"Beyond" TGSL ....

[Monk]

I applaud the efforts of dfbowers and simonbaker in their quest to produce the optimum TGSL. But.....

Are they beating a dead horse? And is it time to move on (at least for some of us)?

I refer to one of the last posts made by Ivconicin the original TGSL construction thread and I quote:

10-03-2009, 04:45 AM
ivconic ivconic is offline
Guru

Join Date: Feb 2006
Posts: 3,384
Default
Older Tesoro models like TGS,Silver,Bandido etc..etc.. do have pretty limited area of detection. If you manage to make good 27cm DD coil to reach 1e coin up to 34cm in air than that is it. Only much larger item it will detect a bit better. So it will stay limited to those 34cm and less (soil conditions). Same thing with smaller diameter coils. Recently i made 21cm DD coil and it is performing pretty good - 1e coin at 27cm in air. No false signals, very stable, excellent disc. Excellent performing on soil. And very loud and sharp audio!
But... larger items it is also detecting up to that limit...maybe 3-4cm more. No matter how large is the item - similar "depths" in air.
So...it is limited. It is not about TX power here. TX power at TGS/TGSL is quite enough. It is much more about RX stage and coil.
I had some doubts about that, but than i got original Bandido II and i noticed same occurrence and behavior. So...it is not about my (our) handmade - it is general.
Once i realized that i just took it as reality. TGS/TGSL is good as it is. Best you can do is to put more effort to make and balance better coil. With my previously balanced coils (in the past) i achieved some 36 to 45cm in air for 1e coin, but than i realized - no benefit from those on real soil, depths in soil were same as with latest coils (1e at 30cm). Only differences between those coils (in the past and recent) were in residual voltage and audio strength. In the past i kept residual voltages under 1mV and got more inches in air and...very weak and mum audio. Nowdays i keep residual voltages around 4mV and benefit is loud and sharp audio. On soil those coils performs, as i said, pretty same.
So it is all for-each-other very dependable! Very important to know this!
"Chain reaction". Coil dictates all the rest in detector.

Conclusion;
not much sense in modding the rest of schematic if coil is not done well.
So i suggest to you to also take 4mV residual voltage at RX as MUST default. I made over 35 coils so far and that's how established this - pure rough experience, statistic and practice.
That was also the main reason why i insisted to get any similar original Tesoro - to see what is situation at those, originals.

I think those at Tesoro also realized this long time ago. Newer series (Tejon,Vaquero,Cibola etc..etc..) acts much different than older series.
I had chance only to check Tejon and Vaquero on soil. I hadn't chance to open one of those and trace out some stages. Pity. But also i measured coils which came with those. Much different coils than those we make here!
Inductance at newer coils are much different - looks more like Minelab coils for Musketeer and RH....some ~1mH for TX and ~14-16mH for RX!
Newer Tesoros now are not looking "limited". Meaning large items those detecting at much more distances!
I am looking forward to see some traced out schematic from newer Tesoro. Mostly am interested to see TX stage and RX front end.
Regards!

As would Ivconic, I would appreciate if anyone has information (partial schematic) of the front end (coil configuration and front-end amplifier) end of the latest Tesoro design (Cibola, Vaquero, etc.).

Thank you in advance.

[simonbaker]

I'm still just learning the ropes, and TGSL is perfect for beginners. I like to see how things work, and need to do the experiments for myself. Many more ideas for me to try to learn. Wish I was faster, but TGSL is my home for now. I will try to keep an eye on other schematics too, because I like the Lobo, and wonder what goes on inside the Vaquero.

Cheers,

-SB

[mikebg]

I applaud the efforts of dfbowers and simonbaker in their quest to produce the optimum TGSL. But.....

Are they beating a dead horse? And is it time to move on (at least for some of us)?

I refer to one of the last posts made by Ivconicin the original TGSL construction thread and I quote:

As would Ivconic, I would appreciate if anyone has information (partial schematic) of the front end (coil configuration and front-end amplifier) end of the latest Tesoro design (Cibola, Vaquero, etc.).

Monky,
Why do you mean that the best front end should be designed by Tesoro?
Did not know what company uses the best designers?
For now Pebe is about to design more sophisticated front end and I'm sure he will succeed because the more reason and considering the proper block diagram, rather than wasting time in blind experimenting.

However, 8 months along, none of the participants in this forum is not paid attention that in November I reinvented, redesigned, experimented and published the perfect front end or in ham terminology, most sensitive RFA (Radio Frequency Amplifier) for narrow band metal detector:
http://www.geotech1.com/forums/showt...eferrerid=2910

IMNHO (in my not humble opinion :-), over the next few years, can anyone come up with something better than my design, but it will hardly not only an ordinary professional designer. He should have in addition a hobby to design RXs, TXs and antennas for QRP amateur radio, to know how to receive nanovolt signals.

In January I reinvented and published the best operating (for now) TX for narrow band metal detector. Ham radio designers found theoretical shortcomings in my design and showed what could be improved, but I'm so happy with the measured parameters that will soon not do improvements.

Currently I'm about to rediscover and and redesign most sensitive BFO metal detector. Many believe that this primitive principle can not be upgraded more, but soon will be pleasantly surprised. Our Noise Induction Metal Detector is so sensitive and so discriminative that beats most complicated VLF metal detectors.
In conclusion I would say:
If a hobbist designer not lose time in the blind copying of factory projects, in useless testing and SPICE simulation, but first think and analyze how should the perfect block diagram work, then he do not have to make much. He just have to convert the perfect block diagram in an operating circuit diagram.

[dfbowers]

I applaud the efforts of dfbowers and simonbaker in their quest to produce the optimum TGSL. But.....

Are they beating a dead horse? And is it time to move on (at least for some of us)?

I refer to one of the last posts made by Ivconicin the original TGSL construction thread and I quote:

As would Ivconic, I would appreciate if anyone has information (partial schematic) of the front end (coil configuration and front-end amplifier) end of the latest Tesoro design (Cibola, Vaquero, etc.).

Thank you in advance.

No.. not beating a dead horse at all. Maybe for some, but there are other reasons why this thread is still alive. Many of us who are not designers are still learning from what others have done. Plus, a forum for anyone who wants to build this project can be successful. Sorry, we are not all at the head of the class, but that doesn't mean that we cannot contribute.
If we can help anyone else, then we have contibuted. All of our goals are not the same. If one wants to build the optimum Tx circuit, have at it!
Will you find more than me? I doubt it if you spend your life in the lab.

My goal was simply the arrive at the satisfaction that I can produce something with my own hands that equals a store-bought unit, and thoroughly understand how it works to keep it functional.
I commend all who have contributed!

That being said, we might be reaching the point of diminishing returns on some points.. Simon, we need to make you an etched PCB?

Don

[simonbaker]

No.. not beating a dead horse at all. Maybe for some, but there are other reasons why this thread is still alive. Many of us who are not designers are still learning from what others have done. Plus, a forum for anyone who wants to build this project can be successful. Sorry, we are not all at the head of the class, but that doesn't mean that we cannot contribute.
If we can help anyone else, then we have contibuted. All of our goals are not the same. If one wants to build the optimum Tx circuit, have at it!
Will you find more than me? I doubt it if you spend your life in the lab.

My goal was simply the arrive at the satisfaction that I can produce something with my own hands that equals a store-bought unit, and thoroughly understand how it works to keep it functional.
I commend all who have contributed!

That being said, we might be reaching the point of diminishing returns on some points.. Simon, we need to make you an etched PCB?

Don

Never!

Yeah, it's inevitable. Also, I don't think I can handwire one more PCB!!!

Which exact layout is question. I'd like to start with the layout I'm currently using for comparison. Then stock TGSL.

The next (last?) one I'm currently wiring is my "TGSInsane!" model and probably won't work worth a damn, but I'm curious....

-SB

[Monk]

Monky,
Why do you mean that the best front end should be designed by Tesoro?
Did not know what company uses the best designers?

Yes my post is too limiting. Any and all front end designs are more than welcome.

In conclusion I would say:
If a hobbist designer not lose time in the blind copying of factory projects, in useless testing

Factory projects have great value to me because these people make their living in stake their lives on being able to produce products that survive in a very competitive marketplace. My first inclination is to take their efforts over the efforts of some unknown, no offense.

and SPICE simulation,

With respect to simulation tools they are is only as good as the ability of the designer to model, i.e. the designers understanding of the design problem. For myself personally I think they are a very valuable tool and help me in many designs including developing a better understanding the strengths and weaknesses of some of the designs presented here.

but first think and analyze how should the perfect block diagram work, then he do not have to make much. He just have to convert the perfect block diagram in an operating circuit diagram.

Having been to many years in the design world this is in my opinion a typical upper level management view of the design process. In other words it's a good example of The Pareto principle (also known as the 80-20 rule,[1] the law of the vital few, and the principle of factor sparsity) states that, for many events, roughly 80% of the effects come from 20% of the causes.

When you have a field proven design that you have built at least 10 of them show its superiority over existing detectors then I will be more than willing to consider your designs.

That being said, we might be reaching the point of diminishing returns on some points.. Simon, we need to make you an etched PCB?

Exactly the point I was trying to make. Anyone with any type of electronics experience especially with low-level signal processing could tell you that a perf board construction will most likely not perform to the same level as a well laid out at PC board. And if you want to take it to three steps further use a multilayered with ground and power planes, copper pours, guard traces, etc., etc. 98% of recent exchanges between you would probably have been eliminated with a halfway decent PC board.

I understand that most to come here have a minimal amount of electronic design and construction knowledge. And the exchanges between yourself and simon are probably very helpful to many people. But for myself personally your recent exchanges with Simon has been beating a dead horse (his board construction).

My goal was simply the arrive at the satisfaction that I can produce something with my own hands that "equals a store-bought unit", and thoroughly understand how it works to keep it functional.

You have the same goals as mine. The only difference may be the interpretation of "equals a store-bought unit". I would like a store-bought unit takes advantage of the latest technology and is considered a top-notch detector.

I commend all who have contributed!

So do I. but unfortunately not all can contribute equally as defined by individual perceived value.

[Qiaozhi]

With respect to simulation tools they are is only as good as the ability of the designer to model, i.e. the designers understanding of the design problem. For myself personally I think they are a very valuable tool and help me in many designs including developing a better understanding the strengths and weaknesses of some of the designs presented here.

I agree. Simulation is very useful to aid understanding, and allows easy experimentation with new ideas, or exploration of an existing design. One of simulation's greatest strengths is the ability to run Monte Carlo and Corner analyses, which is something that's virtually impossible to do on the workbench. Also, the ability to "blow up" a circuit and simply re-start it without having to fault-find and solder in new components. Of course, it's just another engineering tool, and needs to work alongside the actual circuit construction, not act as a substitute. If you use it in the right way, it's a great asset.

[Monk]

http://www.geotech1.com/forums/showp...&postcount=264

I'll post here in my own thread so not clutter main TGSL troubleshooting thread.

Personally I'm interested in trying to see difference in PCBs because... I feel like it. No, I'm intentionally reinventing the wheel because often the rules become progressively restrictive and while excellent practice (which I wouldn't contest), I want to see what really are the critical factors in this particular case. It is sometimes important to not over-design, especially if you want to save money. Commercial manufacturers often do research (I would guess) to see how far they can cut corners to save cost. So the most beautiful robust designs are not always the end point.

To say a handwired pcb can't succeed may be too dismissive -- much of the TGSL is an extremely low frequency circuit where one might expect board layout effects to be much less important. Also, there is evidence that the dominent noise is from the coil acting as an antenna, not stray pickup because you didn't trim a solder blob low enough.

For me, it's fun to troubleshoot and investigate.

Granted, boring to others.

I write because it helps me think. So I'll post progress here even if not earth-shattering news, and people can ignore this thread. If anything useful pops up, I'll post it in TGSL Troubleshooting thread.

simon,

Please believe me when I say I did not mean to ridicule or belittle your efforts in any way. Your efforts (postings) are informative and entertaining. It takes a lot of time to do what you're doing and I applaud your efforts.

My posting was intended to try and stir some creative "ashes" in hopes of creating some new design fire. From my perspective these foums have lost some of the zip that they previously had. We've had an appear to have lost some very good posters along the way. All entities have a lifecycle. I would say this forum has reached middle age, bordering on old-age. Again that's just my opinion. Keeping things interesting is not easy. It takes time and a lot of effort. The efforts of one or two people can make a big difference.

I would hope that I can add something to these forms as time goes on. I have made a significant investment in materials and most recently time and effort. I am typically interested in maximizing front end performance as I hope to marry a high-level front end with a high-level backend in the form of a microprocessor. Something with some reasonable capabilities starting with dspic33 and a 16 Bit, 1 MEG/sample/sec A/D. To be used in both IB and PI designs. Couple this with some reasonably high level PCB layout and an understanding of grounding and shielding and I would hope we could create something that "equals a store-bought unit".

Peace....

[moodz]

http://www.geotech1.com/forums/showp...&postcount=264

I would hope that I can add something to these forms as time goes on. I have made a significant investment in materials and most recently time and effort. I am typically interested in maximizing front end performance as I hope to marry a high-level front end with a high-level backend in the form of a microprocessor. Something with some reasonable capabilities starting with dspic33 and a 16 Bit, 1 MEG/sample/sec A/D. To be used in both IB and PI designs. Couple this with some reasonably high level PCB layout and an understanding of grounding and shielding and I would hope we could create something that "equals a store-bought unit".

Peace....

.... I agree also ... If someone can point to a functional block diagram I will code it up with 24 bit ADC and DACs.

Moodz.

[simonbaker]

http://www.geotech1.com/forums/showp...&postcount=264



simon,

Please believe me when I say I did not mean to ridicule or belittle your efforts in any way. Your efforts (postings) are informative and entertaining. It takes a lot of time to do what you're doing and I applaud your efforts.

My posting was intended to try and stir some creative "ashes" in hopes of creating some new design fire. From my perspective these foums have lost some of the zip that they previously had. We've had an appear to have lost some very good posters along the way. All entities have a lifecycle. I would say this forum has reached middle age, bordering on old-age. Again that's just my opinion. Keeping things interesting is not easy. It takes time and a lot of effort. The efforts of one or two people can make a big difference.

I would hope that I can add something to these forms as time goes on. I have made a significant investment in materials and most recently time and effort. I am typically interested in maximizing front end performance as I hope to marry a high-level front end with a high-level backend in the form of a microprocessor. Something with some reasonable capabilities starting with dspic33 and a 16 Bit, 1 MEG/sample/sec A/D. To be used in both IB and PI designs. Couple this with some reasonably high level PCB layout and an understanding of grounding and shielding and I would hope we could create something that "equals a store-bought unit".

Peace....

I appreciate that Monk!

You know what? I agree with your perceptions. It's kind of like any specialty niche -- surfing, rock climbing, music, sports -- the "cool" people start it, the pioneers who make up their own rules, the geniuses, the brilliant, the talented -- and then it becomes diluted by the wannabees, the bean counters, the technicians who come in later to enjoy the success and glory carved out by the pioneers and who over-analyze and perfect it rather than creating some new niche of their own.

I know I'm not in the same class as the "hard men" who built this website with the shadowy side of reverse-engineering and midnight hunting in questionable areas. They're cool guys. They go straight for results, grab the low-hanging fruit, take the risks, and keep moving.

If I've got anything to offer it is doggedly trying to dissect what the gurus created and try to pass on anything I learn to other enthusiasts at my level. I also by nature question everything. In this field there is a lot of conventional wisdom which although often true, may be off the mark in a particular case.

But the old crowd has more charisma than I can bring, I admit it's more entertaining to listen to Ivconic lambasting a newbie than me holding someone's hand, so I will try to keep most of my stuff in my thread.

I applaud your efforts to ignite a new spark. You may be fighting natural evolution. Or it may be simply the accident of what personalities are currently inhabiting this board. Ten years from now, it may have a whole different atmosphere because of who's here.

I have a number of projects in mind, some quite similar to yours. I would like to make a "digital" TGSL -- sample the RX waveform and use digital processing from there on. Ideally drive the TX coil with digital output. Aziz really already did that and then became disenchanted and moved on to digital PI detectors. But I'd like to take a crack at it from a microprocessor point of view.

All I know is I'm obsessed with and entertained by messing around with metal detector circuits. Can't help it. Can't stop. (unless life intervenes).

And no offense taken for sure!

Cheers,

-SB

[ivconic]

"...When you have a field proven design that you have built at least 10 of them show its superiority over existing detectors then I will be more than willing to consider your designs...."

Thumbs up Monk! I like the way you'r thinking!

[Nicko]

.... I agree also ... If someone can point to a functional block diagram I will code it up with 24 bit ADC and DACs.

Excuse me for butting in here, but 24 bit? With 5V rails thats an LSB of 0.3 of a MICRO VOLT. Even 16 bits gives an LSB of 80uV - well below the noise floor of most front ends.

That is very very rarefied analogue design territory. Even Jim Williams probably wouldn't go there.

A few years ago, in about 2001, I built a test bed using ADSP2191 DSPs and a 16 bit 1MSPS ADC/DACs, and rapidly came to the conclusion that it was tricky. It worked, but the mix of analogue & digital at those frequencies and noise levels was a combination that was subtle & quick to anger.

Now, it seemed that the correct engineering approach was to let the analogue side do what it does best, and to reduce the requirements on the digital side such that the A/D interface was kept as simple as possible.

Basically, what is of interest is generally subtle changes over a small range - its better to get the analogue side to present only that window, and to work within that only on the digital side. This radically simplifies the technical requirements for the A/D interface and thus the system as a whole.

Cheers

[moodz]

24 bits is OK

Nothing wrong with using a high resolution ADC ...the higher bits means less amplification on the front end. More gain means more noise and less dynamic range. Once the signal is digitised the noise is 'frozen'.
The first amplifier in any analogue signal chain will determine the noise performance for any subsequent analogue processing.
I find it amazing that some of you analogue guys are so attached to your single ended unbalanced noisy shielded frontends. Balanced modulators / demods , amplifiers, impedance matching etc have been around for as long as I can remember in audio and RF ( and thats as long as the little glass tubes with heaters in em ) but this 'technology' just has not made it into metal detectors.
A 'proper' front end for an ADC is not very difficult ... professional RF/audio has been doing since the last century ... with one critical factor ... the first amp or balanced mod input is differential as this is what will set the noise floor of subsequent processing operations.
On single ended inputs why do you need Faraday sheilding ? ... because the input circuit is susceptable to coupled voltage / noise fields and capacitive coupling to other conductors like the ground ... a Faraday shield is an electrostatic shield not a magnetic shield ... so why design a circuit that detects voltage fields when it is supposed to detect magnetic fields ... sheeeesh ... nough said already.

[simonbaker]

Nothing wrong with using a high resolution ADC ...the higher bits means less amplification on the front end. More gain means more noise and less dynamic range. Once the signal is digitised the noise is 'frozen'.
The first amplifier in any analogue signal chain will determine the noise performance for any subsequent analogue processing.
I find it amazing that some of you analogue guys are so attached to your single ended unbalanced noisy shielded frontends. Balanced modulators / demods , amplifiers, impedance matching etc have been around for as long as I can remember in audio and RF ( and thats as long as the little glass tubes with heaters in em ) but this 'technology' just has not made it into metal detectors.
A 'proper' front end for an ADC is not very difficult ... professional RF/audio has been doing since the last century ... with one critical factor ... the first amp or balanced mod input is differential as this is what will set the noise floor of subsequent processing operations.
On single ended inputs why do you need Faraday sheilding ? ... because the input circuit is susceptable to coupled voltage / noise fields and capacitive coupling to other conductors like the ground ... a Faraday shield is an electrostatic shield not a magnetic shield ... so why design a circuit that detects voltage fields when it is supposed to detect magnetic fields ... sheeeesh ... nough said already.

Hi moodz:

question: dfbowers recently experimented with coil connections for the TGSL where the RX coil is not grounded at all, only the shield around the cable is grounded (and one of the TX wires). Does that meet your requirements for differential inputs?

Also: what do you think would be the minimal, easiest microprocessor for trying a digital TGSL, meaning sampling the 15 kHz RX signal directly and doing everything digital after that. Low power would be very important too.

Regards,

-SB

[Aziz]

Hi all,

Hi moodz:

question: dfbowers recently experimented with coil connections for the TGSL where the RX coil is not grounded at all, only the shield around the cable is grounded (and one of the TX wires). Does that meet your requirements for differential inputs?

Also: what do you think would be the minimal, easiest microprocessor for trying a digital TGSL, meaning sampling the 15 kHz RX signal directly and doing everything digital after that. Low power would be very important too.

Regards,

-SB

this task is destined for a digital laptop TGSL version.
Just feed the X and Y demodulated signals into left and right input channel of the sound card. Then a digital lock-in amp can be done on the X/Y-signals.
Hardware:
- LC-oscillator (trivial)
- X/Y demodulator clock generation (trivial)
- (Buffered) RX in-phase signal X (trivial)
- (Buffered) RX quadrature signal Y (trivial)

The hardware and the software isn't much complex. You need only one sound card (internal). The output of the sound card can be used for audio signalling. But the performance will be outstanding!!!!

Aziz