Thoughts on Design time for a switching power supply

AC line powered?

John

What topology are you thinking of? AC or DC input? Isolated or not? That's 43W (call it 50W) in 12 cubic inches, and the 6mm height (including a case?) might present some real challenges.

Compare a stock 50W off-line switcher, for example:

29 cubic inches (single output) with more than 10x the ripple, and 6x as thick.

/

Lithium battery 14.4 nominal 16-11V usable range. At least that is what is written on the battery.

It sounds like he's one of those "better scheduling through shouting" guys.

One of my more cynical rules of happy employment is that if you're bound to screw up in the boss's eyes, you may as well piss him off by doing the job right. Remember that he's your boss this month, but that your name will be on that drawing for _years_.

1/4" thick is going to present an immense challenge; better start digging for _really thin_ transformers.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" was written for you.
See details at http://www.wescottdesign.com/actfes/actfes.html

Just pay very close attention to the layout guidelines. IIRC TI has recommend PCB layouts in their datasheets. Just make sure to calculate the inductor properly for the range of input voltages.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

You could make it out of those Linear Tech LTM bricks, but that wouldn't be cheap. Next bet is to do what you suggested, use the National online design thing. 2 weeks is optimistic, all things considered.

John

Power supplies are one of my favorite pet peeves!!!! A power supply is an IMPORTANT part of any design. You don't have a spec...and I mean a well-thought-out, FIRM, IN WRITING, PASSED AT LEAST ONE SYSTEM DESIGN REVIEW...and on and on...

Anybody who specifies a form factor like that either has no clue or has an overriding constraint and is willing to spend a lot of time and $$$$ doing it. I vote for the former.

Your question suggests that you have no clue either.

Can you design a power supply in two weeks? Of course you can!!!! Only problem is that it's gonna take the other 90% of the schedule to make it work.

Anything you fix BEFORE the first layout is much cheaper than doing it AFTER the first layout...or the second...or the third...

The schematic is only a tiny part of making a power supply work. People typically use the cheapest parts that meet spec. Problem is that a PS is a high-stress environment and many parts don't really meet their specs. There's a lot of "art" involved. Learning that art is EXPENSIVE for your first project. You don't mention whether you're responsible for the care and feeding of the lithium battery. Or the compliance testing by all the regulatory agencies. Somewhere, there's gotta be a way to charge the battery from something hooked to the mains???

I spent most of my career bailing out projects in trouble. The issues often were with the EASY parts that got left until last or assigned to rookie engineers. Power supplies ranked high on the list. I can tell you horror stories about working with power supply design firms who were supposed to know what they were doing.

I had one case where I had to fire a local PS design firm. Purchasing found another firm in another state. Turns out they didn't have staff, so they hired another engineer for the project. Guess who they hired? The local guy that got fired after the first design fiasco. Didn't find this out until the schematic showed up looking a LOT like the first one.

Your boss sounds typical. I predict his opinion of you will be less after this project, no matter what you do. Stuff happens. And you're gonna get some on you.

t /

just finding magnetics that will fit on a pcb and be less than .25" high will be a challenge I think

-Lasse

Lots of small inductors in series and parallel?

John

out

s

his

g it

it

DC/

the

nd

do

are

B o

yeh something like that, or really high frequency. He also said he had to do with what he already have in stock so

.25" is not much, I think a d2pack on a normal pcb is more

-Lasse

Coilcraft has very flat inductors. Bare in mind that there are very high frequency switchers available nowadays which run at 1MHz or more so even with tiny inductors you can handle high currents. And like John said you can always use multiple inductors.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

I know I use some, ~2 mm high together with some 1.2MHz analog devices switchers, but it sounded like he had to do with what he could find in his own stock

-Lasse

Yes, I see some decent candidates that are 0.181 or so high.... you know the old saw.. good/cheap/fast, pick any two... this is quite do-able with expensive parts fast (depending on the other requirements), but you might be able to reduce the cost in moderate volume by better than half by doing 1.5 or 2 x the work. Or put two

60-70 hour weeks in and take a week off, and the time to market is not affected.

Make sure you get a solid written spec on this one- the battery adds another dimension to the system .. you have to protect the battery (not draw current from it when it's discharged), maybe charge it?, it will change impedance depending on charge state, you probably have to be reasonably efficient presumably.. etc. Those ripple specs could cause you problems. If your supply is 80% efficient, at 43W you'll have more than 8W to dispose of, from a rather small package... maybe not a problem, but be careful.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

o

yout

is

this

ng i=3D

it

f DC=3D

a

the

ind

do

care

CB

to

n

is

"The Journey is the reward"

eff.com

Thanks to all for weighing in on this one. This is a great forum to run a sanity (insanity?) check.

Just found out that schematics for this and the uP controlled (three

8051s) optics and fluidics controls are due next Friday. I am going to use off the shelf DC/DC converters from Murata or XPPower and the like and the battery will get pulled out for recharging. Ripple wise I am going to get what I get at this point and if it doesn't work the unit will have to run off of a high quality lab power supply (wish I had one). We are in Alpha development stage on a medical diagnostic instrument and if it falls short on portability at this point we will just have to live with it. I told my boss he needed a senior electrical design engineer in addition to me (20 year electronic tech/ machinist) when he hired me but it does not appear to have sunk in yet;-)

At least I get paid by the hour.

Best, EdV

I seem to recall hearing of some low profile converters where "holes" were cut in the PCB to allow the magnetics to extend down to the bottom of the PCB. If you're counting mm, ...

Serves you right.

RL

Buy, don't build. You're going to need UL and medical device approval, which will cost you more than a buy.

What voltages are relative to what? Could you generate

-12 and +12, and use the difference for the 24V supply?

What do you want -12V for, anyway? Not an RS-232 port, I hope. Nobody does it that way any more. If it's for an RS-232 port, see this Maxim tech note:

"

"

All those voltages can be obtained from most PC-type power supplies, of course. So you can run the prototype off something you probably have on hand, assuming that the

24VDC supply can be taken from +12 and -12.

John Nagle