another current limiter problem

I spend too much of my life worrying about current limiters.

This time, I have a 5 volt power supply, powering some logic that drives a customer load. The load should be 50 ohms, but someone might decide to short it, and I don't want to fry the driver or drag down the +5 rail. So I need a low dropout (under 0.1 volts maybe) roughly

120 mA current limiter. I need two, actually, for two output drivers, and I don't have much room or budget.

I can't find any commercial limiter chips or LDOs that can do this. The limiters mostly work at USB levels, 1 amp roughly, and the LDOs have wild short-circuit current spans, like 3:1 or something.

Looks like I'll have to do it myself. This isn't bad:

formatting link

It's a limiter, not a regulator, so the loop doesn't need to be stable. I think.

(I think I need some new whiteboard markers, too.)

Reply to
John Larkin
Loading thread data ...

LT1120?

Cheers

Reply to
Martin Riddle

An LDO, regulating +5 to +5, is interesting, but that one has a pretty high dropout voltage, and a short-circuit current as high as 400 mA, which would fry my TinyLogic parts.

Reply to
John Larkin

Theres 2 data groups on the short circuit graph, I didnt look into why but it looks like it might be adjustable. 400ma is at 36 vin.

Cheers

Reply to
Martin Riddle

Jbox_Ilim_1.JPG

I've used plenty of TI TPS2560A for similar purposes. They're dual, but only adjustable 250mA to 2.8A.

Perhaps a pair of TPS2551 / 2552 / 2553, which can be adjusted down to < 100mA?. Should be less than $2 each in your quantities.

formatting link

Allan

Reply to
Allan Herriman

Cool parts, but the current limits cover almost a 3:1 range, like 110 to 300 mA for the TPS2551.

If I make my own, I can get within 5 to 10% without difficulty.

I wish someone made a good, low-dropout, accurate current limiter. I guess the shunt resistor would need to be off-chip to get decent accuracy.

Reply to
John Larkin

How about one of the surge limiters (i.e. LT4356). It will give you 10% with

50mV drop.
--
Reinhardt
Reply to
Reinhardt Behm

drives

to

have

Current_Limiters/

stable.

The '2552 isn't quite that bad, with datasheet limits (over recommended operating conditions) of 110 to 150 mA with a 210kohm resistor on the ILim pin.

BTW, The TPS2552 will behave like a current source (until thermal shutdown), whereas the TPS2552-1 will latch off quickly.

Allan

Reply to
Allan Herriman

ox_Ilim_1.JPG

formatting link
?

-Lasse

Reply to
Lasse Langwadt Christensen

Yeah, that's not super complex and the 110-150 current limit would probably be OK. I could parallel some more output driver sections (like, 6 instead of three) so my drivers could stand the 150 mA.

Digikey wants $2 each for the 2552 in small quantities, which seems high for a USB protector. It is nice and small.

Reply to
John Larkin

Fairchild does have some interesting parts.

formatting link

The BGA and leadless packages aren't appealing, but the FPF2125 is similar, in a SOT23, and could be programmed for 110-190 mA, which might work. I suppose I'll have to blow up some TinyLogic gates to make sure.

Reply to
John Larkin

What does the output chip do when its PSU falls - its input thresholds will change, changing outputs, perhaps unstably. Perhaps that doesn't matter.

As pointed out there's various ways to do it. Why not use a series fet on the output driven by opamp? Minimal V drop, no instability or threshold changes producing bad data.

NT

Reply to
tabbypurr

Well, the customer output goes to hell, but that only happens if he shorts it anyhow. At least my box doesn't get blown up.

That would be too slow for this application. The output pulses could range from a few ns to DC, and the edges will be sub-ns. It's conceptually easier to limit the power supply.

Reply to
John Larkin

If this is a high speed output won't there be a 50R series termination resistor that could absorb any possible fault current?

piglet

Reply to
piglet

MCP6561 (comparator instead of an OpAmp), but it'll tend to sing, or chatter no matter what. Maybe some hysteresis would help... or a manual reset. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

I plan to include a switch, so the customer can have 5 volts behind 50 ohms, or 5 volts with maybe six or eight fast CMOS drivers in parallel, no series R.

It's really hard to get some users to understand that, if we provide 5 volts with a 50 ohm impedance, they will only get 2.5 volts if they terminate it. Too many buy the gadget and then complain that it doesn't deliver 5 volts loaded.

So, I need a current limiter, somewhere above 100 mA.

Reply to
John Larkin

Have you seen this?

formatting link

Reply to
papabear546

No. I hadn't, but none of those circuits would work for me. Too much voltage drop and way too many parts. I guess Pease never heard of opamps.

Reply to
John Larkin

I was just about to suggest one of those. Not particularly cheap, and you need an external current sense resistor and FET, but we needed the accuracy (not really achievable with on-chip sense resistors IME) and our supply voltage is 24V which rules out a lot of the monolithic devices. The programmable hiccup mode on overload (to limit dissipation in the FET and whatever has caused the short) is a nice feature.

Reply to
RBlack

The old-school way to make 5v protected into 50 ohms was to make twice that, and add a beefy 50-ohms series termination.

Grins, James

Reply to
dagmargoodboat

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.