BUCK DC-DC converter with hysteretic comparator loop design

[Ayman Essam]

Hi
I am designing a DC-DC buck converter... the circuit works perfectly when I use a vpulse in cadence to control the switching mechanism... but I want to replace it with hysteretic comparator to be more stable and all these advantages.

THE MAIN ISSUE IS that I cannot control the hysteretic window since everything is dynamically changing :/

please I need a help

 

[BradtheRad]

To create hysteresis, feed the output voltage to the non-inverting input of the op amp.

Adjust resistor values so you get the desired range of On-Off action.

The op amp is a 'cheat' which takes the place of several components.



Notice how the op amp turns on when load voltage drops to a certain point.

When load voltage rises to a certain point, the op amp turns off.

It can be tricky to adjust operation so the hysteresis works properly.

 

[Ayman Essam]

thanks
but you say "feed the output to the non-inverting" and the attached image says the opposite

 

[dick_freebird]

Many low-cost buck converters for applications not involving
RF use hysteretic control. I'm sure their block diagrams (in
some cases) would be enlightening. You may not (probably
don't) want much deadband / hysteresis because that is
going to make for more low frequency ripple voltage on the
output. If you have a simple comparator you may end up
at a PFM operating mode that is not hysteretic per se but
not a classical small signal closed loop either (meaning no
laggy / stability-challenged / transient step performance
limiting feedback loop, just "gimme more" or "shaddap!".

 

[BradtheRad]

thanks
but you say "feed the output to the non-inverting" and the attached image says the opposite

Sorry, I guess I was ambiguous. To create hysteresis, feed the op amp's output back to the non-inverting input, through a resistor. This provides positive feedback.

Confusion is easy because the load voltage is labelled 'out'. It goes to the inverting input for comparison with the reference voltage (which is set by the potentiometer). However the non-inverting input is influenced by hysteretic action. It gets to be a handful.

 

[Ayman Essam]

thanks.
I am designing the below circuit.... the hyst comp isn't working properly for a reason i DONT KNOW


and the output is below:

 

[BradtheRad]

* It looks as though your reference voltage is the same as overall supply voltage. Reference voltage should be your desired load voltage.

* To properly turn off the Pmos, your IC needs to send 3.3V, referenced to the Pmos source (more positive) terminal. This means your IC needs to receive its power from the supply labelled V0.

* To properly turn on the Pmos, your IC needs to be grounded to the same ground as V0. One is connected to 'gnd', the other is connected to 'gnd1'. Are these exactly the same ground?

 

[Ayman Essam]

* It looks as though your reference voltage is the same as overall supply voltage. Reference voltage should be your desired load voltage.

* To properly turn off the Pmos, your IC needs to send 3.3V, referenced to the Pmos source (more positive) terminal. This means your IC needs to receive its power from the supply labelled V0.

* To properly turn on the Pmos, your IC needs to be grounded to the same ground as V0. One is connected to 'gnd', the other is connected to 'gnd1'. Are these exactly the same ground?

ok thanks
but how can i get the whole circuit to work and function after editing those ones i got no physical change

 

[BradtheRad]

Does your mosfet model turn on fully when biased at 3.3V? Even real mosfets do not necessarily turn on at 3.3V.

I use a simulator whose mosfet model is inconsistent in its behavior when biased at low voltages. Therefore I often use transistors instead of mosfets. Transistor action is more predictable. It is obvious to see how bias current goes hand-in-hand with collector-emitter current.

It may help if you create a separate circuit containing only the IC (comparator). Create the voltages you expect to be present in your buck converter. Apply them to the inputs. Watch to see if the output produces the waveforms you expect.