- Heat Sinks and the Raspberry Pi 4
- Is the Maker Movement on the Decline?
- How the Latest RPi Could Have Scored a Perfect Zero
- The Problem(s) with Neopixels
- The Problem with Overclocking and the SPI Bus
- Nice Video of Last Summer's KC Maker Faire
- Python Cheat Sheets
- Benchmarking the New Raspberry Pi 2
- Power and the Single Pi
- The Problem with HATs
- The Problem with Relays
- Geekspace Gwinnett Pi Jam
In the process of writing last week's blog we got sidetracked from making our power measurements and ended up benchmarking the new Raspberry Pi 2. Well, this week, we finally got around to those measurements. Here's what we wanted to know: how much power does each of the three generations of Raspberry Pi consume. Specifically, using the same versions of Raspian, the same peripherals, and the same programs running in the foreground, what improvements have the engineers at rasberrypi.org made? Did the power consumption of the Rev 2 increase by 4X to support those new cores? So, here we go with all the gory details:
Table of Contents
Equipment Used
- Voltmeter: Best DT-9205M
- Current Meter: Wavetek 25XT
- Power Supply: Unifive US318-0530
- Wireless KBD/Mouse: Logitech K400R
- WiFi Adapter: Tenda W311M
- Raspberry Pi Models A, B+ and 2.
Software Used
- Fibonnaci script used in benchmarking article. See it here.
- OpenGL demo program for Python found on github.
- putty SSH client for Windows.
While running, the OpenGL demo program produces a spinning cone on the screen without the need of running x-windows:
Procedure
Power up the DAQCplate and measure the power required to drive it alone.
For each Raspberry Pi being tested:
- Run sudo apt-get update and sudo apt-get upgrade.
- Copy OpenlGL demo program to Home folder using FTP.
- Connect HDMI cable that drives monitor with resolution of 1920X1080
- Attach a Pi-Plates DAQCplate for use as a power injector
- Attach power supply to DAQCplate and current meter
- Plug in Wfi and KBD/Mouse fobs
- Apply power
- After login, change directory to OpenGL demo folder and launch cone demo (python cone.py) from the command line
- Launch SSH client
- Execute uname -a and record kernel version.
- Collect data using the following conditions:
- Idle
- Running Fibonnaci program from SSH client
- Running OpenGL cone.py demo
- Both of the above
Results
What Have We Learned
Right away we see the very good news that the four cores on the model 2 did NOT increase the power requirements by 4X. In fact, we see very little increase in power from the B+ to the 2. The biggest difference in power consumption occurred between revisions B and B+. This is when the designers went from linear voltage regulators to switchmode converters. They also added a nifty ideal diode circuit. Thesedesign changes reduced the power consumption of the Raspberry Pi by about 32%!
So there you have it - the new Raspberry Pi model 2 runs no hotter than the old B+. Nor will it require a new power supply, heat sinks, or a cooling fan.
Thanks for doing this investigation. I have been trying to choose a power strategy for my robot upgrade using the Raspberry Pi B+
(Regulated 5v for Pi+peripherals, Regulated servo power, unregulated motor power - with one recharge connection and power on/off)
I am guessing that soon after I get my bot upgraded (from HC6811 at 1MHz 64kb) to the B+, I will want to drop in a Pi 2 for twice the memory and four core "thinking".
I was really hoping that I could upgrade the Pi without upgrading my power strategy, and your investigation suggests I may be OK.
Thanks,
Alan