Novel way of damping?

[nick_f]

Hi All,

I was thinking for some time if there is a better way of damping the oscillation of the coil instead of using the classic circuit, with Rd in parallel with the coil. I thought, what would happen if we would place the damping resistor in circuit only at a very specific moment in time? Could that improve the decay time? The discussion with B^C today motivated me to check my theories.
Here is a screenshot with my circuit diagram. I don't know if it works in real world, but I expect some close enough results. I think the most important adjustment would be to generate the control pulse for transistor U3 at a very precise moment in time.
Here is the circuit diagram and simulation results, also the LTSpice simulation. With this circuit, I expect to be able to obtain sampling pulses very close to the flyback. The coil should be a concentric coil or DD coil, since the limiting diodes would worsen the decay time (or maybe to connect them via another switching transistor for the monocoil circuit).
For LTSpice simulation, place the file nemos3t.mod in C:\Program Files\LTC\SwCADIII\lib\sub, the file nmos4.asy in C:\Program Files\LTC\SwCADIII\lib\sym.


Regards,
Nicolae

[moodz]

Hi Nick
This is a very interesting idea ... I have not tried it on spice but I have tried to do it on a breadboard ... could not remove ringing in the receive response. The spike is like a rollercoaster ... must apply the brakes evenly and constant or you get jerking. There is a patent where they have a circuit like yours but no resistor ... by applying correct bias to gate you can use the mosfet as a damping resistor ... the patent suggests DAC or variable resistor. An interesting variation might be to provide feedback to an error amplifier which compares the decay to some ideal / reference decay signal and controls the gate biasing to vary the resistance of the mosfet accordingly.

regards,

moodz

[moodz]

found it ....
US Patent 7075304 - Variable damping induction coil for metal detection

moodz

[nick_f]

Hi Nick
This is a very interesting idea ... I have not tried it on spice but I have tried to do it on a breadboard ... could not remove ringing in the receive response. The spike is like a rollercoaster ... must apply the brakes evenly and constant or you get jerking. There is a patent where they have a circuit like yours but no resistor ... by applying correct bias to gate you can use the mosfet as a damping resistor ... the patent suggests DAC or variable resistor. An interesting variation might be to provide feedback to an error amplifier which compares the decay to some ideal / reference decay signal and controls the gate biasing to vary the resistance of the mosfet accordingly.

regards,

moodz

Hi Moodz,
Thanks for the information about the patent. One would have expected that other people thought of this long before me. I briefly checked the patent and it is quite interesting, they are using a DAC to gradually increase the damping current, and they don't use the resistor at all. In my case, the resistor also has to have a specific value, like 50 ohm or 100 ohm (I tried with 20 ohm and 200 ohm and the effect wasn't as good)
The paten is obviously a more evolved version than mine.
I am surprised you already tried it on a board, you are very fast! Based on my simulation, the adjustment of the control pulse for the damping transistor has to come exactly at the right time. I think that is at the peak of the negative voltage, and the required precision is like better than 0.1us (100ns!). I am not sure if you can obtain that resolution on your microcontroller. If you go more than 0.1us to the left or right of the right point, the ringing will occur again, with a higher and higher amplitude. I think people without microcontrollers can use a monostable and a pot to finely adjust the point. But I don't know if it works.
I am not sure how your suggestion could be accomplished in practice. How would you generate the ideal reference signal? I think it would require a DAC and some preset data from a microcontroller. I also think the error message is not linear with the required control voltage of the damping transistor. My simulation shows I have to open the transistor completely at a specific moment in time for a short period of time, and in rest, it is better not to have any signal at all on the gate. But, of course, I could be wrong.

Regards,
Nicolae

[Carl-NC]

I tried it a couple of years ago, incredibly sensitive to the exact placement of the timing. Decided it was impractical.

- Carl

[moodz]

I tried it a couple of years ago, incredibly sensitive to the exact placement of the timing. Decided it was impractical.

- Carl

I was using a PIC and coding in C so the timesteps were not exactly what you call precise ... maybe it could be done with a really fast FPGA or state machine.

moodz.

[moodz]

Hi Moodz,
Thanks for the information about the patent. One would have expected that other people thought of this long before me. I briefly checked the patent and it is quite interesting, they are using a DAC to gradually increase the damping current, and they don't use the resistor at all. In my case, the resistor also has to have a specific value, like 50 ohm or 100 ohm (I tried with 20 ohm and 200 ohm and the effect wasn't as good)
The paten is obviously a more evolved version than mine.
I am surprised you already tried it on a board, you are very fast! Based on my simulation, the adjustment of the control pulse for the damping transistor has to come exactly at the right time. I think that is at the peak of the negative voltage, and the required precision is like better than 0.1us (100ns!). I am not sure if you can obtain that resolution on your microcontroller. If you go more than 0.1us to the left or right of the right point, the ringing will occur again, with a higher and higher amplitude. I think people without microcontrollers can use a monostable and a pot to finely adjust the point. But I don't know if it works.
I am not sure how your suggestion could be accomplished in practice. How would you generate the ideal reference signal? I think it would require a DAC and some preset data from a microcontroller. I also think the error message is not linear with the required control voltage of the damping transistor. My simulation shows I have to open the transistor completely at a specific moment in time for a short period of time, and in rest, it is better not to have any signal at all on the gate. But, of course, I could be wrong.

Regards,
Nicolae

Hi Nick ... in CRT devices ( tube tvs and oscilloscope ) they have to control high voltage amplifiers very accurately to do stuff on the screen ... maybe like some high voltage wide band power amplifier that drives the coil like a loudspeaker .... of course you will start reading in the newpapers ...
Man killed by metal detector ha ha.

moodz

[B^C]

Gday Nicholae,

Now a lot of people won't agree but sometimes a simple RCD can eliminate a lot of issues depending on your configuration & your components & the way you look at things.
How you configure this into your circuit can make all the difference.

[Tinkerer]

Gday Nicholae,

Now a lot of people won't agree but sometimes a simple RCD can eliminate a lot of issues depending on your configuration & your components & the way you look at things.
How you configure this into your circuit can make all the difference.

Hey Galileo, now you are shaking at the very foundation of PI belief. In the old time they were burning people on the stake for coming up with such heretic ideas.

Tinkerer

[B^C]

Gday Tinkerer,

Hahahaha, your correct they used to burn people at the stake for being different. In fact they used to burn people at the stake for any silly reason, normally because the masses didn't understand --being different has it's advantages though.

[Tinkerer]

Gday Tinkerer,

Hahahaha, your correct they used to burn people at the stake for being different. In fact they used to burn people at the stake for any silly reason, normally because the masses didn't understand --being different has it's advantages though.

The masses still do not understand. There are forums where you will be prangered and hostigated for coming up with new ideas.
But doesn't it feel good to KNOW?

Tinkerer

[B^C]

Hi Tinkerer,

I'm not one to worry about what other people think, life is to short for that, everybody has there own way of doing things & rightly so.

What may work for some people may not work for others for numerous reasons.
I look at issues differently, firstly i look at the way everybody else has done something & stay clear of that completely. If everybody thinks the same as in ABC 123 then nothing will ever be improved.

I have read some things on the net from years ago with people still asking the same old questions then as they do now about PI. The reason for this is obvious & is because there is no one to give the correct answers.
The one man who knows a bit more than others --Mr Candy--is not going to answer anything other that what he wants people to know--which is what they already know.
Even then, the ML detectors are far from what is really needed, there better than everybody else but still far from the mark.

Science is trying to understand what we don't know not what we do know already.

There a many ways to look at things as history has proven.
People made up language to communicate & science & physics to understand things. The problem is with that is then we all follow along like sheep & take it as Gospel when, just maybe people were looking at it wrongly in the first place which has also been proven in the past with many things?

At the end of the day--to me, if something works better in real world situations, not behind a desk, than something else does, then it is better.
We can all make up things like string theory & then try to understand what we just made up from fresh air or we can make a real difference in understanding what is actually there.

If i can detect a small--say, 1grm nugget in bad ground deeper than i could before then that's an improvement to me. I care not for mathematics or Physics that would suggest it can't work--no such word as can't.

Thinking outside the box is the only way to discover "new" things!

A great man said:

Insanity:
Doing the same thing over & over again &
expecting a different result.

Albert Einstein

[simonbaker]

energy recovery

I'm interested in techniques to achieve fast damping and energy recovery at the same time. I think it's been done to some extent, but not much discussion or circuits here.

All I can think of so far is a coupled flywheel design, like using a large inductor (and maybe capacitor in series) instead of the Rd, maybe switched in and out at the right time. The idea is that the inductor will slam the current down, but capture the energy in it's field (how much?), which it then feeds to the capacitor for storage.

Probably very clumsy and hard to engineer and just easier to use big batteries, but interesting to think about anyway.

Cheers,

-SB

[Altra]

Hello SB,

Here is patent that doing something similar to what you are suggesting. They are storing the flyback, hundreds of volts in a capacitor. Then reapplying it to the transmit coil during another time period.

Best regards
Mark

[simonbaker]

Hello SB,

Here is patent that doing something similar to what you are suggesting. They are storing the flyback, hundreds of volts in a capacitor. Then reapplying it to the transmit coil during another time period.

Best regards
Mark

Interesting, thanks.

-SB