Suspended mode

Andy Green andy at
Wed Feb 13 14:43:24 CET 2008

Hash: SHA1

Somebody in the thread at some point said:
> 於 三,2008-02-13 於 16:19 +0800,Neng-Yu Tu (Tony Tu) 提到:
>> When GSM on with sleep mode (or power saving but still connect to GSM
>> network), the GSM module will _flicking_ the power from 1-2mA to couple
>> uA, could observed on power supply's LCD panel. If not in power saving
>> mode, it will takes about 20mA at idle. But yesterday's suspend
>> experiment result was GSM module still takes 40-50 mA, that is way to
>> high. 
> 	I found the root cause of this phenomenon. According to previous mail,
> the suspend current could be reduced into 5mA~4mA if the GSM related
> GPIOs are configured OUTPUT low. 
> 	But this would affect not only GSM incoming call wake-up behavior but
> also blocking GSM into sleep mode. Once GSM could not get into sleep
> mode itself, its current consumption would become high.  

I'm not sure I understood the explanation properly yet.  How can we get
4mA performance by setting low GPIO which "also block[s] GSM [entering]
into sleep mode"?  You mean we get 4mA performance if we do the LOW GPIO
*and* we turned OFF power to the GSM module?

But if we stay with the forced low GPIO and leave power on GSM, we get
crazy behaviour during suspend from the GSM module power consumption?

We don't need the GSM GPIO to be forced low any more to get lower
current with GSM *OFF*, Allen will remove the resistor that made the
reason to force the GSM GPIO low to hide the current flowing.

Maybe we need to force GPIO for this RTS CTS thing -- is that still
believed to be a problem?

> If you don't have GSM related GPIO configuration, then resuming from GSM
> incoming call works quite well. Unfortunately, the suspend current would
> becomes 18mA. And the current does not always stay at 18mA, it would up
> to 30mA periodically.  

What is the path this 18mA travels when it flows?  It's inside the GSM
module that this current is used?

> 	I try to have GSM related pins as PULLUP configuration. But nothing
> changes. It seems that we might need HW solution such as a buffer
> between GSM and CPU. 

PULLUP is really PULLDOWN now I think.

- -Andy
Version: GnuPG v1.4.7 (GNU/Linux)
Comment: Using GnuPG with Fedora -


More information about the openmoko-kernel mailing list