Just discovered why high side P-channel MOSFETs are suboptimal

[PE9ZZ]

I had a brilliant idea: since I have no stable power source, why not use the PWM modulator to take care of that? I have a small RF deck with two IRFP240s and a CMCD topology (will create a new topic for that sometime soon) with 20 W carrier at 4.8 V supply. My raw DC is 12 V which results in 130 W RF. The only PWM topologies I find online are feedforward designs where a PWM modulator drives some switch fed by a switchmode PSU.

The TL494 is a true PWM modulator and using LTSpice I ended up with this design (below). Jim Tonne's notes were very inspiring and resulted in a clean spectrum. I needed an external error amp because the internal opamp was slewing too slow destroying the bandwidth. But I reached only 71% modulation before the peaks started to show mains ripple. Turns out that my IRFP9240 has a 0.5 ohm R(DS,on). With an 8 A switching current this results in a 4 V drop across the MOSFET. Oops.

Next iteration will be a half bridge with IRFP150s and an IR2103. Same TL494.

This being my first post I changed the title into something potentially less offensive...

Tjerk, 9ZZ

[vk3alk]

Hi Tjerk…

Thanks for your post and interesting read….
Not sure what your power levels are in the Netherlands but maybe 130Watts PEP is it ... or close to it anyway…
You would think that a good 12 volt analog power supply or SMPS would be good enough without using feedforward to cancel ripple etc:
Have never used the TL494 as a PWM as they have limitations one which which is the slew rate etc:
Can offer a couple of ideas which over time would achieve better results for you maybe….
One is using higher voltages and the other using other devices for the PWM generator other than the TL494…. leave it for SMPS or motor controllers etc:
Higher voltages would mean different FETs too as the IRFP240 in your CMCD has a 200 volt rating…
But it’s your call and as long as you get the results and are happy with things then that’s all that matters…

Wayne

[PE9ZZ]

Hi Wayne,

Thanx for your kind remarks. You're right, the TL494 slews slow but that's solved with an external opamp. The PWM modulator is just fine and with the LTC6992 unobtanium the poor man's version of that chip. I am about to start modding my PMOS high side switch into IR2104 NMOS high/low side switches (attached).

As for power levels, right now I am just heating my dummyload as I have no intention of putting a signal on air. Which will be a challenge anyway with me living in a third floor apartment...

Tjerk, 9ZZ

[vk3alk]

Ohhh Ok

If you are just having a bit of fun why not use this Power Supply circuit.....
Will give a lot more control to things etc:


Wayne

[PE9ZZ]

If you are just having a bit of fun why not use this Power Supply circuit.....

Ehrn... that is a high side PMOS switch. Just discovered that they are suboptimal 

Still have the parts for the NMOS switches waiting on my workbench. Hope to pop them in soon...

Tjerk, 9ZZ

[PE9ZZ]

Hope to pop them in soon...

Too many other interesting projects but finally put them onto the board. It certainly makes a lot of difference but I ran into issues that I did not foresee. Like the 33 kHz oscillation that takes place when not modulating. Under certain loads. Also, the LM358 is not a rail-to-rail opamp and really doesn't like to be driven close to ground. And at higher frequencies the PWM modulator starts misbehaving at higher levels. Until 200 Hz I reach close to 100 % modulation but above that it starts to slew in an ugly way.

Other than that it easily drives my 1.1 ohm dummy load at 4.8 V carrier. My RF deck is about the same. It only gets lukewarm.

Will post piccies later.

Tjerk, 9ZZ