Re: Best-shielded SMPS inductors?

Worked just fine here. Joerg must have yet another problem with his "perfect" system that crashes all the time ;-) ...Jim Thompson

So what do you think, Jim?

Cheers

Phil Hobbs

I'm about to head out to California to pick up a granddaughter and bring her back to Phoenix for the month-long Valley Youth Theater summer program... she is quite the thespian ;-)

But I did take a quick look, and your snubber approach is one of my favorites... used that in most of the supplies for the OmniComp/GenRad portable testers... lovely load lines that follow the axes!

I won't be back until Sunday. I'll run the simulation then, but first looks, I'd give it a thumbs-up! ...Jim Thompson

On a sunny day (Fri, 01 Jul 2011 12:17:58 -0700) it happened Jim Thompson wrote in :

I looked at both, one had some problem with values. The other one popped up automatically.

Same here. The one posted on Phil's web site works fine. Probably the use of the Greek letter for micro caused that. It hadn't been bailed out yet and was setting the file on fire in protest:-)

An interesting way to keep things quiet is the Push-Pull design. Jim Williams app note An70 describes it.

Basically it keeps things quiet by being mostly a 50% duty cycle switcher and relying on the transformer turns ratio to do all the work. Linear Tech has a newer chip the LT3439 besides the LT1533 in the app note. The app note also has some great pictures of old Tektronix scopes.

I always liked that LT1431/opto isolation method. Very few people know that a TL431 can be used as an OpAmp. I did that trick in most of my OmniComp/GenRad switchers from 1977-1987... predating this appnote by 10 years :-) ...Jim Thompson

That is a really, really old trick ;-)

First time I saw it was in a very old Unitrode publication. That was IMO the company with the best SMPS experts.

Was it more than 34 years ago? ...Jim Thompson

You said "predating this appnote by 10 years" and the app note is from

In the really old days they used other shunt regulator chips such as the LM136.

I can't count, it was 1977 when I first did that ;-) So _20_ years before the appnote ;-)

Can you cite something predating 1977? ...Jim Thompson

I have repaired SMPS from before then that worked in such a fashion. Citing isn't possible anymore because there was no Internet back then and I've chucked lots of old manuals over the years because they were no longer useful. They have become tissue and paper bags :-)

I hate to be smug (sure I do :-), but I think the TL431 was first introduced around 1977... I can recall the TI rep pushing it. ...Jim Thompson

Read again :-)

I wrote that really old SMPS did not use the TL431. In fact, some had a discrete version in the shape of a diff amp and a 5.1V or 5.6V zener. One supply of the Methusaleh category transferred it not via opto but a lamp (!), a piece of cardboard tubing and a Cadmium-based LDR.

What made my knuckles white on that one was the fact that they hadn't reversed polarity. Meaning that if the lamp ever failed the output would rail hard and the thing would most likely have blown its gut out. Oh, and no crowbar so whatever was connected would have gone *POOF* as well. Seems they weren't concerned about that ...

I found this supply on a computer scrap yard in the town of Aachen, Germany, around 1980. From some sort of mainframe that was totally outdated back then and most likely discarded by an institute of my alma mater. Used it for a long time but scrapped it when moving to the US because if was 220V only. Big switchers weren't 90-264VAC back then.

Then there was a similar shunt -> opto path in a 1970 vintage gambling machine. Not in a primary side SMPS (it had a linear one) but in the HV generator for the neon lighting. I was very surprised that it had a feedback path at all.

Thanks. For this purpose the buck switcher is pretty nice since it avoids putting transients back out the input, as well as making the output quiet. There's SMPS quiet and then there's instrument quiet.

Cheers

Phil Hobbs

Careful with those extra LC filters at the output. They form a series resonant circuit (at least that's how the PWM chip sees them) and can mess with the loops stability.

Phil makes it to SF once in a while. We have beer here, too.

Three hours, sac to SF, is worst-case. Unless you have to go through Roseville on the way. We just did. It was 15 MPH and 105F in Roseville. 66F as we crossed the Bay Bridge. The gradient was just about 1 deg F per mile after Red Top.

John

I just got back from Amsterdam after an overnight trip. The view from the air was interesting, it reminded me of the early "sim city" games if anyone remembers those. Everything neat and tidy and *engineered*, put there by design not because it grew that way.

And it is indeed flat. I liked it.

I could live in Germany if language etc. wasn't an issue. Well planned, tidy and the countryside isn't visually polluted by billboards, giant signs and so on (you might argue about all the windmills for electric generation). Cars are mostly in very good shape (even those in former East Germany). And your car can escape idle in a lot of spots on the autobahn and get up to 200km/h+ when conditions merit.