1. Home
  2. Electronics Design
  3. PCB heater

PCB heater

I was thinking that I could thermally isolate a small patch of a PCB ground plane, about one square inch, and add some resistors as heaters

formatting link

to make sort of a local oven for the parts above the patch, and especially for the FR4 capacitance. Layer 2 is our usual ground plane, so the resistor grounds would terminate there.

Thermal and electrical conductivity go together. For copper, it's about 150,000 K/w per ohm, so a few milliohms of grounding spokes (the heated patch still needs to be grounded) could have a healthy thermal resistance to the main board ground plane. I could always add capacitors across the plane-cut gap to get better AC grounding without much added heat loss.

The control could be PWM or simple bang-bang. Since I'll have other temp sensors elsewhere on the board, a little feed-forward tweak could correct for imperfect local control.

The alternative would be to use SOT89 transistors or mosfets as the heaters and go all linear. That's safer - less potential noise - but not as elegant. Maybe use voltage regulators as the heaters? LM317s in SOT89?

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin
Loading thread data ...

I'm not sure how well a ground plane based heater will work as an oven without insulation around the device, but that aside...

I was thinking about the noise issue with a pulsed heater supply. If the circuit was driven by a half H pair alternately switching between power and ground and driving a small LC, the current fluctuations in the load resistors will be much less. The only source of noise is the current change in the switches and the loop they are in.

But then this is no different than a typical switching power regulator. Why not use one of those? The noise into the load can be minimal if the supply is designed with short current loops. Whatever is controlling the loop can force a voltage into the feedback pin of the regulator to control the voltage output. The usual reason to limit how low the output can go is the voltage required on the FB pin. Since you are driving that input you can probably get the output down to zero.

If a switcher is not so good for a variety of reasons, the same thing can be done with a linear regulator. Don't know what you are using to sense the temperature and/or its impact on the ovenized circuit, but if it is something like an MCU or an FPGA this should be simple enough.

--

Rick C
Reply to
rickman

I was thinking that if it's an issue he could use those little RF isolation cans filled with a suitable foam. I would think seriously about milling a moat around the temperature-controlled patch, with just enough bridges to make sure the board didn't break in service.

If you're thinking "why not use those" then you may as well use a pass transistor, with or without a resistor. The reason to use a switcher is to minimize power loss. A heater is turning ALL its power into heat -- so as long as your pass transistor is part of the heater, there's no need for a switcher.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com 

I'm looking for work -- see my website!
Reply to
Tim Wescott

I just got a sample kit from Leader Tech. It has a bunch of tiny surface-mount clips and a couple of pieces of grooved metal that can be folded into boxes. So one can make custom shield box prototypes.

I have found that, with a shallow shield can over a PC board, foam makes heat loss worse. There's not much convection in this situation, so the best (and cheapest) insulator is air.

Fets or transistors or voltage regs would make nice linear heaters. SOT89s aren't a lot bigger than 1206s with thermal vias.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

I think routing the PCB is a bit of overkill. Once you cut the copper, I think the rest of the circuit board is a pretty minimal conduction path.

Yeah, I guess if you are doing your own regulation you might as well use the transistor to do the heating. I was thinking it would be simpler to use a regulator chip, but you still have to condition the temperature signal. But if that is in the form of a thermistor and you want to drive it to a voltage that is above the FB voltage of the regulator, Bob's your uncle, you are done.

--

Rick C
Reply to
rickman

Den onsdag den 28. september 2016 kl. 18.12.19 UTC+2 skrev John Larkin:

linear regulator with enable, constant current grounded output and you get it all in one part and a resistor, the overtemp protection comes free

-Lasse

Reply to
Lasse Langwadt Christensen

I did something simple and inelegant using a resistor and a MOSFET, with a resistor/thermistor ladder tied to the MOSFET gate, and the resistor value chosen so that the MOSFET was conducting a suitable amount of current at the desired "hold" temperature.

In this application (an "outer" oven heater for a GPS-disciplined OCXO) I was more concerned about stabilizing the temperature in the case, and not all that concerned about the specific temperature (as long as it was consistently a few degrees above the warmest ambient temperature in the room),

Reply to
Dave Platt

How about a nice thermally-protected FET, such a BSP75N?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.