opamps in magnetic fields

You can get within an order of magnitude waving around a rare-earth magnet.

It's interesting that a fair number of LT op-amps have magnetic leadframes. This sort of stuff doesn't matter to most people.

Best regards, Spehro Pefhany

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

Probably not, not much ferrous material. I know some products that operate in 1-5T fields with out problems. Like DR panels. I'm sure there's Analog stuff in there too.

Cheers

Most op-amps contain some magnetic materials. The ones with a copper lead frame seem to have a nickel-gold plating on the inside end.

Depending on where the field vector points you can experience loss of beta in bipolar devices... surf back in time to my musings over strange behavior in an alternator regulator chip (1964) that had current flowing in the ground underneath the chip.

And you make Hall effect devices with structures similar to CMOS... current flow in the plane of the chip ;-)

...Jim Thompson

The trace on your analog sampler CRO won't like it?

Tried that; no obvious change.

It would matter to us; it would wreck the field homogenity.

John

Not static, but TL431 shows lower line frequency noise when situated away from 60Hz magnetics. Assumed this was the lead frame.

RL

It matters to us too, although only at a 2nd or 3rd order, but that's bad enough.

Not in the datasheets, and of course they could change packaging at any time, since it's not specified. In an MRI field there might be enough force to rip the op-amp right off the board, which might outweigh any changes in characteristics.

Any vibration is going to cause noise.

greg

if the only problem were the package and its properties, it may be solved with bare die on board, wire bonded to your substrate of choice ( alumina etc.)

Jure

But it won't. ANY current flow is affected by a magnetic field, aimed in the right direction, particularly at the small dimensions found on chip.

Vertical bipolar devices are particularly sensitive to a magnetic field _in_the_plane_ of the die, loss of beta... _personal_ experience.

Not experienced, but I would expect... field perpendicular to a CMOS die would cause deviation in current flow direction... loss of gm?

(Actually used to make Hall Effect devices.)

Remember your school boy right-hand-screw rule ;-)

(And IR will generally go straight thru phenolic packages.)

...Jim Thompson

I've no direct experience with op-amps in magnetic fields, but lots of headaches from field inhomogeneity issues. Hang the opamp from a string and see if you can make it move with your rare earth magnet.

Concerning the V cross B force on the charge carriers and a change in device performance at high fields. Could you simulate this by using the rare earth magnet and upping the device current by a factor of ten?

George Herold

Once the degree is in hand that rule is typically only used for Merlot and Shiraz bottles.

It depends on which finger you use to indicate the Z-axis ;-)

...Jim Thompson

The guys from CERN could tell you a lot about that:

One problem is that an ever so slight vibration will cause a DC field to develop an AC component and that's when trouble can really show up. Depends on how sensitive your stuff is.

If you use electron flow, it's the left-hand rule. ;-)

Know the difference between a connoisseur and a wino?

A: The connoisseur takes the bottle out of the paper sack before he drinks it. ;-P

Cheers! Rich

I recall waving a $1 rare earth magnet* over a bipolar micropower op amp which had a gain of about 5,000 to see if a problem we'd had reported was due to an MRI machine at the customer's location. There was a noticeable signal out of the amp while the magnet moved. The magnet was simply inducing an EMF in tracks; this effect was not (of course) blocked by our electrostatic shielding, ground plane etc. However, the magnet had to be almost touching the PCB to get a 4% full scale deflection; the effect diminished rapidly with distance. Whereas, you are presumably talking of putting an op amp *inside* an MRI machine. I suggest compact circuit layout as the tracks outside the op amp will have a higher surface area than those inside it. If the circuit aim can be achieved in a differential mode rather than single-ended one, and the layout is symmetrical and compact, you may find the offsets are the same in both conductors and can be ignored. Or maybe you can use some optical technique to achieve the aim?

Thanks. That paper mentions 500 gauss; we're talking 150,000 maybe. No relays, I suppose. No transformers or big inductors, no ferrite beads.

Yup, we see that in NMR gradient coils. When you pulse them there's an audible PING and then you see a lot of acoustically-driven voltage across the coil for a while after.

I've asked Analog Devices about their part, a fet opamp I'd like to use. Worst-case, I can build a test board and persuade some contacts that I have to let me dunk it down the bore of an NMR magnet.

Nuisance.

John

And Murphy sez that they'll probably not allow you to use mu-metal and stuff so their field aren't disturbed. Can you work differentially and orient things so that the resulting induced signals cancel?

Sure. OTOH if there never were challenges like this the world would not need guys like us. And then places like Zeitgeist would have to close.