Vintage AM MW radios often had double tuned IFTs .The shape of the frequency response curve is well documented elsewhere .The coupling coefficient of these double tuned transformers is low and the Q is high probably to aim for KQ =1 .These old school IFTs do not have taps and therefore are high impedance devices that are intended for Valve radios.I have rebuilt several valve car radios with a cascode JFET IF amp to match these old IFTs .The sound on AM is fantastic compared to the muffled AM sound on most modern car radios .This brighter sound is because the Double tuned transformers do not sideband cut like the single tuned IFTs and narrow band ceramic filters found in the AM sections of more modern car radios .These old junk box IFTs could be unreliable and I do not have many of them .The tightly coupled single tuned 455 KHz IFTs that are used on transistor radios is easy to find .If I used them then the retro car radio would sound just like a modern one so there is no point in that .How could I configure 2 standard single tuned IFTs to perform like a Double tuned IFT?


Besides Tony Stewart's inductive bottom coupling, you can also couple two single-tuned resonators with a small capacitor:


simulate this circuit – Schematic created using CircuitLab
R1 and R2 are wire resistance estimates of 10mm cup-core tunable inductors - these are not real resistors. Source and load resistance of about 90K ohm give about 18 kHz bandwidth. C2 is the top-coupling capacitor. A smaller value decreases coupling, a larger value over-couples giving double-hump frequency response.
This network has superior attenuation for frequencies less than 455 kHz. Tony's coil-coupling network has superior attenuation for frequencies higher than 455 kHz.
These networks can include more resonators, each coupled to the next. It gets very tricky to tune more than four, but there are methods (see Dishal method).

  • \$\begingroup\$ Good to see top coupling as an answer +1 .It is easy to try and reassuring to see 5pF .I was not sure of the sensible value .Would there be any benefit of using the low impedence tightly coupled base windings in some back to back arrangement perhaps using a larger value of coupling C .Such a concept could allow a series tuned LC with a reasonable L value .I am happy with 5pF and will try this because it is simple . \$\endgroup\$ – Autistic Mar 11 '17 at 1:50
  • 1
    \$\begingroup\$ For AM audio, you want to start with a resonator loaded Q about 35. If your resonators have lower loaded Q, yes, you'll need a slightly larger coupling capacitor for optimum coupling, but the bandwidth will end up being very wide, and you'll hear adjacent stations (if they're strong enough). Most 455 kHz transistor radio 10mm or 7mm I.F. coils have about 620 uH (tuneable). Have done a 4-resonator top-coupled filter having 50 kHz bandwidth. Top-couplers were about 15 pf, 30K-ohm source & load with these coils. \$\endgroup\$ – glen_geek Mar 11 '17 at 2:19
  • \$\begingroup\$ @ glen geek.Cap top coupling is simple .More resonators look interesting but what about the shape factor .Say if I raised bandwidth by placing parallel resistors across the IFTs then the shape would be the same and there would be interstation interference if I wanted good sounding wideband .I must preserve the shape . \$\endgroup\$ – Autistic Mar 11 '17 at 2:26
  • 1
    \$\begingroup\$ In an ideal multi-resonator filter, individual resonators are very high-Q. Once you choose bandwidth, top couplers combine with Rsource & Rload to give flat passband. You don't add resistors to the middle resonator(s). Rsource & Rload do it all. To get great shape and good adjacent-channel attenuation, you'd need higher-Q than most LC resonators can provide. 10mm I.F. transformer unloaded Q is a little more than 100. You'd want even more Q. The temperature coefficient of its iron powdered core is compensated by a carefully-chosen 180pf tempco. \$\endgroup\$ – glen_geek Mar 11 '17 at 3:07
  • \$\begingroup\$ What if both R source and R load are very high? \$\endgroup\$ – Autistic Mar 11 '17 at 3:09

How could I configure 2 standard single tuned IFTs to perform like a Double tuned IFT?

Here is a picture I provided for a different question and it relates to this question also: -

enter image description here

As you should hopefully see it shows how two capacitively coupled tuned circuits create a merging bandpass filter when the coupling capacitor is altered. The centre frequency in this example is 100 kHz but this can be moved to 455 kHz by selecting lower values for the inductors.

Link to original answer

  • \$\begingroup\$ Good picture +1 .So it looks like cap coupling will do it if I select the correct cap .Now Andy would I be able to back to back couple the unused base windings and normalise the cap to N squared ?So if N is say 10 then C goes up by 100 now is there advantage of doing this due to 2 lots of mag coupling which may help far away response? \$\endgroup\$ – Autistic Mar 11 '17 at 11:45
  • \$\begingroup\$ @Autistic I don't have any idea what the unused base windings are or why you are talking about mag coupling. I don't understand what "far away response" is. \$\endgroup\$ – Andy aka Mar 11 '17 at 11:50
  • \$\begingroup\$ @ Andy Aka .Standard IFTs used on normal transistor radios have a second low impedance winding .On the 1st and 2nd IFTs they go to the base of a common emmiter BJT stage that has Z in of roughly 1K ohm .The turns ratio is roughly 10:1 .If I could link couple the tightly coupled base windings with a normalised cap then I have galvanic isolation which is handy for my fullwave envelope detector .By far away there was an old superstition that magnetic coupling was better for rejecting stuff that is way way out of band . \$\endgroup\$ – Autistic Mar 11 '17 at 12:04
  • \$\begingroup\$ Superstitions aside, my answer is about using capacitively coupled inductors each tuned with its own capacitor. I appreciate that for vintage stuff the low turns extra winding was able to drive the base of the next stage without significantly affecting the Q of the circuit. Yes you could capacitively couple the low turn windings so that the transformers operate back-to-back but you basically have the same performance as in my embedded circuit. Yes you have isolation but, in the absense of what your full wave rectifier circuit is I cannot say there will be a benefit.... \$\endgroup\$ – Andy aka Mar 11 '17 at 12:58
  • \$\begingroup\$ ... Whatever method you employ (mine or the back-to-back transformers) you will need a high impedance buffer stage on the output - you lose that benefit with back-to-back transformers so that is something to bear in mind but, these days, 455 kHz isn't a big deal really. Back in the day, using single transistors yes it was! \$\endgroup\$ – Andy aka Mar 11 '17 at 12:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.