SPICE Simulation Challenge

[green]

@green, there are always some influences of a measurement set on the obtained results. I'll have to contemplate a bit on what you got.
What is your measurement set? How do you excite the coil, and did you play with the damping resistor? The copper coin results look to me as if there is a bit of an overdamping messing with your results.

https://www.geotech1.com/forums/show...255#post239255 defines the tester I'm using. Quite long so if there is a question just ask. Coil: Rx(two 20cm round coil side by side connected figure eight) Tx surrounds Rx. Tx 160us on time, constant rate 1A peak. The coins TC is about 500usec so a longer Tx time would effect decay off and it would lower the Rx signal when Tx is on Damping was adjusted. Schematic is correct except I'm using the other log out off pin 12.

[Davor]

@green, there are always some influences of a measurement set on the obtained results. I'll have to contemplate a bit on what you got.
What is your measurement set? How do you excite the coil, and did you play with the damping resistor? The copper coin results look to me as if there is a bit of an overdamping messing with your results.

[bbsailor]

Are you referencing graphs in reply 108? What do you mean by assume?

Green,

These were meant to be general comments as we all tend to use the TX turn off point as the start of the discharge curve. The analyzed variables are the size, shape, thickness and TC of the target being analyzed.

We have had a few discussions about why spherical targets are used, mainly to not have the target orientation as a variable in the target analysis graph.

All I was doing is to identify targets near 2uS needing to assume that it becomes fully charged to do a graphic analysis. This is based on a paper that Eric Foster posed on his forum many years ago that the optimum TX discharge TC slope should be 5 times faster than the target TC.

Eric was also concerned where, on the on-current rise chart, the TX signal was turned off. He said that the optimum turn off rise time was 3 TCs or when the current raises 95 percent of max current. This point assures that the current rise point is more horizontal at turn off where most of the current can stimulate the target unlike if the current would turn off sooner when the rise point is more vertical and some cancellation of the stimulus current takes place.

When I integrated the main point of Eric's article being that the turn off TC, based on the coil inductance divided by the damping resistance should be 5 times faster than the target TC. When I connected these dots I came to realize that targets at 2uS need to have a discharge TC of 0.4us for a 300uH coil with a 750 ohm damping resistor to fully stimulate the target. This is just another understanding of the challenge of optimizing the coil design to achieve these objectives with low TC targets.

I was just bringing this observation into the simulation challenge so we can properly isolate variables. Eric was very creative with his gold seeking PI metal detectors by integrating many RX signals to make his detectors more sensitive on the RX side by being more sensitive to weak signals from small, low TC targets.

I enjoy your input to this subject and only wanted to help extend forum member analysis with these low TC targets.

Joseph J. Rogowski

[green]

The universal current rise and fall charts show the following percent of total current rise with each time constant.

TC Percent
1 63.2
2 86.5
3 95.1
4 98.2
5 99.3

It works by the current rising at the same percentage rate with each time constant. 63.2 percent at 1 TC with the remaining 36.8 percent rising by 63.2 percent. 36.8 X .632 is 23.2576 added to 63.2 and you have 86.4576 rounded up to 86.5. Do that again and again with the remaining current and you will approach 100 percent being 99.3 percent at 5 TCs.

If you have a 2uS target, using the same theory, but in reverse, it's full PI TX stimulus charge will fall to almost zero in 5 TCs. That is why a 10us PI delay has problems detecting targets near 2uS thus motivating people to seek to get their delay down to a lower number to still see lower TC targets.

Targets either have a full charge or a surface charge depending on the metal composition (inductance and resistance), surface area of the target and mixture with other metals or spaces, like in gold nuggets.

Once you create a mental model of the universal current rise and fall charts, you can better appreciate how you can fine tune your TX PI parameters (PPS rate and TX pulse width) along with RX parameters to get the fastest delay that your circuit and coil will allow. Remember, to fully stimulate a target the theory states that the TX discharge current slope (coil inductance divided by damping R value) must be 5 Times faster than the target TC. That means that your 2uS target must have a TX discharge TC of .4uS. This translates to a coil of 300uH having a 750 ohm damping resistor. Your discharge graphs assume that the target is being fully stimulated.

Keep these points in mind when reverse engineering a PI design to detect specific low TC targets and doing analysis.

Joseph J. Rogowski

Are you referencing graphs in reply 108? What do you mean by assume?

[bbsailor]

The universal current rise and fall charts show the following percent of total current rise with each time constant.

TC Percent
1 63.2
2 86.5
3 95.1
4 98.2
5 99.3

It works by the current rising at the same percentage rate with each time constant. 63.2 percent at 1 TC with the remaining 36.8 percent rising by 63.2 percent. 36.8 X .632 is 23.2576 added to 63.2 and you have 86.4576 rounded up to 86.5. Do that again and again with the remaining current and you will approach 100 percent being 99.3 percent at 5 TCs.

If you have a 2uS target, using the same theory, but in reverse, it's full PI TX stimulus charge will fall to almost zero in 5 TCs. That is why a 10us PI delay has problems detecting targets near 2uS thus motivating people to seek to get their delay down to a lower number to still see lower TC targets.

Targets either have a full charge or a surface charge depending on the metal composition (inductance and resistance), surface area of the target and mixture with other metals or spaces, like in gold nuggets.

Once you create a mental model of the universal current rise and fall charts, you can better appreciate how you can fine tune your TX PI parameters (PPS rate and TX pulse width) along with RX parameters to get the fastest delay that your circuit and coil will allow. Remember, to fully stimulate a target the theory states that the TX discharge current slope (coil inductance divided by damping R value) must be 5 Times faster than the target TC. That means that your 2uS target must have a TX discharge TC of .4uS. This translates to a coil of 300uH having a 750 ohm damping resistor. Your discharge graphs assume that the target is being fully stimulated.

Keep these points in mind when reverse engineering a PI design to detect specific low TC targets and doing analysis.

Joseph J. Rogowski

[green]

Charted a deck screw linear-linear, linear-log and log-log at 9 different positions. I like the log-log, fairly straight line decay same as ground. Interested in how you read the linear-log charts. Rx(two 200mm round coils connected figure eight)Tx surrounds Rx. IB allows looking at signal during Tx linear-linear(signal opposite polarity during Tx so no log amplitude)

Charted a 1oz 99.9% copper coin at the same 9 locations. Ferrous target reply#103 charted close to straight line decay at all target locations. The coin charted close to straight line decay log-log or some locations were straight line decay log-log in the beginning, then decayed straight line linear-log.

[dbanner]

Interesting idea.
Please post results.

Ha ha, you are being too kind waltr, no thanks, i am not experimenter like so many here on forum. fishing expedition is for people who know how to fish. I'd rather buy fish in marketplace.
any ways, this array is basis for true hybrid design. The on pulse used for true vlf detection, when pulse switched off, tx cancel field is also timed to switch off in sync, leaving balanced receivers to detect target decay field. although the receivers are balanced, they are not the same. They are designed to respond differently to same conductive target. I dont know. only conceptualizing.
aiming for two slopes for same target, one in time domain and the other in.......

[waltr]

I am having idea of two separate RX coil and preamp for a pulse machine. Balanced with each other, and also both within a tx cancelled field. Each RX would be made to act complimenrary to each other when field is disturbed by target. The complimentary angles would vary according to type of target.
Each RX would be fed through filters then unto ADC then micro, where angles can be reconstructed and interpreted.
Maybe it's not good idea. But I strongly feel that for more to be done with pulse, special attention have to be put on coil array.

Interesting idea.
Please post results.

[dbanner]

Something like this

[dbanner]

I am having idea of two separate RX coil and preamp for a pulse machine. Balanced with each other, and also both within a tx cancelled field. Each RX would be made to act complimenrary to each other when field is disturbed by target. The complimentary angles would vary according to type of target.
Each RX would be fed through filters then unto ADC then micro, where angles can be reconstructed and interpreted.
Maybe it's not good idea. But I strongly feel that for more to be done with pulse, special attention have to be put on coil array.

[green]

log amplitude-linear time. Log-log is linear for viscous responses.

Charted a deck screw linear-linear, linear-log and log-log at 9 different positions. I like the log-log, fairly straight line decay same as ground. Interested in how you read the linear-log charts. Rx(two 200mm round coils connected figure eight)Tx surrounds Rx. IB allows looking at signal during Tx linear-linear(signal opposite polarity during Tx so no log amplitude)

[Qiaozhi]

Now ... in case you're wondering ... I'm investigating this because I wanted to have some target models that work correctly in a known VLF detector, and then use these same models in a PI simulation with a mono coil. from#1

Does the ferrous simulation reply #98 work with a mono coil PI simulation?

No.
As I also stated in Post #98: "Even without running a similar simulation for a PI detector with a mono coil, it should be fairly obvious that the PI will be unable to distinguish between the ferrous and non-ferrous targets, as only changes in amplitude are available. Without a balanced loop system, there is no way to detect the effects of absorption produced by the ferrous target. All you see is an increase is RX signal for all targets."

The ability of a VLF detector to distinguish between ferrous and non-ferrous targets is basically down to the balanced coil system. Non-ferrous targets cause a redistribution of the field, and ferrous targets exhibit absorption. These phenomenon distort the magnetic field in different ways. To a mono coil, these two effects look the same.

[green]

Now ... in case you're wondering ... I'm investigating this because I wanted to have some target models that work correctly in a known VLF detector, and then use these same models in a PI simulation with a mono coil. from#1

Does the ferrous simulation reply #98 work with a mono coil PI simulation?

[Qiaozhi]

George, I see you finally resorted to fiddling like I did. In the end it's all about a change in mu and its effects on the inductances and the coupling

I purposely avoided playing around with the loop inductances, as the effect we're trying to simulate is due to a distortion in the magnetic field which upsets the coil balance. That's why I only adjust the coupling coefficients to account for absorption by the ferrous target. If you adjust the inductance, this would be similar to how a BFO operates, which is not the case in this particular instance. As you can see, the coupling coefficients associated with the targets are quite small, and the adjustment for ferrous is only 0.0035.

However, your method did lead me to this solution. In particular, the use of the table construct in LTspice.

[Teleno]

George, I see you finally resorted to fiddling like I did. In the end it's all about a change in mu and its effects on the inductances and the coupling