chopamps

I usually use TI ones--the OPA2188 dual, which runs up to 32V, or the OPA378, which has the best millihertz performance I know about, but is

Cheers

Phil Hobbs

Some of the little devils have current pulses at the chopping frequency coming out of the inputs. Disclosed in the fine print maybe.

--sp

Hopefully that'll be the only surprise. I just plcaed half a dozen AD8629 in a new design that's in turn-key fab right now.

Speaking of which, do you know a good turn-key place, for higher-end boards in the hundreds per year? Companies that take care of PCB fab, stuffing, maybe some testing or even box assembly.

Oh, yeah, there's lots of buried treasure in all of them. You have to get hold of the apps guy to get the straight poop, but they'll usually give it to you. There's a pretty detailed Powerpoint deck about the OPA2188 and its competitors that they'll give you if you ask. I've posted it at

Cheers

Phil Hobbs

Great reference, Phil! Thanks for posting it! ...Jim Thompson

On Sat, 25 Jul 2015 14:42:46 -0700, Jim Thompson Gave us:

I thought the paint can test set-ups were very cool.

On Sat, 25 Jul 2015 16:30:28 -0400, Phil Hobbs Gave us:

Better not let JL see you write that. Him and krw will accuse you of being "a fetishist", whatever that is. (and that *is* the straight poop).

The poop is the roof of the captain's cabin. Perfectly proper word.

And here I thought it was called that on account of all the old Sears Catalogues there. ;)

Cheers

Phil Hobbs

Hard to open, though. I have a bin full of 70-mm film cans made of nickel plated steel that I use for that. Got them from Surplus Shed a few years back.

Cheers

Phil Hobbs (not the film director)

Nice- explicitly (page 13) shows enormous 2.1uA spikes that make up part of the '160pA' average input bias current.

--sp

Danish butter cookie cans are great for building low-level things in. The problem is, you have to get rid of the cookies first.

Like the one about the guy who drowned in a stream whose average depth was 2 inches.

Cheers

Phil Hobbs

Yup!

Then comes to toughest of all problems, you have to get rid of the extra pounds caused by those cookies.

On Sun, 26 Jul 2015 07:45:30 -0700, John Larkin Gave us:

Tell that to your wardrobe evolution.

Paint cans are thicker, and can usually be wrapped with mu metal easier.

On Sun, 26 Jul 2015 02:21:38 -0700 (PDT), Phil Hobbs Gave us:

Precondition them by nailing them shut empty, then opening them up. Repeat until they are a bit looser. Even better to do it with a copper mesh in the interface. Eventually you will get a slight press fit just near free fit, but not the dang'd "requires several screwdrivers" the new jobs exhibit.

Many times there is a coating on the lid too. What is around the sealing interface lip needs to get abraded off anyway to make a good faraday cage, as it were.

The serious problem is running out of coffee.

Thin mu metal doesn't shield mag fields very well, and the wrap doesn't help all axes equally. I don't seem to have problems with 60 Hz mag field pickup; luckily, most test gear now has switching supplies, and old-fashioned fluorescent ballasts are mostly gone, so there are no big 60 Hz mag fields around. Most pickup is electrostatic.

Our problem is powerline noise in the hundreds of KHz range, nasty bursts from motor controllers probably, and RF from Sutro Tower. The butter cookie cans are good at shielding those.

We are practically in the shadow of this monster:

The RF fields are plenty high enough to cause rectification DC offsets in opamps.

Yup. It saturates really easily, and if you bend it, you have to re-anneal to get the mu back. Even dropping it on the floor has a big effect on the mu. Putting a mu metal shield _inside_ the steel can can be a win--the steel knocks down B so that the mu metal can do its job.

Cheers

Phil Hobbs

What about cutting with shears? Does that mess up its properties?