Diode connected transistor as temp sensor

I don't know anything about the cryo behavior of an AD590. I told you what I do know, in response to your request ...

"I was hoping for for some more detail on how thses parts go bad when operated below the minimum specified temperature"

Why my information set you off I can't imagine.

You could test the AD590 and report its behavior to us. I can imagine a nice little business as a spinoff.

But you won't.

John

the

voltage

e in the

ure.

utliers.

le point

curacy

sted.

test.

egrated

e AD590

your

ial

es

y not

ill

You didn't tell me what you know, but what you guessed might happen - it strikes me as good guess, but it doesn't represent practical knowledge.

You'd have think about the implications of your post to manage that.

Your imagination may be fertile, but it is also a little profligate. My imagination tells me that the AD590 probably dies permanently when submerged in liquid nitrogen, which would be a poor basis for any kind of business. It's much cheaper to ask if anybody has done it than its is to find a dewar flask, sone liquid nitrogen and an AD590.

Correct. The odds of making money on the deal aren't good. And we move house on the 24th March, so I'm not going much free time to expend on wild goose chases.

-- Bill Sloman, Nijmege

e:

as a

th the

rd voltage

one in the

ature.

outliers.

ard

ngle point

accuracy

rested.

ck test.

ntegrated

ice AD590

g your

ecial

akes

o 77K.

nly not

and

will

st

od

nt-

nt

Hey, the uncalibrated AD590's aren't too expensive. I'll add one to an order and check it out. I'll bet a beer it'll work, and survive ten dunkings (I think I saw an AJP article that used it.)

I was interested in what Phil A. said about the different mountings of the silicon and lead connections. (Though, I've pissed him off somehow.) I tortured both to-92's (2n3904/2n4401) and the to-220 pac. I figured the higher power ones would do a better job of thermal coupling to the silicon and also be more robust against thermal expansion/contraction.

Any thouhgts?

George H.

IIRC thermal expansion tends to level off around 100K, so going to 500 K would probably be much tougher on it than 77K. Of course the metal will get brittle at low temperatures.

Cheers

Phil Hobbs

voltage

the

outliers.

point

accuracy

test.

integrated

AD590

Of course it's practical. I know for a fact that the output is 10 mV/K, and I know that there's a silicon PN-junction substrate diode that conducts when the output goes a few hundred mV below ground, and that the chip latches up if it conducts currents on the order of 100 uA.

You don't have to be hit by a truck to know that it would hurt.

John

the

voltage

in the

temperature.

outliers.

point

accuracy

interested.

test.

integrated

AD590

your

not

It'll liklely survive and may well work to 70K. Bipolars quit working around 20K. Jfets and gaasfets work all the way down.

John

voltage

the

outliers.

point

integrated

AD590

Not much of anything, even if you shock things by dipping it right into the LN2 and letting it boil away. I doubt it's any worse than reflow soldering, possibly gentler. LN2 is probably harsher than He because the heat capacity is a lot greater.

Best regards, Spehro Pefhany

No, but!, normally when you get hit by a truck you don't feel anything for a while and sometimes, never again!

Jamie

I am trying to visualize the effects this temperature would have on the body encasement of the component..

Jamie

voltage

the

outliers.

point

integrated

AD590

...and any water absorbed.

voltage

the

outliers.

point

integrated

AD590

I remember dunking a banana lead in LN2, in college physics class. After I pulled it out the plastic on the end shattered (rather loudly) as the metal warmed. The LA wasn't impressed.

voltage

in the

outliers.

point

accuracy

interested.

test.

integrated

AD590

your

not

will

If you freeze a real banana, you can drive a nail with it. (Okay, a sufficiently broad nail and a sufficiently soft board. Stil.)

I used to monogram screwdrivers by freezing the handles in LN2. The polycarbonate crazes in artistic patterns, but is still strong enough to use with no worries. (Sort of like what happens when you fry marbles.)

Cheers

Phil Hobs

Cheers

Phil Hobbs

te:

as a

ith the

ard voltage

one in the

rature.

e outliers.

ward

ingle point

accuracy

erested.

ick test.

integrated

vice AD590

ng your

pecial

makes

to 77K.

inly not

and

t will

ost

ood

en

ent-

int

The assumption embedded in your extrapolation is that something else wouldn't screw up the LM35 at a slightly higher temperature.

-- Bill Sloman, Nijmegen

BTW, a quick google finds a few people have reported trying it, with no hard failures reported. Nobody says it doesn't work, (for some definition of work at 77K), but one report says that it won't start up at 77K but will function if cooled while powered. BJT gain doesn't hold up well at low temperatures - it's probably single digit at 77K.

BTW, I don't think dipping in LN2 is as bad thermally as (say) dipping in molten solder, despite similar temperature differences. The LN2 boils and forms an insulating layer for a while, like a ball of spit on a wood stove. One could (and I'm not suggesting actually trying it, hear me) _briefly_ stick one's finger into LN2 and it will come out pretty much unscathed (a second too long, and painful frostbite appears, then tissue death etc).

Best regards, Spehro Pefhany

te:

the

voltage

e in the

ure.

utliers.

le point

curacy

sted.

test.

tegrated

ce AD590

your

cial

kes

ly not

and

will

ost

ood

en

ent-

int

Yeah that is the 'lore' in low temperature stuff. If the differential contraction doesn't cause something to fail at 77K it's most likely good to any lower temperature. Anyone know how the thermal expansion scales with temperature? I would have assumed linear, but maybe not at low temperatures.

George H.

t- Hide quoted text -

e:

s a

h the

d voltage

ne in the

ture.

outliers.

rd

gle point

ccuracy

ested.

k test.

ntegrated

ice AD590

g your

ecial

akes

o 77K.

nly not

and

t will

post

good

hen

tent-

oint

"The Journey is the reward"

eff.com- Hide quoted text -

Yeah I shock them, by direct dunking. In the application they are bolted to a slab of copper and so can't cool as fast.

George H.

voltage

the

outliers.

point

integrated

AD590

The ionization energy of the dopants isn't quite zero, so they do freeze out. IIRC JFET amps often won't start at 4K, although self-heating is often enough to keep them working there if they're started at some higher temperature.

Cheers

Phil Hobbs

Sure just like water drops skittering around on a hot frying pan. I stick my hand in and pull it out, no problem.

George H.

"The Journey is the reward"

eff.com- Hide quoted text -

Sphero, Can you tell me what didn't 'start up'? Was this a simple transistor or some IC?

George H.

"The Journey is the reward"

eff.com- Hide quoted text -