[OLPC-AU] Mass imaging

James Cameron quozl at laptop.org
Thu Jan 20 18:12:29 EST 2011


On Fri, Jan 21, 2011 at 09:12:31AM +1100, Sridhar Dhanapalan wrote:
> Are you referring to a means to tell the battery level from within
> OFW? Because I'd like to know how to do that.

Yes.  Here's a transcript showing you how to do it.

ok .bat
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 
ok watch-battery
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 

(the output continues to update, press any key to stop)

For remote monitoring over a network at one second intervals:

ok wifi ESSID
ok telnetd
Scan for: ESSID found
Associate with: ESSID
DHCP got 10.0.0.1
telnet://10.0.0.1
Connected

On a desktop or server:

$ telnet 10.0.0.1
Trying 10.0.0.1...
Connected to 10.0.0.1.
Escape character is '^]'.
ok .bat cr d# 1000 ms many
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 
AC:on  Battery: BYD LiFePO4  95% 6.67V 0.00A 26C full 


(Note that the time between starting telnetd and starting telnet the
laptop is vulnerable to another telnet access.  Once you have made a
connection, any other access is ignored.)

"test /battery" is much more typing, but almost effectively the same
thing.

The 95% above is the reported state of charge from the battery gauge
chip in the battery casing.  The 6.67V above is the voltage on the
battery terminals.

> Alternatively, can we assume that an XO with a green battery light has
> sufficient power to complete a NANDblaster receive?

No.

There is the possibility that the reported battery state of charge
differs from the chemical state of charge.  This is likely to occur if:

(a) the battery is in storage for more than two or three months at high
temperature (e.g. 40C),

(b) the battery has experienced quite a high temperature in transport or
storage,

(c) the battery is discharged in a device other than an XO.

If all three conditions can be excluded, then yes, a green battery
indicator should mean it has sufficient power to complete the task.

> We did that with XO-1s in the past, but I understand that the XO-1.5s
> have four times as much to write.

Not just that, but also the NANDblaster operation requires more power
operating the wireless and CPU for a long period before starting the
write.  There are more reasons that you describe.

> If we could do that, the total we could flash at once would not be
> limited by the number we could plug in.

Fresh units from the factory should not be affected by the above.  Try
not to order them and store them for months in a hot warehouse.

On Fri, Jan 21, 2011 at 09:44:49AM +1100, Sridhar Dhanapalan wrote:
> Alternatively, is there a minimum battery level we can use as a guide
> to know if an XO should be plugged in for the nandblasting?

No, because you cannot measure the minimum battery level chemically, and
you will only find out by operating the laptop.

> The people executing this aren't technically adept, so I'm trying to
> put together simple, low-risk instructions.

To handle the possibility of difference between chemical state of charge
and reported state of charge, you might discharge a laptop until it
reaches 90% reported state of charge, then charge it.  After charging,
the chemical state of charge and reported state of charge will be more
similar.

To characterise a batch kept in a warehouse, run ten of them to
depletion measuring the time it takes.

Alternatively, just keep them plugged in during the operation.  Live
with it.

For batteries that have self-discharged, the voltage reported by .bat
may give a clearer indication of chemical state of charge.

-- 
James Cameron
http://quozl.linux.org.au/


More information about the OLPC-AU mailing list