6-gang FM stereo tuner heads

[regenfreak]

Originally Posted by G0HZU_JMR

Thanks. The return loss and smith chart plots looking into that Minicircuits amp look to be very dodgy to me.

At 34:40 onwards the ripple in the return loss and the squiggles in the smith chart response look to be caused by a system calibration issue. The TSP presenter thinks this is to do with the way the amplifier is matched but I doubt it. The regular pattern suggests an incorrectly calibrated system at the end of a cable that might be 35cm long. There does appear to be a cable of that length in the video.

Either that or the Siglent VNA has really poor directivity after calibration. The datasheet suggests >40dB directivity so I'd expect to see much better results than what is shown in the video.

A general rule of thumb states that in order to get < +/-1dB uncertainty ripple in the return loss measurement the directivity of the VNA has to be >20dB greater than the return loss being measured.

So if the VNA works as per the datasheet (if calibrated correctly) then there should be less than +/- 1dB ripple on the return loss plots where the return loss is less than 20dB. But the plot shows up to +/- 2.5dB ripple in the circled region below where the RL is about 16dB. This suggests the effective directivity of the VNA is only about 26dB. I strongly suspect that the VNA hasn't been calibrated properly. The regular peaks and troughs do suggest there is a ~35cm long RF cable at the input and this is a strong clue that there is a significant problem with the setup.

There's no way the Minicircuits amplifier will have regular ripple in the return loss like that so either the VNA directivity is really poor or the operator doesn't know how to calibrate the VNA properly. I suspect there is a calibration issue. Maybe the calibration was carried out correctly but it isn't set to be active. Maybe the calibration is disabled in the menu system somewhere.

I agreed. He said it is a Mini-circuits 10MHz-2.7GHz LNA but I cannot find anything in Mini-circuits web site that matches the spec. Here I attached the broadband AFL-2500 amplifier as a example, its S21 has much less ripples. Also I find it strange that he did not use attenuator at the input to mitigate and "mask" the small mismatch. If I remember correctly, the Siglent SVA1032X 3.2GHz uses broad-band resistive-divider type of directional coupler. Its VNA circuit design is very clever being minimalist in hardware.

Originally Posted by G0HZU_JMR

I let the computer optimiser play with various matching networks and left it to optimise for an hour last night. The result below is quite good but this uses a two stage match on the input and output so it is a bit unrealistic in terms of component count. Also this would be hard to replicate in reality unless tuneable components were used in the matching as the optimiser was set to allow 1% tuning increments on each component in the matching network.

cheers. I am curious what is the toplogy of the matching networks you have tried with your program?

I have learnt about the benefits of T-network (two back-to-back L-networks) from the link below that you can set the desired Q for the matching components but it has a higher component count than L-pads:

https://www.analog.com/cn/technical-...mulations.html

[Radio Wrangler]

Is there any reason why all sections of quoted text in this thread have had the attribution to their originator removed?

It is going to make it very difficult to follow for anyone coming along later.

David

[Station X]

I've attributed the quotes in post #141, but it would be too much work to do that for the whole thread.

[regenfreak]

Originally Posted by Radio Wrangler

Is there any reason why all sections of quoted text in this thread have had the attribution to their originator removed?

It is going to make it very difficult to follow for anyone coming along later.

David

Apology, it is my bad habit. I always click " Go Advance " first and use "Quote" from the top menu as I often got more than one reply. I tried to reply to two posts in one single post. I will need to remind myself not to do that from now.

I replied to G0HZU_JMR in most of my recent posts.

Attached is the measurement with of MAR-1+ up 1GHz using 15db attenuator at the input and 3db attenuator at the output ( with the NanoVNA V2 plus 4). I dont see the same ripples seen in the video.

[G0HZU_JMR]

In my case I used two L networks (one after the other) in order to get a lower overall Q and also to afford some flexibility. It's effectively a low order ladder network. The classic T network will increase the Q compared to a basic L match but if you connect the L matches one after the other like a ladder (rather than mirrored back to back) it's possible to get a slightly better match over slightly more bandwidth because the L match can then match up to 330 ohms in two stages rather than just one.

In the end the computer preferred to mix up the network a bit more in order to squeeze the last bit of bandwidth where the return loss was >20dB. So some caps became inductors and vice versa. However, I'm not sure the slight benefit is worth having twice the number of matching components.

[regenfreak]

Originally Posted by G0HZU_JMR

In my case I used two L networks (one after the other) in order to get a lower overall Q and also to afford some flexibility. It's effectively a low order ladder network. The classic T network will increase the Q compared to a basic L match but if you connect the L matches one after the other like a ladder (rather than mirrored back to back) it's possible to get a slightly better match over slightly more bandwidth because the L match can then match up to 330 ohms in two stages rather than just one.

In the end the computer preferred to mix up the network a bit more in order to squeeze the last bit of bandwidth where the return loss was >20dB. So some caps became inductors and vice versa. However, I'm not sure the slight benefit is worth having twice the number of matching components.

Cheers. I thought of repeating the 4:1 transformer experiment using two stages as well; using a 200 to 330 ohms minimum-loss matching pad. But I thought it wont worth the hassles. The best solution seems to be the simplest in this case.

PS: Now I am convinced he didnt do calibration properly because those ripples are tell-tale signs of the effect of the cable before calibration.

[G0HZU_JMR]

See below for the result with a basic Lmatch on each side. This matching network consists of a series 120pF cap followed by a shunt 2uH inductor across the filter inputs. You can see the return loss is good but the bandwidth where there is a good match is slightly narrower now.

This 10.7MHz ceramic filter is a Toko part and it was used in a VHF/UHF comms receiver I designed at work back in the late 1990s. The receiver had various bandwidths for FM reception.

I think the Toko part number was SK107M1-A0-10X but I've had to look through an old company database to get this number. It is described as having 280kHz bandwidth.

I've also taken VNA data for SK107M3-A0-20X but this is described as having 180kHz bandwidth. The receiver also had an option for SK107M5-A0-10X and this had 110kHz BW according to the old database at work. However, I've not got any of these here to measure.

I also measured a Murata SFE MS3 and this has E10.7S stamped on it. This is a much newer part and was in a bag of samples at work. This is quite a narrow filter with a rounded response and it is a bit lossy with just over 6dB loss when matched.

[regenfreak]

Originally Posted by G0HZU_JMR

See below for the result with a basic Lmatch on each side. This matching network consists of a series 120pF cap followed by a shunt 2uH inductor across the filter inputs. You can see the return loss is good but the bandwidth where there is a good match is slightly narrower now.

This 10.7MHz ceramic filter is a Toko part and it was used in a VHF/UHF comms receiver I designed at work back in the late 1990s. The receiver had various bandwidths for FM reception.

I think the Toko part number was SK107M1-A0-10X but I've had to look through an old company database to get this number. It is described as having 280kHz bandwidth.

I've also taken VNA data for SK107M3-A0-20X but this is described as having 180kHz bandwidth. The receiver also had an option for SK107M5-A0-10X and this had 110kHz BW according to the old database at work. However, I've not got any of these here to measure.

I also measured a Murata SFE MS3 and this has E10.7S stamped on it. This is a much newer part and was in a bag of samples at work. This is quite a narrow filter with a rounded response and it is a bit lossy with just over 6dB loss when matched.

Cheers. I know very little about the design of VHF/UHF comms receivers and I thought they are mostly narrow-band FM. In fact I have very little interest in Ham radio or QRP stuff, but I have to force myself reading the relevant technical materials in order to demystify the RF black art. I am more into the the design and evolution of high-end FM stereo tuners.

The standalone measurement of the ceramic filter has limited value and that's why I measured the bandwidth of the two ceramic filters with the CA3053 cascode amp to quantify the bandwidth reduction in the synchronous tuned cascade filters. Also the bandwidth of the IF starts to widen when the amp is limiting with strong signals. So it is difficult to predict how many stages of ceramic filters are required for the best audio sound quality without compromising the selectivity. There are many variables..

Here I attached two random examples of interesting designs:

SAE Mark VI- digital tuner: It uses 10-pole toroid filter in front of the CA3053 IC IF amp followed by 5-pole toroid filter and limiter ICs.

Kenwood KT6040-It is a low-cost tuner, having a 6-gang varactor front end tuning to save costs. The typical Q of the varactor diode tank circuit is about half the value for the air variable gang capacitors. The 6-9 air gang capacitors are the most expensive components in high-end tuners. The diode switching allows the option for either the narrow-band IF with 6 ceramic filters or wide-band with 2 ceramic filters.

[G0HZU_JMR]

The comms receiver was a bit strange in that it was designed mainly for receiving stuff from low cost digital transmitters that might use FSK or OOK. It was mainly a downconverter ahead of a narrow digital IF but it also needed to monitor crude (and drifty) transmit devices that had a wider bandwidth that transmitted FSK/FM. There were two wideband FM bandwidths and these both used a conventional limiter/detector instead of doing the demod in the digital domain in the narrowband IF.

It also had a wider bandwidth output that could feed to a digital IF of about 1MHz bandwidth but this was done at a previous (higher) IF frequency. All the digital IF/DSP stuff was in another box.

Quite a few of these receivers were made, probably over a hundred. I've still got some sample parts from the dev work including some very exotic KVG crystal filters. Eg I think one of them is a 21.4MHz crystal filter with 50kHz bandwidth for example. It had preselection crystal filters ahead of the narrowband digital IF. I think the selectable IF bandwidths were 10kHz, 20kHz and 50kHz and this was done using crystal filters.

It also had a hybrid DDS/PLL synthesiser that ran up at UHF that could tune in tiny steps. This was quite exotic in those days. This used the AD 9850 DDS as a tuneable reference and this DDS was a brand new part in those days. I still have one of the hybrid synthesisers here but it hasn't been used for over 20 years.

[Radio Wrangler]

KVG were an exceptional quartz company. Last I heard of them they'd undone a takeover and were an independent private company again. They had been gobbled-up by Vectron at one point.

I had some inverted-mesa 100MHz fundamental VCOs done by Vectron.... they were no slouches either, but they did go on a buying spree of other quartz firms.

The most exotic crystal filters I was ever involved with were special low distortion jobs for NPR testing, where intermodulation was the big trouble. We found only C.E.P.E part of Thomson-CSF could make them and the up-front engineering charges.... were in the jeepers! league.

David

[regenfreak]

Originally Posted by G0HZU_JMR

The comms receiver was a bit strange in that it was designed mainly for receiving stuff from low cost digital transmitters that might use FSK or OOK. It was mainly a downconverter ahead of a narrow digital IF but it also needed to monitor crude (and drifty) transmit devices that had a wider bandwidth that transmitted FSK/FM. There were two wideband FM bandwidths and these both used a conventional limiter/detector instead of doing the demod in the digital domain in the narrowband IF.

It also had a wider bandwidth output that could feed to a digital IF of about 1MHz bandwidth but this was done at a previous (higher) IF frequency. All the digital IF/DSP stuff was in another box.

Quite a few of these receivers were made, probably over a hundred. I've still got some sample parts from the dev work including some very exotic KVG crystal filters. Eg I think one of them is a 21.4MHz crystal filter with 50kHz bandwidth for example. It had preselection crystal filters ahead of the narrowband digital IF. I think the selectable IF bandwidths were 10kHz, 20kHz and 50kHz and this was done using crystal filters.

It also had a hybrid DDS/PLL synthesiser that ran up at UHF that could tune in tiny steps. This was quite exotic in those days. This used the AD 9850 DDS as a tuneable reference and this DDS was a brand new part in those days. I still have one of the hybrid synthesisers here but it hasn't been used for over 20 years.

The modulation of analogue carrier with digital data is alien to me. Then they have the modulation of digital carrier with analogue data or digital data...the variants of frequency phase-shifting technology go on and are permutated. I read about DSP, I and Q theories but soon I lost interest in DSP transceivers and radios quickly.

I regress to the 1970s to 1980s FM Hi-Fi tuner technology that i have found most fascinating. In addition, I am also looking at the designs of early FM transistor radios using discrete components. Previously I was into 1920s valve regens and then valve AM superhets.

I can't afford to own any of the 6 to 9 gangs sought-after Kenwood or Pioneer FM tuners that easily cost £800 to 1000 each. But there is nothing to stop me building bare-bone, primitive versions of them and learn about the RF stuff associated with it. Also I dont own a TV and I listen to FM radio everyday. I spend many hours listening my homebrew stereo tuners and i really enjoy it

[G0HZU_JMR]

One thing I spotted in your CA3053 test circuit is that you have used it in cascode mode back in post #116. The SAE tuner uses it in differential mode at the 10.7MHz first IF. I'd expect the cascode configuration to have a fair bit of shunt capacitance at the input. Possibly over 20pF. I'd expect this to degrade the matching of the ceramic filter a bit.

When configured in differential mode I'd expect the CA3053 to have much lower input capacitance. It might be as low as 6 or 7pF for example.

It probably doesn't matter that much but the filter matching and the filter response won't be as good as it could be if you feed the ceramic filter into the 470R || CA3053 in cascode mode.

[Radio Wrangler]

Ceramic filters have a number of imperfections. Their centre frequency isn't well-controlled in manufacture, so selection and matching is needed whenever more than one is used. Worse, there are believable reports of them drifting (not always together) over extended time.

The linearity of their phase/frequency response is not as good as can be achieved with L-C filters, though this and their adjustment puts the cost up dramatically. A number of high performance tuners avoid ceramic filters and do it the hard way with custom L-C designs. Passing an FM signal through a filter with non-flat group delay (= non-linear phase) effectively adds distortion to the demodulated audio. Some manufacturers only quote distortion at low amounts of deviation to get better looking figures.

There are plenty of paramaters setting the performance of an FM tuner, not just the number of tuned circuits at the incoming frequency.

Varactor diodes don't just sink the Q of n RF tuned circuit, they are non-linear, with big signals pumping the capacitance which smaller signal see. Varactors can change capacitance very quickly indeed and are the active element in UHF/Microwave parametric amplifiers. So a varactor front end is prone to intermodulation.

FM tuner design has to go down either a low noise path, or a low intermod path. You can't have both at once, there are different architectures to each at the expense of the other. Choosing a compromise is an interesting game.

The BBC used Revox B261 tuners in two applications:

To monitor off-air quality (so the group delay an the discriminator have to be especially good)

And to receive a signal from another site (usually tens of miles away in the UK) for re-broadcast from the site where the tuner is. This provides service should the audio leased lines or nicam distribution fail. This is a hellish environment for a receiver, recieving one signal in the band on the actual site of a high power transmitter radiating in another channel of that bandl.

I have one of these beasties and it does sound rather good. The BBC had them racked-up with other gear, all running 24/7 so they got cooked. I had to go through all the electrolytic capacitors in the thing. The BBC ones have several mods, nothing for special performance, but a standard 3-pole IEC connector in place of the rarer C8 2-pole one found on civilian Revox stuff. The plastic side-pieces are usually missing for rack mounting.

David

[regenfreak]

Originally Posted by G0HZU_JMR

One thing I spotted in your CA3053 test circuit is that you have used it in cascode mode back in post #116. The SAE tuner uses it in differential mode at the 10.7MHz first IF. I'd expect the cascode configuration to have a fair bit of shunt capacitance at the input. Possibly over 20pF. I'd expect this to degrade the matching of the ceramic filter a bit.

When configured in differential mode I'd expect the CA3053 to have much lower input capacitance. It might be as low as 6 or 7pF for example.

It probably doesn't matter that much but the filter matching and the filter response won't be as good as it could be if you feed the ceramic filter into the 470R || CA3053 in cascode mode.

thanks. Very interesting points. I also noticed the use of differential configurations in IF stages of other tuners. Often they don't use 330 ohms termination resistors at the I/O. I have attached Nelson Jones's paper that he explained the CA3053 has much higher I/O than 330 ohms. In fact i repeated the test in post 116 and replaced the 330 to 50 L-pad by 390-50 L-pad, the result on the Smith chart did not show any improvement in matching.
I am a bit confused about your comments that the differential configuration has lower input capacitance than the cascode configuration? Probably the explanation is very complicated. I read some theories on BJF amplifier design analyses..nothing is simple.

I thought the whole point of having cascode is to reduce the feedback capacitance( and also input capacitance), setting the gain for the lower common-emitter BJT about 1. The upper common-base BJT is not subject to Miller effect due to the grounded base shielding the collector signal from feedback to the emitter input. Now the input capacitance of the cascode is the capacitance across the base and emitter of the lower BJT .but then the story may be more than that.

I have completed the Nelson Jones If/quadrature demodulator TAA661B stage. Now i need to build the MPX decoder.

I am going to build the IF stages based on the mighty Kenwood KR-9600. This should outperform the Nelson Jones' design. It uses only four ceramic filters (two in back-to-back cascade) and LA1222 IF Amps. I wont use HA1137W demode chip as it requires double-tuned coils for the quadrature detection, meaning i would have to wind the coils myself.

PS, I came across this test method using op amp:
http://earmark.net/gesr/cf.htm

Originally Posted by Radio Wrangler

Ceramic filters have a number of imperfections. Their centre frequency isn't well-controlled in manufacture, so selection and matching is needed whenever more than one is used. Worse, there are believable reports of them drifting (not always together) over extended time.

The linearity of their phase/frequency response is not as good as can be achieved with L-C filters, though this and their adjustment puts the cost up dramatically. A number of high performance tuners avoid ceramic filters and do it the hard way with custom L-C designs. Passing an FM signal through a filter with non-flat group delay (= non-linear phase) effectively adds distortion to the demodulated audio. Some manufacturers only quote distortion at low amounts of deviation to get better looking figures.

There are plenty of paramaters setting the performance of an FM tuner, not just the number of tuned circuits at the incoming frequency.

Varactor diodes don't just sink the Q of n RF tuned circuit, they are non-linear, with big signals pumping the capacitance which smaller signal see. Varactors can change capacitance very quickly indeed and are the active element in UHF/Microwave parametric amplifiers. So a varactor front end is prone to intermodulation.

FM tuner design has to go down either a low noise path, or a low intermod path. You can't have both at once, there are different architectures to each at the expense of the other. Choosing a compromise is an interesting game.

The BBC used Revox B261 tuners in two applications:

cheers. Revox B261, what a tuner! It was made like Swiss watches! The synthesizer input goes to a push-pull BJT mixer and then to a 6-pole filter in front of the IF amp. I guess the 6-pole filter gets rid of the intermod and harmonics junks from the varactor frond end and synthesizer(attached).

A lots of the ceramic filters display asymmetry in their responses. I came across this history about the ceramic filter:

https://citeseerx.ist.psu.edu/viewdo...=rep1&type=pdf


Interesting, as a young man Murata admired the legend optical maker Carl Zeiss greatly. I used to collect and use Carl Zeiss SLR lenses in 35mm format for many years. I was obsessed with Carl Zeiss too.

[G0HZU_JMR]

regenfreak: I am a bit confused about your comments that the differential configuration has lower input capacitance than the cascode configuration? Probably the explanation is very complicated. I read some theories on BJF amplifier design analyses..nothing is simple.

I thought the whole point of having cascode is to reduce the feedback capacitance, setting the gain for the lower common emitter BJT about 1. The upper common base BJT is not subject to Miller effect due to the grounded base shielding the collector signal from feedback to the input. Now the input capacitance of the cascode is the capacitance across the lower the B-E...but than the story is more than that.

Sadly, the cascode configuration will still show some increased capacitance at its input. This can be demonstrated via simulation of a cascode circuit using discrete BJTs. It's also possible to see the difference by looking at the extended datasheet for the CA3053 as it gives plots of input admittance in both cascode and differential configurations.

The b curve (susceptance) for cascode shows just over 1.5mmho at about 10MHz. This can be converted across to an equivalent parallel capacitance at 10MHz. This online calculator shows that this would be about 24pF

https://www.everythingrf.com/rf-calc...cra-calculator

In differential mode it is about 0.5mmho on the datasheet. This would be about 8pF. This is slightly higher than I was expecting but still much lower than the capacitance in cascode mode.

What isn't clear from the datasheet is how the cascode circuit is configured. I'd expect to see a bypass cap at pin 4 but the value of this cap will affect the b11 and g11 curves in the datasheet plot below. The g11 of 0.6 at 10MHz seems a bit lower than I expected especially if the bias resistors are factored in. Maybe this plot was taken with a very small value of bypass cap more suited to 100MHz operation. Otherwise, I'd have expected to see g11 at about 1 to 1.5 at 10MHz in cascode mode with something like a 10nF bypass cap.

The other minor niggle with operating the CA3053 in cascode mode is that it won't deliver symmetrical limiting with large signals. Only one side of the waveform will show flat topping. The differential configuration should give fairly symmetrical limiting so it ought to produce a classic square wave output when driven hard. This is probably better for FM reception but I don't know how significant this would be in reality.

[regenfreak]

Thanks. I will try to think about it. I have not been able to wrap my head around how to understand and read admittence parameters or the y-matrix. I stumbled across them when I looked at the design calculations of dual gate mosfet FM tuners in two IEEE papers. I was like Dante stumbling across the wild beast in the Inferno.

Correction: The Revox B261 mixer is not in pull-push.

[Radio Wrangler]

Originally Posted by regenfreak

I guess the 6-pole filter gets rid of the intermod and harmonics junks from the varactor frond end and synthesizer(attached).

I'm afraid that can't happen.

Say there are two stations one on frequency A, one on frequency B, and the 3rd order intermod products they create will be on ABS(2*A-B) and ABS(2*B-A).

Let's say the tuner is trying to listen to a station on one of these frequencies. Because the wanted signal and the unwanted intermod are both on the same frequency, they both get converted to the IF frequency of the tuner, and appear there both on the same frequency, both in the passband. You hear interference.

No amount of filtering in the IF will separate two signals if they're both on the same frequency.

Once that intermod product has formed, it is too late. It is on the tuned frequency and uit will go straight down the receiver through all its filters, just like a real signal on that frequency.

Unfortunate, but we're stuck with this. The only solution is to get the RF and mixer good and linear.

David

[regenfreak]

Originally Posted by Radio Wrangler

I'm afraid that can't happen.

Say there are two stations one on frequency A, one on frequency B, and the 3rd order intermod products they create will be on ABS(2*A-B) and ABS(2*B-A).

Let's say the tuner is trying to listen to a station on one of these frequencies. Because the wanted signal and the unwanted intermod are both on the same frequency, they both get converted to the IF frequency of the tuner, and appear there both on the same frequency, both in the passband. You hear interference.

No amount of filtering in the IF will separate two signals if they're both on the same frequency.

Once that intermod product has formed, it is too late. It is on the tuned frequency and uit will go straight down the receiver through all its filters, just like a real signal on that frequency.

Unfortunate, but we're stuck with this. The only solution is to get the RF and mixer good and linear.

David

I see, the third order intermodulation products can be too close to the fundamental signals to be filtered out within the passband. What a pain!

[Radio Wrangler]

Close, but not there yet. The two signals which are intermodulating don't have to be close to the wanted signal

Let's say I have two strong signals on 95 and 100MHz. Their 3rd order products will be at 90 and 105MHz. If I'm trying to receive a third signal, on 90MHz, the intermod product can be smack on the centre of the channel I'm trying to receive.

In this case with the intermodulation-causing pair are 5MHz and 10MHz off of our tuned frequency. So RF selectivity (if we have some, not all radio receivers do) will help by attenuating the causes. However earlier stages in the RF get them stronger than the mixer does. Also the varactors in the first RF tuned section gets the full thing.

Now let's try something closer... if the problem signals are 90.0 and 90.2MHz representing typical channel frequencies, and we are listening to 90.4MHx, then the big signals are close enough to the tuned channel that there won't be much help from RF selectivity. The RF amp and the mixer get hit hard. One intermod product will again lie exactly on our wanted channel. It is the proximity of the interference causing signals that is of interest and relates to benefits from RF selectivity. The intermod product doesn't have to be spaced from our wanted channel, it can hit it dead centre.

For each channel I could tune to and for each value of spacing of intermod causers, there will be two pairs of causers which will hit the channel. The RF selectivity can't do the channel filtering job, so a few channels on either side of the wanted channel will be the worst for intermod performance, only when you get further out does RF selectivity start helping.

You can't handle or test all possibilities, so you have to sample several cases to indicate that the structure covers the possibilities.

This is a big problem for aviation radios. The ext band up from Band II FM broacast is where instrument landing system localisers live. If you ever fly anywhere, your life depends on this in bad visibility or for autolanding (Cat 3A and Cat 4) So there are pairs of FM broadcast channels which will plant an intermod on each localiser frequency. The aviation ILS receivers have to be remarkably well protected from intermodulating broadcast channels. Filters can't do it all. They help, but there has to be remarkable toughness in the receiver front-end design. I've spent the past several years on this! FM receivers are much easier in comparison, lives don't depend on them.

For many years there was a gentlemen's agreement not to populate the FM band above 100MHz to give a safety margin and give some space for filtering in aviation radios to start getting some attenuation in.

So now you know why older FM radios usually didn't tune above 100MHz.

But governments discovered that MegaHertz could be transformed into MegaBucks, and the part above 100MHz got sold off to commercial broadcasters. Plane owners were forced to scrap their radios and buy ones hardened against the problem.

Spectrum management agencies look at the localiser frequencies assigned to each aerodrome, and look at frequencies assigned to broadcasters in that area to try to minimise the risk. Minimise isn't quite the same as eliminate.

David

[regenfreak]

Originally Posted by Radio Wrangler

Close, but not there yet. The two signals which are intermodulating don't have to be close to the wanted signal

Let's say I have two strong signals on 95 and 100MHz. Their 3rd order products will be at 90 and 105MHz. If I'm trying to receive a third signal, on 90MHz, the intermod product can be smack on the centre of the channel I'm trying to receive.

In this case with the intermodulation-causing pair are 5MHz and 10MHz off of our tuned frequency. So RF selectivity (if we have some, not all radio receivers do) will help by attenuating the causes. However earlier stages in the RF get them stronger than the mixer does. Also the varactors in the first RF tuned section gets the full thing.

Now let's try something closer... if the problem signals are 90.0 and 90.2MHz representing typical channel frequencies, and we are listening to 90.4MHx, then the big signals are close enough to the tuned channel that there won't be much help from RF selectivity. The RF amp and the mixer get hit hard. One intermod product will again lie exactly on our wanted channel. It is the proximity of the interference causing signals that is of interest and relates to benefits from RF selectivity. The intermod product doesn't have to be spaced from our wanted channel, it can hit it dead centre.

For each channel I could tune to and for each value of spacing of intermod causers, there will be two pairs of causers which will hit the channel. The RF selectivity can't do the channel filtering job, so a few channels on either side of the wanted channel will be the worst for intermod performance, only when you get further out does RF selectivity start helping.

This is a big problem for aviation radios. The ext band up from Band II FM broacast is where instrument landing system localisers live. If you ever fly anywhere, your life depends on this in bad visibility or for autolanding (Cat 3A and Cat 4) So there are pairs of FM broadcast channels which will plant an intermod on each localiser frequency. The aviation ILS receivers have to be remarkably well protected from intermodulating broadcast channels. Filters can't do it all. They help, but there has to be remarkable toughness in the receiver front-end design. I've spent the past several years on this! FM receivers are much easier in comparison, lives don't depend on them.

For many years there was a gentlemen's agreement not to populate the FM band above 100MHz to give a safety margin and give some space for filtering in aviation radios to start getting some attenuation in.

So now you know why older FM radios usually didn't tune above 100MHz.

But governments discovered that MegaHertz could be transformed into MegaBucks, and the part above 100MHz got sold off to commercial broadcasters. Plane owners were forced to scrap their radios and buy ones hardened against the problem.
David

Thanks for the insightful remarks. It makes lots of senses. I guess the loaded Q of the first 2-pole varactor tuned RF section would not be bigger than 30-40.

The working principles of the overlapping beams of 150Hz and 90Hz in the ISL localizer/glideslope are very complicated, which is based on the mixing of two modulating signals with the creation of sidebands. Pilots and avionics engineers know those kind of things.

Slightly off topics; What would happen if the UK government decide to switch off analogue AM and FM broadcast in 2030 I absolutely loathe DAB radios. Collectors will all have to buy very low-powered transmitters to transmit MP3 music or internet radio to their vintage radios and tuners at home? It would be ridiculous for sure.

Norway decided to shut down all analogue radios in 2017, I dont know how the Norwegian vintage radio collectors copied with such drastic and idiotic decision by politicians. I would be enraged.

In Norway It was dubbed FMExit, cheesing many people off.