GPSMAP 65s - In-depth analysis of power consumption

[ewald.gruen]

I spent the better part of the weekend analyzing the power consumption of the GPSMAP 65s. Here are the results:

Test Parameters:

• Ambient Temperature: 25 ºC
• Multi-GNSS: Enabled (unless otherwise stated)
• Multi-Band: Enabled (unless otherwise stated)
• Compass: Enabled (disabling it does not change the power consumption)
• Bluetooth: Disabled
• ANT+ Sensors: Disabled
• Track Log Recording: 00:00:01
• Memory Card: No (unless otherwise stated)

Power consumption vs. input voltage:

• 65s_power_consumption.png

Predefined voltage thresholds of the different battery type selections superimposed on the discharge curves of common batteries:

• 65s_battery_status_da.png
• 65s_battery_status_el.png
• 65s_battery_status_bt.png
  65s_battery_status_wt.png
• 65s_battery_status_bp.png
  65s_battery_status_wp.png

The discharge curves were taken from lygte-info.dk

Average current consumption under various operating conditions at 2.5 volts:

Code: Select all

                | Multi-Band | Single-Band | GPS only | 
--------------------------------------------------------
Backlight: 100% |     176 mA |             |          |
Backlight:  50% |     118 mA |             |          |
Backlight:   0% |     113 mA |      102 mA |   100 mA |
   Screen:  Off |      81 mA |       70 mA |    68 mA |
   Device:  Off |      69 uA |             |          |

Simulated runtime with different kinds of batteries:

Code: Select all

                | Eneloop | Eneloop Pro |  Duracell Ultra Power | Energizer Ultimate Lithium |
----------------------------------------------------------------------------------------------
Backlight:   0% | 16.47 h |     21.45 h |               22.17 h |                    30.27 h |
   Screen:  Off | 22.76 h |     29.64 h |               30.68 h |                    40.78 h |

On a side note, power consumption with the screen off (battery save: on) is about 70% higher when the compass and not the map is open in the background. This smells like a bug if you ask me. The same is true for the 64s.

[Zoot]

Did you try it with different track logging modes/intervals? The one time I tried setting it to one second caused it to seemingly suck the batteries dry very quickly (making me wonder if they're doing a flash block erase/write operation each second rather than doing something intelligent like buffering the track points and flushing them every 30-60 seconds).

[inu]

Hi ewald.gruen,

Thanks for your time. A couple of questions.

1. The first chart shows that the power consumption with an SD card inserted is a bit higher. Could you provide mfg. part number for this SD card? Better power moding support is where a card vendor can try to differentiate itself. So it'll be interesting to know about the brand.

2.

On a side note, power consumption with the screen off (battery save: on) is about 70% higher when the compass and not the map is open in the background.

The compass can be enabled or disabled. Still no difference?

[ewald.gruen]

Did you try it with different track logging modes/intervals?

Yes, but even at minimum recording interval, track recording increases power consumption by only 0.225 mW (or 0.1% if the screen is off).

wonder if they're doing a flash block erase/write operation each second

As far as I can tell, the shortest interval between two writes is 5 seconds, and the power consumption during a write looks like this:

1. The first chart shows that the power consumption with an SD card inserted is a bit higher. Could you provide mfg. part number for this SD card?

SanDisk Ultra R98 (SDSQUAR-032G-GN6MA), but I also tried it with a Transcend 300S R95/W45 (TS32GUSD300S) and the increase in power consumption was basically identical (+ 23 mW).

The compass can be enabled or disabled. Still no difference?

Yes, still no difference (and it would have been a surprise if it did):

Code: Select all

                | Compass | Trip Computer | Main Menu |
-------------------------------------------------------
   Screen:   On |  147 mA |         95 mA |     89 mA |
   Screen:  Off |  139 mA |         86 mA |     81 mA |

By the way, when you turn off the 65s, it still takes a good 10 seconds after the screen goes completely black for the device to actually turn off. So it might be a good idea to wait that long before actually removing the batteries.

[inu]

On a side note, power consumption with the screen off (battery save: on) is about 70% higher when the compass and not the map is open in the background. This smells like a bug if you ask me. The same is true for the 64s.

BTW the same is true for 62s. So this is a very old bug. Just checked. 2.5V Vbat, room temp, GPS simulation mode, LCD backlight off.

Compass page: 194.5 mA, averaged over 15 min.
Map page (almost the same for the other pages): 97.0 mA, averaged over 15 min.

Enable/disable compass makes no difference. No wonder indeed. So substantial increase can't be caused just by the sensor device. According to the data sheet for ST LSM303DLH (one of the two MEMS part types the 62s can be populated with), it consumes less than 1mA@3.3V when active. The rest may be attributed to increased CPU activity as huge amount of computations is necessary to transform the sensor's raw data to something meaningful. Perhaps Garmin continues with that business when it's not required (compass has disabled by user). Easy to check by monitoring the I2C bus activity, but it would be overkill here.
Or it can be a pure SW design issue. It's easy to observe increased current consumption when on the Map page. When the Map page is clean (i.e. no data fields are configured), the device consumes 97mA. When any data field is there, the consumption increases to 194.5 mA. But only when the data field(s) is visible and for as long as the screen is on. When it turns off, the consumption returns back to 97mA. In contrast to the compass page, where the device withdraws 194.5mA all the way, even after the screen has turned off by the battery saving timeout.

Anyway the advice to increase battery life by switching the device to Compass page instead of Map page may be questioned

[ewald.gruen]

Another trap for young players is that once the device loses connection to a paired ANT sensor, it goes into search mode 20 seconds later, increasing power consumption by 65 mW (or 30% if the screen is off), and the search never seems to stop until you either manually disable it or bring the ANT sensor back into range.

When the Map page is clean (i.e. no data fields are configured), the device consumes 97mA.

Right, I forgot to test that and also did not mention that I had 2 small data fields visible on the map page.

[inu]

Another trap for young players is that once the device loses connection to a paired ANT sensor, it goes into search mode 20 seconds later, increasing power consumption by 65 mW (or 30% if the screen is off), and the search never seems to stop until you either manually disable it or bring the ANT sensor back into range.

Played with Chirp to evaluate its suitability for cat finding. That's right, the search is continuous. I think this is by design. Otherwise Chirp could not be timely discovered. Addition of another configurable item would complicate the menu system further. So they opted for continuous search as it's the simplest strategy

I forgot to test that and also did not mention that I had 2 small data fields visible on the map page.

The fact that little has changed so far suggests that it's rooted deep inside the software architecture so changes are highly unlikely. It's stable. What is missing is a spreadsheet of some kind that specifies the energy cost for each feature. It must be a Garmin internal document. Good research field for enthusiasts to reveal that information

[GPSrChive]

What is missing is a spreadsheet of some kind that specifies the energy cost for each feature.

The GPSrChive > Oregon 6x0/7x0 > Operation > Power Sources pages have something similar.

[Domi93CH]

Anyway the advice to increase battery life by switching the device to Compass page instead of Map page may be questioned

I tried this yesterday and found that the batteries last much longer. when I stay on the elevation profile page and the device goes into sleep mode.

[ewald.gruen]

Yes, but when the screen is off, the power consumption is still 13% higher when the 'Elevation plot' is running in the background instead of the 'Map'.