Boost Converter 2

164

01:58:11

This circuit is an improvement on a Boost Converter is uploaded a long time ago: 
http://everycircuit.com/circuit/5388359727054848
 The problems with it were:
- bad feedback system
- no switch current limit
- no latch
- unrealistically low switching frequency
These flaws have been overcome in this circuit.

This version still might not be the best, but it's better than the previous version. This circuit switches at 100KHz and outputs 12V, taking an input of 9V. 

published 6 years ago

kiani

6 years ago

It seems to work for a little while, and then it starts to turn off and back on again... Noticed you haven't used any kind of snobber network , is it left out for any particular reason '?

PrathikP

6 years ago

What part of the circuit needs a snubber? I never thought of that.

PrathikP

6 years ago

Oh and can you elaborate on "It seems to work for a little while, and then it starts to turn off and back on again".

PrathikP

6 years ago

I think your talking about the absence of pulses for a while. It's because the output voltage goes above the set voltage, so until the output voltage goes below the set voltage, the latch stays off.

PrathikP

6 years ago

The feedback system is slightly different. A DC voltage is not compared to a triangle voltage. In this circuit, the latch has inputs for 2 square wave generators, one turning it on and the other turning it off at 60% duty cycle, if it hasn't turned off already. Then the latch can also be turned of if the current limit or voltage limit is reached, and it stays off (because the comparator output stays high) until the current/voltage fall below their limits.

PrathikP

6 years ago

And the voltage at the drain of the mosfet only goes upto 17-18V, so a snubber network is not needed. Didn't check that before publishing the circuit, but it looks like it isn't needed.

PrathikP

6 years ago

And back to the feedback system, comparing a dc voltage to a triangle wave needs a huge amount of components, and for some reason EC cannot handle error amps (which I would need). And any circuit with a lot of op amps connected together crashes.

PrathikP

6 years ago

This is a sync buck I built a while ago. The dead time control ckt is made up of op amps. The ckt is fine, but the simulation crashes, which makes me sad because the dead time circuit works fine on its own http://everycircuit.com/circuit/5840292138450944

kiani

6 years ago

Yes the output voltage is maintained, when it stops switching periodically,,, didn't realise it is a feature of the supply. Snobber also doedn't seem to be needed. Even current an inductor being switched, ! What i hsve seen done a zene across the Fet, in some designs.

kiani

6 years ago

EC crushing,, have you tried spice ir Lt.?! I don't know whether the crashes are the limitations of EC.., somehow one has to be careful, as a simulator, EC or any,, is forcing designs in a certain way.. Sivour design takes inti account if simething called simulation,,, machine control!

kiani

6 years ago

I tried a smps a long time sgo on EC, the loop was so slowed up by EC, couldn't keep up with calculations..ever increasing matrix multiplication and crashing due to singularity,, ( i think).. The whole thing can not be done, so just blocks.....we had a smps. Designed well working and meeting the spec.. Yet we had to spend 6 months to get it running on a sim..cause the customers insistence.. One of the engineers made changes to the actual circuit ( negatively affecting the performance) to make it work on the customers sim.. Laughably.

PrathikP

6 years ago

I've tried LT spice, it's a nightmare to draw schematics in it. I rarely use it.

kiani

6 years ago

PSPICE is the uni's choice. In recent years they have intriduced input and output live interface for ease of use.

LJ1234

6 years ago

*liked

add comment in editor