Power Supply PWM control chips as PDM generators

Welcome and feel free to ask questions about class E, mosfet's, and AM.
You must be registered to post.

Power Supply PWM control chips as PDM generators

Postby frank carcia » Thu Feb 13, 2003 12:58 pm

I started playing with PWM about 20 years ago and quickly found Current
Mode superior to voltage mode. The feedback is the pulse by pulse monitor of current through the filter. Operating at high current like we
all use would dissipate a lot of power in a sense resistor so I use
Current transformers to monitor pulse current. I actually have a small RC filter to handle the first few hunderd nano seconds of leakage inductance in my filter. The biggest problem I had was scaling all the voltages around the modulator to be within the limits of the SG1846 PWM. This meant my current mode feed back voltage had to stay below 1 volt or an internal OP amp would saturate. The error amplifier is just like a normal PDM generator where you feed in audio and a DC offset. All these voltages are referenced to the internal 5 volt reference. This means the maximum voltage was below 5 volts at the output of the internal error amplifier. I ended up with needing about 1 volt of audio maximum. The 1846 has two outputs each with a maximum on time of 49% so I had to use both phases to fully modulate. The extra 2 % is fixed dead time for magnetics to reset and in PDM case time for the damper diode to do its job. Each phase drives 4 - 9 Amp MOS FET Drivers which in turn drive 4 IRFP340s. With slope compensation I am able to get down to about 96% negative modulation before the filter goes discontinous. I tend to run about 150 % positive peak so the 4% of negative clipped by the negative peak limiter is lost in the noise. I still have not connected negative feedback but it is quite simple to add. A resistor to the error amplifier and
set the Carrier up a bit. Someday I will actually add VSWR shut down.
This chip actually has another input that fires a SCR and latches off until
power is cycled. This modulator also serves as a voltage regulator for
the power supply. There are a number of newer PWM parts out there
worth checking out. I think it is a lot easier than using discrete parts.
The current wave form based on pulse current going through the filter
replaces the triangle generator. This current wave form is converted to a voltage ramp by the current transformers. The slope of the ramp is the
real power going into the filter rather than the triangle wave method which doesn't take into account of power supply voltage sag or filter
resistance. Still a carefully set up triangle wave system will be a great first step in understanding the system. All of us should build a good triangle wave generator once. fc
frank carcia
Senior Contributer
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby Dan Brown » Thu Feb 13, 2003 5:59 pm

Hi Frank:
Your modulator seems to work very well. Sounds good for a long time.

I tried a TL5001 and UC3825 (?) PDM chips. In the TL5001, I could not get the feedback amp to be linear. On the 3846, it was pretty linear, but only went to 50% modulation.

In the interest of getting on the air, I used Steve's Ultimate PDM generator circuit, which works very well. I feel there is an easier way to get the PDM signal that Steve's circuit.

I will study your chip to see how it works. Maybe, if you are interested, you can post a full schematic of your modulator?

Do you run any DC or audio feedback around the recovered audio?

Dan Brown
Posts: 25
Joined: Sat Feb 01, 2003 2:52 pm
Location: Natick, MA

Postby frank carcia » Fri Feb 14, 2003 10:00 am

Hi Dan,
We stopped using the 3825 in '89 when the current mode became popular. I did a switcher design with the 1846 and at one point connected a audio source to the loop and got hooked on the idea.
The 3825 was a real problem for us. The loop had a problem with feed back delay through the filter. I don't know anything about the TL5001.
The 1846 has 2 outputs so you only get 50 % with one. That is why my
modulator is 2 phase. There are a number of single ended components
that work in current mode worth looking into. The whole trick in making
them play is finding the proper internal operating voltages.
I tried DC feedback after the filter when I first built the modulator but
it isn't presently connected. I thought it might be better to rectify the RF output and use that for feedback. Never got that far yet. All the hardware
is in the modulator I just never tried it again after the rig was working.
The triangle wave test went through ok fine so I just stopped playing with it. fc
frank carcia
Senior Contributer
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby Dan Brown » Sat Feb 15, 2003 11:09 am

Hi Frank!

Thanks for the reply.

It is interesting in the limitations that are found with these chips.

I will try the 1846 sometime. Right now I am slowly working on a 75M RF deck.

Dan Brown
Posts: 25
Joined: Sat Feb 01, 2003 2:52 pm
Location: Natick, MA


Postby VK3KRI » Sat Feb 15, 2003 5:31 pm

Gday Guys.
I am currently working on a modulator using a TL494 (as used in nearly every PC powersupply ever made). Even with no feedback - ie feeding audio straight into the comparator with the error amps just as a buffer, the audio is sounds cleaner than any other AM TX I've ever built.
I haven't got a distortion measurement, as I still have a problem with residual switch frequencies being picked up by the N&D meter. I am a bit suspicious of the sawtooth wave form, but as I said it certainly sounds OK so far - certainly for the minimal amount of effort to build it.
The '494 has the choice of single or double ended outputs and also has 3-5% deadtime built in.
Doing the frequncy response runs are fun, - I don't know where the LF rolloff is as my sig gen only goes down to 10 Hz!

Posts: 44
Joined: Fri Jan 31, 2003 8:32 pm
Location: Melbourne, Australia

Postby frank carcia » Sat Feb 15, 2003 8:15 pm

hello Ian,
Good luck using the PWM chip. As you see it is easy to build compared to a board full of discretes. My first one was the board full of parts but provided a good way of
understanding the process. A good test is the triangle wave test. Try a 500 hz or 1 KHz
triangle wave into the audio input and see what comes out. My rig easily passes the
sound of the oil burner running. sounds like a large rocket taking off. My best low frequency note was the washing machine coing through a cycle. Some reported I was
FMing until I explained the source. The strongest low frequency transient comes from
my 7 year old son hitting the flooe directly above the shack. Look at my filter design
for ideas. I like to run about 77 KHz sample frequency with a filter flat out to 10 KHz.
3 dB point a few KHz above that. The best filter is a AIR core single layer. I am able to get away with a 2 section filter. Running open loop you will see your problems below
10 % modulation. Current mode feedback will get you lower. I had to put a small RC filter
in the feedback to roll off the transient caused by leakage inductance and turn to turn
C when the pulse first turns on. It only lasts a couple hunderd nano seconds.
Look at the points of the triangle coming out of your modulator. Sharper the point the better the modulator. (compared to the input signal). Good Luck! fc
frank carcia
Senior Contributer
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Re: Power Supply PWM control chips as PDM generators

Postby lu1jpj » Thu Jun 05, 2014 12:13 pm

Hello Friends, I'm in Argentina and here are building this, here I leave a link, of course thanks

Posts: 1
Joined: Sun Jun 30, 2013 5:46 pm

Re: Power Supply PWM control chips as PDM generators

Postby Gilly » Fri Jun 06, 2014 4:25 am


Have enjoyed your WEB site....

Here is some more....



Class E Enthusiast
Posts: 167
Joined: Tue May 15, 2007 6:59 pm
Location: Mt.Eliza Victoria Australia

Re: Power Supply PWM control chips as PDM generators

Postby Diego-CX4DI » Fri Jun 06, 2014 7:25 am

Hi Frank,

I have some circuit with the TL494 if you like to know.
I do not know how is it post in this web but I can send via email.

best 73's
Class E with PWM modulation
Posts: 43
Joined: Fri Sep 30, 2005 10:22 am
Location: Montevideo-Uruguay

Return to Class E Discussion

Who is online

Users browsing this forum: No registered users and 4 guests