[Chas]

The plate circuit and any B+ de-coupling will add stray capacitance to the primary side. Similarly on the secondary side the grid circuit will add stray capacitance. The Aluminum shield will as well. I suggest have several values of resonating capacitors both up/down around the calculated goal to adjust for these strays.

GL

Chas

[462ron]

According to this online calculator, for your 22uh inductor to resonate @9.1Mhz, you need a total capacitance of around 13.9pf! Capacitance in the uf or mfd range is way too much! Just plug in your inductance in uh and your resonant frequency in MHz and set the capacitance box for pf!

https://goodcalculators.com/resonant-fre...alculator/

Ron

[jrblasde]

Thank you for double checking me! I do appreciate it. I went to bed last night thinking that microfarads just didn't make sense.

Alright, so where did I go wrong? Two days ago I had written that C = [1/( 2*pi*f*sqrt(L) )]^2 for a resonant LC circuit, and that is correct. Mid-day yesterday I was away from my notes and wrote down C = 1/( 2*pi*f*sqrt(L) ) (note the lack of the final square) from memory rather than re-deriving the equation. Rookie mistake! Thus, I came up with 3.729 µF. If I square this, I get 13.905 pF. That's what I was missing! This matches 462ron's findings.

Lesson of the day--always derive the equation rather than recall it from memory!

Attached is the revised schematic. We'll be back to the original issue of potentially touchy adjustments now that we're back in the range of adjusting by just a few picofarads, but I will start with this. A fixed 10pF capacitor in parallel with a 5 pF trimmer. I'll get a better understanding of the parasitic capacitance introduced from the surrounding can, as well as the plate and grid capacitance for the primary and secondary sides, respectively.

[462ron]

Glad you were able to straighten it out. It’s been a loooong time since I learned equations, way before there was an internet. So my old brain is too lazy to do that kind of math anymore so now we have these great quick calculators online, there are a few of them actually! I checked only because I remember seeing some component values on other IF transformers and kind of thought it would be in the pf range. I just didn’t want to see you waste money on buying the wrong value components. It will be interesting to see how this project works out! GL

Ron

[jrblasde]

Thank you, sir! I agree—measure twice, cut once.

[jrblasde]

I've ordered the components. I may have found the most elegant breakthrough yet--a 22 µH coupled, shielded inductor. See the attached data sheet. If this is really what I think it will be, then I can do this entire design on one flat PCB the size of old the 3/4" x 3/4" IF can--and without the can for a shield!

I took Chas' advice, and looked up the inter-electrode capacitance of the primary and secondary tubes. The plate capacitance of the 14F8 on the primary side is 1.4 pF, and the grid capacitance of the 6BJ6 on the secondary side is 4.5 pF. So I have bought some different fixed resistors to use for testing. If the shielding on the coupled inductors proves to be adequate enough, I can bypass any added parasitic capacitance introduced by placing the old shield over top of it all.

Just for backup, I still bought two 22 µH inductors with radial leads.

[jrblasde]

I meant to say that I bought an assortment of fixed capacitors, not resistors.

[462ron]

A neat little package, will be following this project! 

Ron

[jrblasde]

The components arrived today, which was far faster than I expected. I’ve tried prototyping both the coupled inductors and the individual conductors. I believe I’ll have to adjust the capacitance to get things just right, because I seem to have no range of adjustment at 9.1 MHz. I seemed to have better luck with the individual inductors than I did with the coupled inductors, but the resonant frequency was 3.5 MHz. I noticed that I seemed to have roughly a 200 kHz bandwidth (perfect for FM), because the output signal began attenuating whenever I adjusted the signal generator up or down by even 0.1 MHz. I also happened to notice that the output had some variation as I adjusted the secondary capacitance, but not as I adjusted the primary capacitance. This could be some external capacitance (possibly from my oscilloscope) interfering with the total capacitance of the circuit.

I also find that the inductors need to be very close to each other for best results. If the gap gets bigger than about an eighth of an inch then the output signal is attenuated severely.

I’ve attached several photos showing the reduction of amplitude of the output voltage as frequency increases from 3.5 MHz up to 9.1 MHz. By that point I had to adjust the volts per division of my channel 2 input.

I was testing this circuit in unloaded conditions, meaning that I did not have it integrated into the IF circuitry of the actual radio. 

At this point, I have two thoughts on why the resonant frequency was lower than expected.
1) My oscilloscope may be introducing some series capacitance. Recall that capacitance is additive when in parallel.
2) The inductance might be too high. A value of 22 µH is quite small, but it may still be high enough to start choking any signals higher than about 3.5 MHz.

[RodB]

I think you're picking up 10pf in lead length. See if you can put the inductors close to the caps.

[jrblasde]

A keen observation! Thanks, I will give that a try tomorrow. I had started with short leads when using the coupled inductors, but moved the individual inductors further away due to the increased footprint. I will reposition it all.

Might not hurt to get some smaller inductors on order. Using, say, a 2.2 µH or a 3.3 µH inductor will give me some more breathing room to upsize the capacitors. This should help limit the effects of parasitic capacitance.

[462ron]

jr, the reason you haven’t noticed any change when adjusting the primary could be maybe you have the primary directly connected to your signal generator. If that is the way your setup is, then the generator is putting a constant signal amplitude to the primary and any adjustment is getting swamped. If you put some experimental resistance in series with the generator and primary, then as you adjust things the signal is shared between the resistor and the primary and the primary impedance will vary as the frequency you chose varies. Hope this helps!

Ron

[jrblasde]

Ron, you are correct. I was not inserting any resistance. I will do so, and give it another try. In fact, I’m wondering now if the unrestricted current was responsible for the odd behavior I noticed with the coupled inductors. I’ll have to retry both, with leads shortened.

I have no idea how long you all have been servicing these radios, but you all are a wealth of knowledge to me. It’s almost disadvantageous that younger electrical engineers (class of 2018, myself) don’t get the privilege of many analog electronics labs. I recall learning about these filters, briefly, in Circuits I, but I would have loved a lab as well! Most labs were for digital circuits.

Joseph

[jrblasde]

I have reattempted with short leads to the inductors and with some resistance inserted on the primary side (I started with a 15 kΩ as the output resistance of the 14F8 on the primary side is 14.5 kΩ, but lowered it down to 4.7 kΩ for better readings). The results were more stable (with both the individual inductors and with the coupled inductors), but still I felt that the individual inductors worked better. I am still noticing that there’s no peaking or attenuation as I adjust the trim capacitors, so I believe that I’m still facing added parasitic capacitance from something.

My next plan of attack is to reduce the inductance while increasing the capacitance. I should be able to use a 3.3 μH inductor with a total of 92.69 pF in capacitance. I could use an 82 pF fixed capacitor and a 6.8pF fixed capacitor in parallel with each other for a fixed capacitance of 88.8 pF, in parallel with a 10 pF trim capacitor. This should raise the capacitance enough that any minor interference will not drastically affect the resonant frequency.

I thought about determining the theoretical inductance of the existing, broken IF transformer’s coils. There are 15 turns at a diameter of 0.175” and a solenoid length of 0.15”. The formula for inductance of a coil is L = (N^2 * μ * Α) / l, where N is the number of turns, μ is the permeability, A is the cross sectional area, and L is the length of the coil. I calculate 4.61 μH for the inductance with no iron core. I should be closer to this value if I use the 3.3 μH inductors.

Joseph

[462ron]

Joseph, the reason why this site is a great one is there are people from all walks of life with a variety of experiences and knowledge that are willing to help. You say you’re a recent 2018 electronics graduate with heavy digital training, whereas I started in the 1960s with an interest in electronics as a teen. Graduated from electronics trade school in 69. We had training in tube circuitry and some solid state, just barely touched on simple digital. I am not savvy at all in computer so now that I’m retired, I enjoy antique tube radios. Your brain is still fresh on engineering, theory and math where I have become rusty on a lot of that. Anyway, you are to be commended for your experimenting with this project and depending on what you come up with, this could be the solution for many of us that find ourselves needing an unobtainable IF transformer! I know working with a high frequency like you’re doing is touchy, there are so many factors that affect the outcome. I’m following this project with interest!

Ron