need help understanding PWM

Thread Starter

raffter

Joined Feb 28, 2008
113
Hi guys,

here was a simple way of getting a pwm with varying pulse width

http://forum.allaboutcircuits.com/showthread.php?t=49467

now I have read from somewhere that trilevel pwm is better than bilevel

though I dont quiet understand how the switching "sequence" is done.based on the attached diagram, how is the positive half cycle switched? is it like A=PULSES and C=just a POSITIVE of the sine reference? and on the negative cycle, B=PULSES and D just a POSITIVE(negative cycle) of the sine ref?

pls note that the fullbridge is single supply only.

thanks
 

Attachments

SgtWookie

Joined Jul 17, 2007
22,230
That's pretty much it, except due to the way that many high and low side drivers work, switch C would be on and A PWM'ed, then for the negative part B would be on, and D PWM'ed.



In your schematic, B and C are the low-side switches; A and D are the high-side switches. Drivers such as the IRL2110 use a "boost cap" that has to be constantly charged back up by having the low side switch turned on. Drivers like that one are pretty difficult to understand at first.
 

Attachments

Thread Starter

raffter

Joined Feb 28, 2008
113
thanks bertus,
lots of reading to do.

@sgt.
HOW about if the roles are switched(lower switches be the PWMs and upper switches=on/off)? that way the bootstrap caps are sufficiently charged<by the switching of the lower switches>. would this be a better option?

attached is a simulation. still unsure if these are the right signals... I have just added a comparator to have a square output from my sine ref
 

Attachments

Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
Your output waveform:



You asked:
HOW about if the roles are switched (lower switches be the PWMs and upper switches=on/off)? that way the bootstrap caps are sufficiently charged <by the switching of the lower switches>. would this be a better option?
If you can get it to work, go ahead.
Drivers like the IR2110 require the high-side MOSFET to be PWM'ed in order to keep the cap charged.

In the meantime, have a look at Linear Technology's datasheet for the LT1336 Half-H-bridge gate driver that has a boost regulator:
http://www.linear.com/product/LT1336
On page 15, there's an example Class D amplifier, which is similar to what you seem to be attempting to do.
 

Attachments

Last edited:

SgtWookie

Joined Jul 17, 2007
22,230
A and B must never be on at the same time.
C and D must never be on at the same time.
It would be OK for B and C to be on at the same time, but then you couldn't turn A or D on without shorting the supply.

A typical sequence would go something like this:
1) All MOSFETs are turned off.
2) MOSFET B is turned ON.
3) PWM is applied to MOSFET D's gate.
4) PWM is removed from MOSFET D's gate, and MOSFET B is turned OFF.
5) A small amount of "dead time" passes where all MOSFETs are turned off.
6) MOSFET C is turned ON.
7) PWM is applied to MOSFET A's gate.
8) PWM is removed from MOSFET A's gate, and MOSFET C is turned OFF.
9) A small amount of "dead time" passes where all MOSFETs are turned off.
10) Go back to step 2.

You cannot apply PWM to both A and D at the same time, or you will short the supply across the high and low side.
You should not use an inverter between MOSFET B and C's gate signals; as there must be some "dead time" to prevent shorting the bridge on one side.

If the PWM is still applied to the gate while the low-side MOSFET is off, the boost capacitor will discharge and the high-side MOSFET will either become disabled (if the driver is so equipped) or will burn up due to insufficient Vgs (voltage on the gate using the ground terminal as a reference point.)

Logic level enhancement mode MOSFETs have Rds(on) specified when Vgs is 4.5v to 5v.
Standard enhancement mode MOSFETs have Rds(on) specified with Vgs=10v.
 
Top