Coil Saturation in a PI?

[dbanner]

We seem to be in a bad place with this thread's title.
Saturation really does not apply to either the coil or the target and needs to be struck from being used when discussing PI detectors.

What is important is the coil current reaches a 'steady state' so that the eddy currents due to the increasing magnetic field have died out before switching OFF the coil current.

Yes, but switched off where the current curve is more horizontal than vertical.

[Tinkerer]

- Eddy currents propagate at a finite speed within a conductor.
- This propagation speed is defined by the conductivity of the conductor.
- Resistance converts the eddy currents into heath. The eddy currents "dissipate".
- High resistance makes the eddy currents "dissipate" faster.
- In a high conductivity conductor the eddy currents propagate or move a greater distance due to higher speed.
- true? or false?

[Tinkerer]

Next question:

Why is the skin effect less deep in a high conductor?

Hint: does the magnetic field generated by the eddy currents have something to do with that? Lenz's Law?

[ivconic]

...What is important is the coil current reaches a 'steady state' so that the eddy currents due to the increasing magnetic field have died out before switching OFF the coil current.

This correlates closely with my logic too. It is partial answer on my questions too.
Another question is; do the eddy currents always die out before switching OFF, in every possible case?
Or put different; do the current reaches the steady-state in all possible cases?

[ivconic]

...Hint: does the magnetic field generated by the eddy currents have something to do with that? Lenz's Law?

All we know is fact that it opposes the primary field coming from coil.

[dbanner]

I disagree that the coil current should flat top before switch off, it should be switched off where the current curve is more horizontal than vertical, somewhere along the curve there is a compromise because of the target TC in relation to the coil TC. This is where the optimal(note: not maximum) saturation of the target occurs. You want the eddy currents to dissipate sufficiently not necessarily completely.
The point of switch off determines your pulse width. So in determinig your pulse width, knowing what the current curve looks like is key.
So in the end, the coil TC and current curve and pulse width are just as important as the sample delay, when detecting short TC targets
If you look at the current curves and data for the Vallon( it should reveal quite a bit), I'll bet you'll see switch off well before flat topping.( just a hunch) I could be completely wrong.

[ivconic]

There are some interesting statements here:
https://www.minelab.com/__files/f/11...S_&_THEORY.pdf

[green]

Interesting Spice simulation that does show Target response verse series resistance with coil.
The first sim, with coil R = 10mOhm, it takes a long time for the current to peak (flat top) and the target response is greatest with largest coil current.
The last sim, with coil R =10 Ohm, the current peaks (flat tops) much sooner so TX pulse time can be shorter to obtain the same response.

These sims were with a short TC target (US nickle) and as I have stated in above posts a short TX pulse works.
I would like to see these same sims done with a longer TC target, US Quarter.

Also, why are the coil inductance and damping R different in the three sims? They should all be the same then ONLY change the Series Resistance.

To throw in another factor: A larger series R gets the current to peak sooner. This allows shorter TX pulse times. This then allows high pulse rate (pps) which is known to improve over all detector sensitivity IF the integrator stage has the optional values of C & R. Read posts by Eric Foster about this.

Target signal and Rx signal are proportional to Tx current at Tx off if every thing else remains the same. https://www.geotech1.com/forums/atta...1&d=1570599081

Coil inductance and R damping are the same in each sim. Tx current was set the same at 300us(2 TC's for a quarter)by adjusting PS volts. Maybe shouldn't have stepped Tx on time. Included the zip so anyone could try what the wanted. https://www.geotech1.com/forums/atta...2&d=1570599224

Think minimum coil resistance gives the most signal if battery power is the same. .01 ohms coil resistance is difficult, wondering if coil TC 1/2 of Tx on wouldn't be a good choice.

Zero R gets the current to the same current sooner if the PS volts is the same.

Including a sim with the quarter.

[waltr]

This correlates closely with my logic too. It is partial answer on my questions too.
Another question is; do the eddy currents always die out before switching OFF, in every possible case?
Or put different; do the current reaches the steady-state in all possible cases?

No, it can not due to unknown size of targets. Carl mentioned this, a huge target would not have eddy currents die off since it's TC is also huge.
Now we really don't care in must cases about huge targets so we design for the target we desire, coins, etc. where the target's TC are less than a couple hundred usec.

I disagree that the coil current should flat top before switch off, it should be switched off where the current curve is more horizontal than vertical, somewhere along the curve there is a compromise because of the target TC in relation to the coil TC. This is where the optimal(note: not maximum) saturation of the target occurs. You want the eddy currents to dissipate sufficiently not necessarily completely.
The point of switch off determines your pulse width. So in determinig your pulse width, knowing what the current curve looks like is key.
So in the end, the coil TC and current curve and pulse width are just as important as the sample delay, when detecting short TC targets
If you look at the current curves and data for the Vallon( it should reveal quite a bit), I'll bet you'll see switch off well before flat topping.( just a hunch) I could be completely wrong.

The Current rise is exponential and in theory NEVER reaches a Steady state. When we say 'flat top' or steady state we typically mean close the full theoretical current. From experiments with real detectors and with the spice simulation we have seen and determined that when the current is more than ~80% maximum, this can be considered 'flat top'. Now we just need this current for about 2TC of the target.

Yes, knowing the current curve is 'a' key to determine when to switch off TX, the other is the target's TC.
As I stated in a post earlier, going from a 100us TX pulse (12 Ohm total series R and 400uH coil = 33uS coil Tau) to 200us TX pulse increase the response distance of a US silver quarter by 20%.
The response distance for a US nickle did not increase since a nickle has a short TC (low conductivity) verse a longer TC for a quarter (high conductivity).

[green]

As I stated in a post earlier, going from a 100us TX pulse (12 Ohm total series R and 400uH coil = 33uS coil Tau) to 200us TX pulse increase the response distance of a US silver quarter by 20%. reply #84

Tried your example with spice simulation. Signal increased from 3mV to 4mV with 10us delay, 33% increase in signal. Using https://www.geotech1.com/forums/atta...4&d=1570460303 a 33% increase in signal(about 1.3 divisions on chart) predicts maybe a 5% increase in detection distance. Wondering why such a big difference(5% to 20%).

[waltr]

As I stated in a post earlier, going from a 100us TX pulse (12 Ohm total series R and 400uH coil = 33uS coil Tau) to 200us TX pulse increase the response distance of a US silver quarter by 20%. reply #84

Tried your example with spice simulation. Signal increased from 3mV to 4mV with 10us delay, 33% increase in signal. Using https://www.geotech1.com/forums/atta...4&d=1570460303 a 33% increase in signal(about 1.3 divisions on chart) predicts maybe a 5% increase in detection distance. Wondering why such a big difference(5% to 20%).

I would say additional Gain in the pre and post amps, then 1500pps into 100nF/100k Ohm integrator. Plus the use of the Three sample method with a Long second sample (~110us) to detect/disc high conductors.

[Carl-NC]

[ATTACH]47763[/ATTACH]

[ATTACH]47763[/ATTACH]

An Atocha silver bar of about 90 pounds.

Many years ago I held one of these at a treasure show, I recall it was 70+ pounds. It belonged to Carl Fismer and was for sale. I wish I had the money back then to buy it.

[Carl-NC]

To throw in another factor: A larger series R gets the current to peak sooner. This allows shorter TX pulse times. This then allows high pulse rate (pps) which is known to improve over all detector sensitivity IF the integrator stage has the optional values of C & R. Read posts by Eric Foster about this.

Yes, given that you want, say, a 3-tau settling then a higher R allows shorter TX times, but at the expense of a lower peak current. To boost the current back up, you then need to increase the supply voltage. Trade-offs abound.

[green]

I would say additional Gain in the pre and post amps, then 1500pps into 100nF/100k Ohm integrator. Plus the use of the Three sample method with a Long second sample (~110us) to detect/disc high conductors.

Don't know what your detection distance was. If it was 10 to 12 inches(20% increase)with a 12 inch coil, signal would have to more than double according to the chart. Maybe there is something wrong with the simulation or something else I'm missing? Was Tx on time(100 to 200us)the only thing changed?

[ivconic]

No, it can not due to unknown size of targets. Carl mentioned this, a huge target would not have eddy currents die off since it's TC is also huge.
Now we really don't care in must cases about huge targets so we design for the target we desire, coins, etc. where the target's TC are less than a couple hundred usec.



The Current rise is exponential and in theory NEVER reaches a Steady state. When we say 'flat top' or steady state we typically mean close the full theoretical current. From experiments with real detectors and with the spice simulation we have seen and determined that when the current is more than ~80% maximum, this can be considered 'flat top'. Now we just need this current for about 2TC of the target.

Yes, knowing the current curve is 'a' key to determine when to switch off TX, the other is the target's TC.
As I stated in a post earlier, going from a 100us TX pulse (12 Ohm total series R and 400uH coil = 33uS coil Tau) to 200us TX pulse increase the response distance of a US silver quarter by 20%.
The response distance for a US nickle did not increase since a nickle has a short TC (low conductivity) verse a longer TC for a quarter (high conductivity).

My understandings of things seems are compatible with your ways of explaining them, because now i have answers on few questions that i had.
What's even better; correlates with my logic too... other words "i thought so".
Thanks!