PLL based FM Tx

[KamalS]

Hi,

I am actually searching for a nice FM Tx using which I can transmit FSK/ASK WITHOUT worrying about freq. instability over a period of at least 1 - 2 hours of continuous usage.

Would like to sending pure FM and never use carriers above 100Mhz.

I have no problems using, say, a 40kHz carrier, but stability of the frequency is the MAIN criterion.

I chose 40Khz as an example, because this freq. can be had for cheap using normal CMOS digital gates and a XTAL or even a standalone microcontroller.

What do you guys think about a simple 565 PLL based FM Tx for such low freq (40Khz to say 300Khz range)?

 

[mvs sarma]

Hi,

I am actually searching for a nice FM Tx using which I can transmit FSK/ASK WITHOUT worrying about freq. instability over a period of at least 1 - 2 hours of continuous usage.

Would like to sending pure FM and never use carriers above 100Mhz.

I have no problems using, say, a 40kHz carrier, but stability of the frequency is the MAIN criterion.

I chose 40Khz as an example, because this freq. can be had for cheap using normal CMOS digital gates and a XTAL or even a standalone microcontroller.

What do you guys think about a simple 565 PLL based FM Tx for such low freq (40Khz to say 300Khz range)?

32.768KHz crystal gives a clean and stable output. toy can always have a varicap diode,and implement fm modulation.

at 40KHz PLL may not be necessary.

 

[Mikebits]

40KHz is silly, the frequency is just to low. What bandwidth are you needing? Your antenna would be incredibly large and you would require much power.

 

[mvs sarma]

40KHz is silly, the frequency is just to low. What bandwidth are you needing? Your antenna would be incredibly large and you would require much power.

afterwards he might use an uplift with a second modulator , perhaps.

 

[Nigel Goodwin]

Hi,

I am actually searching for a nice FM Tx using which I can transmit FSK/ASK WITHOUT worrying about freq. instability over a period of at least 1 - 2 hours of continuous usage.

What's wrong with all the cheap licence free radio modules?, these work flawlessly - and are legal, unlike your ideas.

 

[Mikebits]

afterwards he might use an uplift with a second modulator , perhaps.

Operating a FM modulating carrier at 40 KHz is just absurd. I have no idea what you mean by uplift. Do you mean up-converting to a higher frequency?

 

[mvs sarma]

Operating a FM modulating carrier at 40 KHz is just absurd. I have no idea what you mean by uplift. Do you mean up-converting to a higher frequency?

we used to have FSK in voice band with 240Hz drift (FREQUENCY SHIFT)
In the voice band of 0.3 to 3.4KHz we used to pack 12 such FSK lines and transport.
Yes, i meant up convert.
Sorry for the wrong term used, Mike.
thus it is not absurd and we worked for decades on those principles. agreed that it was a low speed telegraphy.

 

[KamalS]

Guys, I believe 40Khz based FSK had been used for decades using ultrasonic sensors?

I know that the 433Mhz based remote band is deregulated for hobby use, but the transcievers are too noisy.

Well, forget abour 40Khz - that was just can example, what about 1Mhz etc?

The main idea I am looking for is a PLL based Tx with stable freq. carrier generation.

 

[mvs sarma]

Guys, I believe 40Khz based FSK had been used for decades using ultrasonic sensors?

I know that the 433Mhz based remote band is deregulated for hobby use, but the transcievers are too noisy.

Well, forget abour 40Khz - that was just can example, what about 1Mhz etc?

The main idea I am looking for is a PLL based Tx with stable freq. carrier generation.

Most probably there is no need for PLL to have a 1Mhz frequency. any crystal oscillator would do.

However the recent method of getting a clean and stable clock is DDS, employing chips like AD9835, 9850/51 AD9854, AD9952 etc from Analog Devices. They might be costly, and have to be controlled using a micro or PC. Even these chips need a reference clock input.

 

[KamalS]

Analog Devices and TI have ready made FM trancievers

I am looking for something cheap using which I can send ASCII data at 80 - 100bps

 

[Nigel Goodwin]

Analog Devices and TI have ready made FM trancievers

I am looking for something cheap using which I can send ASCII data at 80 - 100bps

I'll repeat again, what's wrong with the legal option of licence free radio modules?.

Why would you think they are noisey?.

 

[mvs sarma]

Analog Devices and TI have ready made FM trancievers

I am looking for something cheap using which I can send ASCII data at 80 - 100bps

Simply follow Nigel's advise and write coded software on micro to suite the Tx and Rx needs. Once you send few dummy byes and then the main byte (means 3 to 4 bytes every time you need to send a byte), it would work without noise.
the very first byte RX may not be able to receive.
there is one sample on AVR based program written in C
It is here
WinAVR AVR-GCC Tutorial

you can get the philosophy and translate it to PIC.

 

[KamalS]

I'll repeat again, what's wrong with the legal option of licence free radio modules?

I have already written my own protocol based on a crude version of manchester for those 433Mhz trancievers. The problem is that in shows and competitions, the band is simply overloaded and you cannot just ask people to turn off their stuff!

Reliability can be had in such cases using more complex protocols OR by using a different frequency!

Also, there are multiple variants of the 433 on different price points. Almost all are chinese and affect the remaining microcontroller circuits as well - I can almost never use my USB port for debugging with those remotes turned on.

 

[Nigel Goodwin]

I have already written my own protocol based on a crude version of manchester for those 433Mhz trancievers. The problem is that in shows and competitions, the band is simply overloaded and you cannot just ask people to turn off their stuff!

That doesn't mean you can randomly pick your own frequency, which will be illegal.

Presumably this is radio control we're talking about?, why not use the specified equipment, in the specified way - with different coloured crystals.

 

[KamalS]

Well, I believe the 170MHz spectrum can be used for RC as well?

Anyways, my main question is how to get a stable carrier frequency.

I have been doing some reading and saw that I need to use:

1. An XTAL to provide the oscillations
2. A PLL that would run on top of that (yet another IC)
3. A multiplier (yet another IC)
4. A uC to control 2 and 3 (yet another IC)
5. A varicap always there

Went to the library and had a look at a good book and they schematics they had were for stereo FM... quite complicated (I would rather buy a ready to go IC from ROhm than build that)

My second question is, cannot be make a discrete multiplier (3) withotu buying the damn IC?

 

[ccurtis]

I once used a CD4046 PLL to make a stable 146 Mhz FM transmitter. The VCO section of the CD4046 is unused. You divide a sample of the output of a VHF oscillator using a digital counter to get the frequency down to where the CD4046 can operate against a fixed crystal reference (say 1 MHz) into the phase comparator section of the CD4046. You sum your modulating voltage with the output of the CD4046's phase comparator after the low pass (loop) filter and feed that to the FM input to the VHF oscillator. This works for small FM deviations, say up to +/-10 Khz, and has the stability on the order of a crystal oscillator.

edit: But, yes, today I would opt for an off-the-shelf module and try to tweak one if could find none that are 100% suitable.

 

[Nigel Goodwin]

I once used a CD4046 PLL to make a stable 146 Mhz FM transmitter. The VCO section of the CD4046 is unused.

So really you didn't use a 4046 PLL, you only used the phase comparator out of one - you could just as easily have used a standard CMOS gate to do it.

 

[KamalS]

ccurtis, that is interesting.

How did you make the VHF oscillator?

1. I am finding it VERY difficult to generate more than 20Mhz from a XTAL .. an LC can generate in 100s of Mhz but it's EXTREMELY unstable.

2. I was thinking - maybe use the LC to generate twice of thrice the carrier freq. and then use a divider to bring it down - this should reduce variance?

 

[ccurtis]

I made the VHF oscillator from an idea in an old QST magazine. It was a single BJT transistor colpitts type. The key to a "stable" LC VHF oscillator is the Q of the resonant tank, which is mainly determined by the tank inductor. Of course, the power supply must be well regulated too and the oscillator and tank should be only very lightly loaded. It used a short section (roughly 1/2") of .144 semi-rigid coax for the inductor. The outer conductor is grounded at both ends. The center conductor formed the inductor. It has good mechanical stability and is self shielding. Connections between the tank capacitor and tank inductor must be very short. I bent the inductor into a U-shape to bring the ends of the inductor mechanically very close to where the leads came out from the capacitor bodies to ensure the connections were indeed short. It used a varactor lightly coupled to the tank to provide the very slight change in the total capacitance of the tank needed for the phase comparator to tune it. I used the more complex of the three phase comparators in the 4064 since it gives better phase noise performance. The oscillator should be mounted in a metal shielded container, as mine was, to avoid the effects of outside influences.

You can actually build a more stable SHF oscillator, because you can build practically sized, very High Q resonators at those higher frequencies. It would be tough to divide them down with digital circuits though.

 

[Nigel Goodwin]

ccurtis, that is interesting.

How did you make the VHF oscillator?

1. I am finding it VERY difficult to generate more than 20Mhz from a XTAL .. an LC can generate in 100s of Mhz but it's EXTREMELY unstable.

There have been overtone crystals in the past that operated directly at 145MHz, for 2m band transmitters without multipliers.

However, it was never a good idea, and the more normal method was to use a 12MHz one and multiply it 12 times - this also enables you to get decent NBFM whilst using a crystal.

Any remotely modern method though uses a PLL, with a free running LC oscillator locked to a much lower frequency crystal.