Small heaters

How big is the test mass? And what sort of time constant can you tolerate on the sensor?

You may be able to do it with a small thin 2-sided PCB. Use surfmount resistors or fets as heaters, with thermal vias to copper pours on the back side, in contact with your mass. A surfmount RTD, ditto.

This is an OCXO heated by mosfets:

The Minco heaters are good, very low mass. They have some standard products. This gadget is heated by a Minco, with a thermistor on the PCB for feedback.

--

John Larkin         Highland Technology, Inc 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

One problem with duals would be the thermal coupling between diodes in the same package. I'd guess that you'd want the sensor far from the heater, maybe even on opposite sides of your test mass.

Measuring specific heat or thermal diffusion as a function of temp?

--

John Larkin         Highland Technology, Inc 

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com

You could try a small peltier cell. As a bonus when you are not pulsing it, it provides a voltage proportional to the temperature difference of the two sides. Depending on your set-up and requirements that might be accurate enough to be your temperature sensor too.

Something from this family: NJL0281D probably too big

If it's tightly coupled to a significant test mass I'd think the heatsinking would make it safe.

Having seen your application described in the other thread, it's an interesting problem.

How are you going to make sure all the heat you make gets into the sample, accurately and consistently? (E.g., as opposed to being lost down the leads.)

And what happens if accidentally run without a sample. 1J heats

10ug of Cu ~260K--that's fine for copper, but potentially pretty rough on a BJT.

You're right, to use an RTD you'd have to measure resistance change during the pulse. But, OTOH, constant current drive and two simple Kelvin voltage measurements lets you calculate total heater power very well, no calibration required.

I don't know if an ultra-low mass heater or a calibrated mass heater works better. A few parameters and calculations would tell.

(The quasi-humorously mentioned optical methods have some appeal too-- no mass, no contact.)

Cheers, James Arthur

for that topic

So I'm thinking it's about 1 cm^3. But it'd like it to be able to handle all sorts of different samples. (I was looking at 3 cm^3 of copper the oth er day. 1 watt was fine. We want to do powder samples but don't have a co ntainer yet. There are these little scanning differential calorimeter aluminum dishes bu t they are too small for me.

Yeah I'm certainly figuring that the transistor temp sensor will be on double sided thin pcb as you say... for another part of the project I need to make these thermal shorts for all the wires. Again I'm hoping* to use p cb. (*Well I did some calculations, but when you do the measurement you alw ays seem to learn something new.)

Yeah, low mass is good! The thing I hate about the minco heaters is the kapton. I was hoping to solder things together. Fiberglass as the only insulator. (oh here's an idea, how about a pcb with thermally conductive coating.) With kapton, I've got to squeeze it or glue it.

So are mosfets better than bjt's as heaters? (I guess there is less base current error.)

Hey did you know the volume heat capacity of things (solids) is all about t he same.

2.5 to 3.5 J/(K*cm^3) (at 300K) (well except for nylon, and other wimpy plastics.)

Material Heat capacity (J/(K*cm^3)) at 300K.

Aluminum 2.42 Copper 3.4 Iron 3.5 Al2O3 3.0 G-10 2.7 Nylon 1.7

George H.

Hard to measure the energy with accuracy though.

Grin.. Well heat capacity and chemistry might be a lot of fun! (I'm not a chemist.)

Interesting, Thanks. I think I've given up on the idea of using one thing for heater and sensor for this project. (But I'm filing 'em all away for next time.)

George H.

Grin.. nice. I was thinking I couldn't fit it into 1/2 in square with 5 mil traces.

1/2 inch is the size of the smallest heater from minco. It's got a hole in the middle to fit a 2-56 screw, which is perfect... except for the kapton.

George H.

Grin.. nice. I was thinking I couldn't fit it into 1/2 in square with 5 mil traces.

1/2 inch is the size of the smallest heater from minco. It's got a hole in the middle to fit a 2-56 screw, which is perfect... except for the kapton.

George H.

The Mincos come with sticky stuff on one side.

Shouldn't matter much.

Something to do with physics?

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation

Anodized Aluminum ribbon?

You'd want to have two black objects in the beam, then; keep one as a control, and experiment with the other. Hey, it works on my infrared-sensing two-slice toaster!

If the heater's separate, the question is how to build it, then how to thermally couple it, the test sample, and the temp sensor.

Smd resistor(s) on an aluminum pcb as heater? Or nichrome clamped between thin slabs of anodized aluminum?

Cheers, James Arthur

Try 50ohms RF termination resistor on ceramic substrate. Or Tyco MPC series. Or can you use SMD resistors like Welwyn PWC series ?

-- mikko OH2HVJ

On a sunny day (Thu, 13 Feb 2014 10:29:10 -0800 (PST)) it happened George Herold wrote in :

See the circuit diagram of the 'black box', it is quite readable in this case.

I used PWM from a PIC, and filtered that, and divided the obtained DC level so it was in the analog driving range of the MOSFET. The software is a PI loop, no 'D' part, as no fast changes will happen as the thing is isolated in a double box from the environment. the 'I' (integral) part increments PWM or decrements PWM once per unit of time (second? cannot remember). The loop stabilizes in a few minutes (could be minutes then duh, look a the soure code), with little overshoot. Been running now for a year an a half, look at the data on the side; it is within about 2/700 , flippers between a few ADC steps, say fraction of a degree C.

On a sunny day (Thu, 13 Feb 2014 12:38:14 -0800 (PST)) it happened George Herold wrote in :

Reverse diode in MOSFET?

Never heard of it! This stuff?

Is there some advantage over coated resistance wire from MWS? (I guess better thermal contact.)

George H.

Grin.. yup that's the question. And low mass.. or volume!

All the clamping and slabs sounds like more mass.

1/2" 2-56 brass screws are about 0.36 g's each. At the moment I'm picturing this little clamp, like a machinist parallel clamp. one of these but much smaller,

that will be used to grab the sample. Well and only one side will be stiff. The other side will be thin and springy. For a while I was thinking of using 0-80 screws. But that just seemed too wimpy.

The one thing nice about a trnasistor as heater is that I can solder the collector right to the metal surface I care about.. OK maybe to a thin pcb with copper. but still a fairly direct (all copper) thermal path.

George H.

Hey I can read that! :^) So you just controlled the FET current, that's what I was thinking of doing.

Thanks, George H.