Wanted: Voltage mode PWM controller

Checkout:

The UC3825 has two out of phase outputs. I need one output with full duty cycle from 5 - 95%

e .

Well, you said: "An UC3843 would suit if it were voltage controlled instead of current". Read the link to the post again=B4.

ere

e.

nt.

..

text -

Look, I said I want ONE output with full duty cycle (5-95%). The UC3843 does have that, but it's intended for current mode operation. If you take a look at the UC3825 datasheet:

You'll see it's a two output device most suitable for bridge type converters like half and full bridge. Each output provides a maximum of 50% duty cycle and that's too less for my application.

I

where

age.

rent.

t...

d text -

The good old NE5561 would do the job perfectly, if it was still available. Unfortunately it isn't. There must be a similair ic around.

There are lots of controllers like that.

TEXAS INSTRUMENTS - TPS40200DRG4 - Voltage Mode Contoller

SEMTECH - SC1103CSTRT - Pulse Width Modulation (PWM) Controller

For that matter whats wrong with using a UC3842 IT HAS 95% DUTY. You dont have to use it in CMC there are no laws that I know of.From the datasheet typical application VMC Buck.

UVLO is 16V for the 3842,if this isnt a problem.

from this data sheet.

There is a way to use opto feedback using the comp pin,if you already have an EA.This seems more trouble then it's worth but it's doable.If you dont need isolation I wouldnt do it.I think Basso has an example floating around the net somewhere.

If you really want to omit the EA you could search for primary side controllers or isolated controllers.These dont have or usually dont have EA's on board the control signal is sent acrross the isolation boundary by a transformer or optocoupler. Onsemi makes several of these NCP series comes to mind.

You can turn voltage into current with a resistor. ;-)

Cheers! Rich

On Mar 27, 5:52=EF=BF=BDam, powerampfreak wrote= :

sg3525

The thread points out the ease of converting the CMode circuit to voltage mode.

For 3842/3 types, you could use the Rt/Ct waveform, suitably attenuated so as not to exceed 1V peak, as an input to the Csense pin to produce voltage mode control. Your external error signal is easily fed into the 'compensation' pin, if the local error amplifier input pin is grounded, but why not use the integrated part? It's free.

I hesitate to go further, as you claim to have found no suitable parts in two fairly comprehensive mfr's product lines. Obviously it's not just voltage mode that is your stumbling block, or you would have found 78s40, 34060/3, 494/594 etc. Note that parts providing two-phase outputs can often be easily summed, paricularly when these outputs are open collector or other simple structures.

PWM can be achieved fairly simply. An integrated device attempts to provide the bells and whistles that you might find you need in addition to that simple feature.

Perhaps if you make a list of what actual properties you expect from the device, someone can help you.

RL

re

.

t.

.

We all know that the UC3842/3 devices has a maximum Isense voltage of

1 volt, since the Rt/Ct ramp has a decent 1.8V offset, it requires level shifting. If the ramp really was starting at zero (close to) It could be useful for implementing a compact and cost effective voltage mode control. The Rt/Ct ramp is often used to implement ramp compensation in peak current mode converters though. To those of you who claim that SG3525 would do for a 5-95% duty cycle output is wrong. The small deadtime between the two outputs will cause problems. The MC34060 could eventually suit, I'll take a closer look to it. I can't use the internal error-amp since the error signal is fed across a safety barrier, the error-amp is placed on the secondary side to compensate for opto CTR variations. Here's a wish list for the ic:

Voltage mode PWM Duty cycle 5-95% Directly access to the PWM comparator input, NOT through an error amp. Undervoltage lock out Frequency up to 150-200kHz As low current consumption as possible (max 5-10mA) Soft-start possibilities

Well, that's all I guess.

I had to make my own 1Mhz V mode controller cause I couldn't find one.

D from BC myrealaddress(at)comic(dot)com BC, Canada Posted to usenet sci.electronics.design

You are being disingenuous and possibly obtuse.

Internal error amplifiers in a great range of controllers feed high impedance low-current pull-up nodes. When internal error amps are disabled by biasing their inputs appropriately, they leave this node free for external manipulation.

Dead-time in two phase outputs simply affects maximum duty, when summed. A 95% duty requirement is not prevented.

Bite the bullet, pick a good, fast chip and go.

RL

I

where

age.

rent.

t...

Yes - I know. And my requirement included access directly to the PWM comparator, which is the same point as the error amp output node. BUT, many devices doesn't put the error amp output (and pwm comparator input) to a separate pin on the chip. THEN, it's a problem. You're forced to use the internal ea. So, what do you mean by I'm disingenuous, really?

UCC3570 series

or digikey

e:

. I

sewhere

stage.

urrent.

y

d/t...

l n y

ts

se

re

Take the sg3525 for example. The error amp output has a pullup resistor and a current sink, so you can toggle the error amp low and control the pwm from the compensation pin.

(snip)

ote:

5%. I

elsewhere

t stage.

current.

any

ead/t...

ull

pin

ily

t

arts

t

hase

are

o

om

f l e

se

de

p.

sorry, the error amp has an active pull-up, not a pull-up resistor. The rest of my post was correct.

TL494, cheap part.

Cheers

I appreciate all your tips, unfortunately it seems some of you didn't read all the requirements. It should be a ONE output 90% duty cycle pwm. I've used the '3525 chip many times for converters like half bridges. I got two interesting tips here, that is MC34060 which appears to be exactly what I'm looking for. Also the UCC3570 series seems great. Thanks a lot!

Cheers