Re: Best-shielded SMPS inductors?

I didn't seem people driving crazy fast (in 25+ hours of autobahn driving a few weeks ago). 200-220 was about tops, and of course less in tunnels or in the rain. Of course you don't linger in the left lane, that's just rude and as Germans don't ever overtake to the right (a very good habit, I might add) it would certainly earn you their scorn. I don't feel comfortable driving above about 180 for very long anyway, at least in a Ford. ;-)

The differential to cars and trucks in the slow lane(s) is a bit disturbing.. it could easily exceed 100km/h.

The car I had wouldn't go more than 185-190km/h and I often travelled at that top speed for hours. But occasionally clients asked me to use their company fleet and usually gave me one of the more powerful sedans. My preferred time of driving were afternoons on the weekend because nearly every German will either be at a soccer stadium, a sports bar or in front of the TV at home. I've often had others pass me at very high speed while doing around 200km/h myself. There is a voluntary manufacturer limit of about 250km/h but people have cars "re-chipped" and the ones from tuning shops such as AMG or Brabus aren't limited anyhow.

I remember the CEO of one client who drove his various Mercedes S-class sedans at the limit a lot. He said "There is a marked and very noticeable difference between an S500 and and the S560".

Yep, usually a lot more. Got to watch out for Eastern European truckers. They aren't or at least back in the 90's weren't used to the high speeds and would "innocently" pull out in front of you to pass that Russian truck with the wobbly wheels doing 50km/h.

Sepic is nice. The first thing in the input path is an inductor, so the input ripple current is roughly a triangle.

John

Still SMPS-quiet. In the version of my circuit with a bead instead of a base resistor on the cap multiplier (as Joerg suggested), the input ripple is in the tens of nanoamps. It'll just be inductive pickup I have to worry about.

Cheers

Phil Hobbs

If that's still a concern even with the toroid being on the back and a loopy-loop underneath, consider a shield can. Nothing fancy, just snipped and bent out of copper stock. I do that all the time. It kills them magnetics dead :-)

Maybe then it qualifies as instrument quiet.

Polymer aluminum caps are a good tool too. You could put some on the signal processing boards. Their ESRs are really low.

John

Yup, all the switching stuff is going inside its own can. The only worry is the wires picking up junk before they get outside, but I'm pretty confident we can get rid of that with good layout, including putting the input and output wiring on its own layer, on the other side of the ground plane.

I'll report how it works.

Cheers

Phil Hobbs

of

rs,

=A0I

No, we'll be back there in September, but we expect to spend the greater part of our time in Australia from October 2012.

If I had my tools over here, and four extra chunks of scrap wood to protect the chair-legs, I could have done that, but wouldn't have needed to because I already own a belt-clamp - back in Nijmegen. As it happens I did need to buy the belt clamp, which cost me $A19.95, but it did the job, and didn't distress the wood surfaces that took the pressure - I'm sitting on the chair at the moment, and my wife is happy with the way it looks.

-- Bill Sloman, Nijmegen (but in Sydney at the moment)

tes:

cle

se boat

boxes of

well as

the

omputers,

some

e lab. =A0I

while

ual

;-)

s if

ut

w.

The only motorway driving I do in Germany these days is from Nijmegen to the airport at Dusseldorf. Most of the way it's two-lane motorway and restricted to 140 km/hr, which is well within the capacity of both the cars we own. I do get passed from time to time, presumably by people who can either afford the fines or have paid to have their car navigation systems notified of the locations of the speed cameras, but it doesn't happen often.

-- Bill Sloman, Nijmegen

I meant a local can, just around the toroid. Then the wires don't "see" it anymore. Seriously, in one unit I placed a 2W switcher for an isolated power transfer while running microvolt-level RX signals across the same barrier, right next to it. Woiks.

The other thing to do is minimize loop areas on the low-level boards, so they don't see any fields that the power supply may leak.

In a reasonably complex circuit, doing that isn't always obvious. I have a couple of boards that have explicit pickup coils and trimpots for nulling magnetic field sensitivity. I tend to get 60 Hz and harmonics from power transformers, and 120...360 Hz spikies from the currents downstream of rectifiers.

John

I see, That's a good idea too.

I've always been able to do instrument design the paranoid way--linear everything, lots of nice metal boxes, cap multipliers, etc. Normally the small amount of power wasted over the lifetime of the instrument won't pay for a board turn and the attendant schedule slip.

Usually that's paranoid enough, but not always, as in this case--1 uV of pickup from the heater resistors on the flex to the photodiode amp next door will blow the entire error budget for the whole instrument.

Cheers

Phil Hobbs

Just plop down 4-6 pads around it for that purpose, if you want to be extra good with non-covered vias in it. There's also off-the-shelf shield covers but not sure if small enough.

Was the same with our instrument, 1uV of noise would have blown it out of the water because it would have been visible in the image. One of the tricks I often played was to make noise perfectly synchronous to the data acquisition so if there was any it would cancel out.

The worst source I find these days are GSM cell phones. The gigeehoitz RF gets in somewhere, is rectified at the first BJT it finds (usually inside some opamp) and... game's over.

I guess in your case it'll be Sutro Tower :-)

Can you afford to go linear on the heater driver? That could avoid a spin or three.

John

Ayup! Our radios are 2.4GHz with a TDMA "modulation" of 201Hz. Guess what we find all over the audio if we aren't careful (where "not careful" may be using a different board vendor).

Aren't you in the audio biz? TDMA schemes are usually a no-no there, it's as if some dudes with loud Harleys move in next door :-)

Murata has some nice cheap filter and also feedthru caps like this:

Other than that, if you can use a CMOS opamp that makes the problem pretty much go away.

Yup. It gets interesting. It sounds more like a guy in the next car over with straight plugs and copper wires, when you're listening to AM.

We're using them, but once the RF is in, there is no getting it out. Our problems started with a "new" reel of headphone driver amps, a couple of years ago. Evidently they changed the process to improve the 2.4GHz performance. ;-)

Precisely my thoughts, though CMOS opamps that will drive 150ohms from 12V are kinda hard to come by. We have a part now, an old Motorola part now made by some garage in China, that's slow enough to work reasonably well. There is a Intersil CMOS opamp that might work, too, but it'll take a board spin to try it. We didn't find it until after we already had the Chinese part (and it's way cheaper).

boat

of

as

computers,

Or just use a ratchet tie-down. Then you have it to hold the file = cabinet to the hand truck as well as other uses.

?-)