JFET in LN, Second post.

My first post must of got lost? I didn't see it.

Has any one dipped a basic N channel JFET biased circuit in Ln and does it still function or does it completely go off in left field?

I would do this myself however, it seems that for what ever reason, every time I check to see if I can use that part of the lab, it is always out of Ln or it's being used! That is, my schedule just does not align with theirs I guess.

I would like to know the behavior of a JFET in a very cold environment or generate a graph starting from 0c down to where it fails.

I am actually looking for characteristic changes before the fail point.

Jamie

You probably want to use a metal-case transistor for this. The plastic cased parts will probably suffer damage to the bonding wires from the cool-down/warm-up cycles.

In nuclear detector preamps, it is common to have the preamp cooled with the detector crystal. Sometimes only the input JFET is cooled, in others the whole preamp is cooled. There are some shifts in parameters.

Jon

Not a huge sample size, but I've not noticed any failures with plastic parts. The temperature change is 220°C, which is comparable to the temperature change for reflow soldering.

Make a little box from foamed PE packing and use some dry ice. If you use acetone for thermal transfer, better line the box with Saran Wrap or the like first, to contain the liquid, as the foam packing will "melt" rather fast.

You'll probably be ok with the FET. But what else is on the board? For example, some kind of capacitors might not like this, and crystals really don't like to be that cold.

Thanks for the reply. I have an experiment to do and it requires very high input Z with low charge at the probe. It also gets very cold in this area where I need this.

I have some T-92 metal cans JFETS to experiment with and we'll see in a couple of days how that works.

Jamie

Just the JFET and three components will be inserted in a copper tube.

case of the fet will be solder to a smaller tube where it will get inserted into a larger copper tube.

remaining circuit will be in a cozy place..

Jamie

Yeah I've dunked many plastic parts into LN2. (TO-93, T0-220's, transistors and R's) I've seen no problems with any of them.

George H.

Make sure that nothing can crack due to an excessive temperature gradient. That could happen if the requirement spec calls for a fast insertion into the stuff, and extraction. This would require some thinking about the thermal masses and construction of the probe.

It will be interesting to know what happens to the connections. Soldered? What type? Welded? Crimped?

--- Posted via news://freenews.netfront.net/ - Complaints to snipped-for-privacy@netfront.net

there is not abrupt temperature differential to worry about here, this area is slow to get there and slow to come back. :)

Btw, we did a test today, all seem to work out find. It's in use now and we'll see how long it last :)

Jamie

Well, the only thing I can say is I used a silver+tin solder that I had to hit with a micro torch. I think this to bond the metal case to the copper tube.

There shouldn't be any thermo stress on the solder joints. AS for the terminal connections, I did a back loop+solder. These are all lead wire parts and I put a Pi crimp in each one.

Jamie

It's not really that bad.. whatever you stick into LN2 forms an insulating layer of bubbles around it until it gets pretty cold. Helium has almost an order of magnitude less heat capacity, so it's even less harsh even though it's 4K rather than 77K.

Best regards, Spehro Pefhany

Yeah-but, unless you've got money to burn and no time, you don't dunk things into liquid Helium.*

I'm using transistors as 77 to 400K temp sensors. So I'd monitor the forward voltage as I dipped 'em in and out. It takes several seconds to come to equilibrium. Next time I have some LN2 I'll save some 'scope shots.

George H.

"The Journey is the reward"

eff.com

How are they connected? B+C and E? How does it compare with the purpose-built parts like the DT670?

You can put things on dip sticks, and it's a lot faster than cooling down a cryostat for quick tests- so what if it burns up $150 worth of He.. if it's a storage dewar it's gone in days anyway.

Best regards, Spehro Pefhany

Yup, Diode connected transistor. Certainly cheaper than Lakeshore. I'm using pnp's in a to-220 pac. (TIP32's) I like the mounting hole! Here's a plot of voltage difference for three different diodes, (all with the same data code.) A and B were the 'extremes' of the several I looked at

So I do a single point calibration (at room temp.) And get something good to better than a degree over the whole range.

They also follow the ideal diode equation fairly well and so you can change the current and get a ~1% measure of the temperature that way too.

I've got another project coming up where I'm going to need several temp sensors. It would be very nice to have something smaller than the TO-220. I'm thinking maybe a surface mount on a little pcb... but I'm worried about how to make the mounting hole... do I want fiberglass in the screw - sensor - sample stack? So I also wonder if I could 'turn down' the To-220 to remove some mass.

George H.

"The Journey is the reward"

eff.com- Hide quoted text -

On a sunny day (Fri, 16 Mar 2012 06:12:35 -0700 (PDT)) it happened George Herold wrote in :

Perhaps make your won thermocouples? I used a M3 nut to have it sink. http://127.0.0.1/panteltje/pic/th_pic/

On a sunny day (Fri, 16 Mar 2012 14:13:24 GMT) it happened Jan Panteltje wrote in :

or for you

What about something in this kind of package?

Do you know at what temperature the TIPs typically stop working? 50K?