TGSL Experiments

[dfbowers]

If I could get a solid 25 cm with quarter I would be pleased, but I think I'm more around 20 cm at best.

Comparing with your coils is an exciting idea, let's keep that in mind, thanks for the offer. It would really help identify how important the coils are in the performance vs. possible circuit limitations. It will take me a while to feel like I have a stable platform to test with here. Right now, things seem to change somewhat day-to-day and I don't trust my observations yet. It's hard to be rigorous, so many things to play with.

Of course, do we know if coils would survive the journey unchanged? I guess after returning them you can test again to verify.

I will work toward getting ready to try your coils some day; very kind of you.

I'm working on getting a 1 Euro - my wife has friends in England.

Cheers,

-SB

Simon,

You may be able to pick up a euro at a local bank or airport depending where you live. Let me know if you would like to compare coils.

Don

[simonbaker]

Gain vs. Noise

Main goal: improve depth to 28-30 cm for 1 Euro.

I think two main things to check:

1. Gain vs. Noise

2. Are both channels (disc, ground balance) working together correctly.

Plan for 1 (gain vs. noise)

I notice that at output of LM308, in my "workshop", I typically see .1 to .2 volts PP of noise. That's a lot! You would expect constant beeping from detector. (there actually is a fair amout of chatter with higher sensitivity pot. But I would expect even more). The threshold range for detection of LM393 U106 comparator is around 3 to 36 millivolts. What gives?

Well, the circuitry between LM393 U106 and LM358 U107 filters out fast noise pulses, so you don't hear the shorter noise pulses.

But it indicates a problem. We want to detect very small signals. Ideally the noise level should be below the smallest signal we want to detect.

Let me modify that. Because the pre-audio circuitry filters out fast pulses, we want the low frequency noise level to be less than the smallest signal we want to detect. The low frequency is essentially determined by our sweep speed as we move the head over targets. The TGSL circuitry seems to respond in the 10 Hz range.

The sensitivity of the comparator U106 is from 3 to 36 millivolts when you turn the sensitivity pot. At most sensitive setting, it will make a beep when both the Disc and DB channels have a nice long pulse which is greater than 3 millivolts at outputs of LM308.

What does that mean for us? Ideally, it means I'd like to see the background noise to be less than 3 millivolts all the time, even the fast pulses. On my workbench, that may never be possible because of various EMI interference. So maybe we have to do final tuning outside far from EMI.

Here's a weird question: is noise good or bad? Technically, we detect a target when the signal is above the comparator threshold. The minimum threshold is 3 millivolts. We could reduce that to zero by eliminating the 1K resistor R37. In fact, that is one technique to keep in mind for tuning if we can achieve a low noise situation. But otherwise, a little noise can help if the target signal is added to it and rises above the threshold. However, this is really not ideal, since the signal might be broken up and not smooth. But it is just a technical point to consider.

Back to the gain vs. noise. If our detector is not deep enough, the first thought is to turn the sensitivity pot to detect smaller signals. But there is no point in increasing sensitivity if the noise is bigger than the signal. So the first check is to analyze the noise at LM308 outputs. If greater than 3 millivolts, we really can't take advantage of the gain we already have.

My circuit is very quiet when I unplug the RX coil. So it seems most noise comes through the RX coil. It would be helpful to know what the source of the noise is. Is it interference? From what? Is it due to the oscillator instability or jitter? etc.

Let's say we find a way to eliminate the noise to a very low level. Now what? First thing to try is turn the sensitivity pot to maximum sensitivity (3 millivolts). If still can't get the depth we want, then what?

Next I would look at the target signals at LM308 outputs. Typically, the GB channel is weaker than the Disc channel. Maybe the problem is the GB channel is not going above 3 millivolts even though the disc channel is.

A way to test this is to put separate sensitivity controls on the GB and disc channels, and allow the GB sensitivity to go to zero threshold. Or to test the theory, jumper the input of the GB channel LM308 to force it high all the time so only the Disc channel needs to detect the signal. (you may need to be careful how you do that to not damage the LM308 ).

Anyway, we may learn that by making the GB channel sensitivity threshold smaller, we can increase depth. Or we may need only to work on the GB channel gain.

Plan 2 - channels working together

Another thing to check is if the disc and GB channels are working together in a coordinated way for good targets. Discrimination of unwanted targets is achieved when the disc and GB channels move in opposite directions for the target. But sometimes for wanted targets, the two channels may not exactly go up and down together perfectly. That would be a clue of something to fix. If they don't go up and down at the same time, the comparator makes a shorter pulse which might not get through to the speaker.

back to gain

Ultimately, if we still can't get enough depth and the noise is very low, we can think about ways to increase the gain of the system by changing resistors, capacitors, etc. But this is last resort because you have to be very careful not to alter the response of the amplifier channels.

response

That is, unless we want to alter the response. A slower response for example can go deeper, but is more difficult to work with. If the back-to-back capacitors act unexpected like a 10 uF instead of 4.7 uF, it would cause a slower response. That is something to consider that could be making some MDs work differently from others.

Coil

Finally, there is the mystery of the coil. It has been said that if you make the right coil, that is the solution to getting 28 - 30 cm depth. I would like to believe that most coils can be made to work well if you tune the circuit for them -- whatever that means. Still a mystery. Understanding it better is probably best accomplished by building many coils and comparing good ones to bad ones and trying with different PCBs and tunings.

But first I need to find a noise-free location to check gain-noise question.

-SB

[simonbaker]

Simon,

You may be able to pick up a euro at a local bank or airport depending where you live. Let me know if you would like to compare coils.

Don

Hi Don - Maybe I'll get one from Stefano before long, or as you suggest. Need one for sure.

If you have a coil that is proven that you can spare for a while, it would be very valuable for me to test with and report results here, answer some questions for those of us stuck around 20 cm depth. I will PM you about it. I expect I would need to get a compatible connector to put on a cable. Actually, cable may be part of the equation, we need to think about that too.

Regards,

-SB

[dfbowers]

Hi Don - Maybe I'll get one from Stefano before long, or as you suggest. Need one for sure.

If you have a coil that is proven that you can spare for a while, it would be very valuable for me to test with and report results here, answer some questions for those of us stuck around 20 cm depth. I will PM you about it. I expect I would need to get a compatible connector to put on a cable. Actually, cable may be part of the equation, we need to think about that too.

Regards,

-SB

Absolutely. Check for message..

Don

[Vladimir]

It is obvious that a bad relationship L / C in Tx round, I personally think that we should reduce the number of turns at 60. Probably this will reduce the noise level.

[simonbaker]

It is obvious that a bad relationship L / C in Tx round, I personally think that we should reduce the number of turns at 60. Probably this will reduce the noise level.

I recall you have some different ideas for the coil and circuit. Would you describe them in detail, maybe start a special thread for your design? It would be interesting to see what results you get and how they compare.

Tesoro moved on to different designs so I expect there are many variations that will work, perhaps better.

-SB

[lunamay]

hi to all
i achive +2cm of depth by using a scheald for coil from old and big electrolitic capacitor. there is two al foils ...use thicker..good faraday's cage if you don't have a AL tape


best regards

[simonbaker]

hi to all
i achive +2cm of depth by using a scheald for coil from old and big electrolitic capacitor. there is two al foils ...use thicker..good faraday's cage if you don't have a AL tape


best regards

Thanks for good tip. What shield is it 2 cm better than?

-SB

[simonbaker]

bad table

Just discovered that a standard card table with collapsible legs is bad as a worktable for coils. I nulled a coil, then moved it and it required renulling at different point. Both nulls worked fine as long as the coil stayed in its place where it was nulled. So... should we null over the ground where we hunt????

I guess the metal frame of the cardtable really had a powerful effect.

-SB

[Vladimir]

I recall you have some different ideas for the coil and circuit. Would you describe them in detail, maybe start a special thread for your design? It would be interesting to see what results you get and how they compare.

Tesoro moved on to different designs so I expect there are many variations that will work, perhaps better.

-SB

This is my modification TGSL, the sensitivity is 1 EUR coin about 34-36cm in air with 27DD coil, a small gold ring - 30cm, Turkish silver Aspres (5 mm diameter) - 25-27 cm. everything is measured by the DISC. mode (3.5-4).
The device has excellent discrimination and good GEB, on the ground shows the tendency of putting the occasional false signals (peak), but not so hard to work. I am pleased with this modification and the number of finds of Roman coins.
Greeting

[magma_kostic]

Vladimir are you from Montana, Bulgaria? I am confused!? Weeks ago you were signed as "Vladimir, Montana, BG" and now you are from "Serbia" !? Because there is real Vladimir from Montana, Bulgaria. Man who is producing various products related to metal detectors. Real gentleman and pretty educated man. Were you tried to act as that Vladimir in the past? And recently you changed the idea and corrected "Montana,BG" to " Serbia" in your User CP?
BTW those modifications, you are trying to push here, are not your ideas at all, but Orbit's ideas from past 2 years, right? Orbit applied same feedbacks on TGSL channels as those were and still are on "Caesar". Anyway that is bad idea. Those will degrade TGSL behavior in huge percent. Another question for you; why you left your first nick "MAGMA" here? You don't like to be "MAGMA" anymore?

[simonbaker]

This is my modification TGSL, the sensitivity is 1 EUR coin about 34-36cm in air with 27DD coil, a small gold ring - 30cm, Turkish silver Aspres (5 mm diameter) - 25-27 cm. everything is measured by the DISC. mode (3.5-4).
The device has excellent discrimination and good GEB, on the ground shows the tendency of putting the occasional false signals (peak), but not so hard to work. I am pleased with this modification and the number of finds of Roman coins.
Greeting

Thanks for showing these mods, I will study them and maybe try some out. Some I have seen before. The variable gain resistor on the LF353 preamp looks interesting for tuning coils with different gain.

Some of your mods appear to slow the response of the basic design, which may help for deep targets but make pinpointing more difficult; e.g. the 200K resistors for R25, R26 instead of 20K.

I have trouble reading some values for the audio amp because of the image resolution though.

Also please describe the details of the coils that work with those mods.

TGSL is a very good MD and also a good starting circuit for trying mods like these to suit different tastes.

-SB

[Vladimir]

Vladimir are you from Montana, Bulgaria? I am confused!? Weeks ago you were signed as "Vladimir, Montana, BG" and now you are from "Serbia" !? Because there is real Vladimir from Montana, Bulgaria. Man who is producing various products related to metal detectors. Real gentleman and pretty educated man. Were you tried to act as that Vladimir in the past? And recently you changed the idea and corrected "Montana,BG" to " Serbia" in your User CP?
BTW those modifications, you are trying to push here, are not your ideas at all, but Orbit's ideas from past 2 years, right? Orbit applied same feedbacks on TGSL channels as those were and still are on "Caesar". Anyway that is bad idea. Those will degrade TGSL behavior in huge percent. Another question for you; why you left your first nick "MAGMA" here? You don't like to be "MAGMA" anymore?

Ivice (IVCONIC)...Potrazi drugog psihijatra , za tebe nemam vremena.
A o ostalom se moze i orbit izjasniti.

[Vladimir]

.

Also please describe the details of the coils that work with those mods.

-SB

The values for the coil are the same as the originals, I will try soon with other values, after that I will give my comments on this modification.

[Vladimir]

Oops, I made a small mistake ..... this is a tone circuit that I used, I apologize.