Heater needed

IR floodlight? (Too big, power-hungry?) Heated blanket? (Too wide, not hot enough?)

I suppose a mess of resistors would work, too, but then you get a bunch of nearly infinitessimal heat sources, necessitating a whole wad of resistors for some arbitrarily even heat.

Tsk tsk, bad form.

Tim

-- Deep Fryer: a very philosophical monk. Website:

There are resistive heating foils used for car mirrors which are even self regulating. you can get them in many shapes and for many preset temperatures. Check out google.

Minco and others make stick-on flexprint heaters, but they're not cheap.

I make some assemblies that are a thin pcb bonded to an aluminum plate, with a Minco heater (or some power fets scattered around) as the heater, closed-loop temp control based on a thermistor on the pcb. Temperature is much more uniform than a simple heated board, because the thermal conductivity is so much better.

Or why not just a serpentine trace? Figure 1 oz copper at about 500 uohms/square, half-ounce twice that. So each inch of 5 mil trace, 1/2 oz copper, is about 0.2 ohms. Zigzag that all over one layer of the board and apply amps.

Or apply a huge RF voltage between planes, and let the FR4 dielectric loss heat up the board. (Just kidding.)

John

In what sort of enclosure? To what temperature? How uniform?

I'm thinking of something along the lines of a convection oven. A heat source and a fan to circulate the hot air uniformly.

Using a radiation source or contact heat sources might not be as uniform as you'd need.

Agreed. The OP hasn't given any clue as to whether this is a temporary (for testing) or permanent arrangement.

If the former it could be as simple as placing the item in a closed (partially or fully) box and piping hot air into it using a hot air gun. Thats what I do.

Electric iron: Even heat, temperature adjustable, steam if needed!

Luhan

Well, that's what I get for being terse with my question. Ok, here are more relevant details:

Basically, what I'm trying to do is to prevent an exposed PCB (to the air) from getting wet from condensation or dew. The idea is to attach something to the back of the PCB and heat the whole thing up. This is a permanent design. The unit is about a 4 in cube.

The problem is almost like the rear window defroster but in a much smaller space. Does anyone know what those heat traces made out of?

Thanks

More people should realize this....

Ok, I'm still not getting a picture here, my exposed PCB's get put in developer and then in etchant. Also, here in Arizona, we dont get much 'dew'. So you live in the midwest and leave PCB's out over night?

Luhan

generate heat for

The obvious way is to scatter a

and run the current through

like a sheet resistor? or

the back of th board?

to

temperature? How uniform?

convection oven. A heat

uniform

you'd need.

Agreed. The OP hasn't given any clue as to whether this is a temporary (for testing) or permanent arrangement.

If the former it could be as simple as placing the item in a closed (partially or fully) box and piping hot air into it using a hot air gun. Thats what I do.

It's possible you don't need more than one resistor to get fairly even temperature, particularly if the board is in an enclosure and you have a ground plane or copper pour on the board. PCBs are fairly thermally conductive. You could also use (a) PTC thermistor(s) to get a kind of temperature control without additional components.

Best regards, Spehro Pefhany

are more

the air)

something

permanent

The side mirrors for cars have a flexible film with resistive tracks for heating.

The easiest prabably is - as another poster proposed - to make a layer out of thin track and use that as a heater element.

There are different materials. I suppose you want a continous voltage applied, but that will vary temp. whole bunch with different air temps., exposure. If you use a regulator, will switching affect circuit. An analog controller may be necessary. Seems like the back could be a trace of your own? greg

Morning JK, we are a heater manufacturing and distribution co. There are plenty of heater options for your application, some not so expensive in quantities. Since you are only trying to eliminate condensation you need to keep the board at some temp above the condensation point, so 40- to 50 F should suffice. There several ways to do this, warm air, Foil trace Kapton, polyester, or rubber heaters with PSA for easy installation, small T3 Lamp. If you contact me with the size and shape of you board, [photos would be great] the weight / mass that need to be heated we can calculate the watts needed and get a heater designed and quoted for you voltage.

Look forward to hearing from you soon

Mike Zelisko snipped-for-privacy@convectrinics.com

sorry fat fingered the web address snipped-for-privacy@convectronics.com

controller

Even more "lame"; a solid layer of copper and a transistor running in linear mode dumping heat into it! Temperature controller is needed.

I would just find a PTC resistor mounted on the foil.

greg

We did something similiar to this. Bascially a temperature controlled oven to drive moisture out of the PCB. For a 4" to 5" rectangle (3" to 4" high) insulated with 0.5" styrofoam, the temperature gradient could be about 10C peak-to-peak. We used two auto light bulbs and PWM them with an AVR. We set the target temperature to 45C and blink it slower as it reach the target. The blinking slow down in a few hours and keep constant for 8 days. The PCB was reflowed fine, no popcorns on the BGA.

If enough people is interested in this, we could make it a product.

The scattering of surface-mount resistors sure sounds sensible. That way, it's easy to experiment with the proper amount of heat.

The big old Hoffman industrial-controller enclosures sometimes have internal heaters to keep the damp off the gear inside.

John

This is the first time I've heard of a 4" thick PCB. ;-) Anyway, if all you have to do is keep it above the dew point, just design your circuit to be inefficient, and keep itself warm. ;-)

Good Luck! Rich