Another design is a broadband one, LC based no PLL, all broadband binocular ferrite cores except two trimmers in the final stage for 3.5W .
Even these can be removed. No tuning stage anywhere except in the oscillator. Reverse isolation is so high that 6 hours drift was 1KHz.
It outputs half a watt with no tune network in driver or final stage. It was developed in 2011.
A breeze to make as there are no air core coils except on in the oscillator stage.
Good bye
[Bericht gewijzigd door Anoniem op woensdag 13 april 2022 17:16:59 (100%)
Hi Dare4444,
I hadn't followed this thread since last year, I've been out of the running for the past month too, hope you'll be okay soon.
Last year I built your design together with @rob007.
It certainly works well for amateur applications, I also made one with 4 crystals!
I have written a model for Pspice of such a parallel crystal configuration.......
In the past I have used a VFO with an AFC that also works well.
The FM modulator that I developed 40 years ago works with a specially cut crystal due to the offset to the center frequency !
I hope you keep visiting Circuitsonline.....
Greeting,
Henk
Op 13 april 2022 17:17:30 schreef electron920:
Hi Dare4444,I hadn't followed this thread since last year, I've been out of the running for the past month too, hope you'll be okay soon.
Last year I built your design together with @rob007.
It certainly works well for amateur applications, I also made one with 4 crystals!
I have written a model for Pspice of such a parallel crystal configuration.......In the past I have used a VFO with an AFC that also works well.
The FM modulator that I developed 40 years ago works with a specially cut crystal due to the offset to the center frequency !
I hope you keep visiting Circuitsonline.....Greeting,
Henk
Ohh, great Henk!
The 96MHz was okay but the later design which I developed sounded much louder. Even music was loud with good fidelity. IN4007 beat varactor diodes in both designs. Pleasant surprise it was.
The second circuit is the holy grail for homebrewers coz of its simplicity and high fidelity. Anybody here saved the circuits I posted am hour back?
Nice to hear from you, Henk. You're very kind.
BR from Joy
Op 13 april 2022 15:51:22 schreef Brightnoise:
Problem with quartz oscillator core is the limited deviation. For a standard FM transmitter you should reach 150 kHz peak-to-peak frequency deviation (at 100 MHz broadcast band).Your other circuit on Harry's Homebrew, with the PLL on 750 kHz, is almost reaching that for audio frequencies.
Not almost, but 100% . AF BW is 10Hz to 25KHz !!
But leave that circuit. The varactor diode has large RF voltage present as it's connected to the collector of a power oscillator of 25mW. The RF swing distorts the audio. It's barely noticeable but when I compared with a colpitts oscillator in which varactor has low RF present as it's connected between base and ground. Audio was indistinguishable whether it was coming from my phone's speaker or radio speaker.
If I share my circuits here, they are not gonna be used by someone for commercial purposes, right?
Op 13 april 2022 15:23:28 schreef Brightnoise:
Joy, welcome on this forum.And your offer: post it here, or give a link to where it can be found. Many people here are interested in a frequency-stable FM wideband solution!
-----------
Mel: het is geen verkoper, maar een hobbyist.
https://www.electronicsforu.com/electronics-projects/simple-fm-transmi…
Exactly. I found it rude. I sometimes tend to take life way too seriously, haha. No issues. It's okay.
Lucky that you decided this!
Most people here are friendly....
But to answer a question that is enclosed in your system with 750 kHz reference frequency and maximum deviation: your assumption that this system can handle large deviation is wrong.
An FM signal at 96 MHz with FM modulation frequency 20 Hz and peak deviation 75 kHz has (integral) phase deviation of 7400 radians p-p.
After division by 128 (your divider to come to 750 kHz reference frequency) the deviation is 57,8 radians p-p. The HEF4046 on CP2 can handle plus and minus one radian so 2 radian p-p before it will loose lock.
Conclusion: the product of lowest modulation frequency and allowed peak deviation is factor 30 (roughly) too low.
----------
Here is a solution for the problem: divide further down to around 25 kHz.
Then you can have maximum deviation and lowest audio frequency and maintain lock.
My recent experiment was a cheap DDS from AliExpress, the audio type up to 64 kHz or the RF-type up to 55 MHz with modifications, that allowed 1 Hz step around 25 kHz and consequently an RF step of 4 kHz. And then the maximum deviation, lowest audio frequency and maintaining lock can easily be fulfilled. Please give attention to your loop filter. May be I come back on this.
Check this. HiFi audio and volume is as loud as commercial broadcast station.
750KHz reference. Pls check!
Tested it extensively. Same quality audio as other stations, same loudness, hifi.
It's using the loop filter, I'll post schematic here. I eliminated all trimmer capacitors and varactor diodes. 16V 1W zener as varicap gave amazing audio. Pls watch.
The premphasis is crude but four opamps in tl084 should handle pre emphasis and act as a compressor as well!
[Bericht gewijzigd door Anoniem op donderdag 14 april 2022 02:40:20 (11%)
Op 14 april 2022 00:48:36 schreef Brightnoise:
Lucky that you decided this!
Most people here are friendly....But to answer a question that is enclosed in your system with 750 kHz reference frequency and maximum deviation: your assumption that this system can handle large deviation is wrong.
An FM signal at 96 MHz with FM modulation frequency 20 Hz and peak deviation 75 kHz has (integral) phase deviation of 7400 radians p-p.
After division by 128 (your divider to come to 750 kHz reference frequency) the deviation is 57,8 radians p-p. The HEF4046 on CP2 can handle plus and minus one radian so 2 radian p-p before it will loose lock.Conclusion: the product of lowest modulation frequency and allowed peak deviation is factor 30 (roughly) too low.
----------
Here is a solution for the problem: divide further down to around 25 kHz.
Then you can have maximum deviation and lowest audio frequency and maintain lock.My recent experiment was a cheap DDS from AliExpress, the audio type up to 64 kHz or the RF-type up to 55 MHz with modifications, that allowed 1 Hz step around 25 kHz and consequently an RF step of 4 kHz. And then the maximum deviation, lowest audio frequency and maintaining lock can easily be fulfilled. Please give attention to your loop filter. May be I come back on this.
Oh. I re-read what you wrote again.
Oh, sorry, got it. I was opiate medication for the shoulder injury and it left my mind blank. In this PLL true phase lock isn't possible, but the average frequency remains the same. You cannot hear the difference in real life as audio is just as loud as commercial stations and fidelity is excellent. Check the YouTube video I posted.
I've a new greatly simplified circuit, it used 65,536 total divisor value of prescaler so it covers your + - 75KHz audio deviation.
For the sale of simplicity, the 128 divisor rate is more than enough. Difference is not audible. I know I did hours and hours of listening test of my 800mW TX.
[Bericht gewijzigd door Anoniem op donderdag 14 april 2022 02:35:50 (88%)
The YouTube video is a nice one.
The 750 kHz PLL is a kind of limiter detector: too high frequency or too low frequency.
With normal audio the average level of that limiter-like comparator will correspond with the average of the signal and I can imagine that the introduced distortion is low and inaudible.
Please attention to the loop filter: that filter must be slow enough to allow 20 Hz audio in this PLL without deviation degradation and without asymmetric distortion.
All with all: I am inmpressed. I always thought that the deviation should fit within the phase range of the phase detector. And you showed that a PLL that is overdriven in phase still can average correct, or correct enough.
Thanks for sharing your ideas.
----
As Electron920 already mentioned: FM-control loop stabilises the frequency also in a correct way without the addition of distortion or affecting the average frequency.
The RF section. I could never get > 900mW out with the 2N3866.
I've a genuine source for 2N3866 though.
Cost is <$2.
Anybody wants?
I'll post the PLL part later. This is the RF section only. Notice no trimmers! 19AWG = 20SWG , 25AWG = 26SWG, 36AWG = 39SWG.
Yes, the distortion is not audible at all and not even seen on an oscilloscope when the audio sine wave is recovered from the receiver.
I eliminated unnecessary trimmers. 900mW is output power at 96MHz.
Audio is CD quality. I have done extensive listening test and I'm an audiophile, lol.
A LPF is needed at the output.
This circuit is tuned by slightly squeezing or stretching the oscillator coil for a PLL lock. I've tons of other designs.
[Bericht gewijzigd door Anoniem op donderdag 14 april 2022 15:16:11 (26%)