Teensie 3.2 Max Temperature

Status
Not open for further replies.

J-Ro

Member
Hi guys,

I have built a project where I have a teensie 3.2 measuring 2 thermistors, totaling interrupts from a hall effect sensor (measuring a flow meter), and then sends the 2 temps and total counts from the flowmeter over canbus using a TI CAN transceiver.

I included a picture of my setup, the can transceiver is on the bottom of the board otherwise you can see everything.

For power I only have 12VDC available so I am using a 3.3V linear regulator. The whole thing is potted in DP-270 3M Scotch Weld Potting epoxy. I've built 4 of these and they have been running pretty much constantly on my car for several months. Yesterday I had one die on me. By die I mean I stopped receiving data over CAN and error frames existed on the bus. I investigated a little and noticed that if I removed the power and waited a few seconds/minutes and plugged the power back in the teensie would power up and start sending data again for a variable amount of time. It seemed the longer I kept it unplugged, the longer it would stay powered up and send CAN data.

I think I may have gotten the teensie too hot. I don't know the exact temperature the teensie was running at but the fluid being measured by this exact flowmeter was hotter than its ever been (but other 3 not as hot) due to high ambient temperatures (Fluid was 65C) and the whole block of potting compound containing my circuit is potted directly to the brass flowmeter. Additionally the linear regulator is inside the potting compound as well (and it gets really hot...bad design element there I know, plan to use a switching regulator on my next version).

At any rate, between the block of potting compound being heated up to 65C from the fluid and the linear regulator being an additional heat source right next to the teensie in the potting compound I think I might have exceeded the maximum temperature of the teensie...which from searching the forums I believe to be 85C...is that correct?

I replaced the whole power input side of the circuit (regulator and capacitors) by grinding away the potting material and get the same behavior with those components replaced.

MY question is, does this behavior of powering on for a short period and then turning off line up with what might happen if the teensie were heat damaged?
 

Attachments

  • flowmeterBoard.JPG
    flowmeterBoard.JPG
    24 KB · Views: 108
Ill take a crap shot at this but if it was me I would be doing a little more effort on the regulator side. Granted you using a 3.3v regulator but what you need to be careful of is that this regulator becomes very hot and may not provide the appropriate voltage. Plus your feeding whatever it steps down directly into the teensy MKL chip and MK20.. chip. Over time if they do not receive the appropriate clean 3.3v your going to instantly or gradually destroy the chip(s) On Teensy you have a 5v to 3.3 regulator that is worth looking at. I would get a voltage circuit that can tolerate automotive voltages (including voltages that are higher coming from the alternator) and cleanly step them down to 5v for the teensy to step down again to its 3.3v (double filtering and clean voltage reduction).
 
The whole thing is potted in DP-270 3M Scotch Weld Potting epoxy.

Rigid epoxy potting can cause reliability problems when used in contact with surface mount parts, like the ones on Teensy. As the expoxy hardens its volume shrinks slightly, which can apply shearing force to the parts. Likewise when the whole thing heats and cools. The thermal expansion of the epoxy and the electronic parts isn't well matched, which can apply shearing force to the parts relative to the PCB.

There are potting materials meant for SMT parts. Here's one of the many pages Google found when I tried a quick search just now.

http://www.efipolymers.com/applications/electrical-potting/surface-mount-technology/

Usually potting is done in 2 steps, where you first fill just above the electronic part with the soft rubber-like potting material and let it dry. Then you fill the rest with the hard epoxy.


Also, must agree with crees, a linear regulator from 12V to 3.3V is going to dissipate a *lot* of heat. Not good it surrounded by hard epoxy that's a thermal insulator. Maybe using Duff's Snooze library can help you get the current consumption down, if you have more that are already potted with this regulator inside (but USB is still accessible for a code update).

Probably best to use a more efficient regulator, or plan on heatsinking to get that heat coupled to the outside.
 
Status
Not open for further replies.
Back
Top