TINKERERS_TEM-TX

[Tinkerer]

Here is a very powerful TEM-TX.
It is suitable for coils up to 2meters and will still give resolution for small targets like coins.

Attached is the schematics in .jpg format as well as in .pdf format.

Enjoy

Tinkerer

[Tinkerer]

TINKERERS_TEM-TX BOARD_3D

Here is a 3D view of the TEM-TX PCB.

Tinkerer

[Tinkerer]

TINKERERS_TEM-TX_PCB

Here is the PCB, all layers, but without the copper poor.

Then there is the PCB with top copper poor.

Comments, critics and suggestions for improvement are most welcome.


Tinkerer

[Midas]

Here is the PCB, all layers, but without the copper poor.

Then there is the PCB with top copper poor.

Comments, critics and suggestions for improvement are most welcome.


Tinkerer

Hi Tinkerer,

I think it would be good idea to get your tank capacitors as close to your connector as possible. Your asking that chunky-than-the-rest but still relativley narrow trace to carry a lot of current (How much did it end up being, out of interest ?). In fact there is probably a good case to be made for sticking them off the board completely in the actual coil head, increased efficiency, less radiated EMI.

Midas

[Mechanic]

Hi Tinkerer,

I agree with Midas that the tank caps need to be located as close to the connector as possible. That is the 4 rectangular caps.
The other thing I would do is cp1 -cp 4 should be mounted as close as possible next to each mosfet. Keep the trace between each cap and mosfet source the same length too. Then make the Tout trace a copper pour to roughly match the size of the copper pour behind the mosfets 24+v trace. It looks like you are using a trough hole gate resistor, the leads are easily left longer to cross the copper pour. You may want to mount them on the bottom of the board. Doing all of this will minimize the high current loop area on the board and the board will radiate a lot less emi. Also avoid 90 degree bends on high current traces.

Cheers Mick

[WM6]

Here is a very powerful TEM-TX.
It is suitable for coils up to 2meters and will still give resolution for small targets like coins.

Attached is the schematics in .jpg format as well as in .pdf format.

Enjoy

Tinkerer

Excellent work Tinkerer.
Congratulations and thank you to deal your work with us.

Can you give us some more explanations about circuit functionality, e.g. how coil is wound?

[Tinkerer]

Thanks for the excellent suggestions. Below are some numbers about the TEM-TX. I am working on a more complete description, where I will try to answer all of your questions.

TEM-TX description.

The TX pulse can be varied between 50 to 300us or about 20,000 to 3,500 PPS

TX voltage from 12V to 30V

Maximum Flyback is about 780V. This does not limit the power, but it limits the resonant frequency.

The size of the Flyback capacitor is chosen so as to maintain the peak Flyback voltage within the limits of the Mosfet, so that the Mosfet does not avalanche.

TX current 2A to 7.5A. The actual power consumption will be about 10% of the coil Amps.

TX coil about 1000uH, example 100cm x 50cm with 22 turns would give a magnetic field strength of about 1.5 Gauss at the center of the coil and 0.2 Gauss at 1 meter depth.
A 100cm round coil with 22 turns would give about 1 Gauss at the coil and 0.36 Gauss at 100cm depth.

Bucking coil about 50uH in series with TX coil wound in opposite direction

Total coil resistance, TX, Bu coil, cable and Mosfet, about 1 Ohm

Coil diameter 30cm to 200cm round or square or rectangular.

Tinkerer

[Tinkerer]

Hi Tinkerer,

I agree with Midas that the tank caps need to be located as close to the connector as possible. That is the 4 rectangular caps.
The other thing I would do is cp1 -cp 4 should be mounted as close as possible next to each mosfet. Keep the trace between each cap and mosfet source the same length too. Then make the Tout trace a copper pour to roughly match the size of the copper pour behind the mosfets 24+v trace. It looks like you are using a trough hole gate resistor, the leads are easily left longer to cross the copper pour. You may want to mount them on the bottom of the board. Doing all of this will minimize the high current loop area on the board and the board will radiate a lot less emi. Also avoid 90 degree bends on high current traces.

Cheers Mick

Thanks Mick and Midas, for the suggestions.
The peak coil current at 7.5A at 780V, during Flyback, certainly needs to be addressed. Maybe the best thing to do would be to make about 4 large copper poors, each one with a ground plane on the other side.

To enable the large variations in PPS and TX voltage, the Flyback capacitors need to be changeable. On the PCB above, I used large pads and ample space to accommodate a wide variety of capacitors, so I can try which kind is best.
With the maximum coil current, it is probably best to use several capacitors in parallel, to divide the coil current. This also makes it possible to get the total capacitance as close as possible to the desired amount.
The coil, cable and Mosfet output capacitance is part of the resonant tank.

I have been thinking of using jumpers, maybe something like automotive fuses, like 40A, to switch some capacitors in or out. does anybody have a better suggestion?

This is an experimental platform. For a specific, non variable implementation, the Flyback capacitors could indeed be on the coil.

Tinkerer

PS. The schema, 3D and pcb are drawn with the free DESIGNSPARK software. If somebody wants to draw some most welcome improvements, I can upload the files.

[Tinkerer]

How does the TEM-TX work?

I will try to give a simple explanation of how it works. Feel free to ask any questions. (AZIZ gives a very good, more technical description here:http://www.geotech1.com/forums/showp...3&postcount=51

How does the TEM-TX work?

In a traditional PI TX, we charge a coil and then discharge is again. To avoid oscillations, we add a damping resistor. The damping resistor burns off the energy, so we start with a high current, right at the time of switch off, but it gets immediately burned off, with the resulting loss of energy.
With the TEM method, we charge the coil the same way, but, instead of a damping resistor we use a Flyback absorbing capacitor. This capacitor stores the whole energy of the Flyback and immediately sends it back through the coil, continuing the normal discharge.

This produces nearly twice as much di/dt than a traditional PI configuration.

However, there is another benefit. The Flyback energy, (minus ohmic losses) is sent back to the battery, or a storage capacitor bank. This results in a great power efficiency, that allows us to use high TX power, while using only a small fraction of power out of the battery.
But, there is more than that.
We use the power that we recycle, to kill the oscillations.

WE HAVE NO TIME DELAY TO THE FIRST SAMPLE. In fact we can sample at any time during the whole cycle.

Sampling at the time of the peak signal response has always been the dream of PI. Well, here it is.

Tinkerer

[Midas]

This produces nearly twice as much di/dt than a traditional PI configuration.

I don't think that's quite correct. Adding the capacitance actually slows down the rate of current change, that's whats allowing you to regulate the fly back voltage. However because the TEM method allows you to get more current flowing in the coil before 'switch off' you can sustain a given di/dt for a longer period of time. The net result is stronger stimulation for many targets. However for very small targets that already saturate with a standard PI pulse you probably won't see any increase in response.

Midas

[Mechanic]

Hi Tinkerer,

If you post up the pcb file I will give designspark yet another crack.... At least all of the parts will be there, that's the bit I find difficult! Moving them around is the easy bit!

Cheers Mick

[Stefan]

Hi Mick,

A lot of the parts are there but a lot are missing. Spent a lot of time building components for Designspark. Will keep you busy.

Regards,

Stefan

[Mechanic]

Hi Mick,

A lot of the parts are there but a lot are missing. Spent a lot of time building components for Designspark. Will keep you busy.

Regards,

Stefan

Hi Stefan,

When I said "there" I meant on the board already!!!!

Cheers Mick

[Stefan]

Ok no problems you can have a lot of fun and a lot of hair pulling moving them around and the associated tracks.


Regards,

Stefan

[Mechanic]

Diptrace Oh I wish I could figure out a way to export between the 2!

Cheers Mick