Average DC Voltage

lattesurf

Joined Mar 6, 2007
9
Hi all! I have a question regarding obtaining an average DC voltage.

Example, I have a square wave of 10Vdc peak, duty cycle 70%.
Is there any circuit or IC that is able to convert this square wave into a constant flat DC voltage to 7V?

As the duty cycle varies from 10%-90%, I need some circuit or IC that can dynamically vary between 1V to 9V, depending on the input duty cycle.

I was thinking of using a LPF, but seems like the capacitor would still charge upto the maximum 10Vdc after sometime.

Thanks all!

thingmaker3

Joined May 16, 2005
5,083
Would an op-amp integrator work?

Papabravo

Joined Feb 24, 2006
20,621
A single pole RC filter with a corner frequency at say 0.1 times the PWM frequency should do it. It won't be absolutely flat, but if one pole is not enough then you can cascade several stages at the expense of higher insertion loss.

Opamp integrators prefer a bipolar supply unless you provide a virtual Vss at Vdd/2, and they don't really like to operate close to the positive rail. If you want a 0-10V output, your supply should be plus/minus 12 or plus/minus 15

ashokcp

Joined Mar 8, 2007
50
Hi all! I have a question regarding obtaining an average DC voltage.

Example, I have a square wave of 10Vdc peak, duty cycle 70%.
Is there any circuit or IC that is able to convert this square wave into a constant flat DC voltage to 7V?

As the duty cycle varies from 10%-90%, I need some circuit or IC that can dynamically vary between 1V to 9V, depending on the input duty cycle.

I was thinking of using a LPF, but seems like the capacitor would still charge upto the maximum 10Vdc after sometime.

Thanks all!
True rms-DC voltage converters are available, like, AD736 / 737, AD536A, AD636, AD637 from Analog devices, though these are expensive. You will have to check out the frequency range though. It depends on your application whether such expense is permissible.

thingmaker3

Joined May 16, 2005
5,083
Do average and rms values work out to be identical for a pulse train? (My calculus skills suck.)

n9352527

Joined Oct 14, 2005
1,198
Good point Thingmaker. I think they are not equal.

ashokcp

Joined Mar 8, 2007
50
Yes, It is true that RMS and average values are different.
But in this case, it is a square wave with 0 to 10V with varying duty cycle, and, the average value and the RMS values are equal. I should have mentioned it in the first place itself. I started writing the equation, but gave up as I do not have tools to write maths equations.

n9352527

Joined Oct 14, 2005
1,198
According to several sources, the RMS of square wave is Vpk*sqrt(D) whereas the average value is Vpk*D. Where D is the duty cycle (ton/tperiod). They are clearly different.

Did I miss something?

ashokcp

Joined Mar 8, 2007
50
According to several sources, the RMS of square wave is Vpk*sqrt(D) whereas the average value is Vpk*D. Where D is the duty cycle (ton/tperiod). They are clearly different.

Did I miss something?
No, it's me who missed it. You're right, if is Vpk / sqrt(D). I kept the full period also as D (same var name), while calculating. Sorry about this.

lattesurf

Joined Mar 6, 2007
9
Hi all, thanks for all the responses and suggestions.
i tried an RC filter with a large capacitance (4700uF) and managed to get the PWM signal to a relatively constant DC voltage level, which varies with the PWM duty cycle. Smaller capacitance would only give a ripple voltage, which is not what i want.

Yea, and maybe i should be a little bit more specific, the constant DC voltage level i'm after would be the mean(average) voltage, which is D*Vpk. In this way, i'll be able to obtain the relative 0V - Vpk level range with 0-100% duty cycle.

One problem solved, another pops up.