You can cook up an audio compressor. Similar to this

Do like I did and feed it from a computer running Stereo Tools which processes the audio.

Also could get a Raspberry pi 3 and use that as the processor as there's a version of stereo tools that works on it.

Could then build the transmitter on the pi itself perhaps substituting the PLL for an output of the pi that is a square wave and can be varied in frequency.

HI,
A month ago a start design and building a "simple" low power MW Transmitter.
Instead starting from zero, I search for similar projects.
One of the results was this post.
I started mixing my own ideas and that I found on search, and building test circuits on protoboad.
As final project, I used a CD4053 (same of kit on this topic) as modulator, but, I faced a problem with output stage,
when on 50 Ohms dummy load, modulated waveform are distorced (damped). I try a single transistor buffer,
common emitter and common colector (voltage flow), still have a distorced waveform, now I have assimietrical waveform (positive or negative side depending of buffer configuation). Need to "buffer" both sides, after some consideration I decided to use a two transitor on complementary configuration, like on audio amplifier. Yes, Its works, now have a nice waveform on output with a proper load. Distorced when have a missmatch impedance output.


Now, I working on antenna coupler circuit. The photo is one of my test circuit, a led tuning indicator.


Sorry, I an novice here, I can not post link to full schematic. But, I have a plan to post about this project here, if not have a restrictions for that.

Think you had an issue of using such a low impedance load.

The part 15 transmitters must use a higher impedance antenna such as a 10' maximum wire in order to get any range with the limited available power.

Hi,
The 50 Ohms dummy load is only for testing and calibration procedure.

I don't have a RF wattimeter or other calibrated instrument to measure irradiated RF power.
I can only estimate.
On dummy load I have near 7.6Vpp, 76mW pep (147mW rms). Oscilloscope probe connected direct to dummy load.
Another way to estimate RF output power is measuring supply power "consumption".
With no load, LPF and Antenna disconnected, 50mA.
With 50 Ohms load no LPF, 58.7mA.
With LPF and 50 Ohms load, 65mA.
With LPF and antenna conected and tuned, 61.7mA.
With LPF and antenna conected and untuned, 88mA.
With antenna disconnected and no load on LPF, 88mA.

Power Supply is 15V.
With dummy load, have 15mA of difference, or 225mW (rms).
With antenna, 11mA, and 175.5mW (rms).
If consider the power dissipated on LPF, 94.5mW.
On dummy I have, 130.5mW, and with antenna, 81mW.

All measurement was done with 80% modulated carrier at 1620kHz.

On output stage, R21 and R22 control the final power, the are 10 Ohms resistors.

My multimeter is out off calibration.

Cannot check it with a 50 ohm dummy load unless the antenna you desire to use is 50 ohms.

The common method of checking the output is to measure the current the final amp draws when connected tot he proper antenna. Has to be no more than 100mW to be legal.

You know it is great if you are happy with your modifications but it really was not such a horrible circuit as supplied. You don't want to exceed the modulation index shown below. Within that limit it appears that it works well enough. The output of the modulator chip is a PAM signal and it needs a quasi-linear or better amplifier. The factory amplifier is close enough if you do not push the modulation too hard.

A proper antenna for MW band is huge, even licensed station have a mismatch antenna.
For correct operation, between transmitter and antenna have a matching LC network (ATU).
Output of transmitter are designed to match with transmission line (ex.: coaxial cable).



That 100mW confused me for a while. I thinking about RF power. So, the texts of FCC did not make much sense.
I found other text explain bit more clear the FCC part 15 rule.
Its is a 100mW of input on power supply. Or, power draw of final stage.
I add a 1Ohms resistor on 15V side of final stage, now is easy to measure current consumed by final stage.
My transmitter is exceeding 100mW if adjusted for maximum audio gain and <100% modulation.
Its is the reason why I using a dummy load, to avoid unnecessary RF irradiation.

The maximum antenna length is 10 feet and is a relatively high impedance, but that's what you want to get the most range from the transmitter while not exceeding the 100mW limit.

Perhaps switching to a 10 foot wire and redesigning the output will cause it to be 100mW or less and transmit a decent range.

What is the impedance of a 10 foot (3 meter) antenna at 1MHz?
I want to use a resistor (dummy load) to simulate the antenna for testing/tuning.
I've read that a CB 108" whip is about 35 Ohms?

According to NEC2 for a 3.6 m vertical antenna over a moderate ground the resistance is:

#3m_Vertical: R-in (real)
#Freq [MHz] R-in (real) [ohm]
0.5 378.552
0.6 378.227
0.7 377.843
0.8 377.398
0.9 376.893
1 376.329
1.1 375.707
1.2 375.026
1.3 374.289
1.4 373.494
1.5 372.644
1.6 371.738
1.7 370.779

and the reactance is:

3 m Vertical, moderate ground
Freq [MHz] X [ohm] C (pF)
0.5 -15613.7 20.4
0.6 -13015.4 20.4
0.7 -11160.1 20.4
0.8 -9769.08 20.4
0.9 -8687.67 20.4
1 -7822.96 20.3
1.1 -7115.83 20.3
1.2 -6526.89 20.3
1.3 -6028.86 20.3
1.4 -5602.25 20.3
1.5 -5232.77 20.3
1.6 -4909.71 20.3
1.7 -4624.86 20.2


At one time common practice would have been to use an RMA antenna dummy between a 50 Ohm signal generator and the radio you were tuning to simulate the antenna. But the RMA dummy was intended to simulate a ~30 m horizontal wire antenna. A 3 m vertical looks like 375 Ohms in series with 20 pF.

I seen some alignment instructions that call for a dummy antenna that wasn't 50 ohms and maybe a year ago I googled it and found out how to make it.

Often when I align radios which have a tuned RF amp stage I will align it with the antenna I plan on using connected and the RF generator loosely coupled to it.