Linear 5V reg with adj.current limit

...

I know that it works on paper, but to need high wattage resistors is just poor design. That one watt pot would likely be wirewound, and very noisy when it's adjusted.

The devil is in the often unspecified details. If you're only building a few and don't have to deal with component variations from different vendors, the job gets easier.

How many different current limits do you REALLY need? Sounds like there's no need to switch the current limit while attached to the DUT?? Sounds like the current limit is only approximate. So, why the series current limting IC at all? What's the advantage over a resistor and a 7805 equivalent?

My favorite example is a USB hard drive that draws too much current to spin up. Makes quite a racket and can't be good to have the heads banging into the stops.

...

An LM395 pass element, and a bit of foldback on the current limit. That prevents the voltage and current regulation modes from fighting each other. I sort of like the Power One approach, where the current limit latches the output low until you turn off the supply or crank the voltage way down and back up again. (My general purpose bench supplies are a couple of 25-year-old Power One linears--nice and quiet, and stable stable stable.)

Cheers

Phil Hobbs

Polyfuses are good for fire prevention. The radial leaded ones work well; the surface-mount ones less well.

I like to use a polyfuse and a transzorb on DC voltage inputs to my boxes, for overvoltage/transient/reverse polarity protection.

I also have used an LM317L TO92 as a current limiter, with the adj pin floating. It works great, but loses about 1.5 volts. I wish there was an equivalent lower dropout on-purpose current limiter.

...

Yup. If the internal current limit is good enough for the purpose, you can just leave the adjust pin open, and it works fine. The thermal limiting gives you some slow foldback effect as well, but of course you don't know if the limit is 1.5 A or 3.4 A, because the spec is so sloppy.

Cheers

Phil Hobbs

The 78Lxx parts aren't the same as the 7800s--they have far worse ripple rejection. I gave up on them long ago. 317Ls are okay.

Cheers

Phil Hobbs

...

Multiply all the onesies and twosies times the number of Spehros, Joergs, Johns and so on :-)

It's like Milwaukee. Hardly anyone needs more than one Sawzall but a lot of people need one just like I do.

I have a classic example in my portfolio. A unit that I designed in the early 90's, still coming off a production line in Guangdong. Changes: Zero, it's essentially still Rev A. Like with this board I always try to design without boutique parts so unless Armageddon happens they can keep producing it nearly forever.

...

For bench supplies most folks want another behavior though. I sometimes need to run mine in constant current mode and a foldback or latching mode would nix that.

With clean 12V going in, ripple rejection shouldn't be an issue? Line regulation might be with a series current limit.

Another way...

ILIM sets the output current limit without requiring a 2W potentiometer.

Design of ILIM is left as an exercise for the student :-) Caution, a negative auxiliary supply of at least -1.25V will be required to pull output all the way to zero volts.

A similar approach should work with the two-LM317-in-series presentation ...Jim Thompson

Or add some diodes to drop the output, then the TL431 loop.

...Jim Thompson

LM1086 - (I'm not sure how it goes with a floating adjust pin)

Please refer to the figure in the datasheet labelled "Current Source".

Important: it only gives low dropout if you have an independent bias supply. Dropout is about 350mV from the main input and about 1.35V from the bias input.

See also LT3082, '3083, '3085.

Disclaimer: I haven't used these particular parts in any design.

Regards, Allan

On a sunny day (Sun, 01 Apr 2012 15:10:58 -0700) it happened Jim Thompson wrote in :

What I do not understand is why not just put a PNP in front of a LM117, and use the beta:

PNP out b | | | | | === 150 [ ] | === 470 [ ]

Am 01.04.2012 00:54, schrieb Jasen Betts:

One L200 from STMicroelectronics will do the same. Unfortunately it also requires high power resistors for the current limit.

Cheers, Alexander

I'm mistreating the pot. probably enough to destroy it, it probably needs a 12W part instead of the 1W shown.

but if (as it transpires) a switch can be used instead 100mA steps can be done with 1/4W resistors, (after the initial 2W)

if your prepared to program the current limit in binary it's possibly even cheaper, eg: some number of 12R in parallel, each is good for

the brief (as I understood it) was a linear regulator with a 12V supply

every watt that's burned in a resistor isn't being used to warm the heatsink,

use 1 to 10 12 ohm 1/4W resistors in parallel for the current selection.

That's a $4 part...

The trip voltage is 0.45V, so just exactly how would that require a high power resistor? That's only 1/2W at the 1 Amp limit.