Heat from spark core and relay shield

I just got my relay shield today, and noticed the input filter cap got noticeably warm. I then decided to take some measurements of the different warm parts with my Extech multimeter and a thermocouple.

Ambient: 18c

CC3000 can: 44c
USB connector: 48c

Input filter cap: 41c
Output filter cap: 32c
Inductor: 35c
Regulator: 38c

This is after having been powered on less than 10 minutes. Silk screen says 7-15V, and I am powering it with 12V, so it should be within range.

Are these temperatures normal?

Hi @Mikey

The core has a maximum input voltage of 6V so the relay shield is dropping your 12V down 5V for the core.

The USB connector is meant to be a heat-sink for the voltage regulator on the bottom of the board so it makes sense that is the hottest thing you found.

I would say your readings are a bit high but still in the right range, but I don’t have a relay shield to test with. With mine running on USB power, the TI CC3000 runs about 28 C for instance. Power also depends on how far the access point is, how you are using the core IO, etc.

I know how it works, but I still wonder why everything gets so hot. Specially when capacitors also start to heat up almost immediately when power is connected.

This was even without the core doing anything.

My idea was to put this inside a plastic box, but with this kind of heat from even more components (more than from just the core itself, I am not sure I want to do that.

The main heat comes from the CC3000 module from my observation. Not sure if it’s heat transferred since passive components usually don’t heat up so much :wink:

Except I highly doubt heat from it will transfer all the way down through the pin headers, through the pcb of the relay shield, then into the capacitor only, without heating anything else up on the way enough to be noticed…

Hello @Mikey
@AndyW and I are looking into this. The temperatures are a little higher than usual. Do you wanna see if using 9V input gives you better results in the meantime?


1 Like

I don’t have any 9V power supply to try with. I can rig something up with an extra step down module to just do a test with, but it would not be feasible for what I intended to do.

Also, I forgot to put it in the first post, ambient temperature was 18c

1 Like

Okay, I give up on this.

It draws much more current than the other microcontrollers I got, and while it is “only” 240mA it draws, it still means I can almost have 5 of the other microcontrollers I use (also with wifi), while using the same amount of power as a single spark core.

At the same time the CC3300 still gets so hot it heats everything up around it, and would not make me feel good putting into a closed box, as the temperature would most likely get even higher then.

Accessing the API while trying to upload firmware is still a problem, so if something as a temperature is updated with json on a site, there is no way to flash new firmware.

spark.publish seems to block upload of firmware too if it for some reason is repeated too often.

Firmware upload generally fails so often. Most of the times the site goes back to Ready. while the core blinks like firmware is being updated, it can then either go over in not having anything lid at all, or flashing red, or seems to go in panic with a very fast white’ish flash with the led followed by either a red or blue.

For some reason, what I flashed to it a while ago, no longer works. Spending way too much time trying to debug by adding debug messages, followed by attempting to upload the new code for another ~10-15 minutes, just is beyond my patience.

Up to you if you want to continue to look at any of this or not, I no longer need a solution.


I feel your fustration and wished I could help debug your issues!

Which wifi-boards are you referring with lower power consumption?

Your issue with unreliable OTA flash might be due to the code ending up in some loop for too long instead of the main loop ().

Have a break and if you need help please posr back!

We are always ready to help you smooth out those rough edges! :slight_smile:

@Peekay123 I’ll bring this to your attention.

I have a couple of questions on the CC3000 heating up.

I’d been running the Spark Core using a standard USB power supply from my computer, which would be 5V at a few hundred milliamps (certainly less than 1A).
I realized that the heat was substantially high on the module after 1.5 hours. I would guess the temperature to have been about 42 C in an ambient 34 C.
The access point source was 18 metres away and I was using it to test the first two examples described in the documentation.

Is this something exceptional?

Hello Anirudh,
Thanks for those observations. We have noticed that the voltage regulator on the Core gets unstable and heats up if supplied with a noisy power supply. Although rare, this behavior has also been noticed sometimes with regulated power supplies.

The LDO regulators tend to be inherently unstable under transient loads (WiFi module transmitting) and this worsens if the input/output capacitors don’t meet the specs (variations during manufacturing runs/ parts quality control). We are now aware of this fact and working on to make sure parts quality is maintained throughout the run. Substituting these capacitors with higher values (22uF in place of 4.7uF) has also been observed to make the regulator more stable.

The temperature readings of 42C at 34C ambient are still within specs but higher than usual. Can you try placing a 22uF capacitor (with low ESR) close to the 3.3V pin of the Core and report what you observe?