Intel® NUCs
Support for Intel® NUC products
12532 Discussions

NUC crashes on battery power, ideas to support current spikes

SC5
Beginner
137 Views

Nuc 12 Pro WSB here

 

We're looking to power this NUC with batteries. Current battery architecture is 5s li-ion battery (18.5v nominal) to a 16.5v dc regulator to the NUC. 16.5v is selected to power the NUC (12-20v) as well as 2 other subsystems (12-18v and 12-48v), as well as to be able to be charged with 24v.

 

Given physical volume and battery charging circuit constraints, we're looking at batteries with max continuous current draw of 4A with spikes of up to 7A. It has a battery management system onboard which handles cell balancing and prevents overcurrent.

 

Our NUC can get very hungry! It will typically be fine, but when we benchmark high demand it will crash. It will also crash sometimes during booting.

 

We're watching closely with hardware monitor and have the battery power going through an amp meter measuring 1ms max current draws, and it will try to draw over 7A () when it crashes. This tracks as the power supply that comes with the NUC is 120W continuous.

 

We've looked at reducing the power draw in the bios from 33W standard, 65W burst to as low as 20W standard 20W burst. This helps a lot but even with this it can crash on the battery under the right conditions (typically boot will spike the current).

 

For those that have dealt with this type of thing... These current draw bursts seem to be for 1/2 second or less. Is there a circuit we can put with our voltage regulator to supply the extra current this computer is looking for? It wouldn't have to have any actual capacity for powering the system, just a reserve to fill in during these short bursts, that could itself recharge from the battery once the burst is over.

 

Or is there something else we're doing wrong? Again, the system mostly works, but when we benchmark high demand, or during boot or other system processes, the current draw can spike which crashes the machine.

 

Thanks for reading

0 Kudos
3 Replies
Win7ine
New Contributor I
100 Views

You do not mention the current supply capability of your DC:DC converter between Battery and NUC, nor are you mentioning the discharge capability of the 5S LiPo pack. This ranges from typically "C20" to "C100" and has a significant impact on peak current delivery capability. It may just do the trick to select a higher C rated Battery compared to the one you utilise now.

LeonWaksman
Super User
91 Views
SC5
Beginner
28 Views

Thank you for the replies. The battery pack we were testing with was limited to "Max. 4Amp continuous, recommended"

 

We are looking into a larger 5s battery pack with "Maximum 7 Amp, limited by polyswitch".

 

Also the regulator was limited to 4A output. We are acquiring the regulator suggested by @LeonWaksman to continue testing.

 

However, we were also able to crash the computer running directly off a 15v 90W. Technically the 18v battery at 7A should provide 126W at low-ish voltage, however, the particular cells in this pack can go down to 2.5V, so 12.5 discharge cutoff voltage. We're looking at options to force battery change at 18V. We also have about 15W of additional electronics running off the same battery so there is no headroom on this solution.

 

We would be happy to use other batteries but require something that has been UN38.3 certified and performs on-board balancing. High performance R/C cars and drones offer many batteries with higher C rates and decent form factors but they require a separate charger with balancing function (2 plugs - power plug and balancing plug). We'd prefer a battery with just the power plug that can be balanced from an IC in our device.

Reply