What's an effective way to heat an enclosure in a temperature sensitive project.

[m12lrpv]

I'm working on a project that uses an inductive sensor and it's quite temperature sensitive so i'm getting a fair bit of drift as the ambient temperature changes.
The sensitivity is greater than I can measure and mathematically compensate for with a regular linear temperature sensor (MCP9701A) or even with a thermistor despite the thermistor being 3 times better I still need greater resolution. Part of the limitation is the 10bit ADC I'm using to measure it and I'm not really in a position to change that because it's sitting just above the noise floor.

I'm thinking that what I need is a way to warm the enclosure 110x190x70 to a configurable temperature and hold it there.

With that in mind I thought I'd try 150mA through a power resistor with it heating up together with the heating of the LM7805CD2T having to drop down from the 12v input. Obviously I didn't try and calculate what would be suitable because this wasn't. It warmed the enclosure faster but couldn't hold the temperature when the ambient started to drop.

I know that I can change the components to get more heat generated using more current but I'm really wondering if I'm approaching it wrong so I want to get some ideas about what methods there are to manage temperature sensitive devices without controlling room temperature.

Searching for oven circuits tends to return results on kitchen appliances 

[Alex Eisenhut]

https://www.hammfg.com/part/shgm1410

?

[DanielS]

If you have trouble holding temperature in your ovenized enclosure when ambient goes down and want to avoid cranking heater power way up, you could add thermal insulation wherever possible.

[coppice]

I'm working on a project that uses an inductive sensor and it's quite temperature sensitive so i'm getting a fair bit of drift as the ambient temperature changes.
The sensitivity is greater than I can measure and mathematically compensate for with a regular linear temperature sensor (MCP9701A) or even with a thermistor despite the thermistor being 3 times better I still need greater resolution. Part of the limitation is the 10bit ADC I'm using to measure it and I'm not really in a position to change that because it's sitting just above the noise floor.

I'm thinking that what I need is a way to warm the enclosure 110x190x70 to a configurable temperature and hold it there.

With that in mind I thought I'd try 150mA through a power resistor with it heating up together with the heating of the LM7805CD2T having to drop down from the 12v input. Obviously I didn't try and calculate what would be suitable because this wasn't. It warmed the enclosure faster but couldn't hold the temperature when the ambient started to drop.

I know that I can change the components to get more heat generated using more current but I'm really wondering if I'm approaching it wrong so I want to get some ideas about what methods there are to manage temperature sensitive devices without controlling room temperature.

Searching for oven circuits tends to return results on kitchen appliances 

You need to insulate. Its too hard to regulate the temperature unless you have enough insulation to ensure it only changes slowly. I think most ovenised oscillators now use electronic control, but they used to use simple bimetal strip thermostats. This worked well because they designed the strip to have a narrow hysteresis band, and the insulation insured the rate of change of temperature was low. You might was to use three or 4 resistors spread through the ovenised chamber for more even heating.

[m12lrpv]

Insulation!!!     Duh how obvious. Thanks for the tip. I don't have any suitable foam (all too thick) but I can easily add wool felt.

[babysitter]

If you tell us about the drifting part we can give you maybe advice to get rid of (part of) the drift in a green, eco-friendly way  Maybe you extend the inductive sensor with a capacitor to a resonant circuit and selected a sensitive dielectricum, we can only guess?

The stuff inside sleeping bags is a nice, room-filling insulator!

[m12lrpv]

If you tell us about the drifting part we can give you maybe advice to get rid of (part of) the drift in a green, eco-friendly way  Maybe you extend the inductive sensor with a capacitor to a resonant circuit and selected a sensitive dielectricum, we can only guess?

The stuff inside sleeping bags is a nice, room-filling insulator!

Sensor and reference coils are ferrite core based. Can't change that.

Caps in the resonant circuit are regular poly mkt's as I read somewhere that ceramics weren't ideal with respect to temperature. I did read about people making ultra stable oscillators measuring the tempco's of certain caps and trying to select them to counter the expected drift from the other components but honestly that's a bit beyond me and my project as I'd need oil baths and a lot more spare time than I presently have. I do need the caps cheap enough that I can sort through them to mix and match within the value tolerances to make up an "ideal" value match to balance the reference and sensor coils. It's an awkward situation with hand wound sensors but I can live with that.

I used low tempco 0.1%/25ppm resistors in the initial stages and in some of the gain stages to keep down drift but it was too expensive to use them everywhere.

You've got me thinking though that the source of the drift could be anywhere (or more likely everywhere) in my circuit including the crystal oscillator that's driving the sensor.

[acbern]

If you do not want to invest much time to get that solved by doing an own regulator design, you could use a Wavelength Electronics temperature regulator (you find them used on ebay). they can do very precise regulations to tenths of C, you need to be sure though that you have done the case thermals right (isothermal enclosure, thermally isolated, varying internal sources of heat connected to the heatsink so thaty have no internal neg. influence and can be compensated through the regulator...)

[babysitter]

You've got me thinking though that the source of the drift could be anywhere (or more likely everywhere) in my circuit including the crystal oscillator that's driving the sensor.

Use a straw or tubing and a hot air soldering station at the lowest setting, maybe even lower than that  to blow slightly warm air on single parts. That might give you a clue about the part which causes the biggest fraction. Could imagine even the ferrites being the culprit, sadly.

Consider adding a part with a tempco of different sign to the culprit.

If you want to temperature - stabilize just the crystal (or another singled-out part), consider Kuhne electronics QH40A or QH60A, crystal heaters for 12 euro. Preferrably specify the crystals for thermostat operation at the respective ~40°C or ~60°C setpoint.

And please report!