volume difference between the top and bottom of the band

[radionortheast]

Something i’ve noticed with a lot of transmitters, they seemed louder at the top end of the band, can maybe cause a problem for your modulation. I had a max pro years ago, the frequency was changed from the bottom of the band to the top, there was more noise the stereo pilot was too high, had to be turned down, quite interesting though wondering why that is. Is it something that happens in all fm transmitters? Perhaps it was something that was overcome though, later there was one which had stereo on board, so nothing to a just, they must of figured it out.

[jvok]

If you look up datasheets for varactor diodes (e.g bb809, bb301 etc) you'll see the capacitance vs voltage graph isn't a straight line i.e. the slope changes depending on the bias voltage. The slope of the curve gives the modulation sensitivity - its shows how much change in capacitance (and so frequency) you get for a given change in voltage (your audio signal). So as you tune your rig across the band, the PLL is changing the bias voltage on the varactor, moving to a different part of the curve with a different slope and so you get a different mod level.

(Technically you actually want a 1/x^2 curve not a straight line, to cancel out the square root in the f=1/(2*pi*sqrt(L*C)) frequency equation. But it gets even more complicated because the varactor usually isn't the only capacitor in the circuit).

Rigs with separate varactors for mod and pll are better because the pll doesn't effect the bias on the mod varactor. But thats still not perfect because the mod varactor makes up a smaller percentage of the total capacitance at the bottom of the band (when the pll varactor is at max capacitance) than the top of the band (where the pll varactor is at min capacitance). That means the effect the mod varactor has on the frequency is smaller at the bottom of the band than at the top, so the mod level still changes.

Theres a trick you can do to compensate this where you feed some of the pll voltage back to the mod varactor as well. So you're intentionally changing the bias (and slope) of the mod varactor as you tune across the band, but by exactly the right amount to compensate for the effect of the pll varactor changing its capacitance. Its still not perfect but its a lot better.

The other thing you can do is sample the oscillator output, send it to a demodulator to recover the audio, and use that as the feedback to an opamp at the audio input. Then the whole thing becomes self correcting, i.e. the opamp will try to force the feedback signal to look like the audio input, automatically changing the drive level into the varactor to keep the gain constant. This will also help clean up any audio distortion as well. I've never seen it done in a pirate rig though

[radionortheast]

Does make me wonder how these fm chips get on, like the 8066, they have a little computer inside that can do all that, I don’t know for sure, someone measured one had good stereo it had the perfect pilot level, don’t know if they did it at different frequencies repeated the measurement, obviously they can't be used commercially.

Maybe a reason to keep something on the same frequency, better to have a separate encoder you can ajust the level in. On my rf generator the level is much different on 87.5 than on 108, is what got me thinking about the max, not sure how the rf generator generates the rf exactly, the level goes higher and higher even more so if you go up outside the fm band.
Don’t know if something could be set for 98, could be got away with, they would likely have to high or low level at either end of the band. Surely a big expensive transmitter with a stereo encoder inside, would likely have some mpx ajustment, I was thinking of someone getting one at a community station it turns out not to be plug play, their having to ajust the pilot level..they'll likely need to measure for sure

[Albert H]

Northeast - any proper, professional installation of a trnsmitter will include plenty of calibration. The frequency and the deviation are just two of the dozens of adjustments that are made.

Jvok above is exactly right with the description of the varying sensitivity across the band. I have designed exciters for exactly the same deviation sensitivity at any frequency, but the compensation circuits always require accurate calibration, and almost any physical parameter (ambient temperarure, atmospheric moisture, even atmospheric temperature and more) can all have effects on the operation of an oscillator. The compensation circuits can be as complex as the whole of the rest of the exciter!

One design I worked on used the PLL circuit to measure - and adjust - the deviation sensitivity. I can't tell you exactly how I achieved it because the principle is subject to patent, and I sold the rights to the design to the big radio corporation that paid for its creation! Suffice to say, it was sufficiently accurate to tightly control the modulation characteristics of the exciter to much better than FCC requirements.

Contrary to your claims, the modulation sensitivity of the 8066 and similar chips does vary across the band, but it's less apparent because most of them include some kind of compression and limiting. It's also difficult to identify the real carrier amoungst all the spurii!

[radionortheast]

Contrary to your claims, the modulation sensitivity of the 8066 and similar chips does vary across the band, but it's less apparent because most of them include some kind of compression and limiting. It's also difficult to identify the real carrier amoungst all the spurii!

I was not saying it did, wondering if it measured the same across the band, something in the software, why I was asking, wondering if big commercial transmitters that cost 1,000’s were able to calibrate themselves, put out the correct pilot level for the frequency, seems like i’ve not got an answer to this. if there was something that cost £30, already putting out the correct pilot level, to be spending all the money makes no sense. Your always going on about the cost of things but in this case its ok to spend 1,000’s, its also ok to spend 1,000’s on a modulation meter, we’ve not been given a simple way to see that it is the correct level.

it was claims on a german forum about 80 numbered chip, the stereo was good on the 80 numbered chip, compared to a bh chip which had more cross talk, 80 chip had a perfect 9% pilot volume level, i’ve got nothing here to measure these things so I wouldn’t know, obviously the stereo sounded good on it, I listened to both, but it can’t be used.

[Albert H]

Northeast:

If you want quality engineering, it always costs a lot of money!

As far as the modulation sensitivity is concerned, yes - there are ways of compensating for the sensitivity changes across the band, but doing it accurately isn't trivial. Modulation meters are expensive devices (if they're any good), and your avaerage pirate isn't going to bother with one! My "go-to" mod meter was a Marconi one that cost around £4500 new, back in the late 70s. It's a fine machine, and capable of measuring FM deviation up to 140 kHz at 100 MHz, and measures AM modulation depth very accurately too when switched to that mode.

There are networks you can use in the PLL and modulation circuits in an FM exciter. One design I did compensates reasonably well across the band so that PPM 6 (+8dBm) gives 90 % of the full 75 kHz deviation (to account for the pilot tone). If you're interested in trying to get one working like this, let me know, and I'll put the circuits up, but you'll need use of an accurate mod meter to calibrate it!