Help! Coil design?

I'm having a terrible time getting an air core coil wound for my RFID project. The big problem is that I simply don't know how to specify a coil, and the engineers at the magnetics company can't seem to help me fill in the blanks. Can anyone shed some light on how I should procede?

I've told them I'd like a 750uH coil with a max OD of

40mm to fit our physical constaints. Exact DC resistance doesn't matter too much, but I suggested 10-12 ohms or so based on some coils I wound by hand. Exact wire gauge doesn't matter either, so I suggested 34 gauge or so, again, based on my hand wound samples. I told them the coil thickness didn't matter much either, but a max of 5 mm or so seemed reasonable. The coil is to be operated at about 125 kHz.

I'd like to maximize the OD without exceeding the 40 mm spec, as it is acting as an antenna.

(Sketchy specs, I know, but I had to start somewhere.)

So, what did I forget and what do I do wrong? For the last week I've had to take calls daily with question after question..."can the ID be 10mm?" Probably not, if you expect to make the OD ~40mm, but how would I know? "What is the series resonant frequency?" I don't know that I care? "What is the Q?" Isn't that set by the inducatance, operating frequency, and the DCR? Why are you asking me?

I haven't felt this ignorant in at least a month. Should I have iterated through with the approximate coil formulas and just told them ID, thickness, gauge, and number of turns? Am I asking the impossible? Will I end up winding 500 coils my hand on a toilet paper tube form next week?

ARGH!

--
newell
Reply to
Scott Newell
Loading thread data ...

Well, they're probably thinking of using a multi-layer coil, and they want to know how far in they can go from the max ID. Either that or they're not listening.

Yes you do. Think of the series resonant frequency as the absolute, positive, maximum limit to the useful frequency range of the coil. You probably want to stay well below it if you're using the coil to set the frequency of operation.

And the coil size, and the winding arrangement, and the kind of wire (plain, Litz, braided, etc.)

Because you're the customer, and they see you as the source of the specifications that they should build to. They're used to building coils, testing them out on an LCR meter, and shipping the ones that pass.

Have you tried saying "Look, I'm no expert, and there's a lot of stuff I don't think is critical. Why don't you write _all_ the specs, and I'll buy to them".

Then be ready for frustration all around when your circuit doesn't work and you decide you need the coil changed...

--
http://www.wescottdesign.com
Reply to
Tim Wescott

I'd ask for their standard coil form sizes as a starting point. I can't say I've seen a 40mm form. That's nearly two inches. The again,

125KHz is more like audio frequency than RF. Maybe you need to find a company that makes components for speaker crossovers. Ah, but those use overlapping wires. This is quite the pickle you got yourself into. [RFID project?]

I suspect their are all sorts of pitfalls if you don't use one of their standard core sizes. Can their machine handle the size, blah blah blah.

Reply to
miso

Well, i started with the assumption tha a single-layer coil would do what you want, and used the chart in the Allied Radio Shack Electronics Data Handbook 5th Ed. 1970 pg 36. With N=125, (dia/len) is about 2.7; more turns decreases that ratio. The simple formula is L = (radius*number of turns) * (radius*number of turns) / (9*radius + 10*length) where dimension is in inches; OR Number of turns = (squareroot(inductance*(9*radius + 10*length)) / radius. Pick your wire for reasonable fit, as #0000 will no work here and #80 is too resistive..

Reply to
Robert Baer

Of course, but I didn't think that the strays in a 40mm coil would be anywhere near significant at such a low frequency (125kHz). Maybe I should check into that a bit. (Looks like SRF will be above 1-2 MHz, according to "multilay.exe".)

I guess I thought I could get by giving them some electrical specs and a rough physical box to fit in, and they'd work out the exact dimensions and turns to best suit their winding setup.

(Maybe I'll tell them to try 96 turns of 34 gauge, 39mm ID, 0.15mm high.)

That's pretty much what I said from the start. I can tune my circuit to fit the coil, and I'm adding resistance to bring the Q down to 10-15 in circuit at 125 kHz, so I didn't think it would matter if it came out to 700, 750, or 800 uH. Or if the coil Q was 40 or 60. (Consistent would be good, of course!)

The circuit is pretty forgiving--all the random wound coils I've done by hand have worked fine so far.

--
newell
Reply to
Scott Newell

do

tio.

ber

That's what I recall from one of the ancient ARRL handbooks. Well, the numerator was written r squared n squared, but still.

Anybody know where that equation came from?

Also, OP didn't mention whether a flat or solenoidal coil was preferred; I note that industry standards lean heavily to flat spirals, but that may mostly due to packaging considerations. Speak up, Scott.

Mark L. Fergerson

Reply to
alien8er

I was assuming a multilayer coil, but again, I'm not sure that we care. We're going to epoxy the coil/antenna to our housing and then fill said housing with potting compound. As long as it's not too large, the exact size and shape are not physically critical to us.

This is going in an RFID read head, not a tag. (The tags I've sectioned contain multilayer coils as well, but with much finer wire than I would use for my application!)

--
newell
Reply to
Scott Newell

I have had wildly varying success with asking a vendor to go outside of their usual ordering parameters and deliver something that works to the specifications that I care about. Some people just aren't comfortable outside of their little rut. Some people don't realize that there's a whole big world outside of their little rut. Some people, when presented with this sort of problem, go "cool! you want this and this and this, and if you'll bend on this then I can save you money (etc., etc., etc.)".

Unfortunately, I don't know of any way to find out if a particular place is one kind or another without giving them a whirl.

Have you tried a different coil house?

--
http://www.wescottdesign.com
Reply to
Tim Wescott

Change your supplier. They're requesting far too much information. Which leads one too suspect they don't really know what they're doing. I "specified"! a custom transformer Thursday morning as (quote)

"Nominal 240Vac~ primary with two low voltage secondary windings. Transformer is suggested at 30VA rating Secondary #1 is Nominal 6.6V ac at 2Amp ac~ Secondary #2 is nominal 12.7V at 2amp ac~ (Yes. I know the VA things don't add up to "30" but sod it :)"

I receive the prototype on Monday morning and will expect it to work, having had zero technical questions from it's designer. Suppliers should have the experience to fill in the basic stuff themselves.

Reply to
john jardine

They're not? You sure?

d

How far from the tag will the read head be? Will it be centered on any particular axis of the coil? What will the orientation of the tag's coil be WRT that of the read head?

The read coil will have a particular volume of sensitivity in which the tag should be placed, at a specific orientation, for best response. Consider the system working backwards with the read head generating a field in which you must place the tag.

Oh, hell:

formatting link

Mark L. Fergerson

Reply to
alien8er

From what i did using the nomograph, a single-layer coil will do the job rather well. Calculating L/D with the nomograph (or equations) thengives you the length L; from the number of turns, one then can calculate the wire diameter...(coil resistance will be rather low - for a decent Q).

Reply to
Robert Baer

The DIY coil house is best if only a few are needed...

Reply to
Robert Baer

With that info, almost anyone should be able to make what you asked for.

Reply to
Robert Baer

ned

They can read those damn things from 3ft!

Reply to
miso

I understand your frustration. Coil design is still in the dark ages.

I had the boss of a nearby coil house come in and ask me how to measure inductance at a 100kHz, his bridges only did 100Hz and 1kHz. Not my problem.

I've got problems sorting out the design of a transformer for an off line switcher. Nobody I've contacted seems to have a sane design of former. Most of them wobble around on spindly pins waiting to break.

Find a good former and the core material isn't available in less than

5000 pieces. Asymetric gaps? Can't be done.

I've hated coil houses since I got badly caught out with a 2A, 100uH,

50kHz coil 20 years ago. At 40C external air temperature the units stopped working properly after a few hours. The core of the ferrite material was overheating and causing the inductance to crash. As I hadn't specified the type of ferrite the coil house happily changed it between prototypes and production.

Imagine having to specify resistors and capacitors to your local resistor or capacitor house to get them made.

Reply to
Raveninghorde

The Tag Chip I originally designed around 7 years ago, and have updated every few years since, now has a range exceeding 30' ;-)

It pays to hire a pro ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
 Click to see the full signature
Reply to
Jim Thompson

Your wife would have a fit if you hired a hooker.

--
You can\'t have a sense of humor, if you have no sense!
Reply to
Michael A. Terrell

** I have had no problems getting 1-25 of any (nominally available) core shape and type, along with the bobbins - and ground to my specs. Everything was free as long as i paid for the grinding (not expensive considering everything). Asymmetric gaps? Not a problem.
Reply to
Robert Baer

Speak

care.

large,

sectioned

which

And thus rife with the potential for crosstalk.

Reply to
JosephKK

Yeah, you are in trouble bunky! My coil calculator show for the following:

OD =3D 1.5" ID =3D 1.1" thickness =3D .2" (makes a square cross-section)

Wire =3D 34 Ga. Ohms =3D 66.8 ohms Inductance =3D 25 millihenrys turns per layer =3D 27 layers =3D 28 total turns 756

Not so good.

With 27 ga wire and same size coil we get: Ohms =3D 2.7 inductance 1070 uH turns per layer =3D 12 Number of layers =3D 13 total turns156

Closer to what you asked for IF that is REALLY what you need!

Key unknown item is the distributed capacitance of the coil and hence the natural resonance frequency of it. That frequency should be WELL above your operating point. I'm guessing for these coils it will be. For a receiver coil you'd go with a fine as wire as you could get away with and still keep the coil resonance above your operating frequency.

There is more to this than that. If this is a receiving coil the number of turns is crucial to sensitivity because of a noise floor. As a transmitter you just run more current. But if the number of turns is too high and the wire too fine, it takes too much voltage to get the current you need.

Good Luck.

Reply to
Benj

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.