50mA lost during TOH charging

Just for understanding:

• You connected a solar panel to the charging contacts exposed to the TOH?
• You measured the current running between the solar panel and the charging contacts exposed to the TOH? (I.e. your wires run directly between solar panel and Jolla, one of the wires you opened to insert your amp meter?)
• The results of your external hardware amp meter differ from the indications provided by the Jolla internal measurement circuits by 50 mA?
• Did you get the readings of the Jolla internal measurement circuits with and without the external hardware amp within the charging circuit? Has there been any difference in the values?
• You checked the voltage provided by the solar panel while charging (measure directly at the charging contacts exposed to the TOH) and while idle (measure at the pins you connect to the charging contacts but have them removed from these contacts): What did you measure?
• Which readings of the Jolla internal measurement circuits do you get when you connect the solar panel to the USB charging contacts?

Hi, the panel is connected to the two pins called "External charger 5V DC VIN" in TOH SDK pdf.

The current was measured as you describe (opened one wire and inserted amp meter).

Yes the the difference seems to be always around 50mA.

Without panel or when panel is covered, jolla internal measurement circuits inidicate very low power consumption. E.g. on wifi when screen is off you can read values ~6mA. This quite corresponds with jolla's standby time (~7 days).

Voltage is 4.5V while charging.

As for the last point - i'd need to hack some spare USB cable for that. I can try at home if really needed - but i guess the results will be quite similar - i tried charging jolla with USB hand crack and the results were IIRC very similar.

If your reading charge_now then your not in suspended state (which eats more mA's).

"Voltage is 4.5V while charging.": USB 2.0 specs require max. voltage 5.00±0.25 V, i.e. your 4.5 V is well below specs. I do not know the charging circuitry of the Jolla, but it may be heavily optimized and not work well on voltages too low (even the battery voltage is still lower than your charging voltage).

In addition: When you insert your amp meter into the charging line, the voltage available at the charging contacts drops even further.

Instead of using the solar panel you should set up your tests using a power supply providing stable 5 V output (independent of the current drawn) and allowing to adjust the current from minimum to 100 mA. Then try to verify whether the charge_now is correct for different charging currents (e.g. by measuring the battery charging time from almost discharged to 100% – hoping that this indication is quite correct – and comparing this value with the battery's capacity in mAh). With these data you can estimate what you can get from a solar panel providing min. 5 V during charging at currents <100 mA (it may be that the charging circuitry not only requires the correct external charging voltage but also a minimum charging current to work sufficiently).

rainisto: well i am quite sure you can read current_now even in suspended state. To me it looks that the value is average for last few secs. So connect wifi, SSH to the device, write cat /sys/class/power_supply/battery/current now, wait a few secs and press ENTER. The phone sleeps before ENTER is hit and the displayed value is often from suspended state.

One of the problems is that while charging the phone wont enter suspend - there are wakelocks in kernel which prevent the phone from sleeping. I wrote about it here: http://talk.maemo.org/showthread.php?t=92317&page=7