TGSL : TX frequency - what's the real deal?

[lolbou]

Hi,

I've did extensive search and reading on the site (what a huge resource!!) but can't really find the final answer I'm looking for, so here's my question.

I'm working on a DIY TGSL and i've custom made my two tones PCB, and i've came up to the point of winding my own DD coil after a first try at concentrics some time ago.

I aimed (and succeeded) at 6.0mH and 6.5mH for TX and RX. I've measured these using a Peak LCR meter (which is measuring L at 1kHz frequency) and by using a signal generator to measure the resonant frequency using a given capacitor (using a scope as well)
Both measurements give the expected 6.0/6.5mH.

Now, when i connect my TX to the colpitts oscillator in my circuit, i get the 13.9kHz sine wave that is sometimes descibed instead of the 14.5kHz.

I get poor detection. Most test points give the correct voltages, GEB is probably to look at, the sampling FETs too, as well as trying in a low EMF spot, but still : TX frequency is a problem.

Does it need to be 14.5kHz AND 6.0 mH exactly?

Or just 14.5kHz? Or just 6.0mH? And in any case, what can (or can not) be tweaked to get the circuit taylored? Once set, i will try and get it to work, but now i'm stuck thinking that it fails from the very begining...

Thanks for any solid input on this, and please feel free to point me on any thread over here i may ave over (or never) looked...

Laurent

[Carl-NC]

13.9kHz is only 4% low, which might be caused by the resonant cap being 8-9% high. In any case, it should not cause a problem. The first obvious step is to double-check R & C values, and check the demod timing.

[pito]

First check how much power is coming out from Tx, frequency is not important can be +/- 10% as long as Rx is tuned to it.

Post #245
https://www.geotech1.com/forums/foru...white-s/page17

[ivconic]

Most of the Tesoro coils are actually 5.7mH for TX and 6.2mH for RX.
Although we made through the years with various values.
Coil balance makes the difference and it is important.
Phase shift at good balanced coils is approx 20-22 degrees.
TX amplitude ~12Vpp and rarely up to 16Vpp.
I noticed that S/N at pcb design makes the difference.
Golden Sabre, the original, when you look at it; looks like a total mess.
Nothing special at pcb desing, wires all over the place... any yet it is working alright.
But it is enclosed in metal enclosure, maybe that's the answer?
I made probably 100 of those so far, can't tell exactly.
Success ratio is 70/30 perhaps.
Those with original components works better.
2N2907 and 2N2222 from older series and in metal case works better (I guess newer series are made in China and far from original specs. and yet I might be wrong).
BF245 is not most happy replacement. J107 is better. Tough to find those these days.
At such old analog design every little detail can make the difference, but the coil is most crucial.
The way you balance it.
Frequency is of less importance. It can work good in a wider range from 10 to 15kHz.
...
Even the CD4024 pick can make the difference!? I had cases that some 4024s gives poor signal levels for the charge pump and that results with not good negative supply voltage.
Is it due bad 4024 series or the poor signal fed to it? Good point to investigate.
Supply rails should read +8 and -6.2 volts.
As you can see; after so many years and so many diys made; I still have no universal answer. I deal with it case by case.
Once is done; I investigate each point and change the components if needed, until it starts to work good.
But in most of the cases, more than 90%, is about the coil.

[lolbou]

Thanks for the replies guys, I really appreciate it!

13.9kHz is only 4% low, which might be caused by the resonant cap being 8-9% high. In any case, it should not cause a problem. The first obvious step is to double-check R & C values, and check the demod timing.

Caps and resistor were hand selected when i built the circuit, so i'm quite confident about it (and i've re-checked the TX caps to ensure these were not damaged due to soldering). I'll give a closer look to the demod circuit as you mention though.

I'll give a look at Pito's remark regarding TX power as well, i think i can make some measurement.

I may give another go at coil winding to reach the 5.7/6.2mH combo as well. I get good amplitude on TX. 8V is spot on in my circuit, but i reckon i did not rechecked the "-5V" voltage, so i'll do it again.

I use NOS metal case 2N2907/2N2222 already, metal case LM308H too and... BF245s. I'll try other ones (i have a bunch of metal case 2N4093 to try, seems Vgsoff is matching the original TIS75.

Will keep you updated, and hopefully share my work and troubleshooting once verified and working as expected!

Thank you again!

[sinclairuser]

I use NOS metal case 2N2907/2N2222 already, metal case LM308H too and... BF245s. I'll try other ones (i have a bunch of metal case 2N4093 to try, seems Vgsoff is matching the original TIS75.

I am really impressed by you sourcing metal can packages, the nos metal stuff just works well, dont get me wrong they are not easy to find but if you can they are not in comparison to ebay crap.
there was a seller on ebay in ukraine who sold these plus some ex soviet stuff too, i managed to stock my parts drawers with lots of metal can transistors, and ceramic ic's of the sorts we use here, i even found some lf 357's both ceramic and metal can which i snapped up.
sadly when the war started he vanished, i hope he is ok.
anyway sellers of old radio components is a good start point.
good luck.
aly.

[lolbou]

anyway sellers of old radio components is a good start point.

Most of it comes from an old retired tech once met a car boot sale who gave me most of its stock hoping it will serve some more... and it does! The metal LM308s have been luckily salvaged from old biology lab equipment... I hope luck is going to strike and help indeed!

[ivconic]

... The metal LM308s have been luckily salvaged from old biology lab equipment...

Mine too!
4-6 of those, still have them.

[sinclairuser]

i got the parts to build an idx, but never got round to it.
its the only detector i sourced the board and parts for but never built.
maybe this winter i should give it a go.