OK, let's see here.
This guys 'transformers' aren't really transformers, they are LC circuits that are tuned to (hopefully) the IF you need. He talks about excessive distortion on weak signals, oscillation that can only be corrected with changing the cathode resistor of a tube (which tube?), and the bandwidth being too narrow.
Kinda sounds like the reason for using real IF transfomers.

Back to square one?

If you read that guys article carefully you will see all the drawbacks to his design; probably why engineers have NOT used this method for the last 90 years.

A Commercially Made IF Transformer.
I began with a NIB NOS IF transformer that I have several of on hand. It was made by Meissner and the part number is either 16-5740 or 276841-26. Personally I'm betting on the shorter number. The box is clearly marked input transformer. A photograph of the resultant sweep is shown below. 455 kc is at the center of the screen and the calibration is 10 kc per div.

As you can see the bandwidth is the narrowest of all the transformers tested. In IF response the -6 dB points are used rather than the -3 dB points as in audio work. Two transformers should add together so for a single transformer the -3 dB points should theoretically be used. Since the detector is seriously loading the transformer in the plate circuit of the IF amplifier the one in the grid circuit is probably providing most of the selectivity.

I decided to drop a couple into my 5 tube superhet. Narrow bandwidth, higher Q, and higher gain all go together. The amplifier oscillated. I had to increase the cathode resistor to 1 k ohm to tame it. A well laid out and properly shielded IF amplifier would probably not need this gain reducing feedback. The resistor is not bypassed.

The tuning was quite narrow as you might expect. These would be excellent for DXing on the AM band or short wave listening. The Meissner company was known for making ham equipment so these transformers may have been designed for use in communications receivers rather than broadcast radios.

The one thing I noticed was excessive distortion on extremely weak signals. This may be due to the fact that I was using an input transformer to drive the diode detector. An impedance buffer might help. An infinite impedance detector might also be used but I don't know how to derive AGC from such a detector. The bandwidth is too narrow to make a high-fi tuner with these transformers.

Transformers Made with The Original 1.2 mH Pie Wound Chokes.
The first picture below is the one I was using for the input transformer and the second was the output. I thought I had swept them out but perhaps the flicker on the old oscilloscope made it hard to see what was going on. As you can see I had them seriously over coupled. I wonder what they would do if critical coupling were used.

I didn't make any changes except to slightly readjust the trimmers to account for the difference in capacitance between the radio and the test circuit. When used in the radio, rejection of adjacent channels was quite good. Weak signals were not distorted and the amount of noise did not seem excessive. Although I did not connect it to a high fi system, its frequency response must have been as good as the station was transmitting.

Below are the sweeps of the OSE coils. The first one is the input transformer and the second one the output. Channel 2 of the scope is the Y axis and you can see the progressively lower Q as the input has to be set to a more sensitive range to keep the display the same height. The output of the sweep generator was not changed throughout these tests.


Adjacent channel rejection was quite good and there was no audible distortion on weak signals. One of the things I would like to do is build an AM broadcast receiver with delayed, amplified AGC so I could tune the dial without having the other hand on the volume control. I also want to use the PLL VFO from OSE to make a digital local oscillator so I could know exactly what channel I am listening to. So much to do, so little time.

Meantime remember, life is short so eat your dessert first and always listen to tube radios

If you really want to make your own IF transformers.....

OK, try this guys method and see how well it works. He admitted to having to change components to make the radio work properly; in my book, re-engineering a radio to make it work means you are not using a proper substitute part.

He said he had to change a cathode resistor because of distortion. Then he said something about an 'infinite impedance buffer' (whatever that is) to make a circuit work.

Chuck,

I had seen this article before. It looks like it would work very well -in fact it is exactly like an IF transformer works,

Not sure why the panties are in a wad?

This guys method is the same as any IF transformer - a loosely coupled set of inductors with resonating capacitors?

The old IF cans are still easy enough to get - but in the future it would be a good way improvise.

Pat
N4LTA

He was referring to using a detector with infinite impedance, but we know that's impossible. I think he was saying that in jest.

The Infinite Impedance Detector is well known and documented as an exceptionally good one for AM.
A benefit is that it does not load the last IF as does a diode even with the proper IFxfmr. In fact an input transformer is ideal for it.

Carl