Rockbox mail archiveSubject: Re: ATA - USE_POWEROFF vs USE_SLEEP
Re: ATA - USE_POWEROFF vs USE_SLEEP
From: Andrew Jamieson <ajamiesn_at_optusnet.com.au>
Date: Tue, 10 Sep 2002 20:00:24 +1000
If you can send me the code, and indicate the CVS version you modified, I
will measure the current draw of each so that we can get an accurate view of
how much this helps. I would hazard a guess that this would be worth
sacrificing the HDD cache for, as it would deliver more benifit (in terms of
battery life) than the added MP3 buffer would.
BTW, I have a AJB, not a recorder, so hopefully you have the code for this.
----- Original Message -----
From: "Stuart Tedford" <stuart.tedford_at_piresearch.co.uk>
Sent: Tuesday, September 10, 2002 7:28 PM
Subject: ATA - USE_POWEROFF vs USE_SLEEP
> A while ago, I did some investigation into using poweroff for the hdd,
> rather than sleep or standby.
> It is important to note that the way it is in the code atm is very bad.
> not simply #define USE_POWEROFF because it will apparanty shorten the life
> of your hdd. This is because poweroff is called when the disk is still
> spinning. I have documents that say the drive will typically fail after
> this is done 20k times. This is what happens when you manually switch off
> you jukebox while the disk is spinning - it lets out a short "whine"
> also happens sometimes when disconnecting USB. Beware.
> I have an old cvs version of RockBox (circa early August) which I modified
> to poweroff 1.5 seconds after calling sleep - ie. after the disk has
> spinning. It works really well and I have used it extensively over the
> month or so. To give you an idea of the implications of this, I ran some
> tests at the time, the stock firmware ran for 10h50m, the unmodified cvs
> Rockbox I based my version on ran for 9h25m, and my modified version of
> code ran for 12h50m.
> I also found that it is best to call poweroff at least 1.5 seconds after
> sleep has been called. If you call it imediately after, then the disk
> into some sort of weird "freewheel" mode where it takes maybe 7 seconds to
> finally stop spinning. This caused some problems when the disk was
> up again during that time. It also increases the risk of damage if it is
> dropped during that time too. And I don't know if this freewheeling is
> for the disk, but I have to assume it is.
> What I would like to know is if there is any progress being made on using
> the onboard hdd cache? Becasue if that becomes usable then we can't use
> poweroff because the cache data will be lost. But at least this method of
> poweroff is a good backup plan to seriously reduce power consumption if we
> can't use the cache.
Received on 2002-09-10