Hey I was mucking around in multisim and decided i'd have a crack at designing my own buck converter entirely of discrete components, and to simplify things, opams.
Basically i started by doing your basic pwm thing with the triangle wave generator and a reference voltage, i added a feedback path so the supply self regulates etc. It all works nice.
I wanted to make a triangle generator rather than using the function generator on multisim.
I found a basic circuit off the net that uses an integrator and a schmitt trigger to make linear up down ramps (triangles!!).
Problem is when i increase the operating frequency about about 1khz this thing starts to lose linearity.
This does affect the performance of my supply a bit, i notice most
pwm chips have a built in "precision triangle generator" to ensure the best performance?
Here it is.
What can i do to make it more linear at higher frequencies?
Your help is appreciated many thanks 😀
Basically i started by doing your basic pwm thing with the triangle wave generator and a reference voltage, i added a feedback path so the supply self regulates etc. It all works nice.
I wanted to make a triangle generator rather than using the function generator on multisim.
I found a basic circuit off the net that uses an integrator and a schmitt trigger to make linear up down ramps (triangles!!).
Problem is when i increase the operating frequency about about 1khz this thing starts to lose linearity.
This does affect the performance of my supply a bit, i notice most
pwm chips have a built in "precision triangle generator" to ensure the best performance?
An externally hosted image should be here but it was not working when we last tested it.
Here it is.
What can i do to make it more linear at higher frequencies?
Your help is appreciated many thanks 😀
First, I'd use better opamps, probably jfets. As the frequency goes up, the current through C2 and R3 goes up, so you need to reduce C2 to keep this reasonable. You might need to keep U4 from saturating, as that will slow it down. Ultimately, you might have to use a real comparator there, if you want really high speed. With a good FET opamp and styrene caps, this circuit can achieve linearity in the sub tenth percent region with little trouble. I assume you've got decent bypassing on the opamps? (Oh, never mind, I just noticed that it was a simulation!)
Use faster op-amp models. Reduce C2 value to reduce current demand on U2. Reduce R12 to reduce the amplitude of the triangle wave and the current and slew rate demand on U2 (5Vpp typical).
Thanks for the help guys!
Ok i changed the models to the 'comparator virtual'
Initially multisim 10 has a whine when it simulates but it has some debug that unscrews its head so its all good.
I tweaked the passive components to get a nice clean 55khz triangle wave. That cap in the integrator is now 50pF.
Using multisim's models of real comparators and opamps i noticed some bizarre things happening as well as the program having a fit (literally).
Considering the circuit works well now i might as well attempt to prototype it, I will post it up here when i do.
Basically it runs off 10vDC and drops it to 5vDC.
One question, what is the difference between an opamp and a comparator, is it mainly the specs or am i missing something?
For this circuit what would you recommend in the way of a 14/16 pin dil package with 4 opamps/comparators in it. eg 'lm324' or whatever.
thanks 😀
Ok i changed the models to the 'comparator virtual'
Initially multisim 10 has a whine when it simulates but it has some debug that unscrews its head so its all good.
I tweaked the passive components to get a nice clean 55khz triangle wave. That cap in the integrator is now 50pF.
Using multisim's models of real comparators and opamps i noticed some bizarre things happening as well as the program having a fit (literally).
Considering the circuit works well now i might as well attempt to prototype it, I will post it up here when i do.
Basically it runs off 10vDC and drops it to 5vDC.
One question, what is the difference between an opamp and a comparator, is it mainly the specs or am i missing something?
For this circuit what would you recommend in the way of a 14/16 pin dil package with 4 opamps/comparators in it. eg 'lm324' or whatever.
thanks 😀
An op amp has a dominant pole in it to be able to easily achieve gain without instability in a closed loop. A comparator does not have this artificial bandwidth restriction because it is assumed to be running faster. By adding a capacitor to a comparator you can make it act like an op amp and by removing negative feedback from an op amp you will make it like a comparator. There is no real fundamental difference but of course there are practical differences in real parts - for instance, some op amps have distortion reducing circuits inside them, some comparators have complementary outputs, etc. But op amps and comparators are virtually the same when you boil it down.
A LM324 is a decent choice if you just need a general purpose comparator that can run well off a single rail. An LM339 is another good, general choice for comparators. For op amps, I recommend the TL074 and MC33079/179. There are many other parts available but you can get excellent performance, both in instrumentation and audio applications, using the above parts in most circuits. I think any of the above parts would work for you, but since you're running the parts off a single supply, the LM324 you mentioned is probably your best starting point.
A LM324 is a decent choice if you just need a general purpose comparator that can run well off a single rail. An LM339 is another good, general choice for comparators. For op amps, I recommend the TL074 and MC33079/179. There are many other parts available but you can get excellent performance, both in instrumentation and audio applications, using the above parts in most circuits. I think any of the above parts would work for you, but since you're running the parts off a single supply, the LM324 you mentioned is probably your best starting point.
Here's a neat triangle-wave generator:
http://www.national.com/an/LB/LB-23.pdf
And there's a simpler one in Figure 25, here:
http://www.national.com/an/AN/AN-20.pdf
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
http://www.national.com/an/LB/LB-23.pdf
And there's a simpler one in Figure 25, here:
http://www.national.com/an/AN/AN-20.pdf
- Tom Gootee
http://www.fullnet.com/~tomg/index.html
- Status
- Not open for further replies.