TGSL Tuning & troubleshooting

[simonbaker]

Hi all, I just put together a SilverDog kit IGSL TGSL and while running some initial checks I found that my -5v rail was only -3.2V.

I desoldered the anode of D4 and it measured -6.2v.

I resoldered it and started looking for shorts - found none.

I measured the resistance to gnd and other rails and there was nothing that indicates a short or semi-short.

I then started lifting IC's to see if one of them was dud and found that when I removed U4, the supply rail changed to -6.2V so I figured it was the IC.

To confirm this, I swapped U4 with U7 (they're the same) and the supply rail went back to -3.2V which proves that the IC is fine and there is some other reason.

So my question is this... Is this behavior normal? I'm suspecting that the -ve voltage generator that is supplied by the 7.3 kHz signal isn't capable of providing the necessary current, which may or may not be a fault.

Is -3.2V while the board is operating normal?

I'm working off this schematic FWIW:

http://silverdog.co.uk/shop/image/data/IGSL/IGSLTGSLcoils.JPG

Thanks

p.s. The board is drawing about 100ma from the 12V supply which I figure is normal.

I would suspect transistors Q3 and Q4, make sure they are correct (not leads reversed) and are turning on and off fully (saturating). You could even try replacing in case one is out of spec. If the transistors look good, double-check resistors R25 through R28, make sure they are not wrong. If not wrong, you could try using smaller resistors to see if that pumps a little more current out.

I think 100 mA sounds about right for IGSL, at least near the upper range.

-SB

[Farside]

Hi all, I just put together a SilverDog kit IGSL TGSL and while running some initial checks I found that my -5v rail was only -3.2V.

I desoldered the anode of D4 and it measured -6.2v.

I resoldered it and started looking for shorts - found none.

I measured the resistance to gnd and other rails and there was nothing that indicates a short or semi-short.

I then started lifting IC's to see if one of them was dud and found that when I removed U4, the supply rail changed to -6.2V so I figured it was the IC.

To confirm this, I swapped U4 with U7 (they're the same) and the supply rail went back to -3.2V which proves that the IC is fine and there is some other reason.

So my question is this... Is this behavior normal? I'm suspecting that the -ve voltage generator that is supplied by the 7.3 kHz signal isn't capable of providing the necessary current, which may or may not be a fault.

Is -3.2V while the board is operating normal?

I'm working off this schematic FWIW:

http://silverdog.co.uk/shop/image/data/IGSL/IGSLTGSLcoils.JPG

Thanks

p.s. The board is drawing about 100ma from the 12V supply which I figure is normal.

[simonbaker]

OK.

The point I was trying to make was that the DC resistance will ALWAYS be the lowest as the voltage and current are always in phase in a DC circuit. Therefore even a frequency of 1 Hz will demonstrate an impedance higher than DC resistance (albeit only a tiny bit higher).

Luckily for us here were work our machines at quite low frequencies therefore many of the effects you mentioned do not come into measurable play....and we can usually safely ignore them.....(Helpful to me with my amateur level of analog electronics!)

Furthermore, as the current "lags" the voltage in a mainly inductive circuit, it also leads in a mainly capacitive circuit, but they are fully in phase with each other (current & voltage) when the circuit is at resonance, which gives us also maximum current for that capacitive/inductive circuit combination.....something needed for best search depth in a Tx coil. (this also gives optimum battery life with best power output.)

So it is VERY important for us all to keep this in mind.

Here is a web site that shows the above in a simple animated fashion:-

http://www.kwantlen.ca/science/physics/faculty/mcoombes/P2421_Notes/Phasors/Phasors.html

Regards

Andy

PS. I love all your posts Simon, my posts are there to "add a bit on" to some of yours and to make things clearer for myself and others here.....never to upset you personally. Not as criticism, only discussion for all......

Good thoughts, and never worry about questioning any of my ideas, it is always helpful to discuss and debate -- I have a tendency to always rebut, but I also believe in totally open mind, correcting mistakes, realizing theories can change and evolve and only represent a way to look at experimental data and aren't the "truth". So I enjoy discussions no matter how lengthy. Your contributions are always welcome and valued as are everyone's (IMO).

Cheers,

-SB

[der_fisherman]

Inductive and capacitive loads

I see what you are saying, but there are ways to estimate both the resistance and impedance components of a coil at a given frequency (in other words, estimating the reactance X) by looking at the response. The AC resistance and the impedance should make up the reactance.
The AC resistance can be different from the DC resistance as people have discussed, partly due to radiation, partly due to skin effect (not much at VLF freq), and partly due to "proximity effect" (bundle of wires), the last being the one I'm curious about.
I'll try to get back to an experiment to see how much the AC resistance really differs from the DC resistance. It is mainly important to those who are interested in hi-Q TX oscillators.
I think that was what I was trying to get at.
-SB

OK.

The point I was trying to make was that the DC resistance will ALWAYS be the lowest as the voltage and current are always in phase in a DC circuit. Therefore even a frequency of 1 Hz will demonstrate an impedance higher than DC resistance (albeit only a tiny bit higher).

Luckily for us here were work our machines at quite low frequencies therefore many of the effects you mentioned do not come into measurable play....and we can usually safely ignore them.....(Helpful to me with my amateur level of analog electronics!)

Furthermore, as the current "lags" the voltage in a mainly inductive circuit, it also leads in a mainly capacitive circuit, but they are fully in phase with each other (current & voltage) when the circuit is at resonance, which gives us also maximum current for that capacitive/inductive circuit combination.....something needed for best search depth in a Tx coil. (this also gives optimum battery life with best power output.)

So it is VERY important for us all to keep this in mind.

Here is a web site that shows the above in a simple animated fashion:-

http://www.kwantlen.ca/science/physics/faculty/mcoombes/P2421_Notes/Phasors/Phasors.html

Regards

Andy

PS. I love all your posts Simon, my posts are there to "add a bit on" to some of yours and to make things clearer for myself and others here.....never to upset you personally. Not as criticism, only discussion for all......

[simonbaker]

Hi Simon, in that previous post you appear to have forgotten that at DC you have a DC resistance. Any AC signal in a coil means that the coil does not have resistance, it has impedance. You mentioned a 400% difference if I remember correctly.....
The impedance is also given in Ohms which causes some people to forget that these "Ohms" are slightly different!!
The impedance varies with the frequency, a higher frequency gives a higher impedance with the same coil......
Hoping that I have just "jogged" your memory a bit.......this is why changing frequency changes many other things as well as the null of a search coil head for example.

Have fun.

Have a great day.

Regards
Andy

I see what you are saying, but there are ways to estimate both the resistance and impedance components of a coil at a given frequency (in other words, estimating the reactance X) by looking at the response. The AC resistance and the impedance should make up the reactance.

The AC resistance can be different from the DC resistance as people have discussed, partly due to radiation, partly due to skin effect (not much at VLF freq), and partly due to "proximity effect" (bundle of wires), the last being the one I'm curious about.

I'll try to get back to an experiment to see how much the AC resistance really differs from the DC resistance. It is mainly important to those who are interested in hi-Q TX oscillators.

I think that was what I was trying to get at.

-SB

[joop]

Thanks for sharing Golfnut!

Maybe will give it a try anyway, would be an interesting experiment. Get myself a good coil and make the TX power adjustable, see what happens when you crank up the power.

Okay on the RX frontend. Will look into that...

[golfnut]

Getting the power is no problem - Li Ion or Li po have plenty to spare.


Keeping it out of the Rx front end is though. You cant null all of that away.

So then you may req a higher dynamic range Rx.

Plus if you throw masses of energy into the soil - you magnetise it and it blocks your 'Vision' for small things deeper in the soil - for want of a better description.


A heap of trade offs concatenated.

S

[joop]

To double the depth, you would need to increase the TX power by 64 times. But this brings with it a whole host of other problems.

Thanks! Judging from some of my simulations I can get a 10 percent depth increase by optimizing the TX oscillator in the TGSL. Not much, but worth trying.

64 times the power gives about 2 Watts for the oscillator. So we need some kind of large, hip-mount battery pack (to heavy to mount on the detector stem). Also, we need to re-design the transmit coil. Not impossible, quite a challenge.

[Qiaozhi]

Short question about the oscillator power delivered to the TX coil. In the current TGSL is see about 30-40 milliWatts of power in simulations. What kind of power levels do we need to double the ground penetration depth?

To double the depth, you would need to increase the TX power by 64 times. But this brings with it a whole host of other problems.

[joop]

TX coil power considerations

Short question about the oscillator power delivered to the TX coil. In the current TGSL is see about 30-40 milliWatts of power in simulations. What kind of power levels do we need to double the ground penetration depth?

[golfnut]

Null change over soil

All, we generally null in a garage or whatever.

When using these detectors over AVERAGE soil is the null level the same. ?

OR

should we null over AVERAGE soil. - As the fresh air case is of little consequence!


S

[der_fisherman]

Resistance & Impedance

Looks very nice, similar to experiments habitbraker and I were playing with. http://www.geotech1.com/forums/showp...&postcount=356
I wonder if it is a little difficult to get L2 and L3 with those low resistance values.
Also the low inductance of L2 implies fewer coil turns, and since magnetic field is proportional to current * turns (I think??), are we really getting as much bigger field as we think even though the voltage/current is a lot higher? Probably something we need to calculate, the magnetic field for these high-voltage coils.
Also watch out for phase shift across your "dropper" divider circuit. Not much, but little things can add up...
I like it! Hope you build it and show us some experiments.
-SB

Hi Simon, in that previous post you appear to have forgotten that at DC you have a DC resistance. Any AC signal in a coil means that the coil does not have resistance, it has impedance. You mentioned a 400% difference if I remember correctly.....
The impedance is also given in Ohms which causes some people to forget that these "Ohms" are slightly different!!
The impedance varies with the frequency, a higher frequency gives a higher impedance with the same coil......
Hoping that I have just "jogged" your memory a bit.......this is why changing frequency changes many other things as well as the null of a search coil head for example.

Have fun.

Have a great day.

Regards
Andy

[golfnut]

Very nice feature to consider.

In the future our MDs will have "scanners" to look for clear frequencies and hop there automatically!

Some quality Radio control kit for models does just this.

Andy, there are some big beacons at 12k from russia, plus the SAq Tx for swedish subs on 17.2kHz (but only twice a year!)

I guess you use Winrad, Rocky, digital crystal set, Saq Rx.??


If you wind a big coil 1m dia 50turns thin wire and attach it to an audio cable LH input into soundcard - download Saq SD Rx you can see stations in the 1 to 22kHz band live, using your soundcard as a spectrum analyser.

You could offer your IGSL up to this to get the actual transmit frequency too !!!!!!

Steve

[golfnut]

As the Tx oscillator is the Rx Local oscillator - or a direct conversion Rx.

The only Rx frequency available is that of the LO.


Offsetting the tuned front end further makes the Rx more deaf - like an off tune radio.


Id guess the selectivity of the coil / cap on these is fairly loose.. So we can blagg these offsets.

Again as stated, if the coil / cap was 'on channel', or peaked for reception on the Tx frequency, nulling would be more difficult for starters.

As mentioned, in use the chances of proximity effects pulling the Rx coil / cap AND the Tx / coil cap ---- by the exactt same amount are low.

Or another way to keep the two tuned circuits phase aligned to introduce no falsing is a little fanciful - unfortunately. Although nothings impossible right..

S

[Jerry]

I'm building a IGSL-TGSL right now and once my standard coils are made and everything is working as it should, I intend to modify the Tx and Rx frequencies exactly as you propose.

I used this calculator to work out the necessary capacitor values (especially the Tx Colpitts oscillator):

http://www.zen22142.zen.co.uk/Theory/tunedcct.htm

I was thinking about why 1.6KHz was specifically chosen as a delta between Tx and Rx. why not 1KHz or 1.5KHz?
I then realised that 1.6 is a close approximation to the golden ratio (1.61 and I suspect this was the original motivation for choosing this number.
If I'm correct then the golden ratio has been mis-applied IMHO. I am intending to apply it correctly.

Now the golden ratio delta for 14.5 KHz is 8.96 KHz which seems way too big of a frequency difference to me. However, the golden ratio is fractal in nature and I noticed that the 4th order golden ratio (1.6 of 1.6 of 1.6 of 1.6) is 2.116KHz and a 5th order is 1.308KHz.

I'm personally choosing the 4th order because the 4th order pops up a lot in 3 dimensional mathematics (like electromagnetics of 3D space:
http://www.physlink.com/education/askexperts/ae174.cfm)

Maybe you think I'm a bit loopy... but here is an interesting link about the orders of the golden ratio in nature:

http://jwilson.coe.uga.edu/emat6680/parveen/fib_nature.htm

The 5th order looks promising also.

Anyway, I'm building the winding jig right now so I'll keep you posted once I'm ready to do the mods.

I do not think too much should be read into the 1.6 Khz offset. This has been found to work by experimentation and other offsets work well also. The receiver will get more sensitive as you move resonance closer to the TX however it also gets much more unstable as you move closer. Move too far and you loose too much sensitivity so 1.6 Khz or thereabouts seems to be a good compromise. Picture the Rx resonance point being located x Db down on the skirt of the transmitter selectivity curve.

I see the temptation to compare with the "Golden Ratio" since it is a number used pretty much everywhere. However, I think in this case it is a happy coincidence.

But it is fun playing around with the numbers.

Looking forward to seeing your results.

Jerry