Understanding Double Tuned Transformers

[HankMaurer]

I'm trying to understand the way a double-tuned transformer works in the ratio detector circuit of an old FM receiver. In such circuits the primary and secondary both have capacitors in parallel and are both tuned to a common resonant frequency. I've found several sources that say things like "a characteristic of a double tuned transformer is that the voltages in the primary and secondary windings are 90 degrees out of phase at resonance, and that the phase shift changes as the frequency changes from resonance". I'd like to be able to derive this relationship or find a source (textbook) that can give me some help.

 

[ronsimpson]

https://en.wikipedia.org/wiki/Ratio_detector
https://www.radio-electronics.com/i...reception/ratio-fm-detector-discriminator.php

 

[HankMaurer]

I've already seen both of these articles but neither one gives any analysis of how the phase shift happens, just that it does. Although the second one does imply that loose coupling between primary and secondary is a factor. To simplify the circuit I'm interested in just throw away everything but the transformer and the capacitors across the primary and secondary and add a source and a simple load.

 

[ronsimpson]

I built this in LT Spice. Primary to Secondary coupling is 0.9.
L1 and L2 needed to have the Q reduced a little. I added 1 ohm inside L1 & L2.

The dark blue line is gain for the primary. Light blue dotted line is phase for the primary. Note the phase reverses near resonance.
The dark green line is gain for the secondary. Light green dotted line is phase for the secondary.
Right in the center of resonance it does appear there is about 90 degree shift.

A loosely coupled coil is molded by adding "leakage inductance". So there is another inductor that is not on the schematic. I think that inductor adds phase shift.

 

[HankMaurer]

Many thanks Ron, this will help a lot. I will try this with my own version of Spice and learn a lot by fiddling with component values. I still hope to find some theoretical basis for all of this. There must be a good textbook somewhere that covers this stuff.

 

[ronsimpson]

I found that if the Q of LC is too high it does not work right. And the coupling effects things to.
I was aiming for 10.7mhz of the FM-IF frequency.

 

[HankMaurer]

As I said, I will play around with component values to see if I can gain more insight. This all started because I'm restoring an old vacuum tube Crosley AM/FM table radio from 1953. The schematic I have doesn't give things like inductances and coupling factors but perhaps I can measure some of the relevant components and plug those values into Spice.

 

[Nigel Goodwin]

I found that if the Q of LC is too high it does not work right. And the coupling effects things to.
I was aiming for 10.7mhz of the FM-IF frequency.

The point of double tuned IF transformers is to increase the bandwidth, while still retaining high Q and steep sided filtering. You simply tune the two cores slightly apart, one above 10.7 and the other below.

 

[ronsimpson]

The point of double tuned IF transformers

Yes, I agree but... This is about decoding fm.

double-tuned transformer works in the ratio detector circuit

The ratio detector is about getting audio out of fm.
Many years ago some very sharp person noticed that the phase of a LC tuned circuit changes as you shift through the resonant frequency. In a double tuned transformer the primary and secondary do not change together. You can add the primary and secondary together and the phase will add/subtract. The results is that as the frequency moves from (below-> right on->above) the resonant frequency, the DC output will go from negative to zero to positive.

If you put a dc meter on the output of a ratio detector you will get a "tunning meter".

 

[Nigel Goodwin]

Yes, I agree but... This is about decoding fm.

I was responding to the post about too high a Q.

 

[unclejed613]

Yes, I agree but... This is about decoding fm.

The ratio detector is about getting audio out of fm.
Many years ago some very sharp person noticed that the phase of a LC tuned circuit changes as you shift through the resonant frequency. In a double tuned transformer the primary and secondary do not change together. You can add the primary and secondary together and the phase will add/subtract. The results is that as the frequency moves from (below-> right on->above) the resonant frequency, the DC output will go from negative to zero to positive.

If you put a dc meter on the output of a ratio detector you will get a "tunning meter".

i think the simplest way i've been able to explain it, is that you have two slope detectors with one tuned above the center frequency, and the other below center frequency, and the diodes appropriately are opposite, so one side gives a positive output, and the other gives a negative output. that may be oversimplified, but it's essentially what's going on there.