Super TGSL With H Bridge TX Diff RX Quad Sampling

[Davor]

...I thought to ignore salt we would need to notch 20 degrees out of the top and bottom of the sine?...

No, it is just a little bit more complicated than that.
Ferrites and the easy ground are very near the Tx phase, so in order to get zero response from ground/ferrite you need a GEB that is 90° against Tx (or in close vicinity). Salts are a completely different animal. They respond at about 90° against Tx, so you need GEB that is 0/180° against Tx. Mixed type soils like salty terra rossa, mineralised soils etc. are somewhere in between.
Now, there are no beach detectors other than PI amongst DIY enthusiasts right now, but it can be remedied.

The remaining problems are the discrimination criteria. GEB provides a sound output for any metal target, but unfortunately it also provides one axis of the two axes criteria for discrimination. Every coil has a centre area responding as + and beyond that as - and TGSL is built to respond to + only. But in case GEB is shifted to remove salts, the two axes are near to alligned, and ferrous targets from the 4th quadrant appear as if the coil was in counterphase, by means of "-" areas being reversed. This could be remedied with a channel dedicated for discrimination purposes that would be different from a GEB channel. It would thus provide accurate discrimination even with GEB adjusted to salts, and even the ferrous objects would not creep in - unless Disc cranked full open.

[Silver Dollar]

I often wondered why we shift a range instead of using a variable width, I guess it would be hard to
figure out where you are. A switch for ignore salt would be fine and I would try to use a one shot somehow
to make an ignore area...

On a different note, I wound a new TX today and am going to try running at 12.5 Khz. I was planning on using
this preamp, would I need to tweak those input caps for the higher frequency? I wasn't able to figure how
they are tuned for the noise at 8 Khz....

I am going to try a 34.4 mh centertapped RX with the 12.5 Khz TX

[golfnut]

If you mean C1 R1 you can get the -3dB point using 1/ (2 PI CR) Hz - this high pass topology may accept higher frequency too things like the HF band and up.

S

[Silver Dollar]

I calculated it to 1.29 khz but don't know how that works for a 8 khz TX...

Series LC = 11368 hz maybe resonant tuning?

[golfnut]

Hi Davor looks interesting, decent Mixer, Tx etc..


I dont understand some of it - but Im sure you will have simulated it all...




One bit Im unsure about is the cmos switch on the comparator output - this will be gated with up to +8volts - is that ok?

S

[golfnut]

Yes, everthing above ~1.5kHz would get to the transistor, So no need to retune. The Rx coil and cap form a reasonable band accept filter.

What was the deal with the cmos gate on the comparator - I realise it would require less time to activate it, as no CR delay like the original.
it may also be on for a very short time too! I modded the original cap on mine to 10n and made it snappy but with useful tone duration.


I did model that bit around the comparators with the feedback diode - but never understood hoe it worked Im ashamed to say.

i often threatened to do that stage in real logic. but never did.

If I get a rich handout Ill stop mucking around with this old toot and get a few Deus machines and have done with it !

S

[Silver Dollar]

Your right the biasing on the 4053 switch at the comparitors might need looking at. They use Gnd to +8V and that comparitor
stays at -6V a lot. I found that on an old Compass schematic as a FET (see below). It's supposed to modulate the audio a bit to
help with IDing. I think the diode there keeps the 358 from seeing the rail (as it doesn't like that) you can put a resistor there
instead.

While most of the circuits were designed by Davor I'm the one who glued them all together and am trying to make them into a
functioning detector. I'm not an engineer so there might be some mistakes...

[Silver Dollar]

Maybe something like this? 7.6n for 13960 hz resonance and 15k for 1.4 khz filter?

I'm just not sure what I did to the input impediance...

[golfnut]

Apologies I misread the original schematic. The cap and the coil series resonate, The R takes the base to approx 0v . The emitter is negative wrt to base at 0 to bias it on. Looks good.

[Silver Dollar]

I tested the new H bridge with a different cap and my new 1.0 mh Tx coil/

I got 12.5 Khz and 17V! Looks good. Now to layout that preamp mod...

Oh here is the current schematic. I tried simulating the 4024 mod and it wouldn't
drive the push pull transistors. So I went back to my first inclination and paralleled
the gates. The sim says it works and can put out 5 ma. I hope that is enough!

[Davor]

I often wondered why we shift a range instead of using a variable width, I guess it would be hard to
figure out where you are. A switch for ignore salt would be fine and I would try to use a one shot somehow
to make an ignore area...

Yes, switch, or even a button for momentary switching could be beneficial. Searching a shore line can be demanding. you could have, say, 3 presets. It will add a few more pots though.

On a different note, I wound a new TX today and am going to try running at 12.5 Khz. I was planning on using
this preamp, would I need to tweak those input caps for the higher frequency? I wasn't able to figure how
they are tuned for the noise at 8 Khz....

I am going to try a 34.4 mh centertapped RX with the 12.5 Khz TX

Apologies I misread the original schematic. The cap and the coil series resonate, The R takes the base to approx 0v . The emitter is negative wrt to base at 0 to bias it on. Looks good.

The caps are forming a series resonance. It has a low Q so it doesn't influence phase even if you let those caps the same, but chosing values at resonance improves noise.

I tested the new H bridge with a different cap and my new 1.0 mh Tx coil/

I got 12.5 Khz and 17V! Looks good.

Most probably the transistors are now deeper in saturation. with higher frequency comes a higher Q of the Tx tank, so less current is required to saturate the oscillator. If you are not able to reduce the amplitude to about 15Vpp with a trimmer, you'll need to increase emitter resistors by 20% or so.

[Silver Dollar]

Oh ya I forgot about the trimmer. I needed to trim it to 50 ohms in the sim. I don't have any 200 ohm trimmers here so
have to sub fixed resistors. I am running it off 9v here but measured .432 v across the 33 ohm resistor 11ma a bit high.

Those caps on the preamp, you have them tuned to 12 khz for a 8 khz TX? So I should tune to 14 khz for my 12.5 khz TX?
I'm wanting low noise for this rig!

[Davor]

This preamp is very different from a common detuned parallel tank, and it operates at resonance. So for a coil with 34mH and operating Tx frequency at 12kHz, the best values are 10nF each.

Detuning is necessary only for parallel tank front ends, and only because phase transition at resonance is very wild and difficult to tune. Those achieve low noise by transforming action, and with this preamp we go there directly. The equivalent noise source resistor of this preamp is ~100ohm or less, so there is no real need for transformer action.

[Silver Dollar]

Nearly done wiring. Most of the mods fit on one PCB about 2.5" x 3". I'm not
sure about noise and cross talk. I just have to finish up the power connections
and wire the pots. Maybe tomorrow we can try it out!

[golfnut]

Hi the diode on the long white socket looks like an old germanium type - is that right?? I wouldn't worry too much about xtalk as leaked isolation issues from Tx to preamp wont get into the IF as the mixer wont pass it maybe 25dB isolation.. A nulled coil will give you 60 plus dB - most layouts on here have the Tx and preamp 2mm apart 15Vpp next to 30dB preamp which can use 20uV signals - scary innit S