I have a sealed enclosure 6 in x 6 in x 4 in to be kept outdoors in temperatures as high as 110 F and all components except ATSAMD51 cpu are low power; I had a similar case with a raspberry pi zero 2 w and at 110 F it overheated and shut off; The raspberry pi has the same ambient temperature specs as the ATSAMD51 (-40 C to 85C = 185 F); so the questions are:
(1) how can I tell if the arduino will overheat without waiting all year for a heat wave;
(2) could it be some of the other components overheated instead?
(3) I got the raspberry pi project working stably at 110 F by ventilating the case with a fan but it is not as waterproof (vent is open but should not let in water); anybody have a good tiny vent?
Have you tried putting the inlet in the bottom which has a shroud around it? IE the bottom is recessed from the side by about 1/2 inch.
https://www.amazon.com/gp/product/B07RNM73P3/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
This is the junction box (enclosure). Not sure what you mean; drill a hole in the bottom ???? Sounds like a possibility; The junction box has feet which raise it off the ground and inside it has a grid for the circuit board which is 1cm above the floor so any water would come out the bottom; but the question is how would I know if it works based on the specs and others' experience before waiting for a heat wave; the same circuit worked with adafruit metro minis (arduino uno similar) in 110 degree weather; is this a similar wattage? I will measure it asap;
could it be the adafruit itsy bitsy ATSAMD51 uses only 10mA at 3.3V which is less than the adafruit metro mini (similar to arduino uno) ... ... and the metro mini worked so I should be okay ???
Thermal design suggests heat generated inside the enclosure must exit somewhere.
#1 might be a little overkill as the board is very low power. However it there is a 5V to 3.3V regulator in the enclosure that will generate heat as well.
-
Solder all the connections on the top of the board, leaving the bottom smooth and flat. Use some thermal interface (heatsink) compound and fashion some clamps to hold the board against a to be vertical wall.
-
Mount the board so it is vertical inside the box and allow air to freely pass by the board and circulate in the enclosure (by convection, not fan).
If you have an oven you can set to that low a temperature you could test it in that.
You can build a simple test oven with a box, incandescent light bulb, and cheap digital temperature controller.
I read from other blogs that it helps if you put a white or silver substance on the top lid to reflect heat and not let light into the enclosure; on the raspberry pi version I built this helped significantly; direct sunlight is a significant heat source;
We use to figure out what the package could dissipate, it was easy but I had a thermal person who used ICE Pack to simulate it. Expensive but here is a cheap way. Put a heat source at least 2X your worse case load (this is to guard band against conduction from the leads). For your heat source you can use resistors, light bulbs, whatever is available. Measuring the voltage and current of your heating device will give you the heat in watts. Then add a thermal probe, close the box and heat. Hopefully you can place it in an oven to simulate the ambient. I expect in your case you will not get more then 2-3 degrees rise. Hint Metal boxes conduct heat much better then plastic.
If you plan on doing this often run a series of tests and record the data. This is how we validated our simulations however we had an environmental oven to generate worse case ambient. Since "heat rises" mount the electronics near the bottom with the termination at the top. You will need to place at least a small hole in the bottom to compensate for temperature changes over time. Be sure to shield it from direct sun or you could still fry it.
all sounds good thanx
Now that I realize that the box will be in direct sunlight, it will get a lot more heat input than just being in a 110 degree ambient. Maybe that condition could be adequately simulated with an infrared heat lamp.
Calculate the ambient temperature as the interior of a similar test enclosure under direct sunlight for an hour or with blacked aluminum or some old heatsinks inside. A grill thermometer (cheap) or the temperature probe from an indoor-outdoor thermometer is adequate.
The solar load derives the initial high ambient then add the Volt-Amp Wattage increase approximation from the linked Internet resources.
First link returned by search:
Calculate the temperature rise in an electronics enclosure (powerstream.com)
There is no need to simulate the max ambient temperature. Within limits (which your application does not exceed) thermal rise is linear.
If you measure the internal temperature with a suitable internal heat generation. And measure the outside temperature you can calculate the temperature at what you feel is the max.
for example:
outside temp = 25 °C
internal temperature is 35 °C
We have a internal RISE of 10 °C
You estimate your max outside temperature is 45°C
Then, at 45°C you internal temperature would be 45 + 10 = 55°C
I used four of these air vents on a large, black gate controller enclosure. One at each corner. Then I covered the side facing the sun with wide aluminum tape. Stopped the overheating and venting by the two AGM batteries in the box.
You are not accounting for the temperature rise due to direct sunlight on the enclosure.
the calculator in post #11 predicts the temperature inside the box will be nowheres near 50 C so I should be totally safe; this agrees with the previous behavior for using adafruit metro mini (arduino uno is similar) but is ambiguous with regard to the result for raspberry pi zero 2 w; EmilyJane you are right but it helps if you shade the top of the enclosure with laminated white paper (inside the box) or aluminum foil; for aluminum foil zero sun heat will get through;
Yes, of course, I was just trying to envision some way of simulating actual conditions without actually having to wait for a 110 degree day.
I know, have to start somewhere. Also not considered is outside wind movement and different free air convection.
And:
- Is it near a building?
- North or South side?
- Phoenix Arizona or Denver Colorado?
- Thermal radiation cooling (admittedly small at these temperatures).
So how far down these paths we go depends on the results of initial testing.
John
Each box has a temperature sensor inside the box as well as the main "payload" sensor; e.g. there is a rain sensor box and a light sensor box and a "client" box ...... and each box has an ambient temperature sensor for the interior of the box to detect overheating; I tried it and the light sensor box (enclosure with arduino and sensor) cut out (stopped working) at around 100F ambient inside the box when the air temperature was 70 F outside the box; but it was during a spike in direct sunlight (as reported by the light sensor; it's a weather station); so the next thing to do is to block the sunlight with a laminated white piece of paper on the lid of the box; and if possible get those vents in post #13 (Paul); excellent vents; I looked at Digi-Key they have lots more like that; any others please???? I could not find any; or just drill a hole in bottom of box though I will have trouble cramming those vents into my tiny boxes; note the other two sensor boxes with arduinos (rain and client) kept on working under those conditions (100 F) so maybe it was specific to the light sensor box; be right back ...
the box lid is 5.9 in x 5.9 in = 6 x 2.54cm x 6 x 2.54cm = 15cm x 15cm = 225 * 0.01 * 0.01 m^2 = 0.0225 m^2; solar radiation peak was 500 Watts/m^2;
500 * 0.0225 W = 11.25 Watts from sunlight on each box
measured current got peak 130 mA @ 5V = 0.65 Watts from electrical heating for all three boxes;
so about 0.22 W per box; this does not include a powered usb hub in a fourth previously unmentioned box;
so it seems the electrical power is nothing compared to the heating from the sun;
I use these from Polycase in my outdoor Raspberry Pi projects but they may be available from elsewhere.