Hi, I'm seeing frequent reboots when running the dev kit from a battery. The battery starts around 12v but when under load (it's a drone, when taking off it can drop quickly by a volt or more) it seems to reboot. The voltage might drop to 10-10.5v momentarily, then probably settle about .5-.75v below starting voltage. Current should not be a problem, it's running from a LiPO battery that can provide very high current (50-100A). Is the power system of the dev kit, or the module itself, sensitive to voltage drops?
I'll run it off a buck boost regulator if I have to, but the only 12v regulator I've found that can deliver the necessary current is a bit heavy for a drone.
Thank you for interest in the Intel® Joule™ Products.
I understand that you want to know if the momentary voltage drop may be causing the rebooting of the board.
I think the following information may help us understand what is going on with your board.
In the Intel® Joule™ Expansion Board – Design Guide ( http://www.intel.com/content/dam/support/us/en/documents/joule-products/intel-joule-expansion-board-...) you can see the following in page 34/34:
Intel® Joule™ Expansion Board Minimum Requirements: 2. Method to provide +VSYS power to the module: • Voltage minimum: 3.6 V, typical: 4.2 V, maximum: 5.25 V, with a current capability of 4 A
But perhaps, the most clarifying information is found in Intel® Joule™ Module – Development Kit Hardware Guide ( http://www.intel.com/content/dam/support/us/en/documents/joule-products/intel-joule-dev-kit-hardware...), where you can see the following statements in page 27/61:
The Intel Joule expansion board is designed to accept power from either the DC barrel jack (J4) or the USB Type-C connector (J7). When both the DC barrel jack and USB power are applied, the board takes its power through the barrel jack by default. It is recommended to use the DC power input jack while developing on the platform because it is easier to ensure input power quality over a direct DC connection. Any source connected to the J4 input must meet these specifications: • Voltage: 12.0 volts, ±3% • Current: 2.7 amps, minimum – the expansion board contains a 3.5 amp fuse
What I infer from the previous paragraphs is that the voltage drop, even if it is momentary, is what seems to be causing the rebooting, since the board receives less voltage than what is stated in the required specifications.
Please let me know if you find the previous information useful.
If you need further assistance, don't hesitate to contact us.
Thank you for sharing that information.
Since there hasn't been official tests regarding the use of LiPo batteries, replicating the behavior of a wall adapter power supply should be the target.
I'll suggest you to use a buck boost regulator to counteract the voltage drop under load that is causing the reboot.
I hope you find the previous information useful.
Thanks for the response, but this isn't really the information I was looking for. The power conditioning on the expansion board looks like it should be doing buck boost internally at various stages and should cope with voltage transitions, and was confirmed informally by platform_skater in the link above.
As I said, it's difficult to find a buck boost regulator with sufficient power to drive the requirements of the expansion board. I guess one option might be to use a step down regulator (ie. BEC) to say 9v - they are much smaller and lighter components, but they are typically locked to 5/6v or 12v and I don't trust the variable ones. I can't use a 12v because they will drop out entirely once they reach 12v+overhead, so no good for a 3s LiPO battery.
In the meantime, I'm rebuilding a larger drone with a 4s battery (16.8v charged, 13.2 discharged) and using a 12v 3a step-down BEC (which is a tiny component). I'll report back if that works under load.
In order to provide better assistance, could you please answer the following questions?
1. How are you setting up the device in order to measure the voltage drop? What is this measurement?
2. The 7 V disclaimer is associated with lightly loaded configurations, could you please tell me your power load?
If you could share your voltage measurement method, and a picture of your setup, it may help me understand more about your issue.
I'll be waiting for your response.
Here are are some telemetry traces from the power module. The joule expansion board was connected directly to the power distribution board (so directly to the battery). You can see the power drops when the current increases (when the motors are throttled) but it doesn't drop below 10v at any time. There are a couple of very short sudden drops down to 10.1v (they could be a measurement anomaly). The joule is under full load at this point with all the CPUs at 100% (doing vision processing).
I've taken the joule off this mini-drone to put it on a larger drone with a 4s battery running of a 12v BEC. This has proved to be very stable for the joule (which therefore indicates that the supplied voltage is indeed the problem), but the larger drone has had some 'teething problems', embedding it itself in the ground during testing several times! (Fortunately the joule wasn't onboard during testing).
But on the mini-drone (3s battery), the expansion board is connected directly to the battery (via a distribution board), so a picture of the setup isn't really needed.
Thank you for sharing your experience with the community. We really appreciate it.
From what I read, thanks to your testing, you came to the conclusion that the supplied voltage is indeed what is causing the rebooting.
Now that you we know that, what kind of assistance do you need from our technical team?
We already discussed the possibility of using a buck boost regulator, but you mentioned that it seems quite difficult to find one that suits your needs.
I'll be waiting for your response.
Well I'd like to know why the expansion board is unstable, and what are the likely causes/tolerances? As far as the information I've got so far, it should be perfectly stable under these conditions. It is being provided voltage and current that should be sufficient. Regulators aren't magic wands, they will also exhibit jitter and drops according to the input.
Have you got any further information on this?
In the meantime, I seem to have a stable solution by running the dev kit off an adjustable step down DC regulator set to 9.5v. This gives enough headroom for the regulator even when running the battery down to 10v, and it seems to provide a stable enough voltage to the dev kit that it continues to run even when punching full throttle and inducing a large voltage drop. I haven't yet tried this when the joule is under load, so that's the next step. But for anyone else looking to run on a drone, this is one option.
Well I'm talking to myself at this point, but for anyone else looking to do the same thing, I can confirm that using a step down dc converter like in the amazon link above works a treat, even under load. I've managed to mount the dev kit and a realsense r200 camera on a mini drone, run intensive computer vision code saturating the cpu, and don't see any dropouts or reboots even when the drone is under full throttle. Seems a good solution.
I'm glad to hear that you could solve the issue.
Thank you for sharing your experience with the community, we really appreciate it.
I still would like to inform you that I will ask the engineering team in case they have any updates on the issue you experienced.
Any new information will be posted here.
Thank you for your patience.
Have a nice day.
I just posted the following in another thread, but thought it might work for you too. . .
We had a similar issue with remotely installed systems. Our custom solution combines main-power with a battery system that can communicate status to the SBC and initiate a clean shutdown. .