Low power FM transmitter chip.

[John Drew]

G'day all,
I just read Les's post on Facebook about his work on a receiver using a si4730 chip. He's talking about a future project for a TX but sounds like it will be AM.
Does anyone know of a chip that will generate an FM transmission of up to 10kHz deviation with a carrier frequency up to 450MHz?

Some background, I've used the si4351 to generate programmable frequencies but they don't do what I want as the 4351 is very rich in harmonics (square wave) and I haven't found a way to put FM audio on a si4351 even assuming I can tolerate the harmonics.

The project is a very simple battery powered signal generator that can be taken to a repeater site for calibrating deviation levels. It should have enough output to be heard without plugging into a receiver antenna socket, perhaps 10dBm output. I'll use a programmable pot to change deviation levels with a 1kHz modulating tone which is my standard.

I have a nice Marconi sig gen but it weighs a ton and I don't want it bouncing around in the back of the 4WD on the rough tracks up the hill.

Any ideas?
Thanks
John

[shantanu@india]

John,
What is the compilers Facebook page?
Can you pm me?

[rick.curl]

Hi John-
Far be it from me to deprive you of the thrill of constructing this yourself, but did you see this?: 470 Mhz FM signal generator

-Rick

[John Drew]

G'day Rick,
I had a look at it, not bad for the price. The description is a bit strange. My application needs enough oomph to get into the antenna of the repeater 30m away and vertically above. A null in the pattern so I reckon a minimum of +10dBm output, preferably 20dBm. By feeding via the antenna I include the whole system in the test.

A friend has suggested feeding a tone into a handheld. Could be a workable idea if I arrange for differing levels of audio to match certain deviations.

Thanks anyway for the generator suggestion Rick.
Cheers
John

[david]

Hi John,
Was that device you used a Si5351 rather than the Si4351 you mentioned?   The Si5351B has facility for VCO input and indicates a modulation bandwidth of typ 10kHz and a pull range of +/-240ppm.  It won't hit the output frequency you want but with CMOS output the 3rd harmonic will only be about 10dB down from fundamental so you could use 150MHz output and pick off the 3rd.  Sounds like you may need an output stage anyway.
All the Silabs/Skyworks synthesized devices will be digital outputs with high harmonic levels and are aimed mainly at telco applications with differential outputs.  If you want clean outputs with digital resolution there are always DDS chips from Analog Devices but they get expensive and power hungry.

Cheers,
David

[trastikata]

Any ideas?

Hi John,

I remember couple of years ago I "played" with AX5043 which is Narrow-Band 27-1050MHz Transceiver, supporting FM modulation and the frequency can be programmed in 1 Hz steps but for the love of God I couldn't figure out all the registers there.

Regards

[Yasin]

Hello John. If 1.5mW would work. The old U2740B could have been used. Maybe boosting the RF output can be the solution.

[John Drew]

Greetings all, thanks for your helpful comments.

Was that device you used a Si5351 rather than the Si4351 you mentioned?   The Si5351B has facility for VCO input and indicates a modulation bandwidth of typ 10kHz and a pull range of +/-240ppm.

David, you're right it was the si5351. I became dyslexic didn't I.
I used the code written by Les. Worked a treat. At the time I bought a board from somewhere. If I can find a "B" version on a board it would be good.

Some background on this project which is working but not to my liking. I now want to improve it. I used a si5351 for the RF source which was fine to replace the crystal in an old exciter module from a transmitter. This provided output on the 2M band (144-148MHz) and the 3rd harmonic for a signal on the 70cm band. The exciter had provision for modulation so it all worked OK. I had a touch pad graphic display to enter and display the required frequency and deviation. Math sorted the actual crystal frequency. Audio from an oscillator in the PIC passed through an I2C controlled pot to manage the modulation (all adjusted for multipliers etc). I'm not really happy with the exciter module operation and it's harmonic on the 70cm band is not strong. I also had problems with stability in the exciter module.

So, I thought if I dispensed with the exciter module and produced the signal directly it would be better. Because it's already made up in a nice die cast box with a LiIon battery and graphic display a simpler solution would be a winner.

My first move now (thanks to the info on the "B" version) is to see if the si5351B is available on a board - I'm hopeless at soldering the SM stuff.

Again, thanks to you all for your replies.
John

[top204]

Excellent details John. I've been caught out quite a few times by harmonics. :-) I designed a multi-mode mobile phone detector and it used a 26MHz crystal oscillator on the PCB, and one perticular frequency in the  low GHz range had a noise on it? It had me puzzled, until I worked out that the frequency was a multiple harmonic of the 26MHz crystal oscillator feeding the local oscillator chip. :-) A change of PCB routing was needed to shield the crystal oscillator better, and it worked well.

I still remember, as a young-un, learning about frequency multiplication using tuned amplifiers at a harmonic of the original frequency. At the time it baffled me, until I created circuits myself and it all fell into place. :-)

The si5351B versions do come on boards, but they are quite expensive, for what they are:

https://www.ebay.co.uk/itm/284774469080?epid=22051667314&hash=item424de18dd8:g:S30AAOSw0PtdOqNB

For modulation, you might be better using an SPI digital resistor as a DAC, or a DAC itself, because SPI is a lot faster that I2C.

[JonW]

TI Produce some really nice VCO/PLL that can do FM modulation, I have used the LMX2572 and the LMX2595 in many applications for test equipment.  These devices have exceptional phase noise and good output power (can get > +10dBm @ 500MHz). They generally incorporate high frequency VCO's and sub divide the output.  For sub 500MHz the next harmonic can easily be filtered with a very simple LPF, frequency step size can also be sub Hz if required.  In band spurs are very low if you have issues with fractional or integer boundary  spurs you can add dithering or adjust the sigma delta modulator parameters to reduce them.  .  I have not used the modulation function (Yet) but others have.  Check out the ERA Micro Synth, these guys have implemented FM with an external MOD input or internally generated waveform, I think this is open source so you can get access to the design info. 

Getting these going is relatively simple and all you need is a decent TCXO as the reference and a MCU at sub 6GHz FR4 on a 4 layer controlled impedance process (Something like JLCPCB 7628 is perfect). Don't be put off with the sheer number of registers with <20k of code space I managed to get them dynamically tuning to 1kHz step size over the full bandwidth by sending a simple string to the serial port.  For a fixed frequency setting you can use the TICS Pro Gui to provide all of the register info and just bang these into a lookup table.  You can then manipulate the individual registers associated with the modulation to get the required deviation etc.  Or maybe simple FSK or stepping the source +/- 10kHz can be used

https://erainstruments.com/

LMX2572

Check out Hope RF Modules as well, these are pretty flexible little devices that can be controlled easily https://www.hoperf.com/ic/rf_transmitter/CMT2119AW.html.  Many are available on modules for a few dollars

Many years ago I developed some ASK and FM transceivers and we used Saw resonators and single BJT (BFP420 from Infineon).  You can create FSK of FM fairly easily adding a varactor.  If you have any old crystals lying around you could look at building a simple oscillator using crystal at a sub divided frequency, use a varactor to pull the crystal and then pick off the harmonic, amplify and filter the response.  Be very low cost for a fixed frequency.

[JonW]

If you don't fancy designing the PCB, here is one already done https://www.ebay.com/itm/233317598013

[david]

My Chinese is not too good but this might be an option-
https://www.aliexpress.com/item/4001115066657.html?spm=a2g0o.productlist.0.0.7ad746dfxcDzxn&algo_pvid=3b47d516-275f-4a4f-bf96-c558796ca88a&algo_exp_id=3b47d516-275f-4a4f-bf96-c558796ca88a-22&pdp_ext_f=%7B%22sku_id%22%3A%2210000014500406546%22%7D&pdp_npi=1%40dis%7CNZD%7C%7C23.68%7C%7C%7C%7C%7C%402101e9d216515868979411133eb2bf%7C10000014500406546%7Csea

I think that equates to about US$14

Cheers,
David

[top204]

That is an excellent price. And free shipping to here in the UK, so it should also be to Australia because it is closer (I think. LOL)

Ebay is actually getting more and more expensive, and Amazon is the king of "Rip Off Greed", and always has been. Even when it was just selling books, they were always stupidly priced for second hand books. But the "sheep" follow advertising and promotions. :-)

[John Drew]

Thanks David, I'll order one today and the good part (aside from price) is that the existing code will work.
I've really appreciated all the time people have put into their research.
Cheers from Oz,
John

[david]

Good luck with the project John and watch out for the fine print......  Make sure it is the chip you think it is.

Cheers,
David

[joesaliba]

and Amazon is the king of "Rip Off Greed", and always has been. Even when it was just selling books, they were always stupidly priced for second hand books. But the "sheep" follow advertising and promotions. :-)

No wonder why Jeff Bezoz is wealthiest person on earth!!

Joe

[John Drew]

Board ordered, it says the B version. Working on trust!!
So it will either be a good result or a lost modest sum. They reckon a fortnight to deliver - if it's that quick I'll be amazed freight free.
I'll keep the group informed of whether it works out.
John