attached example I have a great detection , but no discrimation at all .


regards

Dennis the Mennis

Hi DtM:

The sawtooth is the ALL METAL mode signal that goes to the DISC JFet gate when you switch to ALL METAL. It should never go to GB JFet.

Double-check which JFet you are mentioning and let us know, to make sure you don't have a wiring problem.

If wiring is OK, then sawtooth at DISC JFet gate just means your switch is in ALL METAL mode and then DISC pot has no effect by design.

Flip the switch to other position and observe the same waveform and let us know how it looks.

Cheers,

-SB

[simonbaker]

Perhaps also another way to determine the coil value ( also for those who haven’t got a scoop , L meter or frequency counter ) is with F osc , the oscillation frequency of the coil itself .

The first set of coils I've made were complete according the specs : 6 and 6,5 mH ..... I thought so .
Only the L meter on my Digital Multi Meter was a little bit very much out of range , I should have bin more suspicious while removing the big layer of dust of the DMM.

Anyway I went to the electronic shop , measured the coils over there and re-calibrated my DMM according the values given by the expensive L meter from the shop.

But to be sure I double check the coils with F osc. :
In the attached examples is a Function Generator connected to the coil ( Lx) in serie with the capacitor ( 22n for Tx or even better remove the 22n from the pcb in case of any tolerance issues ).

Now use the SweepGenerator or a similar Tone Generator Program on your pc or laptop instead of the function generator

As Don stated :
begin with a very low input sinus voltage round 14 KHz / 0,2 – 0,3 Vac and slowly raise the frequency :

In case of the voltage meter over the coil the voltage will peak at the oscillation frequency of the coil

In case of measuring over the coil and the capacitor the voltage will reach a minimum voltage at Fosc.


Take the value for Fosc ( for example 15,6 KHz and 22n for C ) and use the next formula to calculate Lx :

http://www.whatcircuits.com/lc-resonance-frequency-calculator

The calculated Lx values are pretty close to the values of my re-calibrated DMM J .

I don’t know if this can hurt your pc or laptop so be careful …
I haven’t tried the software SweepGenerator or Tone Generator way myself yet
( it was on my list in the “preliminary” ) .


regards

Dennis the Mennis

Yes, I've used that method too, I think it's pretty accurate (assuming your sine generator and DVM don't contribute much capacitance).

A lot of DVMs have miserable bandwidth on AC voltage I've found, but probably won't affect your method too much.

In some ways, we don't care too much what the inductances are, more interested in the oscillator frequency and the RX resonant frequency. So we can just adjust the oscillator (with capacitor) until correct frequency. Then I think use your method to adjust RX resonant frequency by trimming capacitor C6. I guess we'd calculate the oscillator and RX resonant frequencies based on the "ideal" specs (ideal TX, RX inductances and ideal capacitors).

The question is: what is the ideal oscillator frequency and what is the ideal RX resonant frequency? People who use 6 mH for TX coil and 6.5 mH for RX coil make nice detectors. dfbowers coils were a little higher each and still made very nice detector.

Regards,

-SB

[dfbowers]

Testing a new coil..

For your entertainment.. I have posted a new video of some tests from the basement on my latest coil (#7) . This one is coming out a little better than average.. but keep in mind, my basement is quiet!!

This one has the coil shields grounded only and is set up to minimize wet grass per some of my earlier posts. Also, I used conductive mylar for shielding material and paid attention to the null phase.

Enjoy!

Don

http://www.youtube.com/watch?v=Nx4Oxac0Ufg