Inductors and magnetic cores anyone? Practical problem

[bwinter]

Hi there!
So here's the deal.

I'm restoring a friend's old Philco radio for his 1968 Mustang. This includes changing out the internals to tune FM instead of AM. The old AM internals include a set of three inductor coils with magnetic cores that move in and out of the coils as you turn the tuning dial, which change the inductance of course. I wanted to continue using this mechanism, so to this end, I've designed and sent out for some custom PCBs for a whole new FM circuit which uses an external LC circuit to tune the desired frequency. This way, I can reconstitute the inductor coils and retain the clever mechanical preset buttons as well. Now, onto the problem.

I've calculated that the inductor(s) must vary 42% in value to tune from 88MHz to 108Mhz, the FM spectrum. This is regardless of the capacitance. The problem is, the coils as they are now are machined to cover the AM spectrum, and thus vary quite a lot more (more than 200% percent from lowest to highest needle position). The magnetic cores, if you are wondering, are lowered into and out of the coils as a fixed set and go from being completely outside the coils, to being inserted completely inside them as you turn the dial.

What I have to do is come up with a relation that will tell me just how many turns I must rewire these inductors with so that:

[tex]0.42 = \frac{L_{max}-L_{min}}{L_{mid}}[/tex]

All I have are the min (air-core) and max (magnetic-core inserted fully) values of the current inductors, and the geometry of the inductor tubes (think slightly smaller than a cigarette).

Any ideas other than trial and error? Luckily I do have a fairly accurate LC meter.

 

[vk6kro]

You would probably know that cores that work at 1 MHz would be useless at 100 MHz, so what type of cores are you thinking of using?

I suspect that a better approach would be to buy a AM / Stereo FM radio, Stereo CD player and mount it in the space previously occupied by the radio.
This would have preset channels and a digital readout.

 

[bwinter]

You would probably know that cores that work at 1 MHz would be useless at 100 MHz, so what type of cores are you thinking of using?

Could you expand a little on this? The cores appear to be ferrite so I don't know why they couldn't be used for higher frequency applications.

My main concern is the lack of accuracy with these coils...a slight touch results in a change of 1 uH and with a 5pF cap that's halfway to the next FM station...I need to be looking at much, much fewer turns.

I suspect that a better approach would be to buy a AM / Stereo FM radio, Stereo CD player and mount it in the space previously occupied by the radio.
This would have preset channels and a digital readout.

This was my first approach to the problem but it's actually much more difficult, and the preset buttons would lose their mechanical functionality. The owner wants all functionality and the original look preserved.

I know this can be done because there are step-by-step pictures of a man named Bob Sponsel doing this exact conversion with a near-identical radio...sadly he is not around anymore.

http://www.mustangcountry.com/articles-mustang/radioman/home.html

 

[vk6kro]

There are dozens of different ferrites and one that worked at 1 MHz would be ineffective at 100 MHz. It would be very lossy and unable to tune a coil.
Ferrites are specified for particular frequency ranges.
One well known maker of ferrite is Amidon, but there are excellent ferrites coming from China now.

At that frequency, even a variable capacitor would not be used. A commercial design would use varicap diodes and phase locked loops.

On this page of free software, there are a couple of programs for designing with ferrites:
http://www.tech-systems-labs.com/freesoftware.htm
One of them has a list of available ferrites.
You need a VHF ferrite suitable for resonant circuits.


I have an L / C meter that works from 1 to 30 MHz and it is amazing how much the inductance of a coil with a ferrite core changes over that range. And the inductance is nothing like that I measure on an instrument that measures at 1 KHz.

Maybe you could mount a modern radio under the dash or in the glove box?

Maybe you could hinge the front of the old radio so that you lift it to operate the new one?

Seriously, though, you should walk away from this now.

 

[bwinter]

At that frequency, even a variable capacitor would not be used.

Uhh, now I know this isn't true. I'm looking at a commercial clock radio sitting right in front of me that uses a variable capacitor to tune FM.

So I might need new cores. That doesn't seem like a big problem compared to completely re-doing the internals and wiring up a modern radio to fit and flip out and do all sorts of crazy things with this old interface.

My LC meter measures at 4 MHz and is still seeing a huge change in inductance on each coil--from 37 uH to 327 uH for example. I only need a relative change in resonance, not a specific value. I'm still not convinced this can't be done (in fact I know it can, the only question is whether he switched out the cores or not, I don't think he did).

 

[Averagesupernova]

VK6KRO is correct. Walk away from this. Have you ever actually designed, built, or worked with RF oscillators that work around 100 MHz? What are you incorportating for AFT on this tuner?

 

[eric0524]

Why not use the face and put a new radio behind it. It seems it would be much easier to make the buttons on the face plate operate the new radio. If he wants to preserver the original radio he is already blowing that on the FM conversion. Kinda seems like a waste of time.

 

[sophiecentaur]

By heck. You seem to be expecting to design a VHF FM radio from scratch. That is a tall order for anyone and, for someone who doesn't know about the different ferrites available, I should say impossible. Would you consider modifying a steam engine to run as a four stroke petrol engine? It's the same sort of thing you're talking about.

I could suggest that you might modify the attractive mechanical interface to operate a set of switches / digital encoder with which to control a modern fm receiver. Now that would be a possible starter . . . .

 

[bwinter]

I could suggest that you might modify the attractive mechanical interface to operate a set of switches / digital encoder with which to control a modern fm receiver. Now that would be a possible starter . . . .

Between you and me, the interface is ugly as heck IMHO. I don't see why he doesn't just do away with it completely. But, I think he wants the originality of the car to be preserved, so I'm looking at it as a purely technical challenge.

By heck. You seem to be expecting to design a VHF FM radio from scratch. That is a tall order for anyone and, for someone who doesn't know about the different ferrites available, I should say impossible. Would you consider modifying a steam engine to run as a four stroke petrol engine? It's the same sort of thing you're talking about.

My circuit is not technically from scratch--as Carl Sagan said, "If you want to make an apple pie from scratch, you must first create the universe." :) It's based on an older chipset (TDA7040 and TDA7021) from the 80's that was used in car radios, with most of the tricky bits already laid out internally with just a few external components. A complete FM stereo radio from scratch would admittedly be far beyond my purview, and probably wouldn't fit inside the box with the amplifier. Anyway...

It seems from your revelations that I'm rearranging the deck chairs, so to speak. I've spoken with a professor and she seems to think it's possible. I've found a few examples of this exact principle being used to convert old AM car radios to FM by preserving the old coils (AAR made a conversion kit in the 50's). Regardless, I think the endeavor will be fruitful educationally and I will let you know how I fair. I have not given up hope yet!

Thanks.

 

[sophiecentaur]

I have another solution. You get a bank of cheap FM receivers and basic AM modulators. You tune each FM receiver to a separate channel and then produce a comb of AM signals (at MF) of each channel he may want to listen to. Disconnect the AM antenna and use your frequency multiplex as input. then use the receiver as normal. No mods to the receiver but some rather repetitive work for you and a big grey box in the glove compartment.

Go on. I dare you.

 

[vk6kro]

One thing that hasn't been mentioned. Ferrites vary their permeability quite badly with temperature, so where frequency stability is important, as in this case, ferrites are not used in oscillator coils.

Incidentally, "trial and error" is the way things get made at VHF. It is great fun, but fairly unpredictable.

I like the multiple radio idea. I was checking something similar yesterday.

I was looking at some MP3 players on EBay yesterday. These cost $8 each and throw in an FM tuner with a digital frequency readout.

Suppose you got 4 of these and set each to the frequency of a popular FM station. (Forget the 1 GB MP3 players which could hold 200 or 300 songs each.)

These could be switched by putting micro switches behind each of the buttons on the front panel of the radio and using these to select which MP3 player was being used.
This one could then receive power via a suitable 3 volt or 1.5 volt voltage regulator. They take very little current.

I think they get RF signals from the headphone wire.

They would need a decent audio power amplifier but you probably already have one of these on your PC board.

It still only gives you 4 stations and involves some pretty complex switching.Building an FM tuner from scratch is a bad idea and trying to use an old mechanism from a car radio is pretty hopeless, but you are probably in too deep to back out without trying it.

We understand that.

But be aware that you should not commit your friend or yourself to extra expense for a project that has almost no chance of success.

I would much prefer to encourage you with promises of good results, but many years of VHF experiments suggest that this is going to end badly.

 

[sophiecentaur]

A more 'intelligent' system could involve just one FM receiver and frequency measurement of the mf local oscillator of the car radio. The (not necessarily accurate*) frequency measurement would then be used to tune the fm receiver and an am modulator to provide am signal at an appropriate frequency and with the wanted programme. Every part of the existing am radio would be used with just a tiny search coil placed near the local oscillator coil.

Edit * an accuracy of a few tens of Hz in 1MHz is enough and easy to achieve.