PI time constants and coil dynamics

[Reg]

Hi Carl,

The term "saturation" was about the best term I could think of to use to try to explain that increasing the coil current on time past the target TC has little effect on the target response that can be seen or used. I am not sure what else to use in less then highly technical terms to say the same thing. So, saturation seemed to fit.

My interpretation of what happens after the pulse on time has reached the target's final time constant is any additional rise in target current is because of its resistance.

I agree that what would be seen in a target or the result of a target does require the current to be changing. Stop any current increase or decrease and target signals stop. I guess one can think of the target and coil relationship as a loose coupled transformer with the target having a finite load.

You are correct, the coil's actual TC doesn't change, but increasing the pulse on time will lengthen the discharge time since the current has risen up higher on the charge TC curve. So, it isn't changing the actual Time Constant but it is changing the coil's discharge time or maybe stated another way, the effective portion of the TC curve.

In a nutshell, I probably should have written what I did a little differently. Instead of saying "effective coil TC", I probably should have said something like the effective level on the coil discharge TC or simply referred to the fact that higher coil current will increase the coil's discharge or decay time. It is difficult to try to explain just what is happening and keep all information in very basic terms. In this case, I probably could have done better.

As for what happens with small gold signals and the ability to detect this type of target, the coil TC has a major effect. To sample early, which is necessary to detect real small gold, requires the coil decay signal occur as fast as possible. Once a person has reached what they feel is the ideal coil winding technique, then the only way to further reduce this decay time is to simply reduce the coil peak current. This will in turn will reduce the time to decay to 0V (when no target is present) which will allow for early sampling. That is the basic point I was trying to get across.


If one observes the output of the preamp and adjusts the pulse on time with no target present, they will see the decay curve change as the result of the increasing or decreasing of the coil current. This change is small, but when trying to detect very small gold such as a small gold nugget or a small gold chain, it does make a big difference.

This change has an influence on how soon one can sample. Now, introduce a small gold target and the signal from the gold does not change proportionally as the result of the same increase of coil current. As such, the gold signal gets lost, or stated another way, is or can be overwhelmed by the coil decay signal as the result of an increase in coil pulse on current. Again, this is why I used the term saturated, which I thought would better help a person visualize what happens and help a person understand that target signals will not simply take longer to decay in proportion to the coil decay signal. If it did, then it would be much easier to detect small gold.

I hope this help clear up my previous posts. I am also sure later I will see something differently and probably would have stated something differently. Hindsight is great but on the other side of the coin, if self defense is needed, it is simply easier to simply not post anything since I have a tendency to try to not post the absolute, but try to post what I feel might help one better understand or visualize what is happening..

Reg

[Carl-NC]

The term "saturation" was about the best term I could think of to use to try to explain that increasing the coil current on time past the target TC has little effect on the target response that can be seen or used. I am not sure what else to use in less then highly technical terms to say the same thing. So, saturation seemed to fit.

I'm not sure this has anything to do with the target. If you look at coil current (plot 'b' above) it exponentially rises to a plateau, after which additional 'on' time does nothing but waste power. B-field (which is what really matters) follows the same curve. 'tau1' is TC of the rise time, which is much much slower than the TC of the target, so the TC of the target will have very little effect on the turn-on, and the turn-on time will have to be much longer than the target TC. I guess you could say that past ~3*tau1 the coil has become "current-saturated".

My interpretation of what happens after the pulse on time has reached the target's final time constant is any additional rise in target current is because of its resistance.

What do you mean by "target current"? The target eddies during the turn-on period?

I agree that what would be seen in a target or the result of a target does require the current to be changing. Stop any current increase or decrease and target signals stop. I guess one can think of the target and coil relationship as a loose coupled transformer with the target having a finite load.

Yes, that's right.

You are correct, the coil's actual TC doesn't change, but increasing the pulse on time will lengthen the discharge time since the current has risen up higher on the charge TC curve. So, it isn't changing the actual Time Constant but it is changing the coil's discharge time or maybe stated another way, the effective portion of the TC curve.

Yes, the turn-off TC (tau2 in my plots) is independent of the turn-on current, so it will always take ~5*tau2 to reach <1%. But if you have an absolute threshold (say, 10mV in flyback voltage) then I agree a higher current will create a higher flyback and you will have to wait longer to reach 10mV.

As for what happens with small gold signals and the ability to detect this type of target, the coil TC has a major effect. To sample early, which is necessary to detect real small gold, requires the coil decay signal occur as fast as possible. Once a person has reached what they feel is the ideal coil winding technique, then the only way to further reduce this decay time is to simply reduce the coil peak current. This will in turn will reduce the time to decay to 0V (when no target is present) which will allow for early sampling. That is the basic point I was trying to get across.

Yes, I agree.

If one observes the output of the preamp and adjusts the pulse on time with no target present, they will see the decay curve change as the result of the increasing or decreasing of the coil current. This change is small, but when trying to detect very small gold such as a small gold nugget or a small gold chain, it does make a big difference.

Agreed.

This change has an influence on how soon one can sample. Now, introduce a small gold target and the signal from the gold does not change proportionally as the result of the same increase of coil current. As such, the gold signal gets lost, or stated another way, is or can be overwhelmed by the coil decay signal as the result of an increase in coil pulse on current. Again, this is why I used the term saturated, which I thought would better help a person visualize what happens and help a person understand that target signals will not simply take longer to decay in proportion to the coil decay signal. If it did, then it would be much easier to detect small gold.

OK, I think we're talking the same language, but different dialects! Maybe some more plots will help de-obfuscate this issue, I'll work some up.

Thanks, I enjoy these exchanges, makes me think more about the whats and whys.

- Carl

[Carl-NC]

increasing the pulse on time will lengthen the discharge time since the current has risen up higher on the charge TC curve. So, it isn't changing the actual Time Constant but it is changing the coil's discharge time

Here is a plot demonstrating what Reg is saying. Instead of changing the current level by altering the pulse on-time (as I did with plot 'd' in the post above), I simply changed the peak current level. This is easily done by changing the series resistance. It doesn't really matter.

The rise time of the current (and B-field) is again shown as tau1. For the low-current case (½*I), a higher series resistance will give a little faster tau. It doesn't really matter.

The fall times are what matters. In both cases, the turn-off taus (tau2) are identical; they are not affected by the 'on' current level, only by the L and off-mode R. So both decays will reach a 1% level at the same time. However, the 1% levels are not the same; the ½*I decay will have half the current at that point.

Instead, what's important is when the curves reach some particular threshold level, because that's what we're looking for at the output of the preamp. Let's switch from current to voltage, by looking at the flyback. The TC of the flyback will be the same tau2, and (assuming the turn-off times are identical) the ½*I current will produce a flyback that has a peak of ½ the normal flyback. Let's take an arbitrary threshold voltage of 1mV; when will both curves reach 1mV? These are marked as t1 and t2 on the plot. The lower current (and lower flyback peak) gets there faster.

As Reg points out, the faster you can get the PI coil response out of the way, the sooner you can look for weak target responses. So there is a tradeoff between pulse power and sensitivity. More power will go deeper, but might miss small shallow gold.

So what is the difference between t1 & t2? Mathematically, for a 2:1 ratio in the peaks (as I used for this example), it is tau*ln(2), or about 0.7*tau, regardless of the actual threshold you choose. That can be a microsecond or so for many PI's. Reducing current by a factor of 10 will speed things up by tau*ln(10), or 2.3*tau.

- Carl

[Tinkerer]

Eddy currents, some observations

Eddy currents, some observations

When I look at my old welding transformer, I see that the iron core is laminated. The reason for using many thin sheets of steel instead of one solid block for the core, are the eddy currents. The thin sheets have a short TC, as opposed to a large block that would have a TC so long, that the induced eddy currents would be of a duration that would only allow for an extreme low frequency. The alloy of the steel sheets also contains a specific amount of silicone (SI) so that the TC is closely matched to the 60Hz sine wave input.
When I switch on the transformer, it makes a slight hum, a little amount of current flows in the primary coil, corresponding to the resistance and eddy current losses.
In the secondary coil the same.
Now, when I short the secondary coil with the electrode/resistor, I get a very powerful current in both the primary and the secondary coils.
Looking at the PI from the same angle, when we have a target in the B field, it is like the electrode/resistor above although much weaker due to the weak coupling.
Even so, it has to increase the current in the primary coil.
Questions:
1)Does this increase of coil current due to the target have any influence on the “signal”?
2)Does a large chunk of gold, representing a long TC, compare with the solid block core above?

Skin effect.
We know that with the VLF method, frequencies above about 30Khz the skin effect becomes an important factor. With the PI, the skin effect is less obvious, but if you look for it can be clearly seen.
The “saturation PI theory” claims that a target with a TC of say, 100uS has to be “saturated” with a TX pulse of 300uS or more to generate maximum eddy currents. The current also needs to be raising for the whole time period.
However, if we rely on the skin effect for detecting the large chunk of gold, we do not need to “saturate” it. A short TX pulse would be enough.
Now, when I look at the ON and OFF transients on the PI with a fast coil, I see that the rate of change of the B field is much higher than the rate of change of a 30Khz sine wave. Does that mean that we should look for the skin effect eddy currents? Or does the skin effect get much absorbed by the ground?
************
An unexplained observation
I made a fast coil with TX, RX and bucking windings. Then I shielded the coil with a graphite paint, leaving a thin gap (a strip of tape) in the shield. However, I forgot to pull the tape off and connected the coil and run sensitivity tests.
I immediately saw a superposed noise sine wave of about 1Mhz on the signal, but the response of the coil seemed to be OK, so I continued the full test sequence.
The results of the tests came out quite different. On some targets/metal, the amplitude was enhanced. On other targets much reduced, as if the PI now had “preferences” for certain metals.
After removing the tape strip, with the shield now having a gap, these “preferences” disappeared.
***********
Reg has explained well about the importance of early sampling to detect small targets. Reducing the gain on the coil amp is an easy way to make early sampling possible.
Question: Can the loss of gain be recovered later?

Tinkerer