IF alignment with damping of an adjacent coil

[Alistair D]

Thanks for that.

Al

[G0HZU_JMR]

In my opinion, the main advantages of the damper method are:

You only have to tune each inductor once and you tune for a peak on an AF voltmeter (easy for the operator for every slug).

You can loosely couple the AM sig gen to the mixer output via a tiny capacitance so there is minimal loading of the IF circuit by the signal generator.

The output connects to the AF output of the receiver so there is no loading/pulling of the IF circuits here at all.

You can tune all of the dual IF transformers in one pass through.
The main disadvantage is that you have to know where to fit the damper and you have to do it without shorting/zapping anything if the damper falls off. You may also have to move the damper like this quite a few times during the process.

It should be a much more reliable and repeatable method compared to a dodgy homebrew wobbulator. The wobbulator user will have to continually swap back and forth tweaking each slug until they decide it is OK to stop. There is always the risk of circuit loading/pulling with the wobbulator method and the wobbulator operator may stop the process too soon or they may decide to settle for the wrong response especially if they run the wobbulator with a sweep speed that is too fast for the filter response.

[Argus25]

Originally Posted by regenfreak

Tehman's model in the screen shot is not applicable to parallel inductive coupled LC tank resonance. In Henney's equivalent model the mutual inductance is added to L1:

Yes that is correct, however it is useful for a couple of reasons.

Firstly it helps to understand how mutually coupled circuits affect each other but there are a number of cases in radio work where its useful. One was predicting band-pass characteristics of audio transformers and knowing the factors that affect that.

If there is any primary tuning capacitance of any significance you can transpose it into the secondary as you might a resistance or a leakage inductance, without too much of an error when the K value is close to 1.

The Terman equation is particularly good for high turns ratio coils, say ignition coils, Tesla coils etc because you have a large inductance resonant secondary and a relative small inductance & low self capacitance primary. So for any equations to be of any value they do have to be applied to the right scenario. But even if not, there is insight to be gained from the relationships of L,C,R & M in coupled resonant circuits.

There are many IF coils of course, typically some Japanese transistor types, which have a single tuned circuit and the secondary feeds the base of the next stage for an impedance match with an un-tuned shorter winding, but the load there affects the resonant frequency to some degree.

For example in a radio where the AGC alters the conditions at the bases of the IF transistors, it alters the IF tuning and in some radios this is a noticeable effect, that the tuning peak appears to change between strong and a weak carrier on the same frequency. The conditions at the transistor's collector affect it too.

[Argus25]

Originally Posted by G0HZU_JMR

It should be a much more reliable and repeatable method compared to a dodgy homebrew wobbulator.

I certainly agree with that. For an IF/RF sweep to be of any use it has to be with precision test gear, a proper generator designed for it and a good scope too.

Over the years, for both TV and radio work, I have gone over to sweeping the set at its antenna input with a very loosely coupled input, transmitted signal or via a matched dummy antenna/resistive padder for TV.

For some of my my radios now I transmit the swept signal to them from a far off generator that acts as a radio station. This way there is no loading at all by coupling in signals to the circuits. And you get to see the complete response which includes the bandpass properties of the antenna & RF stages.

Often in TV work, when the sweep generator was coupled into the video IF, it altered the tuning there, and the response the technician ended up with altered when the generator was disconnected. So I avoid that issue.

This was also partly the idea of the H-feild transanalyser generator, to couple signals into a typical ferrite rod transistor radio without altering any of the tuning conditions within it.

[G0HZU_JMR]

Yes, I've seen quite a few youtube videos where people demonstrate a wobbulator in action. I've not seen an impressive demo yet. Some of them are very poor.

If the radio is designed to be aligned with the damper method I'd expect it to be the correct method to use. I think it will be hard to beat this method with a sweeper/analyser unless the objective is to tune the IF stages to achieve a custom IF response.

It's a bit like trying to beat the Dishal method by adopting a random manual approach to BPF tuning. The random/manual method is only going to beat the Dishal method if the goal is to deliberately mistune the system to achieve a custom frequency response that the Dishal method can't reproduce.

[Argus25]

Originally Posted by G0HZU_JMR

If the radio is designed to be aligned with the damper method I'd expect it to be the correct method to use. I think it will be hard to beat this method with a sweeper/analyser unless the objective is to tune the IF stages to achieve a custom IF response.

I think that is quite right where there are clear enough instructions and its a radio, not a TV.

The big problems I have found are more in the TV area, especially when the manufacturer recommended setting up the array of tuning slugs to peak on some particular frequency.

Also, if a sweep is used, it requires the instrument doing it, or another have a very accurate frequency generator to impose the markers, otherwise the sweep just shows the shape.

When I did this with my 904, aligned according to the manufacturer instructions, it became very clear they were much more interested in gain than bandwidth. The response was only 1.4MHz wide. This attenuates the video HF response that it is even noticeable on a 5" screen.

Without the sweep/marker generator I would never have been able to get it right and get a combination of a good HF response in the video and still quite reasonable gain. Even a 1/4 turn on some of the slugs makes a big difference. (In the 904 they are brass and reduce the inductance of the coils they tune)

I have attached a copy of what I had on the scope at the time of the final adjustment. This is one reason why, along with the fine spot focus of the 5FP4 crt, that my 904 has a much better image than is seen on most 904's.

[G0HZU_JMR]

I had a go at demonstrating the damper in action in an RF simulator in a youtube video. Sorry there's no sound, I don't have a microphone here.

The idea was to demonstrate how it is possible to only look at the graph of the signal generator frequency in the lower left corner when aligning the filter using the damper network. The lower left graph has a marker that represents the strength of the RF signal at the centre of the passband at 460kHz. This is going to be similar to displaying the strength of the demodulated AM waveform. Both represent the signal amplitude. The graph on the right shows a swept display.

I start the video with the filter quite well aligned on 460kHz and all the resonators are at 479.5uH. I'm not sure if this is the ideal value but it is close enough. Then I detune the filter and attempt to realign it using the damper method and I only look at the marker level of the cw signal in the lower left graph when tuning. This gets tuned for maximum level at every position of the damper.

It does work quite well and I tried it several times from various starting points. The response does look good at the end and I think it would be better if I started from a less detuned state. Also it would help if I had set Genesys to finer tuning steps. However, my PC is very slow to tune so I had to opt for grainy tuning step sizes for each inductor.

The video is a bit slow and jerky because my PC is having to capture the video and do all the number crunching. The coupled inductor model does seem to use a lot of processor power.

https://www.youtube.com/watch?v=FsAq...ature=youtu.be

It's a bit of a boring and slow video so probably best watched at 2x speed

[regenfreak]

Many thanks for the RF simulator demo video. it is very informative and interesting. I am sold.

Nothing is better than seeing it visually

[Radio Wrangler]

Once you've seen the method of tuning a filter as a set of isolated stages, there are several variations you can play, and some additional tricks for practical application.

If you totally kill the resonators other than the one you are working on, rather than just damping them, then you no longer need data for the resonance of each stage... they all should be resonated at simply the centre frequency.

The shorted resonators, of course, won't work for passing signals into and out of the stage you're working on, so you need one probe to apply your test signal, and a second one to measure the response to your test signal. These probes will tend to de-tune your resonator, and will create an error. You could calculate a correvted centre frequency to allow for their capacitance, or you could make them very low capacitance.

If you're using a spectrum analyser or a radio receiver as the detector, then you have an awful lot more sensitivity than really needed. You can exploit this. Inject your signal through a high value resistor, and couple your analyser/receiver via a second high value resistor. Just peak the resonator on the wanted frequency, then move your shorts and probe resistors around to do the next stage. You only need to work your way through the lot once, but this method will align even terrifyingly complex filters in even one pass. There seems to be a lot of faffing around, but it works and cuts through a lot of confusion that happens to anyone trying to twiddle the whole lot at once.

It may feel like a sledgehammer to crack a nut on just a simple pair of tuned circuits in an IF transformer, but at this simpler end of the world of filters it still works and you're not out of the range of complexity that can be adjusted by eye, given some experience-based skill.

An analogy is reversing a trailer.

Stick a trailer on your car and you can reverse it where you want to go, with a bit of skill it tries to diverge. Driving forwards is trivial things converge on a single path.

Stick a trailer on your trailer and try to reverse it, and it's far far trickier to reverse, each one is divergent. Drive forwards and it all converges.

High order systems, a trailer, pulling a trailer pulling a trailer.... pulling.... are chaotically impossible to control in reverse, but obediently converge to a predictable path going forwards.

Low order systems give you some choice in method. Tuning short-cuts work. With high order systems, short cuts get you lost but the long way round gets you there reliably.

Things like FDM telephone systems which put up to thousands of phone calls onto a single coax cable as a stack of SSB signals were heavily reliant on good and complex filters. A lot of the filter design techniques and the companion techniques for tuning them up came out of places like Bell Labs. With the sheer number of big filters such companies had, there were massive savings to be had from optimising ways of twiddling them.

The method of damping stages, rather than completely killing them is a useful compromise that simplifies the probing business and allows the evaluations to be done using the main signal path. It needs someone to have pre-computed the peak frequencies to allow for the coupling to the damped stage, but it saves messing around trying to lightly probe things.

What you've seen is a window into the world of people who have to tune filters that look like nightmares.

In the amateur radio business, there are two filters that really scare people. Both of them are in a Racal RA17 receiver (and RA117, RA217, RA1217 have them too)
People who just wade into them even with a wobbulator and scope often wind up in a right mess. There are some people who can do them without getting hot under the collar and can sort out other people's muck-ups. So now you know how thy do it.

I think the oath of secrecy comes next....


David

[Argus25]

Originally Posted by Radio Wrangler

Once you've seen the method of tuning a filter as a set of isolated stages, there are several variations you can play, and some additional tricks for practical application.

I think the oath of secrecy comes next....


David

I go along with the whole notion of it, doing it with damping and tuning a complex filter as isolated stages.......but, once you have done that you still, in my opinion at least, need to "check it is correct"...not just because its Christmas.

So you have to have the equipment at hand anyway, to do a proper sweep with frequency markers, or, for all you know, you have created Frankenstein's Monster.

But, I'd have to agree, that doing it the way you suggest, could avoid a tail chasing exercise which could make you go around in circles initially, if you were just observing the band-pass characteristic only in the first place and had 5 to 7 inductors to adjust.

So I still cannot see, how you can escape having to have the equipment for a frequency sweep to ensure the amplitude vs frequency profile is correct, no matter how you got there.

[Radio Wrangler]

Indeed. It's irresistible. You HAVE to do a final sweep of the thing to enjoy your handiwork and as a final check... You just have to resist the urge for a final tweak.

David

[G0HZU_JMR]

Originally Posted by regenfreak

Many thanks for the RF simulator demo video. it is very informative and interesting. I am sold.

Nothing is better than seeing it visually

Thanks! Obviously, the simulation represents a fairly ideal IF path with high isolation amplifiers with stable I/O impedances and no AGC issues.

Some receivers will be far from ideal and won't be suitable for this damper method and I suspect that some will have very touchy neutralisation networks and AGC issues as Argus mentions. So at the very least these receivers might need two or three passes through using the damper. But it still might not give the right result.

Hopefully, the receivers designed to use this method will give reliable and repeatable results. I deliberately detuned the inductors quite a bit in the youtube video to show that the damper method can still work even when only adjusting each inductor once. With a real radio it might not work that effectively on the first pass through if the start point is so far off...

[turretslug]

Originally Posted by regenfreak

https://www.roehrentechnik.de/html/zf-bandfilter.html

You would need to order the adjustment tool as the slug takes triangular tool. I have both chassis and PCB miniature versions.

I remember vaguely their minimum order is 40 euros

It's a pleasant surprise to find that they are still actually being made! I suppose if a few folk here were after IFTs for cherished sets or new-old-build (as it were), then 40 euros for a handful wouldn't be such a hit. After all, a mains transformer re-wind can be pricey, but if it gives a worthwhile set a reprieve then fair enough. I wonder what the current equivalent price of old-time IFTs like the Denco IFT11 would be nowadays?

[regenfreak]

It's a pleasant surprise to find that they are still actually being made! I suppose if a few folk here were after IFTs for cherished sets or new-old-build (as it were), then 40 euros for a handful wouldn't be such a hit. After all, a mains transformer re-wind can be pricey, but if it gives a worthwhile set a reprieve then fair enough. I wonder what the current equivalent price of old-time IFTs like the Denco IFT11 would be nowadays?

Winding your own IF transformer by hand is not as hard as it seems. In fact, it will be one of my projects in the homebrew to-do-list. I wound my own MW superhet oscillator coil using scramble fine Litz windings. It has the same size and design specs as generic stock oscillator coils. The homebrew coil works as good as factory made one and has much higher Q.

I don't know about Denco IF price, but the Denco oscillator/RF/antenna coils are frequently sold for £20-£25 each on ebay. Pairs of old If cans frequently go up for sale £15-18 on ebay.