PWM controller vs switching regulator

[dyunasing]

I understand this question has been raised before. But sorry, i did not really understand the answers.

https://www.edaboard.com/showthread.php?t=236402

http://www.analog.com/en/analog-dialogue/articles/applying-dc-to-dc-step-down-buck-regulators.html

It mentions Switching regulator integrates oscillator, PWM control loop, and switching FETs.

But i thought the oscillator is already built into a PWM controller? FET of course not.

And let me get an example of current mode PWM controller, http://cds.linear.com/docs/en/datasheet/1619fa.pdf

To me, it is as good as a switching regulator. What are the exact differences? Thanks in advance.

 

[FvM]

FET of course not.

Yes, that's exactly the difference. Hard to understand?

 

[dyunasing]

Yes, that's exactly the difference. Hard to understand?

Dear moderator,
I am just confused how ppl market their switching regulator. If FET is the difference, can I interpret the External Power Switch Boost as a PWM controller? And does a PWM controller alone necessarily have a poorer load regulation?

 

[FvM]

can I interpret the External Power Switch Boost as a PWM controller?

Linear at least does, you find the mentioned LT1619 pwm controller in this category...

And does a PWM controller alone necessarily have a poorer load regulation?

A PWM controller alone doesn't work as switching regulator, why do you think it has a "poorer load regulation"?

 

[crutschow]

You are sort of mixing apples and oranges.

A switching regulator typically uses some form of PWM to regulate the voltage.
The regulator includes other functions, such a feedback loop compensation, soft-start, and current limiting.

PWM is a technique used to vary the average DC voltage to many things, such as DC motors, heaters, and lamps.
A PWM controller may run closed-loop or open-loop (no feedback).

So a switching regulator is a specific application of using the general PWM technique.

 

[dick_freebird]

The controller ought to comprise ramp, reference, error amp,
modulator and logic level (or FET driver) PWM outputs. Some
ancillary features are likely as well.

The "regulator" would have all of these plus the power path.
The "controller" would have the power path off chip.

However some "controllers" with >1A FET driver outputs
are perfectly (or acceptably) suitable for lower power
regulator use - UC18xx / UC15xx parts are sometimes
used "bare" (direct driving the buck inductor).

 

[dyunasing]

Thanks all for the replies. Let's say, If I am to build a boost regulator of 24V to 50V(1A) based on a current mode PWM controller, what am I lacking?

 

[KlausST]

Hi,

based on a current mode PWM controller, what am I lacking?

Instead of discussing about this vague phrase....why don't you show a circuit?

Then we all are able to talk about the same "specified" item and don't need to "assume".

Klaus

 

[dyunasing]

hi, KlausST,
You are right.
I have this ST1843 passed to me to build this boost regulator (24 to 50V).

https://www.st.com/content/ccc/reso...df/jcr:content/translations/en.DM00034310.pdf

What i have in mind is something similar to LT1619 reference circuit.


Is it that simple? What are the drawbacks? Thanks.

 

[dyunasing]

Looking at the PWM controller datasheet, it is fairly complete to me. Error amp, oscillator, fet driver, etc. I do not need those soft start, UVLO features anyway.
Anything i have overlooked?

 

[dick_freebird]

It contains the regulation functions you need. The ones
you say you don't, are "play nice" kind of features - for
example, preventing power FETs from being so weakly
driven that they burn up if you turn on your converter
before the upstream source has gotten right (UVLO).
Or keeping the converter from sucking every bit of juice
out of said source while it tries to charge up its filter
and load (soft start). Give some thought to startup
modes and limitations of the local environment.

What it does not have, is a low loss power switch(es).
Only a modest sized totem pole driver meant to drive
power MOSFET gate capacitance, or throw a couple
of hundred mA at a BJT base (not so popular anymore).
So if you want it to be a regulator you need appropriately
low performance expectations.

 

[dyunasing]

Thanks, dick_freebird. it gives some thoughts. :thumbsup:

 

[dyunasing]

hi,
If i have to stick with this PWM controller, what are the improvements i can make to boost its performance?
I have an separate 5V power path for an MCU. The MCU can fulfill the UVLO and SS function. For the totem pole, no much i can do, right?
It is better to get a FET instead of BJT based driver? Anything else?
Thanks.

 

[dick_freebird]

We have no idea what you have to do, what you think
you have to do, what you really want or why. There
are many decent boost converter ICs (but do you want
boost, or need some other features such as isolation
that make things much more complicated?). The ICs'
value proposition always includes lower BOM cost than
an equivalent piece-parts assembly. A question which
bears directly on choices, is your current level (bet on
more than a couple of amps needing discrete power
switches, if you're north of maybe 12V input). Do not
worry about the "ancillary functions", they are almost
certain to be present in any IC (controller or complete).