78xx tolerance Q

It very much depends on the application. For testing op-amp circuits it probably doesn't make much difference at all.

As Tim Wescott says, it probably isn't going to matter in your application.

The fact that the bunch of parts you bought all test out at 8.79V to 8.80V = probably just means that they came from the same batch.

Many years ago (1980s) I'd had to clean up the design of a National Semicon= ductor LM305-based voltage regulator, which ended up needing a select-on-te= st resistor to get the current limit to limit at the right current.

For the next couple of years, production would drop by every six months or = so to complain that value that they'd been selecting for the past six mont= hs didn't work any more. Purchasing was buying roughly six month's worth of= LM305s at a time (about 100, which got them the 100-up price) and each tra= nche would come from a single production batch, so when that tranche ran ou= t production would have to adapt to the new batch.

After a few years, all the board test technicians had been told the story, = and I didn't have to tell it any more.

--=20 Bill Sloman, Nijmegen

ply

Positive and negative supplies don't have to be the same volts?

n.

As I hope.

Powering homemade effects pedals hooked up to homemade musical instruments. In the finest tradition of circuit-bending, imperfections aren't flaws, they're character.

V probably just means that they came from the same batch.

onductor LM305-based voltage regulator, which ended up needing a select-on-= test resistor to get the current limit to limit at the right current.

r so to complain that value that they'd been selecting for the past six =A0= months didn't work any more. Purchasing was buying roughly six month's wort= h of LM305s at a time (about 100, which got them the 100-up price) and each= tranche would come from a single production batch, so when that tranche ra= n out production would have to adapt to the new batch.

, and I didn't have to tell it any more.

upply

it

Certainly not. You can run (say) -5V and +25V. Of course the output will clip at different voltages. But it's useful if you have a signal that's mostly of one polarity.

George H.

That's unusual (though, with eBay anything is possible). Are they brand name parts? Are you sure your meter is accurate?

The two ones I see for sale now are Fairchild and ST, and I'd expect them to be within around 1% typically.

Nope. Op-amps are remarkably tolerant of different supplies. In fact, for a lot of uses they don't even have to be regulated (although supply variations do bleed into the output signal -- look for power supply rejection ratio (PSRR) in your data sheet to get an idea of how much).

The biggest impact that supplies have on the op-amp is the highest and lowest voltages that the op-amp can drive to. The second biggest (if you're doing a DC circuit) is that the bias voltage is generally specified when the inputs are riding at the point halfway between + and - supplies. The third biggest is the PSRR, unless you're mostly concerned with making a Really Quiet circuit, in which case it may be the biggest thing that you care about.

Unless someone's improved things while I was napping, typical data sheet values are +/- 5% for 'vin ordinare', 2.5% for slightly premium parts.

This being eBay, you don't know if you're getting on-spec parts, factory rejects that went out in someone's pocket, or some totally different part that's been relabeled.

I must be getting grumpy in my old age, but stuff like this is so cheap from reputable distributors I see no reason to muck around with mystery parts.

That 'vin ordinare' will give you a headache ;-) ...Jim Thompson

Sure, but 'typical' parts are usually much better than the worst-case tolerance. Actually, I just a couple minutes ago asked for the 5V rail (78M05CDT) to be checked on a new board that a fellow is bringing up and it was 4.99V (0.2%). An error in the region of 4% is something I've never observed personally on that class of part, and if I ever saw it I would probably trash the lot. Field failures are too expensive in $ and reputation to risk.

do you expect them to be that good across the whole temperature and current range?

Certainly not, but for low currents they're pretty good. Typical drift is stated to be around -100ppm/K, really not bad for a cheap part.

Best regards, Spehro Pefhany

Yup. And you being you, you don't just sell them on eBay. But other folks do, and that's why I buy from DigiKey.

No. Notice that op amps don't have a ground connection, only a positive and negative rail.

The 7909s I just received test even closer, 8.98 - 9.02 volts.

Mise en bouteille dans notre garage? :-)

Not in the garage. I worked in the garage with all bays open today... about 3 hours was all I could take... going inside was like entering a refrigerator :-) ...Jim Thompson

Done job. Output checked out at -8.74 / +8.79 volts. Perfect enough.