Solid state AM transmitter modulator

Here is an AM solid state transmitter and I would like to use the TDA7360 to modulate the transformer. The transformer is the 12v side of a 220 to 12v mains transformer. Which of the two configurations of the TDA7360 should I use the left or the right one?

The one on the right but you can only use one channel. In your schematic, remove the topmost 470uF capacitor completely. The 470uF below the transformer is the same capacitor as C8 in the TDA7360 schematic. Ground the input side of C2 and feed the audio in at pin1.

It is a stereo amplifier with an option to make it a bridged mono amplifier but in bridged mode both speaker outputs are driven in anti-phase and you only have a single entry point for the modulation so you can't use it. Beware of high voltages on the transformer 220V winding!

Brian.

When the two outputs are combined in a BTL config, you cannot obviously use the stereo mode. When you have a stereo input, you should derive the L+R and use it for the mono channel. The L-R can be safely discarded.

If you want to use both channels, you need to code the L-R also. You need a signal power regulator to control the modulation level. BTL is good because it can output greater power.

With 5-8W of input audio regulator power you can perhaps have about 100W of transmitter power. But your circuit looks not so powerful. What is the designed power output level?

You must always use a dummy load when you are not having the antenna connected.

To get a proper AM modulation you have to use a 1:1 transformer as in the picture below:
**broken link removed**

You can drive the primary of the transformer in single-ended mode or in bridge mode of TDA7360.
In bridge mode, may get too much audio power than an IRF510 needs, so a single-ended would be better and simpler approach.

I trued a series modulator to see how easy or not could it be. Here is the circuit so far, but the LPF will change to a 160m LPF when I finish with the rest of the TX. I removed the final stage and kept just three bjts to experiment at lower power first. Without any audio and without any LPF, I measure 1.7W (square wave) when the output is loaded with a 50R resistor.
I found this modulator in another circuit on the net and since it used the bd139, I thought it would be powerful enough to modulate a bd139 RF final.
Now it does not have another 1k from the base of the bd139 to the ground, to bias the bjt at 6v but sine it worked for him, I tried to test it as it is.

It did not work ok for me though. I can get some modulation only on tones, but on music only a highly distorted drum is heard, no music. I believe the 1M makes the bjt starve, so I will try a lower value tomorrow. But let me know if you have other ideas or things to comment.
Thank you

The 1M probably isn't the problem but the overall bias is. The voltage at the PA collector should be about half the supply voltage with no modulation present. To achieve that reliably you will almost certainly need some DC feedback to stabilize the current or it would be highly dependent on the characteristics of individual transistors.

Experimentally, bias the AF input transistor from the wiper of a potentiometer, adding a 1K series resistor in the base to protect it. It will not be reliable but it should prove the principle.

Brian.

betwixt said:

Experimentally, bias the AF input transistor from the wiper of a potentiometer, adding a 1K series resistor in the base to protect it. It will not be reliable but it should prove the principle.

Brian.

Click to expand...

like this?
I thought to bias the final transistor with another 1k to ground and then capacitively couple the two stages. It will need another 1k from the collector of the small transistor to the VCC

I would take the top of the potentiometer to 12V. You want the emitter of the modulator to have around 6V on it so the base should be ~0.6V higher than that. Bear in mind that if you assume it will pass around 1A, the base current will be as much as 40mA (based on minimum HFE) so its bias resistor can be no higher than (12-6.6)/.04 = 135 Ohms. With 1K it's no wonder it doesn't work well!

Brian.

betwixt said:

I would take the top of the potentiometer to 12V. You want the emitter of the modulator to have around 6V on it so the base should be ~0.6V higher than that. Bear in mind that if you assume it will pass around 1A, the base current will be as much as 40mA (based on minimum HFE) so its bias resistor can be no higher than (12-6.6)/.04 = 135 Ohms. With 1K it's no wonder it doesn't work well!

Brian.

Click to expand...

Like this Brian?

That is correct, you can probably change the 100R to 150R to save a little current. I based the maximum value of the data sheet minimum HFE but in all probability the HFE will be higher than that. Note the polarity of the input capacitor, the base will never have more than +0.6V on it so be careful of any DC voltage present on the audio input.

Brian.

betwixt said:

That is correct, you can probably change the 100R to 150R to save a little current. I based the maximum value of the data sheet minimum HFE but in all probability the HFE will be higher than that. Note the polarity of the input capacitor, the base will never have more than +0.6V on it so be careful of any DC voltage present on the audio input.

Brian.

Click to expand...

The circuit worked Brian
I have not optimized it yet or measured anything, but it worked right out of the box using the output of the PC sound blaster with enough modulation.
The 100R goes quite hot, proving the inefficiency of this kind of modulator probably.

They say there is something about solid state finals and class C that keeps them from 100% modulation.
With Bipolar transistors, a percentage of the input drive power appears in the output.
Which means that the driver stages must also be modulated.

Is it enough to connect the 330 om of the driver to the collector of the final to increase the modulation level?

They say there is something about solid state finals and class C that keeps them from 100% modulation.
With Bipolar transistors, a percentage of the input drive power appears in the output.
Which means that the driver stages must also be modulated.

Click to expand...

That isn't strictly true but in practice it sometimes helps to modulate more than one stage. The danger is that by feeding the driver stage from a modulated supply, it starts to change the load on the oscillator and introduce FM, the overall result may be worse. You could try it and if the carrier frequency shifts by more than say 100Hz, add an additional buffer stage between oscillator and driver, powered from the regulated supply. Checking for FM should be easy, just monitor the carrier frequency and turn the pot from end to end to simulate minimum and maximum PA load.

Most of the non-linearity you encounter will be from the transistors operating at the ends of their ratings, keeping the modulation well below 100% will help. I'm not sure what commercial broadcasters use but I would guess they aim for around 75% at audio peaks.

Brian.