Searching \ for '[EE] PCB or air core inductor?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/pcbs.htm?key=pcb
Search entire site for: 'PCB or air core inductor?'.

Exact match. Not showing close matches.
PICList Thread
'[EE] PCB or air core inductor?'
2007\05\10@103759 by Gerhard Fiedler

picon face
Hello,

I'm working on a transmitter and receiver in the 400MHz range. The
inductors in that frequency range are really small (tens of nH) and I
thought that buying such an inductor may not make much sense. So I looked
into PCB and air core inductors. Here's what I found so far:

<http://www.educypedia.be/electronics/electroniccalculatorcomp.htm> Lots of
links to all kinds of calculator sites.

<http://emcsun.ece.umr.edu/new-induct/> Inductor calculations for all kinds
of geometries, but not PCB spirals.

<http://smirc.stanford.edu/spiralCalc.html> Calculator for integrated
spiral inductors; may work for PCB inductors also, but I don't know what
they assume in terms of what's around the conductor.

<http://tinyurl.com/2rn89n> A paper by Marc T. Thompson listing a number of
formulas for PCB and wire coils.

<http://www.freelists.org/archives/si-list/11-2001/msg00109.html> Steve
Rogers, who seems to have done some work with PCB inductors, lists a number
of sources of information and describes some of the potential problems with
PCB inductors.


Has anybody experience with PCB inductors? Or maybe even with a
formula/calculator that yields realistic results? Are simple air core
inductors much easier to get right?

Thanks,
Gerhard

2007\05\10@123117 by Dario Greggio

face picon face
Gerhard Fiedler wrote:

> Has anybody experience with PCB inductors? Or maybe even with a

IMO, the easiest way would be "copy an existing PCB inductor" :-)

Seriously, I tried some calculations, time ago... and actually it was a
kind of complicate.

--
Ciao, Dario

2007\05\10@131533 by Peter P.

picon face
Imho your problem will be the board material not the inductor proper. At that
frequency on FR board inductors are more or less transmission lines that need a
little more capacitive loading at the ends to get them into range. The 'normal'
way to get over this is to approximate an inductor (one turn or part thereof)
and a place for a SMD cap at its end (closure) and build it and measure it.
Rinse & repeat until it starts working. Things like trace proximity (including
to iself - as in spiral or meander) can change the frequency by 1-3% (10MHz off
at 400MHz).

Peter P.


2007\05\10@143307 by Vasile Surducan

face picon face
You didn't say if the inductors will be used for LC oscillator or just
for filtering.
At 400MHz and FR4, designing inductors is not an insurmountable problem.
All links you've provided are interesting, but for practical reasons
two configuration are widely used at such frequencies: for
oscillators/antennas spiral inductors and for filtering accordeons
inductors. Some software are allowing you simulation (microwave office
if I remember well).
Both types above are require adjustment methodes by short circuiting
spirals or accordeons part at one end (the way is to add top/bottom
stop surfaces in CAD which will create islands in the mask layers so
further you may solder anywhere.

Russians have used such PCB inductor tehnique back in the 75 designing
an 1Ghz vobulator using varactor multiplication and a bunch of
inductances starting from a 10Mhz crystal oscillating with overtone. I
still have one which works but have a defective frequency meter.

A nice software here:
http://rfic.eecs.berkeley.edu/~niknejad/asitic.html

Vasile



On 5/10/07, Gerhard Fiedler <spam_OUTlistsTakeThisOuTspamconnectionbrazil.com> wrote:
{Quote hidden}

> -

2007\05\10@192845 by Gerhard Fiedler

picon face
Vasile Surducan wrote:

> You didn't say if the inductors will be used for LC oscillator or just
> for filtering.

I plan on using it in a Colpitts SAW oscillator. So it is kind of
determining the oscillator's frequency, but only approximately. The SAW
resonator will determine the frequency.

> [...] for practical reasons two configuration are widely used at such
> frequencies: for oscillators/antennas spiral inductors and for filtering
> accordeons inductors. Some software are allowing you simulation
> (microwave office if I remember well).

I probably can't spend the time to learn how to use that tool, even though
I could probably do something useful in the 30-day evaluation period :)

> Both types above are require adjustment methodes by short circuiting
> spirals or accordeons part at one end (the way is to add top/bottom stop
> surfaces in CAD which will create islands in the mask layers so further
> you may solder anywhere.

Good idea, this.

> A nice software here:
> http://rfic.eecs.berkeley.edu/~niknejad/asitic.html

Ah, thanks, this looks like it's worth a shot.

Maybe I just buy a kit of nH inductors... :)

Gerhard

2007\05\10@193009 by Gerhard Fiedler

picon face
Dario Greggio wrote:

>> Has anybody experience with PCB inductors? Or maybe even with a
>
> IMO, the easiest way would be "copy an existing PCB inductor" :-)

Hm...  I don't really have a cheat sheet of inductors of the size and type
I need, so that doesn't work for me :)

Gerhard

2007\05\10@193505 by Gerhard Fiedler

picon face
Peter P. wrote:

> Imho your problem will be the board material not the inductor proper. At
> that frequency on FR board inductors are more or less transmission lines
> that need a little more capacitive loading at the ends to get them into
> range. The 'normal' way to get over this is to approximate an inductor
> (one turn or part thereof) and a place for a SMD cap at its end
> (closure) and build it and measure it. Rinse & repeat until it starts
> working. Things like trace proximity (including to iself - as in spiral
> or meander) can change the frequency by 1-3% (10MHz off at 400MHz).

Does this matter a lot for a SAW stabilized oscillator?

Also, without being able to actually measure the parameters of this
transmission line, how do I know whether the inductor itself is off or the
capacitor?

Thanks,
Gerhard

2007\05\10@202324 by Peter P.

picon face
Gerhard Fiedler <lists <at> connectionbrazil.com> writes:

> Does this matter a lot for a SAW stabilized oscillator?

You can't peak the output if the inductor is off. Iow, use a trimmer cap and
keep changing the board until you get maximum out tunable with the cap (not at
the end of the tunable domain - there should be two peaks per cap turn). After
that you can put in a SMD cap and 'tune' the inductor as Vasile said. The
'inductor' at 433 MHz is a J or U shaped one turn of copper trace. You can
calculate it yourself, take 5nH/cm of coil wire length and Thomson's formula.
The tuning cap will be mostly the transistor collector plus the trimmer plus
stray capacitance on the board, i.e. you can't go below 5 pF probably. So for
433 MHz and 5 nH/cm and 5 pF it works out to about 5 cm length. If you look in
a keyfob TX you will see a trace about this long. The width of the trace
changes both the 5 nH/cm parameter and self capacitance. So: as I said, take
some values, cross your fingers, build it, and see if it works. Then rinse and
repeat. 'Calculating' this without concrete data for the board material and what
is near the circuit (battery, user's hand etc) is useless.

> Also, without being able to actually measure the parameters of this
> transmission line, how do I know whether the inductor itself is off or the
> capacitor?

RF without instruments is like photography without light. You need 'some'
experience to get things done, but you can never be sure that your transmission
is not mostly on the 3rd harmonic or worse without looking with something. At
the very least a Schottky diode detector hooked to a scope, and an absorbtion
tuner, but that's fairly lame as 'instruments' go. A ham radio receiver for 432
MHz with working S-meter and a converter (10 MHz xtal + mixer) is an asset. If
this is a serious project (more than a one-off), rent a specan or rent hours at
a lab that has one. You need a specan that covers the 3rd harmonic at least
(better the Ft of the used transistor).

Peter P.


2007\05\10@202950 by Peter P.
picon face
> buy nH inductor kit

Yeah. Get the kit where you also get a complimentary bridge in the desert
included. On the other hand, I have a piece of tested nH inductor material here,
it is known to work, going cheap, only $0.5/nH + P&P. FYI there is a file on the
web called 'aircoil.htm' or 'aircoils.htm'. It is a table of inductors for the
tens to thousands of nH range suitable for FM and 144 MHz work.

Peter P.



2007\05\10@212334 by Peter P.

picon face

Fun reading on 'short wires'

http://www.emcesd.com/tt2001/tt010101.htm

Peter


2007\05\10@220032 by Gerhard Fiedler

picon face
Peter P. wrote:

>> buy nH inductor kit
>
> Yeah. Get the kit where you also get a complimentary bridge in the desert
> included.

I didn't understand this... I guess you mean it's not worth it?

> FYI there is a file on the web called 'aircoil.htm' or 'aircoils.htm'. It
> is a table of inductors for the tens to thousands of nH range suitable
> for FM and 144 MHz work.

You mean <http://hem.passagen.se/communication/aircoil.htm>? (Many hits on
"aircoil.htm".) But I think the aircoil calculations are rather
straightforward -- if they work :)

If I understand you correctly, you seem to say that using self-made
aircoils is a lot safer (in the sense of fewer "rinse and repeat" cycles :)
than using PCB coils?

Gerhard

2007\05\10@224719 by David Bengtson

picon face
It really depends on whether you plan to make a bunch of these or if you are
just making one. If you are just making one, I'd suggest a sample kit of
inductors as the easiest thing to do.

PCB coils are straightforward to do, but they take up board space and they
will need some tweaking to get the right dimensions (Translates to PCB
spins). If you factor in the additional board size, it may be cheaper to use
a coil.

The other problem with PCB Coils in production is that if you get in a new
batch of PCB material, your inductance could well shift, since the
inductance is dependent on the board dielectric constant.

Dave

On 5/10/07, Gerhard Fiedler <.....listsKILLspamspam@spam@connectionbrazil.com> wrote:
{Quote hidden}

> -

2007\05\11@051246 by Peter P.

picon face
Gerhard Fiedler <lists <at> connectionbrazil.com> writes:

> >> buy nH inductor kit
> >
> > Yeah. Get the kit where you also get a complimentary bridge in the desert
> > included.
>
> I didn't understand this... I guess you mean it's not worth it?

Yup. It would be a box with pieces of wire and instructions to avoid bending
them or putting them near any metal or dielectric ...

> > FYI there is a file on the web called 'aircoil.htm' or 'aircoils.htm'. It
> > is a table of inductors for the tens to thousands of nH range suitable
> > for FM and 144 MHz work.
>
> You mean <http://hem.passagen.se/communication/aircoil.htm>? (Many hits on
> "aircoil.htm".) But I think the aircoil calculations are rather
> straightforward -- if they work :)

No, it's not that file. The file I had in mind is off the net. My copy was
downloaded in 1998. The calculations are as useless as the ready made inductors:
they cannot predict the real situation. As I said, small details can pull the
frequency off 1-3%. That's 10 MHz at 433 MHz (way too much).

> If I understand you correctly, you seem to say that using self-made
> aircoils is a lot safer (in the sense of fewer "rinse and repeat" cycles :)
> than using PCB coils?

No, it's the same, but they don't cost extra money or waste of time. If you
cannot 'see' what you are doing it would take a lot of experience to get things
done like that and you will never be sure of what comes out. Try very hard to
locate a place where they have a specan. Any lab that does cell phones or handy
phones (UHF) should have one. Or try to rent from an equipment supplier.

Peter P.


2007\05\11@060910 by Alan B. Pearce

face picon face
>Fun reading on 'short wires'
>
> http://www.emcesd.com/tt2001/tt010101.htm

Humph - I fair took the mickey out of a colleague who was attempting to look
at a 5V digital signal, and couldn't see anything on the scope. His ground
lead was about 4' of hook-up wire from the ground post on the scope front
panel, and no ground lead at all on the probe. The picture was that swamped
with 50Hz that there was absolutely no way he was going to see a typical 5V
digital signal that was sub-microsecond long.

2007\05\11@082613 by Gerhard Fiedler

picon face
David Bengtson wrote:

> It really depends on whether you plan to make a bunch of these or if you
> are just making one. If you are just making one, I'd suggest a sample
> kit of inductors as the easiest thing to do.

I'm planning on making a bunch of these, but I still am tending towards the
kit of inductors, for the reasons you (and Peter) mentioned.

> The other problem with PCB Coils in production is that if you get in a
> new batch of PCB material, your inductance could well shift, since the
> inductance is dependent on the board dielectric constant.

I thought that PCB coils would give me good repeatability, at least after
it has been adjusted, and for free in production, but I didn't think of
this. I guess it's not what I want, after all... :)

Thanks,
Gerhard

2007\05\11@085321 by Gerhard Fiedler

picon face
Peter P. wrote:

>>>> buy nH inductor kit
>>>
>>> Yeah. Get the kit where you also get a complimentary bridge in the
>>> desert included.
>>
>> I didn't understand this... I guess you mean it's not worth it?
>
> Yup. It would be a box with pieces of wire and instructions to avoid
> bending them or putting them near any metal or dielectric ...

I thought about SMD inductors, and about repeatability. Keeping a tight PCB
layout with short traces to reduce dependency on the PCB and using
inductors that come with reasonable tolerances would make it more probable
that the next device works like the one before -- I thought. Did I think
wrong?

> [...] they cannot predict the real situation. As I said, small details
> can pull the frequency off 1-3%. That's 10 MHz at 433 MHz (way too
> much).

I have to continue discussing this... When I look at a typical keyfob, they
seem to use standard SMD capacitors and inductors -- with tolerances
probably more than 3% (at least 5%). And yet they seem to work fine,
probably without individual adjustment. From this, I gather that the
frequency of the LC resonance being off 3% can't be that crucial... Again,
am I wrong?

> Try very hard to locate a place where they have a specan. Any lab that
> does cell phones or handy phones (UHF) should have one. Or try to rent
> from an equipment supplier.

Good idea; I'll see what I can find.

Thanks,
Gerhard

2007\05\11@101123 by Peter P.

picon face
Gerhard Fiedler <lists <at> connectionbrazil.com> writes:

{Quote hidden}

There is a difference between the radiating resonant circuit and any other. The
radiating circuit is relatively heavily loaded and forced-driven with the SAW
setting the frequency, so its Q is low and the caps don't matter so much, it
goes more for maximum area to achieve field strength at a distance. That's why
no SMD inductor is used for the radiating element (however a correctly chosen
stripline should work). Still, there is such a thing as a 'keyfob tune-up'
which consists in replacing one of the SMDs with a trimmer and tuning the
device for maximum output. Daring people (who have access to a specan) can also
modify the tap point for the transistor collector. This could increase the
range by 30-50%. The  manufacturers design with wide tolerances, and the
circuit is far from electrically optimal when designed for manufacturability.
I think that you need to understand the tradeoffs between Q, distance,
fieldstrength and sensitivity to parts and labor before embarking on this
project.

Peter P.


2007\05\11@142339 by Peter P.

picon face
The incredible electric chair ... :

 http://www.emcesd.com/pdf/eos93.pdf

Peter P.




2007\05\11@154227 by Peter P.

picon face
The incredible copper strip ... :

http://www.emcesd.com/tt2001/tt060101.htm

Peter P.


2007\05\11@220451 by David Bengtson

picon face
On 5/11/07, Gerhard Fiedler <listsspamKILLspamconnectionbrazil.com> wrote:
{Quote hidden}

Dave

2007\05\11@224947 by David Bengtson

picon face
On 5/11/07, Gerhard Fiedler <.....listsKILLspamspam.....connectionbrazil.com> wrote:
{Quote hidden}

You are correct. It may be that  you can get 5% tolerances with a PCB coil,
but you'll have to do some
calculations to figure it out for sure.

{Quote hidden}

Resonant frequency  is F=1/(2*pi*sqrt(L*C)), so if  the inductance goes up
by 3%, the frequency goes down by ~1.5%.  For a key fob transmitter, that's
pretty good, they are really simple and sloppy.

> Try very hard to locate a place where they have a specan. Any lab that
> > does cell phones or handy phones (UHF) should have one. Or try to rent
> > from an equipment supplier.
>
> Good idea; I'll see what I can find.
>
> Thanks,
> Gerhard
>
> -

More... (looser matching)
- Last day of these posts
- In 2007 , 2008 only
- Today
- New search...