PI coil capacitance

[Chris]

Hi everybody

What is a typical inductance value range for a PI coil? I saw a few coil values in some designs but there were big difference between the designs.

And does someone know the typical capacitance range of commercial PI's coil between the windings and the shield? I measure something between 2.6nF and 3nF at my DIY coils. I measured between the shield and one end of the coil. Could it be correct (in range), wrong or ..... I don't know .... I'm not 100% sure. Need some help on this.....thanks in advance.
Another question is .... my coils are a shielded mono coil design (34wdg/10inch dia/AWG24). Where should I join shield and coil (GND). In front at the search head or at the electronic (PCB). Which point is better and why?

Regards
Chris

[Carl]

I've seen as low as 100uH, up to a couple of mH. Most commercial detectors with the standard front-end design, seem to be a few hundred uH at the most.

I measured my 8" coil, it's about 600pF, so you're in the ballpark. You didn't touch the probes or the coil while measuring, did you? That'll throw it off a bit.

Good question on the shield connection... I assumed that everyone ties it inside the coil. Probably doesn't make much difference, but I don't know for sure.

- Carl

[Reg]

Hi Chis,

I am guessing your coil has an inductance of something between 750 and 850 uh which is somewhat high for high pulse current applications. Are you using magnet wire or stranded hookup wire for your windings?

The higher the inductance value, the harder it will be to reduce the delay because it will take longer for the coil signal to decay.

The figures I have seen for PI coils are inductances between 300 uh and 1 mh. Most of the higher powered units operate with a lower inductance.

Without knowing the application or what you want to search for, I would not want to guess what yours should be.

Now, the capacitance between the coil windings and the shield of your coil seems a little high. What are you using for shielding?

My guess is your shielding is right on top of the windings with no space at all between the windings and the shield. Eric Foster mentioned he uses spirawrap around the windings as a spacer between the windings and the shield. This seems to work just fine. Other people have also indicated a spacer is a good idea.

I measure about 100 pf or so between the shield and the windings with my BK LCR meter. have the shield spaced out from the windings using spirawrap.

I can't say where is the best place to connect the shield, but I connect the shield with one lead of the coil to the shield of the coax used to connect the coil to the control. This effectively ties the shield to the common of the electronics. Doing it this way simplifies wiring between the coil and the control unit and it seems to work just fine.

I hope this helps.

Reg

[Reg]

Hi Carl,

What are you using for shielding and how are you spacing it from the windings?

Reg

[Carl]

>What are you using for shielding and how are you spacing it from the windings?

Adhesive foil, what's used for A/C ductwork. I did not add a spacer.

Before I shielded the coil, I measured the inductance, and put the decay curve on the o-scope, and air-tested a couple of targets. I did the same after shielding, got no change in inductance, decay looked the same, and air tests seemed the same. Except for measuring the inductance, this is all subjective. Next time I wind up a coil, I'll try to look at this a little closer.

- Carl

[Chris]

Hi Reg & Carl,

thanks for your kind reply. I want to make a good general purpose PI coil. No special application in mind. I'm using magnetic wire (enamelled copper wire) and for the housing the coil shells from Hayes Electronics.
I don't know the inductance of my coil. I don't have a LCR meter.
I measured the coil capacitance with my digi. multimeter and compared the reading with a second one (diff. model) but with the same reading without touching the probes. The value of 2.6nF includes the 1.8m cable as well. Maybe the cable I used is not the best to use as a search coil cable. The cable is a shielded twisted pair cable used for S-VIDEO application (US$4/meter).

For the shield around the coil I use lead tape from 3M (suggested by Eric Foster). My coil is nicely wounded (layer by layer) and there is no spacer between the windings and the shield. The shield is connected together with "Signal-Ground" at the PCB side.

Is stranded hookup wire better than enamelled copper wire to use in PI coils? If yes, why? And what is with HF-wire (high frequency wire)?

Chris

[Eric Foster]

Hi Carl,

What metal is the foil? What I have seen over here is usually aluminium. This is OK as long as the foil is thin enough, otherwise you generate eddy currents in the shield which mask those of the target. Comparing the coil recovery times of identical shielded and unshielded coils, shows to what degree this is happening.

Eric.

[Chris]

Hi Reg,

I measured the inductance of my coil with a scope today.....your guess of 750 to 850uH was quite good. I measured 870uH with the scope.
Is this to much for a general purpose coil? Should I reduce the number of windings?
Somewhere I read that are special search coils for gold around (coiltec). What is so special or different? Different (higher) frequency range, more or less inductance than general purpose PI coils, or lower coil capacitance?

Regards
Chris

[Reg]

Hi Chris,

Since you do not have a special application, I would suspect the wire you use would be ok.

Now, if you want to try to use the PI for detecting small gold, then some adjustments probably should be made. If coins and general metal objects are the main objects to be detected, then things like a very short delay isn't that critical.

As for your question about enameled wire versus stranded wire, the answer to that lies in the application also.

One of the things you can do to get an idea of how the coil is effected by using different wire is to simply build two and see how they oscillate. This is a simple process.

Just build the coil but don't connect it to the PI. Instead, have a coil already on the PI. Now, what you will do is to place the new coil close to the operating PI coil and look at the oscillation that occurs by connecting your scope to the new coil leads. This new coil will have a spike imposed upon it from the operating detector, which should ring the new coil at its natural frequency. This frequency is determined by the inductance internal capacitance.

If you know the inductance of the coil and measure the frequency of oscillation, then you can calculate the capacitance. However, I am lazy and do not have a spreadsheet set up with the formulas to calculate the capacitance so what I did was to simulate the LC circuit using a spice program. I knew the frequency of oscillation, I knew the inductance, and I had an idea of what the capacitance might be so I just plugged in values for the capacitor until I got the same frequency. This is simple, sloppy but effective.

I wound a couple of coils using stranded insulated wire on one and basically enameled wire on another. The oscillating frequency was quite different between the two indicating there was a significant difference in capacitance. The stranded hookup wire oscillated at a much higher frequency indicating the capacitance was much less.

This lower capacitance really only becomes important when you want to have the coil decay as fast as possible so you canoperate with a very short delay. This is necessary for looking for very small gold objects. I would think it would be a non issue for coin type objects.

I am not sure of your question about HF-wire. What I have found is some of the so called special wire built for high frequency operation has quite a bit of capacitance built in. In fact, when I was trying to find a very low cap wire for my DD coil applications, I tried a few different SVHS cables and found the cheaper ones to be worthless for my application. The Mogami cable seemed to be the best I could find.

For your application I would simply use RG 58 AU coax or something similar. This should work quite well. As for your coil shield, I would simply tie it to the coax shield along with one lead from the coil. I doubt you will see any improvement by taking a separate lead back to the pc board. If you do, let me know.

Now, I haven't given much thought as to how important the capacitance of the shield is on the operation of the PI. I will have to think about that one for a while.

I also have not taken time to see just what the impact of having the shield closer or farther away from the windings. I really should do some experimenting to see. I can see where it might be more critical to be sure the shield has a small gap due to the coil's ability to detect the shielding. What I don't know is just what the capacitance of this shielding will have on the detection of the shield or the over all decay time.

Reg

[Chris]

Hi Reg

A spacer between the coil and shield reduce the capacitance for sure...as everybody know C is proportional to 1/distance (between the capacitor plates). You can read this in every electronic reference book (capacitor formula).

Like you suggested I'll do some experiments to see the difference in stranded and enmalled wire in coils.

I looked also at the data sheet of the cable that I used for my coil. The capacitance of my cable (57pF/Ft) is much higher than RG58AU coax cable (100pF/m). I'll replace it as well.

And another question....you wrote....

"As for your question about enameled wire versus stranded wire, the answer to that lies in the application also."

What do you mean? Which wire would you take for which for application if you want to design a coil?

Chris

[Reg]

Hi Chris,

I will try to give you and example based upon my PI which isn't drawing a lot of coil current during the pulse on time. This makes a difference also.

In my application I wanted to detect very small gold which which produces a weak signal that decays very rapidly.

Now, I had two choices, first to increase coil current which should produce a stronger signal or I could try to reduce the delay before sample. This would allow me to sample when the signal was stronger.

Obviously, one could do both but that was not an easy option for me due to other constraints.

Also, the size of gold I wanted to detect required the sample delay be short simply because the signal would diminish to nothing very quickly.

So, I needed to reduce the coil inductance and capacitance to allow me to sample sooner. The capacitance becomes important when trying to squeeze the delay down a couple of usec's which I needed to do to allow for the fastest decay possible.

In this case I tied to minimize coil and coil cable capacitance and pick an inductance value that would also allow for a faster decay. The reduction in capacitance in the windings and in the coil cable would allow for a higher resistance value damping resistor which would dampen the coil without oscillation. Together, the decay time would be shorter which then would allow for sampling sooner.

All of this is a balancing act to try to determine the right inductance, etc for the specific application.

So, for applications where the you want to reduce the decay time, you probably should select the wire that provides the least capacitance.

Now, for applications where it is not necessary to sample extremely early, then any wire should probably work just fine. I can't tell you the critical point where the wire becomes a problem. My guess is as long as the sampling is delay is more than 12 to 15 usec or so, then the wire is not a main issue. Inductance and coil current become a problem then.

For applications such as beach hunting where rings and coins are the primary objective, then a longer delay should be ok to a point. If thin gold rings are to be detected, then the minimum delay becomes reasonably important again.

One should also remember that as the delay is shortened then all low conductive items become targets that can be heard. Things like tin foil in gum wrappers can produce a strong signal.

If coin hunting is the objective, then the minimum delay isn't as critical. Pulse on time may become a factor for maximum depth.

If you are looking for iron objects, then the short delay is of little value. In this case, a short delay will cause more problems because there will be more ground signal to contend with.

So, as you can tell, it is difficult to give you a specific answer to your question. Everything is a tradeoff. Reduce the delay time, you can detect small gold better but you will also have stronger ground signals, detect more junk, etc.

Lengthen the delay and the junk can be ignored but some other objects that may be desireable will be ignored also.

Reg

[Reg]

Hi Carl,

Thanks for the input. I have never tried the foil you mentioned. I do have a couple of small pieces of the stuff and find there are different thickness of the foil. At least, there is on the ones that I have.

I will have to try the foil also if I can find the thin stuff. I have tried most everything else. One of the strangest things I tried was a plain old decorative ribbon I got from a Wal Mart store. The ribbon has a thin aluminized layer and is very conductive.

It seemed to work but I really never did do any extensive testing. Instead, I went back to my special shielding tape, the 3M 1190 copper coated polyester foil. This seems to work fine for my applications but it is a little expensive and hard to find.

Also, when time permits, I need to see just what obvious difference there is to having the shield right on the windings and spaced out a little.


Reg

[Allan]

Carl, Eric, Reg, Chris and others on this great site. I'm learning heaps thanks to all your help. Here is my first maybe dumb idea, for a low capacitance coil.
Remove the pvc outer from however many metres of coax required for your coil. Then pull out the inner wire with insulation or unravel the braid. The inner conductor with its thick insulation (air cored for low weight) would wind a low capacitance coil with the end connected back to the shield. You then could even use the braid wrapped around the coil as the shield.
Second maybe dumb idea, after the discussion about ground balance on the PI site,
and target masking. Pertaining to Carls PI-1.
Tap off the 555 pulse output, divide with a 4017, and flip-flop to toggle the spare 4066's and another resistor across the sample delay and pulse width resistors. I realise this will increase noise in the integrator requiring a slower time constant and therfore a slower sweep speed.Probably need a dual 555 timer circuit to eliminate the pulse width/pulse speed interaction but this could fit in with a active fet turn off to help with the quick sample.Does this make sense? regards ,Allan.

[Carl]

It's aluminum, a bit thicker than household foil. I did check decay and detection depth before & after the shield, but I did not do a side-by-side comparison. I need to do this.

- Carl

[Reg]

Hi Allan,

Sorry I missed your post. One could do what you recommend to try to reduce the capacitance of the coil but the coil would get pretty fat in a hurry. I am not sure it would be as much of a help as taking more care in the winding techniques.

Besides, I have found that just using hookup wire reduces the capacitance significantly. It seems to be enough for most applications and I don't have to be that touchy in winding the coil.

Now, as for your idea of switching the pulse delay, I am assuming you mean to do this for a quick check of the target by increasing the delay. Your idea should work, but I am doing something like that now, but simply use a simple toggle switch to change the delay. Thi circuit I am using is a little different so I am not sure a simple switch wouldn't introduce just a little noise, though. My change allows a quick check of a target to see if it is gold or not.

I suspect that if you do not slow everything down, you will not have to adjust the sweep speed.

One of the more unique ways to change things would be to build types of coils such as the figure 8 receive coil. This way you would have noise cancelling and ground cancelling. Recitify the outputs so you eliminate the half positive and half negative signals. Then you could use the second sample normally used for subtracting the earth field effect, to check a target at a later delay.

I am planning on doing more testing using a little different technique on my DD coils. I also plan on trying a concentric type coil again. I tried this once but really wasn't set up right at the time. This time I will try a couple things that are a little different.

As you can tell, there are lots of ideas to play with.

Reg

[Allan]

Reg, What I meant was to alternately have a long pulse and later sample / short pulse and early sample; to find the big bits / little bits. eg 7short / 3 long. Does that make sense now?
regards, Allan.

[Reg]

Hi Alan,

I am sorry about my misunderstanding of why the pulse change also.

I am using a very short pulse now and all testing I have done indicates I lose very little depth ,if any, due to pulse length on my gold nuggets I have which range from about 1 grain to a little over a half oz.

I suspect any depth loss due to a short pulse would be reflected in signal loss on very large nuggets ranging in the multiple oz range. However, a longer pulse should work better for increasing the depth on copper and silver objects. Unfortunately, I have done no testing in this area so I can't tell you just what to expect since I hunt strickly for gold nuggets with my PI.

Your idea of using two pulse lengths is a good one but one should build a second differential amp for amplifying the long pulse signal. Summation of the long pulse/short pulse received signals should be summed at a later point for best results.

Now, to keep the project simple, I would start out by selecting a reaonably short pulse length, maybe 50 usec to 100 usec, concentrate on reducing the sampling delay, reducing noise, maybe increasing the overall circuit gain, and refining the coils first before trying something really new like multiple sampling.

I would also recommend a person spend a little time experimenting with different sampling widths. Don't be afraid of trying widths as low as a couple of usecs.

Once these areas are maximized, then I would try a dual sampling project. This way, one will have a better means of measuring or noticing any side effects that will occur from dual sampling or any other idea one might try.

Reg

[Russ_NY]

Hi Reg
What type of hookup wire are you using? Number of strands and gauge? Tinned or untinned? Type of insulation? How about pre-Tinned wire where the strands are already soldered together, I suspect that is probably similar to solid hookup wire. Any recomendations on which manufacturer to use? I need to restock before I can do any serious coil winding and I really want to get the right wire so any help would be greatly appreciated.
Thanks,
Russ

[Norbert]

Hello Chris!

It's easy to measure the coil inductance.
U = L * I / t
so..
L = U * t / I

L in H
U in V (voltage you switch to the coil)
t in seconds
I in amps (peak-current)

[Chris]

Hi Norbert

yes, I know.... there is also another easy way. See Link.

Chris



Inductance Measuring Technique (http://www.coilgun.eclipse.co.uk/inductance_measuring_technique.html)

[Reg]

Hi Russ,

First, I am not the great expert on PI's so don't take my word as golspel. Personally, I like to try different things and measure the results. This give me a better feel for how things interact.

Rather than buy a large quantity of any particular wire, I would probably try different types and see what happens if all other things are held constant. You may find that magnet wire works fine for your application. Magnet wire is definitely cheaper and if it works then I would use it. The key is to set an objective and then set out to reach it.

For me, that objective was to reduce the delay so I could detect smaller gold with a stronger signal. This required reducing the inductance and then the capacitance to be able to sample very quickly. Building a DD coil helps in this regard and careful alignment allows sampling even at shorter delays than with a mono coil.

Now, I found I could reduce the delay down to 12 to 15 usec without great difficulty. It did take some doing but nothing like dropping below 12 usec. The big factor was lowering the inductance to somewhere between 300 uh and 400 uh. The next key factor is the damping resistor. Also, don't use any compensation caps on the preamp.

Going to a DD coil helps a lot also, since the preamp doesn't saturate as bad. Also, the nulling of the coil really helps.

Now, to answer your question, I use a fairly small wire, about 29 awg stranded tin wire that has a teflon coating. But, I am making coils to use on Eric's Beachscan/GQ design and this PI doesn't draw as much coil current as Carl's project. I happened across this particular wire and got it cheap. That is probably the main reason I tried it.

I really need to build Carls PI and I will build it, once I am satisfied with the mods I am making on my present PI. I am getting close to that point now.

Teflon insulation is generally a little thinner than pvc insulation for the same voltage rating so the coil windings won't be quite as bulky. Unfortunately, Teflon is also quite expensive so I look for places that sell it surplus.

As for whether stranded, tinned, etc is better, I really can't say. I just noticed the coil capacitance was much lower when the wire has a little spacing between windings. In fact, the capacitance was about 1/2.

I have also tried Litz wire and it didn't work that bad. I didn't try served Litz which probably would have worked better for reducing the capacitance than the unserved type.

As for finding surplus wire, I found that Surplus Sales of Nebraska has some decent prices on teflon and Litz wire. The sizes are limited but that is ok if there is a size that works. I would probably try their served Litz on Carl's project and maybe either a 26 awg or 24 awg teflon.

World Wide Wire has a large inventory of wire with reasonable prices. The down side is they require a $100 minimum purchase.

[Russ_NY]

Reg
Thanks for the response. I have been using magnet wire as well as stranded PVS coated hook up wire, keeping the inductance in the 300uH range. No compensation caps on the amp. Reduced the gain to about 500 on the amp. I also use a series resistance between the coil and Mosfet of about 2.2 ohms to reduce the coil current and have the prf running at just over 2kHz with a 100usec on time. I spaced the shield away from the coil using spiral wrap, and still have trouble getting the delay under 20usec. I would like to try a DD coil but am not really clear on how to implement it into the circuit. I assume that one coil goes to each input of the amp, and that both coils are switched by seperate Mosfets at the same time. I have only seen one schematic for a dual coil PI and it was not documented at all. Any guidance on how to set up a DD coil into what is esentially Carl's circuit would be greatly appreciated.
Thanks again

Russ_NY

[Reg]

Hi Russ,

I am wondering about your decay time. I would think you should be able to get down to 15 usec without a major problem.

What do you have for a damping resistor and did your try higher values?

As for the DD coil setup, it is quite easy. You will need at least a 3 pin connector though. The third pin will be for the receive coil connection back to the pc board. Now, once the project is complete, the xmit will be isolated from the receive circuit.

1. You will have two separate isolated elliptical coils in the coil housing. One will be the xmit coil and the other will be the receive coil. Normally they overlap a little less than an inch or so. The actual overlap is determined by coil size, etc. One coil, the xmit coil, connects just like the mono coil. The receive coil will connect to the preamp circuit via R 12 (on Carl's PI).

2. You have to cut the trace between the xmit coil connection and the receive circuit to isolate the receive from the xmit. This means, cutting the trace from R 12 to the xmit coil connection. Now, R 12 will be open ended on the xmit side.

3. The new receive coil will now connect to the open end of R 12. If a plug is used, then a third pin will be used for this purpose to connect R 12 to the receive coil.

4. The commons of both coils should tie to the pc board common. They can tie together at the coil cable plug but I found it best to not tie them together in the coil.

5. You will need a separate damping resistor for the DD receive coil and that can be connected inside the coil housing or in the coil cable plug.

6. I found shielding both coils just like is done on the mono coil works fine. However, I am using a different shielding material that is not detected or if it is, is very slight.

7. The easiest way to to make the mono coils work with this setup is to add a jumper inside the coil cable plug to connect the xmit back to the receive. So, lets say you have a 3 pin plug with the following; pin 1 is the receive pin, pin 2 is the common, and pin 3 is the xmit pin. Now, on the mono coil, the coil connects to pin 3 and pin 2, and there will be a jumper between pin 3 and pin 1.

On a DD coil, the xmit connects to pin 2 and pin 3 in the plug. The receive connects to pin 1 and pin 2 in the plug. The receive damping resistor can connect to pin 1 and pin 2 or be mounted in the coil.

Remember, the plug should have at least 3 pins. One pin for common, one pin for the xmit and one pin for the receive. You will need a dual coax. I use a couple of different types. One that would probably work best for you is a Mogami cable. Look for the dual coax that has a low capacitance. The cheapest way to buy it is to buy the Mogami S video cable. Other SVHS cables I have tried have too much capacitance, so I have stayed with the Mogami.

I hope this is enough info to get you started.

Reg

[Russ_NY]

Reg
Thanks so much. You cleared up all of my questions concerning how to set up for the DD coil. I am currently using a 480 ohm damping resistor. I have tried higher values and the 480 was the highest value I tried where I did not get any ringing. If I remember correctly I tried a 520 and had what looked like about a half cycle of ringing. I will go back and check it to be sure. I don't recall having seen this Mogami cable but I will check around. I am pretty sure I have a few SVHS cables in my collection. I do have a fairly good supply of different types of coax cable. I pretty much cut 1 foot lengths of several and measured the capacitance and selected the lowest one. It happened to be some RG 59 I had on hand. I don't recall at the moment what it measured but I will check that out again also. As to the coil shielding I used the self adhesive aluminum tape. It is fairly heavy and not particularly easy to use. I had to use a drain wire spiraled around it to get a decent ground connection. I would like to try some sort of spray or paint on conductive coating but haven't tried to find any as yet. I well definitely try out the DD coil as soon possible.
Thanks again

Russ_NY

[Reg]

Hi Russ,

You might try this website.
http://www.hometech.com/video/interconnects.html#MO-PPSV403

I purchased a 50 foot cable and just use the amount I need for each coil. It was easier and cheaper to do it this way than trying to find and then buying the cable by itself.

Reg

[Reg]

Hi Russ,

I suspect the heavy aluminum shielding is part of the problem. I know it will be if you build a DD coil.

What happens when you try a coil on the bench without the shielding? Do you still run into the same problem?

Reg

[Norbert]

Is it important to switch off the coil very fast to get a very short delay time?

HH
Norbert

[Reg]

Hi Norbert,

Normally, yes you want to switch the coil off as quickly as possible.

Reg