Tem question for Aziz

[6666]

Hi Aziz
on the other forum apparently you published some TEM schematics in the members only section, would you be interested in publishing them here please ?

[Aziz]

Hi 6666,

you probably mean the following simple principles (see attachment).
All the Tuned EM transmitters (TEM transmitter) do resonate at two different frequencies f1 and f2. In it's basic configuration, there will be half period of f1 and half period of f2. And I remember, that I have given a simple modification of the cross-coupled VLF transmitter to become the basic version of TEM.

The TEM2.0 and Class-E-TEM can be configured to resonate longer at frequency f1.
In terms of period time: T_pulse_cycle = (n + 0.5)*(1/f1) + 0.5*(1/f2), where
n=0 -> basic version
n>0 (n=1, 2, 3, 4, 5, .. integer number) -> advanced version

The use of "magic choke" made it all possible.

Cheers,
Aziz

[6666]

Hi Aziz, thanks for the info
would you mind posting the .asc files
thanks.

[Aziz]

Hi Aziz, thanks for the info
would you mind posting the .asc files
thanks.

Hi 6666,

the transmitter examples are really very trivial. There is no rocket science behind them.
I have to look at my hundreds of LTspice schematics to find something. And this is going to take me a lot of time, which I haven't at the moment.
Aziz

[6666]

Thanks Aziz.

[Aziz]

Hi all,

I just checked all my spice files and I didn't find an appropriate example showing die different TEM architectures. I have to make a new nice LTspice file for you so you can compare them together. I'll try to post it next week. Be patient please.
Cheers,
Aziz

[6666]

Thanks Aziz.

[scrungy_doolittle]

Aziz your last gif VLFOscMultiFreq01 shows a particular .asc file. Could you at least post that file? Thanks.

[Aziz]

Aziz your last gif VLFOscMultiFreq01 shows a particular .asc file. Could you at least post that file? Thanks.

Sorry for being very late.
You will find the file in the rar-file below. Just unrar the attachment and run LTspice.

Cheers,
Aziz

[IBGold]

+Good to see you are back with us.

Regards, Ian.

[Davor]

Aziz, I'd say using resonant coils is a colossal waste of time. Especially if you are following it with ADC that splits the Rx signal into multiple cyclostationary phases.
With just about any modern MOSFET you'll get better efficiency than any resonant oscillator can provide.

[Aziz]

Aziz, I'd say using resonant coils is a colossal waste of time. Especially if you are following it with ADC that splits the Rx signal into multiple cyclostationary phases.
With just about any modern MOSFET you'll get better efficiency than any resonant oscillator can provide.

Yep, it isn't much elegant to demodulate the signal from a free running oscillator. But it is possible. I could lock into an external sine wave using the reference TX signal, deriving the in- and quadrature phase of it and do the very stable lock-in amplifier demodulation for the RX signal. It worked for the single frequency very fine. The modification of the oscillator by adding an additional capacitor to form a wide band oscillator makes it more complex of course. But can also be tamed with tricks.

Nevertheless, it was very funny to explore different ways to skin the cat.

Aziz

[Davor]

Actually I meant free running sinewave oscillator being a complete waste off time. I realised it has a few drawbacks that are easily remedied by a pure square wave.
First off, 1/ff noise. If you use a BJT, you need to apply a serious amount of emitter degeneration, otherwise it is quite noisy in 1/f region.
- square wave drive will have exclusively power rail noise, and that one can be kept reasonably low.
Secondly, the overall amplitude stability. If you do not press it hard into saturation, a LC oscillator will have amplitude variation due to soil proximity.
- at square wave it is pressed to the rail voltage.
Thirdly, energy efficiency with sinus works fine, but at a cost of assuming an odd frequency.
- Square wave bridge is an energy recouperation device by its own right, and with nowadays FETs' small Rds, the efficiency is better than any sinus. And you may pick whichever frequency you want.

Digital filtering is simple enough, so there is no point insisting on sinus.

[Aziz]

Actually I meant free running sinewave oscillator being a complete waste off time. I realised it has a few drawbacks that are easily remedied by a pure square wave.
First off, 1/ff noise. If you use a BJT, you need to apply a serious amount of emitter degeneration, otherwise it is quite noisy in 1/f region.
- square wave drive will have exclusively power rail noise, and that one can be kept reasonably low.
Secondly, the overall amplitude stability. If you do not press it hard into saturation, a LC oscillator will have amplitude variation due to soil proximity.
- at square wave it is pressed to the rail voltage.
Thirdly, energy efficiency with sinus works fine, but at a cost of assuming an odd frequency.
- Square wave bridge is an energy recouperation device by its own right, and with nowadays FETs' small Rds, the efficiency is better than any sinus. And you may pick whichever frequency you want.

Digital filtering is simple enough, so there is no point insisting on sinus.

Hi Davor,

ok, I have misunderstood you.
But the 1/f noise was never be an issue with the BJT implementation as the operating frequencies were well above the corner frequency of 1 kHz (where the 1/f noise gets irrelevant). Ok, the BJT implementation can not achive the same efficiency as the mosfet implementation but can be operated with very low voltages.

Well, I am thinking of a new transmitter circuit with the switch-off phase time for the TX coil.
I've to play around with some circuits first.

Now look below, how the TX losses can be compensated prior before the current reversal of the constant current bipolar transmitter occurs. The boost mode converter is switched on before the current reversal happens. But it isn't efficient due to forward biasing diodes however. R5 (5 ohm) and the switching duration of the U7 determines the losses to get a flat-top current. I think I have posted a much simpler version of the transmitter (boost mode with shorting a low value resistor).



Anyway, I am looking for an another transmitter with off-time.
Aziz

[Davor]

If you introduce a lead pulse half the duration, and tail pulse also half the duration, the coil remains at zero amps average. That's the way FBS does it. It is simple and effective.

The 1/f noise is a problem because it is modulated on top off the carrier, and thus injected back to the Rx via air signal. For that very reason rail voltages must also be properly quiet, or otherwise you are noise limited by coils imbalance.