Induction Balance Stuff - Single/Multi Frequency Response, GB, Disc, Measurements, Ideas, Fun, etc.

[Aziz]

Hi all,

I'm really really impressed.
This µPower free running LC oscillator in Off-Resonance mode works really good and nice.
I had to do a windowing to the samples to minimize the spectral leakage effect before decoding.
As the LC oscillator is running at its own frequency, I am not locked in to the signal to measure the faint amplitude changes.
So it locks in into the same signal. The reference is generated by the signal by phase shifting it 90 degree. Phase measurement is not important at the moment (it will be always 0 as I am locked in into the same signal).

SNR seems to be really good. At resonant frequency, the oscillator amplifies the EMI noise of course.
Frequency stability and change does not really matter. I am measurung pure its amplitude at any frequency.
The reactive response (X) is much lesser than resistive response (R).

Let's start the "hunger" games.

Increasing the impedance to the battery source so the amplitude change will be larger. Much larger.
Cheers,
Aziz

[Aziz]

Hi all,

I was interested in the offf-resonance principle with a local LC oscillator. And how it is performing regard to noise generation and detection sensitivity. Want to check, whether the previous project makes sense or not.

This time, I'm using an ultra low power and ultra low voltage design. It is running from a single flat battery cell (0.7 .. 1 V). It's power consumption is approx. 140 µA.
The LC oscillator gives a nice sine wave form, which can be decoded on the left input channel of the sound card.
The right input channel is used as battery control. I'm already using flat batteries.
Well ok, this can be controlled up to the total battery flatness (approx. 0.7 V).
Below approx. 0.7 V, the oscillator stops running. This is the base-emitter voltage drop of the BJT transistor.

Just finished it. It's running.
Testing will be done next day or so.
Let's look, how much we can get out of the simple local LC oscillator.
The LC oscillator is a Franklin oscillator. Not well known type. But this is outperforming Hartley, Collpits, Clapp, .. It is known to be quite stable.

The circuit:




Cheers,
Aziz

[Aziz]

Hi all,

I haven't been able to look at noise sources yet. (The flat tire of the dad's car made me busy. A nail was sticking to the tire. ) .

Well, if we can reduce the overall noise by 10 - 20 dB, we can make a decent sensitivy detector with a mono coil.
Note: A VLF detector, which works on pure mono coils. Simple round coil design. Only TX-coil. No IB-nulling. No problems at all.

Even if we can not get the optimum the X-free^TM mode, we can at least use the X-min^TM mode (less reactive response). The simple detecting threshold is being set by ground balance procedure, not to detect the reactive response with a ferrite for instance. And on difficult ground, which has little resistive response components (salt or other mineral caused ground response) must be suppressed by setting the detecting threshold appropriate. This is an easy procedure.

I like the Off-Resonance principle.

So, where is all the transmitter noise coming from? Clock signals (lineout) ? Clock detectors? Mosfet driver? Mosfet switching? EMI noise?
So many noise sources to look at.
Cheers,
Aziz

[Aziz]

Hi all,

regardless of my gut feeling and Mr. Occam,

I want to reduce the transmitter generated noise. Note, that the transmitter produces a wide band noise. These happen on switching transitions.
On the combined resonance frequency (during switch-on period, LTX and Ct1+Ct2 are forming a resonant circuit), these switching noises can be seen on the FFT spectrum, which is quite near my operating frequency. This is not good.

I will begin with the n-ch mosfet.
Aziz

[Aziz]

Hi all,

during the week-end shopping, my gut feeling said to me: "There is no free lunch"
1. It must be the linear combination subtractive GB method (X cancellation).
2. It will lose some target response doing this.
3. No one has used the critical energy balanced double tuned TEM2 yet. Perhaps.
4. Due to low impedance of the energy source of the transmitter, no one has noticed this effect (signal on TXref). I have used the high impedance source.
5. No one has modified the operating frequency of the double tuned TEM2, to notice the X cancellation effect.
6. It's all the same. Physics do not change. There is no free lunch.
7. I had a lucky day, to stumble upon it by chance.
8. Mr. Occam would agree with my gut feeling too.

What do you think?
Aziz

[Aziz]

If we answer most of the questions, we are going to do the Occam's Razor to reveal, what is really going on in this process.
Maybe a mathematical simulation of this process will help too. Made in the complex plane. Why the imaginary part (X) becomes zero (phase shift = 0, only the real part is changing). The real part represents the resistive response (R). In an energy balanced system, it represents the energy losses of the system caused by targets.

[Aziz]

Hi all,

I'm going to test the TEM2 transmitter further. No RX channel added yet.

The following questions must be answered soon:
- Does X cancelling works with low Q coils too? (low inductance and/or high resistance coils)
- How can we reduce the transmitters noise? Does a high dampening resistor over TX coil gives any advantage? Does X cancelling work then?
- How low can be the transmitter energy? Does more energy gives really more sensitivity?
- What are the depending parameters on X cancelling feature?
- Has the on-time to off-time relation an effect to the X channeling. Does it defines the linear combination factor K to chancel X response?
- and many more questions...

When I plug the RX coil, more questions arise.
So there is still much work to figure out.

And the most important question:
Why does the X cancelling works? Does it do kind of a subtractive GB?
I don't really know it yet.
Any ideas?

Cheers

PS: more questions added..

[Aziz]

Hi all,

BTW, the simple mono coil mode can be realised with a simple analog circuit too. You don't need any fancy µC/DSP and coding at all. And you have the world best ground balance (WBGB) too. BTW, this is my second WBGB^TM invention.

Following parameters should be adjustable by the user:
- Operating frequency: to adjust the X-free^TM mode (no X response mode).
- Threshold: detecting threshold and GB (for resistive ground responses or not achieved the X-free^TM mode)

Signal processing path:
TXref -> Capacitor (AC-coupled) -> rectifier -> lowpass filter (averager) -> highpass (changes) -> threshold detector -> oscillator -> beep
That's all.

Who is going to make this simple detector possible?

Cheers,
Aziz

PS: Forgot the highpass stage. You want to detect the changes only.

[Aziz]

Hi all,

I will check the transmitter with fully DC blocking soon. Just replacing R2 with a large capacitor. Let's look, what happens.

Congratulations to me. I have re-invented the former TEM mode transmitter.

No, no, no, R2 and L1 are so important.
And energy is relative.
We don't need much power to the TX coil. We are just measuring the pure faint resistive losses on the TXref channel (note: it's free of reactive response and hence no phase shift occurs, only amplitude change).

Therefore, we can make R2 a bit larger to get much more voltage swing at TXref signal. This means increased sensitivity. Enough to operate the TX-coil in mono coil mode. No induction balance coil required in this simple version. You don't even need fancy GB. Simple threshold adjustment is enough to cope with.

With a second RX coil in induction balanced configuration, we get much much more target signals (high frequency wideband response). Also huge ground signals, which can be eliminated by ground balancing scheme processing both TXref and RX channels.

You may leave the magic choke L1, if your R2 is high enough.

Cheers,
Aziz

[Aziz]

I will check the transmitter with fully DC blocking soon. Just replacing R2 with a large capacitor. Let's look, what happens.

[Aziz]

Hi all,

measured the new power consumption for two NiMH cell operation.
Standby: 2.8 mA
Transmitter active: 7.8 mA
R2 = 100 Ohm (The Heater series resistor)
L1 = 15 mH (The Magic Choke)
TX-coil current: +/- 185 mA, this is a 370 mA current span over the TX coil.
A standard 2500 mAh NiMH cell will last for 312 hours (= 2500 / 8 ) or 13 days.

The third test board is not required anymore. The modifications are really little. The transmit mosfet will be switched on/off hard now (fast). No soft switch-on anymore.
Do I get the Trump Nobel Prize now?

Cheers,
Aziz

[Aziz]

Hi all,

what a lucky day!

Good news:
I have finaly found a logic level n-ch mosfet in my box. It's a IRLU 2905 (55 V, 42 A).
I have changed my TEM2 transmitter test board to operate with mosfets. Only minor changes are required.

Good news:
The IRLU 2905 is working fine with 2-cells battery voltage.
The power efficiency went up and the power consumption down.

I can test the transmitter for flyback voltlages up to 55 V now until I get other logic level high voltage mosfets.

Good news:
The transmitter noise can be improved by precisely setting the on/off timing of the transmitting transistor. At least 1 decade less noise. BTW, the noise is picked up by the TX-coil due to EMI noise sources. But precise switching timings do not add more noise.

Good news:
The Heater resistor R2 got a mate: The Magic Choke L1.
So the Heater won't heat anymore. They are working together and doing a very good job now.

Now fixed all parameters to get consistent the zero X-response.
We really need the series resistor R2 to have DC impedance to the transmitter (DC-blocking). And we need the high inductance choke L1 to have a high AC impedance at the operating frequency (AC-blocking). The critical energy balance works fine if the transmitter is decoupled well enough from the power source, to get big measureable effects.
We can also push more TX current through the TX coil.

Cheers,
Aziz

[Aziz]

The Heater heats up my brain.
We need the high impedance path of the LC network to the battery source. It must be good decoupled from each other. And this is best done with a high inductance choke. No heat, no problems, high current to the TX coil.

[Aziz]

Hi again,

I have checked the other harmonic frequencies of the wideband response signal (2x, 3x, .. of the fundamental operating frequency). Some harmonic responses pass through very low X-response. So I have not reached the real optimum operating point yet.
But the multi-frequency responses are all consistent with the fundamental frequency.

As it is independent of the heater resistor R2, we can push more current through the TX-coil later.
Oh yes, you probably don't know, why my resistor R2 is called the heater resistor.
When I power up my test board, it takes approx. 20 - 30 seconds to warm up this resistor. During this time, I can observe/measure the temperature rise up by the decoded response value.
It is heated so far until the temperatur of R2 becomes constant.

Aziz

[Aziz]

Congratulations to us.

We have invented a novel method to cancel X-response.

The "Heater Resistor" R2 just moves my optimal operating point (and therefore my optimum operating frequency).
It's all about critical energy balance in the L-Ct1/Ct2 network.
So it is independent of the resistor R2.
But the operating point is critical. There is a point of level (the optimal), where all the X-responses cancel out.

I adjust it with selecting the optimum operating frequency. At the optimum operating frequency, all X-responses cancel fully out. Only R-responses remain in the TXref channel.

Cheers,
Aziz

PS: I hate my heater resistor again.