TGSL Tuning & troubleshooting

[satdaveuk]

coil checker

This one here, put it all together with a 2n2907 with a TGSL working coil 6mh with 15nf cap and its not oscillating, the transistors were ones that were pulled out of the TGSL a while back while doing tests, although they did read to spec before installing them, but never can be sure with old bits.
Can you see any mistakes on the circuit, the pcb is perfect double checked it.
The only difference was C1 and C6 fitted 47000pf instead of 40,000pf.

Thanks in advance for your opinion

Regards

[simonbaker]

This one here, put it all together with a 2n2907 with a TGSL working coil 6mh with 15nf cap and its not oscillating, the transistors were ones that were pulled out of the TGSL a while back while doing tests, although they did read to spec before installing them, but never can be sure with old bits.
Can you see any mistakes on the circuit, the pcb is perfect double checked it.
The only difference was C1 and C6 fitted 47000pf instead of 40,000pf.

Thanks in advance for your opinion

Regards

I for one don't know that circuit. Of course I'd just use the TGSL oscillator since I know that one, although it too has trouble oscillating as the coil inductance gets smaller.

Regards,

-SB

[simonbaker]

This one here, put it all together with a 2n2907 with a TGSL working coil 6mh with 15nf cap and its not oscillating, the transistors were ones that were pulled out of the TGSL a while back while doing tests, although they did read to spec before installing them, but never can be sure with old bits.
Can you see any mistakes on the circuit, the pcb is perfect double checked it.
The only difference was C1 and C6 fitted 47000pf instead of 40,000pf.

Thanks in advance for your opinion

Regards

Hi satdaveuk:

For fun I hacked an LTSpice sim of your circuit, leaving out the second JFet stage which I don't think affects the oscillator.

It did oscillate, albeit a little ugly.

Maybe the transistors are "duff", or your JFet has some extreme parameters?

-SB

[satdaveuk]

Hi satdaveuk:

For fun I hacked an LTSpice sim of your circuit, leaving out the second JFet stage which I don't think affects the oscillator.

It did oscillate, albeit a little ugly.

Maybe the transistors are "duff", or your JFet has some extreme parameters?

-SB

Thanks simon
ITS actually called a universal coil checker so should be good for all are coils in metal detecting, by all accounts if fitted with the bfr99 which is tr1 its takes you from a few hz right up to the uhf bands 100mhz or so, or at least thats what i was told.
Where i got this from god knows but looks as if it came from a old mag at some point.
While you have it on spice can you check with C1 and C5 at 40,000pf, it may make it less ugley lol
Where you put c3 it should be c1 simon.
As i said early though ive fitted 47,000pf in both c1 and c5 because i didnt have 40,000pf to hand
Regards
Dave

[dfbowers]

Thought of a small addition to these documents -- revision tracking. It would be useful to put a small section with revision descriptions and a revision number on each revision of this extremely useful (TGSL 101) document near the front of the document.

I realized this would be useful because I just spent a half hour trying to convince myself I had the latest version (and locating the latest -- the document releases seem to be divided between two threads, "Tesoro Golden Sabre" and "TGSL Tuning & Troubleshooting").

I had remembered a recent revision of the grounding diagram for the RX cable. If my documents had mentioned that revision, I would have been at peace that I had what I needed.

Cheers!

-SB

Good idea. I think that the only significant change was to fix the error you mentioned. I will see if I can post a new copy with the revision history.

Thanks
Don

[mikebg]

INCORRECT SPICE
The circuit can't oscillate if the transit frequency of Q1 is high, stray capacitance C is small and coil resistance r is large. Pay attention that transit frequency of BFR 99 is high relatively to 2N2907.

[satdaveuk]

universal coil checker now up and running

Hi Simon and Mike
I replaced the 2n2907 with the metal type has a slight differences on the spec sheet against the T018 package but not convinced that was the source of the problem, it was proberly a duffer because as explained it had been in my TGSL at some point.
Anyway the low down on this oscillator is its solid as a rock from switch on far as frequency drift, checked a few different coils with it changing CX etc on them and at this point in time seems very accurate which means its going to be a god sent when making or checking coil frequencys.
I know exactly what im going to do, shall fit a 5 pin regular socket on one end switched to be able to go from TX to RX without picking up a iron, then along side it sticking out will be a vero board to be able to perfect each coil far as L/C, On the other end a BNC socket so I can feed directly into the scope/frequency counter via flylead instead of messing about with probes.
Running at 16.750khz picture on the scope in my untidy workshop plus the PCB .
Note how ugley the waveform is, im going to try and get hold of the proper oscillator TR and order some 40,000pf caps maybe that will correct it.
Anyway once ive boxed the project will post it under tools for everone ok.
Sorry about the qualityof the pictures and spelling but it is 6am not been to bed yet

All the best
Dave

[der_fisherman]

Folling revisions properly

Thought of a small addition to these documents -- revision tracking. It would be useful to put a small section with revision descriptions and a revision number on each revision of this extremely useful (TGSL 101) document near the front of the document.

I realized this would be useful because I just spent a half hour trying to convince myself I had the latest version (and locating the latest -- the document releases seem to be divided between two threads, "Tesoro Golden Sabre" and "TGSL Tuning & Troubleshooting").

I had remembered a recent revision of the grounding diagram for the RX cable. If my documents had mentioned that revision, I would have been at peace that I had what I needed.

Cheers!

-SB

I could not agree more, its something I brought up in the IGSL blog some months ago that nobody was interested in......

You really need to start with a schematic design software that allows revision marking, that is fully compatible to a PCB CAD program. One that can produce either SMD or "normal" parts PCBs.

I personally use Eagle, but any program(s) that are either free or low cost would be acceptable to me. The free version of Eagle, which is absolutely complete with ALL the fantastic libraries, except limited to a particular sized PCB, would be great.

Eagle takes a time to get used to, a few days of reading the help files will do it, but once you are over that hurdle, I can heartily recommend it.

There are other programs around that do similar things, KICAD is supposed to be really good (I've never used it), open source and therefore totally free.

There is even a great program for Breadboards (which some of us use for testing) that converts the Breadboard into a proper schematic and/or PCB (and has a company that specializes in making the PCBs for you). Which incidentally are Eagle compatible, a further reason to use them both as it will fit in even better with our work here.

The various softwares can be viewed and downloaded at the following websites:-

http://fritzing.org/

http://www.cadsoftusa.com/?lang=de

Up to now we appear to use (for IGSL at least) ExpressSch for the schematics and Layout for the PCB, neither of which are compatible with each other, so we often get updated (un-numbered revisions) PCBs weeks before the schematics get corrected......IF they get corrected!!!

If I can be of any further help just ask, but this is the way to go to control revisions of schematics and keyboards in the future, and to be able to integrate Breadboard testing as well as PCBs for either normal or SMD components.....

Regards

Andy

[simonbaker]

Thanks simon
ITS actually called a universal coil checker so should be good for all are coils in metal detecting, by all accounts if fitted with the bfr99 which is tr1 its takes you from a few hz right up to the uhf bands 100mhz or so, or at least thats what i was told.
Where i got this from god knows but looks as if it came from a old mag at some point.
While you have it on spice can you check with C1 and C5 at 40,000pf, it may make it less ugley lol
Where you put c3 it should be c1 simon.
As i said early though ive fitted 47,000pf in both c1 and c5 because i didnt have 40,000pf to hand
Regards
Dave

Sorry I was too lazy to re-label the components! I modified an old circuit I had.

Your waveform looks similar to the Spice sim voltage taken at the corresponding point.

Perhaps the purpose of th oscillator was simply a stable frequency, not a clean sine wave. Usually "over-driven" oscillators are more stable.

I tried the 40,000 PF value -- still hideous. I guess that's how it was designed.

Regards,

-SB

[simonbaker]

I could not agree more, its something I brought up in the IGSL blog some months ago that nobody was interested in......

You really need to start with a schematic design software that allows revision marking, that is fully compatible to a PCB CAD program. One that can produce either SMD or "normal" parts PCBs.

I personally use Eagle, but any program(s) that are either free or low cost would be acceptable to me. The free version of Eagle, which is absolutely complete with ALL the fantastic libraries, except limited to a particular sized PCB, would be great.

Eagle takes a time to get used to, a few days of reading the help files will do it, but once you are over that hurdle, I can heartily recommend it.

There are other programs around that do similar things, KICAD is supposed to be really good (I've never used it), open source and therefore totally free.

There is even a great program for Breadboards (which some of us use for testing) that converts the Breadboard into a proper schematic and/or PCB (and has a company that specializes in making the PCBs for you). Which incidentally are Eagle compatible, a further reason to use them both as it will fit in even better with our work here.

The various softwares can be viewed and downloaded at the following websites:-

http://fritzing.org/

http://www.cadsoftusa.com/?lang=de

Up to now we appear to use (for IGSL at least) ExpressSch for the schematics and Layout for the PCB, neither of which are compatible with each other, so we often get updated (un-numbered revisions) PCBs weeks before the schematics get corrected......IF they get corrected!!!

If I can be of any further help just ask, but this is the way to go to control revisions of schematics and keyboards in the future, and to be able to integrate Breadboard testing as well as PCBs for either normal or SMD components.....

Regards

Andy

It would also really help if this Geotech bulletin board had file areas where project files could be located and most recent versions maintained. Then you just go there and know you get the latest. Qiaozhi maintains one sticky thread for TGSL with Ivconic's documentation, but a file repository would be even better. The problem would be: who is allowed to upload?

-SB

[satdaveuk]

coil checker

Hi
All files have been put into tools section, does the job, i just got back from the workshop after nulling a pair of coils with it, copied the 400mv null from the output with factory coil and to my suprise it was bang on considering I havent tweeked the frequencys yet, more than impressed.
When i got time im going to feed this oscillator through the TGSl and see what results it holds.
All the best lads
Dave

[satdaveuk]

not so sensitive and noisy

Hi everyone
As you all know ive been pleased the way the TGSL has turned out but today i desided to do a bit more tweeking using a factory coil 5.7mh/6.2mh
Up till now i been running the TX at 4.389khz using .22-.022mfd as per ivonics circuit, which the components were measured with what im sure is quite a accurate LCR meter.
Anway playing about with the same stamped capacitors but reading lwr capacitance I ended up with 14.67khz.

Now my findings are its less sensitive by over 1inch with a english £1 coin and its picking up all the electrical noise which is one problem I dont normally have with this machine

Just got me thinking because could this be the problem that many builders of the TGSL have experienced? Maybe the TGSL in general operates better on the lwr frequency of say 14-14.3khz.
I do know that in the old days thats why I settled with the lwr frequency because of better and more stable results, but at that time put it down as proberly had other problems else where, or maybe in fact I still do.

Something else to feast on Simon, when time permits shall do other tests if no luck shall try droping TX even to say 14khz and see what happens.

All the best
Regards

[golfnut]

Hi satdave, You will probably find that a uk 10p gives a great clear signal with more distance. !!



Give the non ferrous GB pot a small tweak and you will get more sens on the UK pound, it will still GB on soil, not so good on ferrite. Which is the right way round as we dont search ferrite for gold!


Steve

[satdaveuk]

Hi satdave, You will probably find that a uk 10p gives a great clear signal with more distance. !!



Give the non ferrous GB pot a small tweak and you will get more sens on the UK pound, it will still GB on soil, not so good on ferrite. Which is the right way round as we dont search ferrite for gold!


Steve

thanks golfnut
Took it outside on the test bed
I tryed adusting ground balance etc but still not right, tomorow ill get the scope on it and see if its effected the nulling etc because thats the typical symptoms im getting although its a factory coil theres other factors that can offset it.
Also ive got Ivonics preset mod on mine so if that is the problem maybe a tweek of that will do the trick.
Far as picking up and disc my TGSL was working as good as any detector I have here, just wanted to bring it to ivonics spec to see if there was depth in the soil to be had.
The frequency it was on was picking up siver sixpence 19cm £1 coin 21cm £2 coin 24cm thats in soil it was the same in air test.
So going by the 30cm thats been mentioned alot on here then there loosing 50% in the ground which gives them average of 15cm, in real terms mine is a good performer although still feel theres more to be had.

All the best

[simonbaker]

Hi everyone
As you all know ive been pleased the way the TGSL has turned out but today i desided to do a bit more tweeking using a factory coil 5.7mh/6.2mh
Up till now i been running the TX at 4.389khz (assume you meant 14.389) using .22-.022mfd as per ivonics circuit, which the components were measured with what im sure is quite a accurate LCR meter.
Anway playing about with the same stamped capacitors but reading lwr capacitance I ended up with 14.67khz.

Now my findings are its less sensitive by over 1inch with a english £1 coin and its picking up all the electrical noise which is one problem I dont normally have with this machine

Just got me thinking because could this be the problem that many builders of the TGSL have experienced? Maybe the TGSL in general operates better on the lwr frequency of say 14-14.3khz.
I do know that in the old days thats why I settled with the lwr frequency because of better and more stable results, but at that time put it down as proberly had other problems else where, or maybe in fact I still do.

Something else to feast on Simon, when time permits shall do other tests if no luck shall try droping TX even to say 14khz and see what happens.

All the best
Regards

Hi satdaveuk:

First, what kind of factory coil? D-D or concentric? What size?

Did you say using lower capacitors you got a higher TX frequency? That sounds backwards, but maybe you meant same caps but different coil.

I have always believed that the relationship of RX resonant freq to TX oscillator freq is more important than actual values. But there may be another factor. If you have local noise sources that are right at 14.67 kHz, then dropping the TX oscillator can be a good way to dodge the noise.

Changing the TX freq without changing the RX resonant freq will slightly change the gain, and even more slightly change the phase shift.

Raising the TX freq should increase the gain, so that could also be why you hear more chatter.

But then wouldn't we expect better depth with more gain (you ask)??? Yes, except noise is a total bugger (pretending I'm a Brit) with the TGSL and can very efficiently reduce depth -- it seems to be due to the final "noise gate" filter just before the audio section.

Whether increased gain can actually reduce depth due to the "chatter filter" effect I'm not sure. But certainly if your TX frequency goes nearer to a noise frequency, you'll pick up more noise and get less depth.

Makes me think that maybe we should consider "tunable" oscillators for our TGSL MDs so we can evade certain noise sources.

Regards,

-SB