The Agile Manifesto relies on 12 pillars (or principles). The more you intend to be in line with those pillars, the more agile your team (or family !) is.

This week, we are focusing on the following 3 principles:

• Pillar #2: Welcome changing requirements, even late in development.

Yes, we’re getting ready for our trip. We are leaving in 4 weeks, but we need to keep being flexible. Yes, we do need to get things done (quickly), but let’s not close the door to late changes, let’s not refuse to consider a new idea because it’s a little bit of extra work. Any little extra work leading to a better product (our trip) is worth it.

• Pillar #9: Continuous attention to technical excellence and good design enhances agility.

Yes, we have a pretty decent electrical system on board ; but there is a difference between feeling that the system is fine, and being sure that you made the best decisions. Once again, your product is only as good as the design decisions you made… and decisions are only valid when based on true facts.

• Pillar #10: Simplicity–the art of maximizing the amount of work not done–is essential.

Let’s keep every thing simple: after all, every spared Watt is a Watt we do not need to produce and store. Let’s be lazy.

So… this is a short post about the measured power usage of various electronic devices we are using on board. After the feedback we got from our previous posts (here and there), something had to be done to:

1. ensure we are really doing good as far as electricity/electronics are concerned (Pillar #9).
2. see if we can do better (yes of course.) (Pillar #2 and #10)
3. have some real numbers to show when we get challenged (warning: we like being challenged)

Warning!

Some of you will prefer to stop reading here:  no worries, it was our pleasure to get you informed on those 3 pillars! Feel free to read along, there might be some non overly-technical stuff hidden between the lines 🙂

For the rest of you who came explicitly to get numbers:  jump right now to the next section!

Power usage, measured in real conditions

When trying to improve the overall electrical efficiency, one key element is to measure things in order to collect facts about stuff. Do not believe the labels, stickers and the marketing guys.

To obtain these figures, we used a volt-meter and an amp-meter. Using the classic formula :

Power = Voltage x Intensity

… we gathered the power usage (by measuring Voltage and Intensity, and then multiplying the two). We prefer to use power (Watts) instead of intensity (Amps) when comparing devices because the voltage can change a lot between the low batteries state (around 11 V for dying batteries) and the “charging” state (right around 14 V). That’s a -8% to +16% variation around the nominal 12 V and this leads to the same changes in the measured intensity.

So, to compare apples to apples, we’ll stick to power which is constant for any given device (a 5 W device will need 5 W, whatever the batteries’ voltage is).

To get an idea of the electric charge those devices are using, we will convert it to Amp-hour, using the formula:

ElectricCharge = Power / 12 * Duration

… where “Duration” is the number of hours the device is used and “12” corresponds to the nominal battery voltage. This is useful to calculate how long you can power up your various devices from your batteries.

Instruments and VHF/AIS

Measured on board while batteries where showing 12.34 V. The intensity was measured right at the exit of the batteries by the Victron Energy BMV602 (now replaced by the BMV702). No other loads or generators were connected on the batteries.

Measurements:

• VHF Antenna spliter (Navico NSPL 400)+ AIS (Vesper Marine XB-8000): 5.9 W
• VHF (Garmin 200i): 5.6 W
  • Total #1 (spliter + VHF + AIS): 11.5 W

• Instruments (wind sensor + DST + GPS25 + rudder feedback + gyro-compass + 5 cockpit instruments): 9.5 W
• Auto-pilot computer: 1.7 W
  • Total #2 (instruments + auto-pilot): 11.3 W

Total for the basic navigation pack (VHF + AIS + Electronics + Auto-pilot): 23 W

Note: of course, this does not include the power to actually steer the boat (electric actuator).

So, that is 23 W. Let’s say that this corresponds to 2 A. This amounts to 48 Ah for a full day, in continuous use.

 Computer and NAS

Measured while powered by a 12 V generator (providing 14.07 V). The intensity was measured right at the exit of the generator, so those measurements include the 12 V voltage stabilizer and the UBEC to power up the 5V computer (0.5 W in total for these 2 elements).

Measurements:

• Minix Z64 A (running Kubuntu 16.10 + OpenCPN 4.4): from 5.6 W to 7.8 W
• Synology EDS14 + USB3.0 Hard Disk Drive: 7.0W (typical) to 8.5W. 6.2 W when drive is sleeping.
• Synology EDS14 (no USB drive): 5.0 W
• Synology DS116 (with WD Red 2.5″ drive): 5.6 W (typical) to 9.2 W (disk, CPU and fan working hard: during start-up)

Note #1: This does not include the computer screen (usually turned off while sailing, unless needed), nor the stereo (that we use to listen to the radio/music from the NAS/sound of the films). From what we remember, the stereo uses quite a lot of power (2 to 3 Amps maybe?). We will update this article later on.

Note #2: Western Digital Red series drives are amazing: check these benchmarks out.

On board, we are using the Minix Z64 A (24 hours) and now the Synology DS116 (8 hours):

• 6W @ 12 V is 0.5 A. Over 8 hours it is 4 Ah
• 6W @ 12 V is 0.5 A. Over 24 hours it is 12 Ah

Let’s round this to 16 Ah for a complete day.

Total

• Electronics + VHF/AIS: 48 Ah / day
• Computer + NAS: 16 Ah / day

Total for electronics + computer + NAS: 64 Ah / day

As mentioned previously, we have 3 batteries of 90 Ah (= 270 Ah) on board. Being lead-acid batteries, this means that roughly 50% of the electric charge is actually usable (= 135 Ah) before damaging the batteries. With 135 Ah available, we can power up those devices for 2 days.

Missing data

Next time we are on board, we will measure the following important data to complete this article:

• Auto-pilot actuator
• Fridge
• Computer screen
• Laptop running on the new 19V circuit
• …

About Raspberry Pi

In most forums, we found someone saying: “With a Raspberry Pi” you can “do the same” or “even better“, etc. While we clearly like and value all the possibilities offered but such a fantastic piece of inexpensive electronics (it is a tiny computer, ranging from 5 to 45 €/$/£): it needs very little power , it has a great community behind it, it’s expandable beyond imagination, it’s fully compatible with linux, etc.

A “raspi” also has (had?) some major drawbacks for using on board:

• It’s rather limited resources (CPU, memory) made difficult to run applications like OpenCPN  – latest versions of the raspi are way better in that regard. Still, it’s not a powerful platform.
• some applications (OpenCPN again, among others) had not been designed from the start to run on ARM architecture instead of the most common Intel architecture (and it has been a challenge to port the software to these less common architectures). However, it looks like latest versions of OpenCPN run fine on ARM.
• To use a “raspi” in an aggressive environment like a boat, you need a good case (waterproof if possible), you also need to take care of the cooling. Check this page.
• To get away from the poor performances of using a SD card, you need to add external (USB 2.0 only?) hard disk drives (which in turn needs power, and a proper casing, etc.)

From this page, you can expect your Raspberry Pi 3 (alone) to draw around 0.5 A (under 5 V), that is 2.5 W. With wireless (WiFi and/or Bluetooth) on, a USB hard disk to power, an active HDMI socket, a keyboard, etc. the RPi3 would surely get close to around 0.75 ~ 1.0 A. So, this is around 5 W (@ 5V). Don’t get us wrong, this is still quite impressive! But still, the competition is not doing terribly bad : 6 W to run a decent/recent Intel mini PC, that is not a load more power.

These are the reason why, after taking all these positive and negative points into account, we went for something else than a Raspberry Pi for our navigation computer and for our NAS.

Has anyone got measurements done on a Raspberry Pi with OpenCPN? We’re curious to see how it compares. We’re looking forward to reading from you !