**EMRFD Message Archive 5083**

MessageDateFromSubject5083 2010-09-12 05:42:38 g3zoh Coupling Coefficients Hi all, I am using Elsie 2.36 (Student Edition) to look at HF bandpass filters for a homebrew receiver project. These filters are nothing elaborate - just top-coupled double tuned circuits with link windings for low impedance input/output.

The inductors are wound on T50-6 toroids and consist of between 12 and 24 turns depending on the band. The link windings are 2 or 3 turns wound tightly over the earthy end of the primary.

Now the question - Is there a "rule of thumb" for the coupling coefficient (K) that should be input to Elsie when using the "manual entry" screen to set the inductance values?

Many thanks in advance for any advice.

Brian

G3ZOH5084 2010-09-12 05:56:29 Leon Heller Re: Coupling Coefficients On 12/09/2010 13:40, g3zoh wrote:

> Hi all, I am using Elsie 2.36 (Student Edition) to look at HF bandpass filters for a homebrew receiver project. These filters are nothing elaborate - just top-coupled double tuned circuits with link windings for low impedance input/output.It'll be high for toroids; I usually use 0.9. You can try different

>

> The inductors are wound on T50-6 toroids and consist of between 12 and 24 turns depending on the band. The link windings are 2 or 3 turns wound tightly over the earthy end of the primary.

>

> Now the question - Is there a "rule of thumb" for the coupling coefficient (K) that should be input to Elsie when using the "manual entry" screen to set the inductance values?

values, of course, and get a feel for the effect.

Leon

--

Leon Heller

G1HSM5085 2010-09-12 08:05:36 chuck adams Re: Coupling Coefficients On Sun, 2010-09-12 at 13:56 +0100, Leon Heller wrote:

>There is a way to measure the coupling coefficient. Measure L_p,

> On 12/09/2010 13:40, g3zoh wrote:

> > Hi all, I am using Elsie 2.36 (Student Edition) to look at HF

> bandpass filters for a homebrew receiver project. These filters are

> nothing elaborate - just top-coupled double tuned circuits with link

> windings for low impedance input/output.

> >

> > The inductors are wound on T50-6 toroids and consist of between 12

> and 24 turns depending on the band. The link windings are 2 or 3 turns

> wound tightly over the earthy end of the primary.

> >

> > Now the question - Is there a "rule of thumb" for the coupling

> coefficient (K) that should be input to Elsie when using the "manual

> entry" screen to set the inductance values?

>

> It'll be high for toroids; I usually use 0.9. You can try different

> values, of course, and get a feel for the effect.

>

> Leon

> --

> Leon Heller

> G1HSM

>

inductance of the primary, with the secondary open-circuited.

Short-circuit the secondary leads and measure the inductance

again. Call this L'_p.

k = sqrt(1 - L'_p/L_p)

If you want to see the derivation of this, look in Electronic

Communication Techniques (4ed is my copy) by Paul H. Young pages

20-34 with the technique on page 31.

There are probably several hundred other EE books that also

show this tidbit.

The mutual inductance, M, is given by M = k * sqrt(L_p * L_s).

L_p being the primary of the transformer and L_s the secondary.

This should work reversing the primary and secondary. There will

be some frequency dependence due to the

permeability of the core material for something like a toroid

and must inductance meters use a relatively low frequency for

measurements.

I hope this helps.

FYI

chuck

--

chuck adams, k7qo

http://www.k7qo.net/

chuck.adams.k7qo@gmail.com

/* no comment */5086 2010-09-12 09:12:34 g3zoh Re: Coupling Coefficients Many thanks Le5087 2010-09-12 09:50:16 Leon Heller Re: Coupling Coefficients On 12/09/2010 17:12, g3zoh wrote:

> Many thanks Leon and Chuck for the helpful advice. I thought the bandpass filter bit of this project would be boring but am finding it fascinating, trying different configurations with Elsie - a very fine piece of software.How did varying it affect the characteristics? SPICE is useful as well,

I design the filter or whatever with ELSIE then simulate the whole

circuit with SPICE.

73, Leon

--

Leon Heller

G1HSM5089 2010-09-12 10:44:27 victor Re: Coupling Coefficients I have read in a number of places that calculating K by measuring the primary inductance while shorting the secondary is theoretically right but in practice is not very accurate (the loss at the shorted inductor causes a significant error).

A preferred higher accuracy calculation of K is recommended by doing this:

Measure the inductance of the primary and secondary in series, and then interchange the connections of one winding for a second reading. Apply the equation below:

M=1/4 * (Lseries+ - Lseries-)

For coupling, measure the primary and secondary separately then apply the equation below:

k = M / (sqrt(Lp * Ls))

Victor - 4Z4ME

>

> There is a way to measure the coupling coefficient. Measure L_p,

> inductance of the primary, with the secondary open-circuited.

> Short-circuit the secondary leads and measure the inductance

> again. Call this L'_p.

>

> k = sqrt(1 - L'_p/L_p)

>

> If you want to see the derivation of this, look in Electronic

> Communication Techniques (4ed is my copy) by Paul H. Young pages

> 20-34 with the technique on page 31.

>

> There are probably several hundred other EE books that also

> show this tidbit.

>

> The mutual inductance, M, is given by M = k * sqrt(L_p * L_s).

> L_p being the primary of the transformer and L_s the secondary.

>

> This should work reversing the primary and secondary. There will

> be some frequency dependence due to the

> permeability of the core material for something like a toroid

> and must inductance meters use a relatively low frequency for

> measurements.

>

> I hope this helps.

>

> FYI

>

> chuck

>

>

>

> --

> chuck adams, k7qo

> http://www.k7qo.net/

> chuck.adams.k7qo@...

>

> /* no comment */

>5093 2010-09-12 14:06:43 g3zoh Re: Coupling Coefficients 5094 2010-09-12 14:32:06 Leon Heller Re: Coupling Coefficients 5096 2010-09-13 07:41:07 Glen Leinweber Re: Coupling Coefficients Was curious about this too, in doing some

tapped Hartley coils on toroids, so measured

mutual coupling on a T50-6 iron-powder core.

Mutual Coupling came out at 0.3, which

seems kinda low.

I believe permeability of this core is less than

ten - for higher permeability cores (like ferrites)

the coupling is much closer to one.

When turns-ratio is large, with a link of only

2 or 3 turns, you can only get matching

approximate. I believe Wes did a pdf note

on link coupling in the the "files" section on the

EMRFD website.5097 2010-09-13 10:46:19 w7zoi Re: Coupling Coefficients Hi Glen, and group,

Yea, I discussed this problem in a note on my web page, w7zoi.net. It was posted on 19April09 in the "designs and experiments" section with the title "Transformer Coupled LC Bandpass Filters." The measurements presented were done with a vector network analyzer. However, a VNA is not necessary. A return loss bridge with a homebrew signal source and either an AD8307 based power meter or a 50 Ohm terminated oscilloscope will provide enough information to do a good job with the matching. Be careful to avoid generator harmonic problems.

I've also measured some coupling coefficients on T50-6 power iron cores and sometimes they are as low as 0.3. I was surprised at the time. I found that k depended upon the position of the link with respect to the main coil. Coupling coefficient is a parameter that should be measured, for a "rule of thumb" is often in severe error.

In the "app note" of April 09 mentioned above, a parallel resistance parameter was specified. This is one of those things that is buried within a filter design and the fellow or gal using an existing computer program never sees the parameter. A point of the posted note was that this parameter is experimentally useful. I now display this parameter in the programs that I have written for my own use. For example, Rp is in the lower left corner of DTC08, which is part of the most recent versi6117 2011-04-13 17:59:29 KK7B Coupling Coefficients For broadband transformers, coupling coefficients above .95 are expected and useful. But old air core networks often used coupling as one of the network variables, and it was supposed to be less than 0.5. Some of the more efficient vacuum tube amplifiers used variable link coupling with large air core inductors. They often achieved efficiencies higher than many of today's broadband amplifiers. It's a different way to design a network--fundamentally narrow-band--but that's how folks used to put clean signals on the air with relatively simple transmitters.

This is a great topic...I'm enjoying the discussions.

Rick