detection depth

[green]

For a target of a given diameter, if you're thinking only of relative voltage induced in the receiver, it's a fairly straightforward (when you know how) axial air coupling calculation. However a target that's "just detectable" is most often limited by a number of factors other than mere voltage induced in the receiver. Also your question as you worded it doesn't make sense: "best coil size 40 inch diameter, target just detectable" means that any other size searchcoil won't detect that target.

So, in the real world your question has no straight answer. The non-straight answer is that if a searchcoil of a given radius has a weak response on a particular target buried at a depth equal to that of the coil radius, an electrically equivalent searchcoil half that radius will kinda sorta more or less usually give a better response. Your mileage may differ.

[EDIT] now the question about a cache 2 feet deep, what size searchcoil will be best. There is no straight answer especially since there is no "given" regarding the effective diameter of the cache or what size of searchcoil is known to detect it. You might do fine with an 8 inch searchcoil and not need "best".

It's been awhile so I started from the beginning to see what had been covered. Second paragraph. Using (same inductance) chart generated with Hyper physics. At target distance of 200mm the 400mm coil has 2 times the target signal strength of the 200mm coil. Signal/noise controls detection depth. Ground and EMI signals originate at the coil. If the combined noise signal is less than half with the 200mm coil the 200mm coil would be better. Starting with the coil(transducer)are there noise signals other than ground and EMI? Am I missing something? At a target distance of 24inches the 400mm coil signal strength is 6.5 times the 200mm coil so maybe the 400mm coil would be better?

[Tinkerer]

It's been awhile so I started from the beginning to see what had been covered. Second paragraph. Using (same inductance) chart generated with Hyper physics. At target distance of 200mm the 400mm coil has 2 times the target signal strength of the 200mm coil. Signal/noise controls detection depth. Ground and EMI signals originate at the coil. If the combined noise signal is less than half with the 200mm coil the 200mm coil would be better. Starting with the coil(transducer)are there noise signals other than ground and EMI? Am I missing something? At a target distance of 24inches the 400mm coil signal strength is 6.5 times the 200mm coil so maybe the 400mm coil would be better?

A coil of 200mm diameter and 25 turns, will have about 278uH of inductance
300mm x 25 turns = 458uH
400mm x 25 turns = 651uH
500mm x 25 turns = 853uH
Or if you keep the same inductance you have to reduce the amount of turns, like 400mm x 14 turns =268uH

14 turns x 1 Amp = 14 amp turns for a coil area of about 2000 cm square 14/2000 = about 0.007 ampere turns per cm square

25 turns x 1 Amp = 25 amp turns for a coil area of about 315 cm square = 25/315 = about 0.079 ampere turns per cm square

Now, with a certain di/dt (Rate of current change), and let's say an incidence of 0.02 ampere turns per surface of the target, the target having a surface of 5cm square, you can just detect this target. With a similar calculation you should be able to find what coil will give the best depth for that target.
Assuming the target in similar ground and with similar conductivity and similar thickness etc.

[Skippy]

Tinkerer: Your calculations are a little bit off, I think.
Green IS keeping the inductance of all his coils constant. From an earlier post:
All coils are 325 uH inductance.
200mm coil : N = 23 turns
400mm coil : N = 15 turns
... so a slighly less extreme ratio than your 25T : 14T figures.

And your area ratio of 2000 : 315 is clearly wrong, the coils are 2:1 diameter ratio, so the areas will be 4:1 ratio, eg. 1257 : 315 cm²
This would make your figures: 15/1257 = 0.012; 23/315 = 0.073. These numbers are 1 : 6.1 ratio.

At the moment, we are assuming the target is a 'unity' target, as long as the same assumption is made for all the simulations, we get consistent results. The reasons why silver/cupro-nickel/ etc coins give different signal strengths haven't been investigated yet.

[green]

More tests today. Wanted to compare 1.5in fig8 Rx vs 8in fig8 Rx. Compared 160usec Tx width(I normally use) vs 5000usec Tx width. Picked 5000 so Tx would be greater than 5 times the target TC Eric suggested for optimum Tx width(don't need to find target time constant before doing test). The longest time constant I've measured is the 1oz copper coin at 500usec. Still plan on ordering a 1.5in copper cube. Tried charting the first 250usec after Tx off, log X log Y since a lot of longer time constant targets have straight line decay on a log-log chart after Tx ends. Was thinking Rx signal shouldn't saturate if wanting to know what target decay would normally look like. Since the decay is straight line on a log-log chart it may not matter. Charted 1200usec charts linear X log Y to get the target time constants. Need to do the stacked quarters with the 8in fig8 Rx. Posting data to see if anyone has suggestions for changing or additions to the charts. I use constant rate Tx with my detector so that is what I'll use for testing unless someone wants constant current for comparison. Can do either.
Need to define recording sequence before testing so the colors would match for comparison.

[Tinkerer]

Tinkerer: Your calculations are a little bit off, I think.
Green IS keeping the inductance of all his coils constant. From an earlier post:
All coils are 325 uH inductance.
200mm coil : N = 23 turns
400mm coil : N = 15 turns
... so a slighly less extreme ratio than your 25T : 14T figures.

And your area ratio of 2000 : 315 is clearly wrong, the coils are 2:1 diameter ratio, so the areas will be 4:1 ratio, eg. 1257 : 315 cm²
This would make your figures: 15/1257 = 0.012; 23/315 = 0.073. These numbers are 1 : 6.1 ratio.

At the moment, we are assuming the target is a 'unity' target, as long as the same assumption is made for all the simulations, we get consistent results. The reasons why silver/cupro-nickel/ etc coins give different signal strengths haven't been investigated yet.

Hmmmm, the legend says: same number of turns, same current profile and peak amplitude.......200, 300, 400, 500mm mono coils. ..http://www.geotech1.com/forums/attac...3&d=1521218105

I assumed diameter.

[Skippy]

Tinkerer:
it is confusing: The left chart is 'constant number of turns' , the right chart is 'constant inductance'. Green has been using 'constant inductance' for nearly all his charts, as presumably that is how he would normally operate his machine, and the test coils he made for experimenting with were also constant L. They were 133mm, 200mm and 300mm diameter.

--------------

I think it would be useful if these hyperphysics decay curves included one for the 133mm coil, then we could compare theory with practice better. Green has posted some measured data for a 10 grain nugget and drink-can square targets, there's probably more I've missed.
http://www.geotech1.com/forums/attac...9&d=1501182674

There is something bothering me about our hyperphys simulations. I'm unsure if we are compensating for the number of coil turns correctly, in one or both of the simulations. The hyperphys model assumes it's a single turn coil, we allow for the multiple turns when transmitting by assuming (15 turns and 1 Amp) = (1 turn and 15 Amps).
But on 'receive' , are we assuming 1 turn, always, and not allowing for the varying number of turns? Should the received signal be multiplied by the number of turns in the coil?
But voltage from an inductor is V = -L * dI/dt , so if we're keeping inductance constant, are we keeping the V vs. I relationship constant? And when we don't keep inductance constant...?

I got tired thinking about this last night ...

[Davor]

It is roughly the same length of wire that dictates the same inductance, with little influence of the coil size (and therefore number of turns), and a bit more influence from density of packing.

[green]

Tinkerer:
it is confusing: The left chart is 'constant number of turns' , the right chart is 'constant inductance'. Green has been using 'constant inductance' for nearly all his charts, as presumably that is how he would normally operate his machine, and the test coils he made for experimenting with were also constant L. They were 133mm, 200mm and 300mm diameter.

--------------

I think it would be useful if these hyperphysics decay curves included one for the 133mm coil, then we could compare theory with practice better. Green has posted some measured data for a 10 grain nugget and drink-can square targets, there's probably more I've missed.
http://www.geotech1.com/forums/attac...9&d=1501182674

There is something bothering me about our hyperphys simulations. I'm unsure if we are compensating for the number of coil turns correctly, in one or both of the simulations. The hyperphys model assumes it's a single turn coil, we allow for the multiple turns when transmitting by assuming (15 turns and 1 Amp) = (1 turn and 15 Amps).
But on 'receive' , are we assuming 1 turn, always, and not allowing for the varying number of turns? Should the received signal be multiplied by the number of turns in the coil?
But voltage from an inductor is V = -L * dI/dt , so if we're keeping inductance constant, are we keeping the V vs. I relationship constant? And when we don't keep inductance constant...?

I got tired thinking about this last night ...

Coil diameter not radius. The wound coils were close to 300uH. The measured data was corrected for same(300uH)coils before charting.

I'll try to generate a 133mm coil with Hyper physics and chart 133, 200 and 300mm diameter coils, measured and calculated on the same chart so it would be easy to see if they are close.

Including the measured data chart

[green]

Coil diameter not radius. The wound coils were close to 300uH. The measured data was corrected for same(300uH)coils before charting.

I'll try to generate a 133mm coil with Hyper physics and chart 133, 200 and 300mm diameter coils, measured and calculated on the same chart so it would be easy to see if they are close.

Including the measured data chart

Chart with measured and calculated values. Hyper physics data was multiplied by a value that made calculated and measured the same at 150mm detection distance for each coil. Looks close at shorter detection distances, planned on using a larger target to detect greater distances but zapped my bench circuit so it might be awhile. The wound coils were flat basket with the mean diameter being 133, 200 and300mm. I'm sure noise caused some error at the longer distances.

I'm guessing the calculated is more accurate than measured and the measured indicates that the calculation method might be correct.

[green]

http://www.geotech1.com/forums/showt...438#post228438

Here are the curves I have used for many years. The range reaches a maximum when it is equal to the radius of the coil. Coils larger or smaller than this optimum will result in less range.(from above link)

That statement works for the same inductance chart and not the same turns chart I generated with Hyper physics.

[Skippy]

Thanks for working out the 133mm hyperphys figures.
The line that is bothering me is this:
"Hyper physics data was multiplied by a value that made calculated and measured the same at 150mm detection distance for each coil."
This means you're 'fiddling' the data three different ways, one for each curve. So any scale-factor that we've got wrong will be 'fiddled out'.
For example: suppose to be correct, we should multiply the hyperphys data by the number of turns in the coil. If you did that, and used the same fiddle-factor for all coils, would all 3 pairs of graphs match? Or would just the one pair match, and the other two pairs showing an offset between calculated & measured?
The fact that the curves match in shape is good, and suggests we've got it 'right', but any small scale-factor error, such as multiplying a number by 23 (turns) instead of 15, will just shift the whole curve 'up' the scale a small amount.
I suppose one question I should ask is: Were the fiddle-factors about the same for all three pairs of curves?

[green]

Thanks for working out the 133mm hyperphys figures.
The line that is bothering me is this:
"Hyper physics data was multiplied by a value that made calculated and measured the same at 150mm detection distance for each coil."
This means you're 'fiddling' the data three different ways, one for each curve. So any scale-factor that we've got wrong will be 'fiddled out'.
For example: suppose to be correct, we should multiply the hyperphys data by the number of turns in the coil. If you did that, and used the same fiddle-factor for all coils, would all 3 pairs of graphs match? Or would just the one pair match, and the other two pairs showing an offset between calculated & measured?
The fact that the curves match in shape is good, and suggests we've got it 'right', but any small scale-factor error, such as multiplying a number by 23 (turns) instead of 15, will just shift the whole curve 'up' the scale a small amount.
I suppose one question I should ask is: Were the fiddle-factors about the same for all three pairs of curves?

All I did was multiply the Y values by a constant to shift the curve up or down to be the same at one point(150mm). That shouldn't change the slope(log scale). I'm still wondering about Eric's curves matching same inductance instead of same turns. Still don't know if my same turns chart makes sense and if Eric's chart(same turns) is correct the same inductance chart doesn't either.

If I multiplied the nugget numbers by 100 the curves would chart close to the same as the 1 inch square foil.

[Skippy]

Quote:"All I did was multiply the Y values by a constant to shift the curve up or down to be the same at one point(150mm). That shouldn't change the slope(log scale)"
That's what I'm talking about. A scale-factor doesn't change the slope, it just shifts every point on the curve up or down by a fixed distance on a log vertical scale.

My feeling is that the constant-inductance theoretical figures are correct. When we used hyperphys, we put in a current, and got a current out.
The equation V = -L * dI/dt means the voltage is proportional to the induced current, as L values are near-constant.
I think the constant-turns chart is the one that's in error.

[green]

Quote:"All I did was multiply the Y values by a constant to shift the curve up or down to be the same at one point(150mm). That shouldn't change the slope(log scale)"
That's what I'm talking about. A scale-factor doesn't change the slope, it just shifts every point on the curve up or down by a fixed distance on a log vertical scale.

My feeling is that the constant-inductance theoretical figures are correct. When we used hyperphys, we put in a current, and got a current out.
The equation V = -L * dI/dt means the voltage is proportional to the induced current, as L values are near-constant.
I think the constant-turns chart is the one that's in error.

I think the constant-turns chart is the one that's in error. Think you are correct.

Think I corrected it. Looks like at coil radius a larger or smaller coil would give less signal(Eric's chart).

[Skippy]

Hmmm.. you didn't say how you corrected it. So I can't assess your reasoning. But I have spent some (lots) time this evening on hyperphys trying to check your charts - including the constant inductance.
I concentrated on the Z = 0 points, as that's easier, mathematically speaking, and we seem to agree the general shape of the curves is correct.
It appears your 'reference' point on both charts is for the 200mm coil, at Z = 0, giving a reading of 12000 signal amplitude. I had to calculate this based on the Z = 25mm data point at 10000 amplitude.

For the fixed inductance curves, I calculated Z = 0 amplitudes of:
200mm coil: 12000
300mm coil: 3270
400mm coil: 1275
500mm coil: 709
These seem to match well with your figures. This assumes 'received voltage' is proportional to 'received current'. (due to fixed inductance).

For the Fixed number of turns chart, the reference curve is the 200mm coil, with 12000 amplitude at Z = 0, as before.
I multiplied the 'recieved current' by the calculated inductance L = 278 uH, then found a scaling factor to get 12000.
For the remaining larger coils, the transmit (current x turns) was kept constant, as turns is kept constant. The received voltage was multiplied by the coil inductance, then by the previously-determined scaling factor. Results are:
200mm coil; L = 278 uH; 12000 amplitude
300mm coil; L = 458 uH; 8787
400mm coil; L = 651 uH; 7025
500mm coil; L = 853 uH; 5892

Notice that these are much closer together - you're gaining with the larger coils because: you're transmitting a constant (Amp-turn) product, and because: the coil inductance goes up roughly in proportion to coil diameter, so the received voltage goes up too.

I would be interested in your thoughts.