TCA0372 troubles

Say, Can I parallel opamps on the output for more current? How well will they share current?

GH

On a sunny day (Tue, 4 Jun 2019 07:34:25 -0700 (PDT)) it happened George Herold wrote in :

Transistor (current drive) or a MOSFET (voltage drive) is much better. Add some opamp for control.

Or use the transistor or MOSFET itself as heater?

Right I might have to do that... it takes more thinking though.

The opa551 looks nice.

fig 29 shows how to parallel opamps.... I might try that with a different opamp.

George H.

How about a ~10 ohm polyfuse?

Fine, if you use separate feedback networks and a resistor in series with each one to limit the current consumed by duelling offset voltages.

I often make 50-ohm outputs by parallelling two sections of an LT1260 CFA, each with a 100-ohm resistor in series.

Cheers

Phil Hobbs

The data sheet is sketchy, if not evasive. It claims that the chip has thermal protections but warns against exceeding 1 amp out. There is nothing that specs actual current capability or current limit.

Is your heater a resistor? That should limit the opamp output current.

What's your supply voltage?

One alternative to using a big opamp is to use a little one and boost its output using a 3t voltage regulator, LM317 or LM1117, as a power booster, essentially a smart, protected emitter follower.

I've used the TCAs without problems, but I'm usually driving a fast output switch that has a final 50 ohm resistor, which takes a lot of stress off the opamp.

Probably too much voltage drop x current can blow out a transistor before the thermal limit kicks in.

Don't short the output!

I'm designing a laser controller that has to power five external energy sensors, 12v at up to 100 mA each. I've considered polyfuses in the 12v outputs, but if the customer shorts things, they can pull a lot of current before they heat up and open, which could reset my box.

I might use some LDOs as non-regulators, just to get their current limiters.

Huh, I've got some poly fuses... I always thought of them as slow.. but I could try. (I was thinking I could add some series R on the output, but inside the feedback loop...)

The OPA552 data sheet showed a master/ slave relationship with two opamps.

George H.

Right, but kids have access to the output and can short it.

Huh.. 317 as power booster? I've used LM395 for that.. but they are spendy... How do I do that with an LM317? (and what's the dropout?)

Then I can't sell it. :^)

George H.

Just like the LM395. IIRC JT said that it's the same silicon with different metal. If you have more than one supply voltage involved, you might need some protection diodes.

Cheers

Phil Hobbs

Drive the ADJ pin with your signal from a small opamp. A series resistor and a zener or something can do your square root.

You'll get the +1.25 volt offset, and the dropout on the data sheet. You'd have to drive ADJ to -1.25 to get zero out.

I tried abusing the ADJ pin of an 1117 all the ways I could, and couldn't break it.

The 1117 has lower dropout than the 317, but it's only rated for 12 volts. The ones that I tried failed at around 60.

OK thanks, I found this,

The LM395 looks like a Darlington... there is the 1.25 V drop in the LM317.. but I guess with feedback it all comes out in the wash.

GH

Sounds sensible as long as they don't short it to someplace nastier than ground, e.g. 240VAC. But if the circuit board is charred, it's a bit harder to blame the manufacturer. ;)

Cheers

Phil Hobbs

Right thanks. Say speaking of the 1.25 V drop, if I added a small R to the output of the '317 and feed that back to the adj pin I could have it current limit when shorted... I guess I'd need a diode between the adj pin and low voltage side of current sense R.

I'll try just the '317 first.

George H.

This customer had a giant ground loop somehow between two of our boxes connected by an RS485 link. After the first box smoked, they tried several more.

tirsdag den 4. juni 2019 kl. 17.52.15 UTC+2 skrev John Larkin:

Well the LM317 was a fail as a power transistor. The output never went below 1.2 volts even with -5V on the 'base' adjust pin. I plugged in the LM395 and it worked fine. I'm going to try the dual opamp (master/ slave) thing.

George H.

Paralleling op-amps works fine, see:

Back in 1980s I used that very successfully with the oft-mocked LM324.

piglet

Makes no sense. How would it know where ground was, to stay 1.2 above that?

Was your load heater resistor connected, with ground on its other end?

LB-44 (Pease) tingled my mind enough to pull down the long forgotten _Linear Applications Handbook_ from my topmost bookshelf and look inside, where LB-44 was found in all its original glory. That means that there's no need to download yet another PDF this time around.

after opamp offsets piqued my interest. As expected, some of the ANs & LBs in the NS Handbook also talk about offsets. There's an interesting, professionally applied, burnt orange semi-peeled label on the Handbook's spine. It reads "1980" and it's intended to cover up a "1982" on a burnt orange background that was originally printed. Weird.

Thank you, 73,