TGSL Tuning & troubleshooting

[Vladimir]

If I understood well, there is a substantial lack of understanding of concepts, Shild the coil and GND on the board are exactly the same and thus should be connecting. Depending on the type of cable used, it can be different connecting from Shild cable. But the basic concept SHILD-CABLESHILD-GND must be respected.
Regards

[dfbowers]

I think then coil shields not directly grounded.

-SB

(which brings up a point. If if cable shield is grounded at PCB side only, and coil shields connected to cable shield only (not TX ground wire in coil head), then easier to switch leads like pebe suggests. But would it work as well? Another config to try.)

One difficulty I am having with trying different grounding configurations is that the coils seem to null at different places depending upon where the shields are grounded, so I don't think that I can cast a set of coils in epoxy with wires exposed. I'm just going to have to make several complete coils and evaluate from there. I may be able to make some sort of quick disconnect to try different cables though. Waiting for epoxy to harden..

[simonbaker]

You could probably correct the situation by reversing the leads of both coils.

I think then coil shields not directly grounded.

-SB

(which brings up a point. If if cable shield is grounded at PCB side only, and coil shields connected to cable shield only (not TX ground wire in coil head), then easier to switch leads like pebe suggests. But would it work as well? Another config to try.)

[simonbaker]

In one of my previous posts I said that I was having trouble with wet grass. I have since discovered that one of my coils was improperly grounded.. I mistakenly tied ground to the non-inverting input (J2-1 instead of J2-2) inside the coil. Too bad it's already cast in epoxy! The coil works absolutely fine otherwise and is a "good" one. I just have to use it when it's dry outside.

Anyway..

I had set out to independently verify which coil grounding scheme is best.
I came to the conclusion that what is documented in the "TGSL coil making" pdf IS best. Most notable, I found that a much deeper null can be achieved when J2-2 is grounded inside the coil. Other ways seem to work OK as well, but just not as good. Even with no grounded RX at all, grounding J2-2 on the coil end and to both shields will give a much deeper null. My advise to everyone testing coils.. Do your testing with shields installed, grounded AND using your final cable, even if you do not intend on finishing the project! It does make a difference.

Don

Even with no grounded RX at all, grounding J2-2 on the coil end and to both shields will give a much deeper null.

Would you clarify that sentence? How can J2-2 be grounded but RX not grounded? What do you mean by "coil end"?

Did you experiment with grounding one coil shield but not the other?

By "both shields" I assume those are the coil shields, not the cable shields (since you typically use a cable with one shield -- but Vladimir for example uses cable with two shields).

Another thing I'm not clear on is the value of shielding the TX coil -- how much difference does that make? What is the principle? It would be interesting to see performance with only RX coil shielded.

One more question -- how did you experiment with the grounding? Did you make a special coil with all wires exposed?

We can stay busy a long time with these combinations!

Regards,

-SB

[simonbaker]

In one of my previous posts I said that I was having trouble with wet grass. I have since discovered that one of my coils was improperly grounded.. I mistakenly tied ground to the non-inverting input (J2-1 instead of J2-2) inside the coil. Too bad it's already cast in epoxy! The coil works absolutely fine otherwise and is a "good" one. I just have to use it when it's dry outside.

Anyway..

I had set out to independently verify which coil grounding scheme is best.
I came to the conclusion that what is documented in the "TGSL coil making" pdf IS best. Most notable, I found that a much deeper null can be achieved when J2-2 is grounded inside the coil. Other ways seem to work OK as well, but just not as good. Even with no grounded RX at all, grounding J2-2 on the coil end and to both shields will give a much deeper null. My advise to everyone testing coils.. Do your testing with shields installed, grounded AND using your final cable, even if you do not intend on finishing the project! It does make a difference.

Don

Thanks for those experiments, they are very interesting to me. When you get time, it would be great if you could go into more detail of each connection and what you observed -- a diagram would be nice.

I really don't have a good understanding of what the shield is doing or shielding us from -- electrostatic gradients; high voltage "static"; EMF, etc., and why the different grounding configs have the effects they do.

I believe graphite type shields are also used as discussed on these forums (fora?). It would be interesting to mix sufficient graphite into the potting resin and see what kind of shield that makes instead of foil.

Regards,

-SB

[pebe]

In one of my previous posts I said that I was having trouble with wet grass. I have since discovered that one of my coils was improperly grounded.. I mistakenly tied ground to the non-inverting input (J2-1 instead of J2-2) inside the coil. Too bad it's already cast in epoxy! The coil works absolutely fine otherwise and is a "good" one. I just have to use it when it's dry outside............

Don

You could probably correct the situation by reversing the leads of both coils.

[mikebg]

Don,
Please write as circuit diagram your mistakenly tied ground and post it in the forum. We will try to diminish the bad effect by adding external components.
Mike-BG

[dfbowers]

More on grounding and shielding.

In one of my previous posts I said that I was having trouble with wet grass. I have since discovered that one of my coils was improperly grounded.. I mistakenly tied ground to the non-inverting input (J2-1 instead of J2-2) inside the coil. Too bad it's already cast in epoxy! The coil works absolutely fine otherwise and is a "good" one. I just have to use it when it's dry outside.

Anyway..

I had set out to independently verify which coil grounding scheme is best.
I came to the conclusion that what is documented in the "TGSL coil making" pdf IS best. Most notable, I found that a much deeper null can be achieved when J2-2 is grounded inside the coil. Other ways seem to work OK as well, but just not as good. Even with no grounded RX at all, grounding J2-2 on the coil end and to both shields will give a much deeper null. My advise to everyone testing coils.. Do your testing with shields installed, grounded AND using your final cable, even if you do not intend on finishing the project! It does make a difference.

Don

[dfbowers]

Yes, I could see that.. The shield should remove most of the undesirable effects though. Nakky, do you observe this at reduced sensitivity?

Don

Both coils are shielded
Reduced sens to zero, forearm still sounds
When disc at 7 pulltabs rejected,there is no sound of forearm

Grt Nakky

Wow.. I donno. The TGSL is not supposed to do that. Mine is dead quiet at near maximum sens and minimum disc. Although my X5 will sound slightly near a body and my Classic I will do the same thing.

Are you sure that you don't have a metal plate in your arms from a past life? (Just kidding).

I can only speculate without looking at it first hand. Grounding would be my first guess since I can reproduce that symptom by ungrounding my coils but I guess we have to rule that out.

Is your arrangement mechanically sound? How about the cable? Is your detector free from noise when you touch or move the cable?

Don

[nakky]

Yes, I could see that.. The shield should remove most of the undesirable effects though. Nakky, do you observe this at reduced sensitivity?

Don

Both coils are shielded
Reduced sens to zero, forearm still sounds
When disc at 7 pulltabs rejected,there is no sound of forearm

Grt Nakky

[simonbaker]

hi
finaly i repeired my md..once i conected +to - and -to+..big mistake made by hot weather and to much sun on head:P:P:P..something has broken..i was looking for broken part and then i read that i should replace parts in audio circuit..now i can say to eweryone who will make this kind of mistake and uses a computer power suply for tests,that he should replace a 7808 part..i just did that and my md works great

best regards

Glad to hear it. I hope you can save up for an oscilloscope soon, because that will allow you to really see how it works and allow you to make interesting mods and also give us data to help all of us with tuning.

Regards,

-SB

[Vladimir]

hi
finaly i repeired my md..once i conected +to - and -to+..big mistake made by hot weather and to much sun on head:P:P:P..something has broken..i was looking for broken part and then i read that i should replace parts in audio circuit..now i can say to eweryone who will make this kind of mistake and uses a computer power suply for tests,that he should replace a 7808 part..i just did that and my md works great

best regards

Hi
I'm glad that your detector works again, it would be good if you present the measurement sensitivity in the air.
Regards

[lunamay]

hi
finaly i repeired my md..once i conected +to - and -to+..big mistake made by hot weather and to much sun on head:P:P:P..something has broken..i was looking for broken part and then i read that i should replace parts in audio circuit..now i can say to eweryone who will make this kind of mistake and uses a computer power suply for tests,that he should replace a 7808 part..i just did that and my md works great

best regards

[simonbaker]

Tuning/Troubleshooting Review

Reviewing recent conclusions regarding tuning/troubleshooting.

1. DC output level of LM308 / U104 chips may be important factor affecting sensitivity. Checking LM308 chips and replacing as needed might help sensitivity and depth. However, what is the best dc output level, and should both chips be the same? dfbowers mod allows more range in sensitivity threshold to compensate for off-spec dc output levels of chips. Someday we can experiment with separate threshold for Disc and GB channels.

2. Noise picked up by coil/cable may be important factor in limiting air test depth, especially if performed indoors where EMI is high. A good place to measure noise is at the output of LM308 / U104 chips. If the peak-to-peak noise is much above 20-25 mV it is probably a problem. It is important to test somewhere as noise-free as possible; not easy I find. I tested dfbowers nice TGSL in a local park recently and got terrible motorboating a few minutes after turning it on, but not at home. At home, I get different noise at different times of day and different days. I will drive into the mountains to test more. I believe noise reduces depth by breaking up target signal into pieces which are filtered out by the next low-pass filter.

3. I'm using dfbowers TGSL as benchmark to evaluate my own PCB and coils. I connect my coil to his PCB or his coil to my PCB. But it is not so easy to compare air tests as I thought. The quality of the signal causes confusion -- how much chatter is present, how long is the beep, how reliable is the beep at a particular depth, how fast do I sweep the coin, how does the loudness of the speaker/headphone affect my judgment, etc. Also, because of parts variations, the oscillator frequency is not identical and the RX resonant frequency not identical -- how do I adjust those for a good comparison.

My impression so far is that my PCB is close to dfbowers' in depth test, perhaps the beep is a little shorter or the depth just a little less. Many more tests needed though.

I do not have a good coil to compare (mine typically use much thicker wire and have too much gain), so my next step is to try to make some coils quite similar to dfbowers'. I have some 30 AWG wire that I'll use, and a modified coil-winder to compress-glue the coil as I wind. I am also working on a mold to cast a final coil out of resin -- not as good as dfbowers vacuum-formed shell, but hopefully solid.

dfbowers uses "Spar Varnish" to stick his coil windings together as he winds. I could not find smaller than a quart, too expensive, so I may try water-based polyurethane (because I have some). I know, water-based may be bad. If the coils stink, I try again with non-water-based polyurethane. I'd like to find something that is tough and permanent but not so nasty to breath.

This morning I was getting target detection probably at around 25 cm indoors with dfbowers' or my PCB connected to dfbowers' coil. But I would have to make a video so you could evaluate what that really means. Eventually I'll try to do that. As mentioned, a trip to the mountains is planned - but I know there are various transmitters in the mountains, and also that ELF (Extremely Low Frequency) transmissions, such as submarine communications, probably penetrate mountains easily. I feel like noise is a poltergeist haunting me right now.

So I'll build some coils, continue testing dfbowers TGSL in various environments, continue testing my PCB with dfbowers coil (trying some tuning capacitors in the oscillator and RX input). I'll also test dfbowers' mod to the sensitivity threshold and effects of different LM308 chips; then eventually try separate thresholds on DISC, GB channels. Also I should keep an eye on batteries in dfbowers' TGSL to make sure I'm not running them down and affecting the tests.

I hope others with problems in their TGSL are trying some of these tests too and will report.

-SB

[Vladimir]

Most common mistakes in the construction of the detector (Part I):
* Badly pinched coils
* Rolled coil without glue
* Al foil broken
* Replaced the polarity of the coils
* Poor connected Shild
* Thermal stability of the coil is bad (hard charging polymers)
* Hydro protection is poor
* The cable has a bad performance (capacity, a small number of wires in Shild, poor choice of cable)
* Bad connector from the coil (an inexpensive, mechanically unstable, corrosive)
* PCB is not well connected to GND
* PCB is not cleaned of welding products (tinol)
* PCB is not protected by varnishing
* System error in setting the parameters
* Poor choice of detector box
* Poor or inadequate battery