RAM Upgrade Impact on Battery Life - Less Than You Think!

[Jason Dunn]

<div class='os_post_top_link'><a href='http://www.ppctechs.com' target='_blank'>http://www.ppctechs.com</a><br /><br /></div>We've seen test results in iPAQ battery white papers where the more RAM an iPAQ has, the worse the battery life gets. I assumed that when I got my XDA back with 4x more RAM (from 32 MB to 128 MB) I'd see a significant drop in battery life. So much so that I ordered an XDA extended battery. Upon doing some tests, however, I was pleasently surprised that with the backlight and GPRS radio on, the battery life went from 8.2 hours down to 7.8 hours - a loss of only 5%. Not bad at all!

[KyleC]

How scientific are these results? Did you just walk around with it until it was dead, look at your watch and say something like: It's been approximately seven point eight hours since this was fully charged! Before the RAM upgrade, it was approximately eight point two hours on a full charge! Astounding! I have an idea for a new post!
Oh, and I love how I can "quote" someone when they didn't actually say it. Example:
We have some Hexacto games to give away. But who should we pick?
I think that we should pick KyleC
That does sound like a good idea. What do you think Marlof and Andy?
No argument here!
Let's go for it!
Hehehe...

[Jason Dunn]

How scientific are these results? Did you just walk around with it until it was dead, look at your watch and say...

Uh, no. :lol:

I have a testing methodology that I'll be posting soon - it basically involves using Scott Seligman's KeepAlive application to run the unit until it dies. The results are accurate to +/- 10 minutes, and I perform each test at least twice, sometimes more if the second test was significantly off in results from the first one.

[toshtoshtosh]

Jason, having more RAM will not significantly affect running time. What it should (theoratically) affect is standby time. Since in standby RAM refresh is a significant portion of the drain (As opposed to when you have the unit on).

MY 2 cents!

[Chris Spera]

I have a testing methodology that I'll be posting soon - it basically involves using Scott Seligman's KeepAlive application to run the unit until it dies. The results are accurate to +/- 10 minutes, and I perform each test at least twice, sometimes more if the second test was significantly off in results from the first one.

I am using this methodology to do tests for The Gadgeteer. My results are on the OTHER END of the pendulum. In my iPAQ 3835 and my Toshiba e310, the battery life fell by about 1/2 after the upgrade. However, my 3900's battery life only dropped by 15%, due in large part to the battery saving functionality of the xScale processor.

The only time this is a problem for me is when I'm in the car, and since I have a revolv Design RoadWriter in the vehicle, its a non-issue. The RoadWriter charges the device while it sits in the cradle.

Christopher Spera

[Timothy Rapson]

So, the manufacturers told us they were not being cheap with RAM, only trying to give us better battery life. And they LIED! What a surprise.

This is an interesting topic. Hope to read more, maybe on the standby time as well as the active run-time.

[KyleC]

How scientific are these results? Did you just walk around with it until it was dead, look at your watch and say...

Uh, no. :lol:

I have a testing methodology that I'll be posting soon - it basically involves using Scott Seligman's KeepAlive application to run the unit until it dies. The results are accurate to +/- 10 minutes, and I perform each test at least twice, sometimes more if the second test was significantly off in results from the first one.
If the unit dies, how do you get the data you need? Won't the RAM be wiped out? Anyway, you said that you'd be posting the process soon. :wink:

&lt;*> EVERYBODY - THIS IS MY 50TH POST! YAY! &lt;*>
Ooh! Just realized that I "ponder" now! :D

[cdunphy]

Jason -

When a unit is on (even in idle), the amount of power going to keep the RAM alive is minor in comparison to the juice necessary to keep the screen lit and the processor running. So it is little wonder that more RAM makes hardly any difference in a "fully charged till drained" runtime test.

But when the unit it "asleep" - the screen is off and the CPU is in standby. This is where the extra memory begins to have a big affect on the total power consumption. Doubling the RAM doubles the amount of power necessary to refresh the RAM, and this can have a major effect on how long your device can keep data while "off".

Since hardly anyone uses a PDA for 4 or 8 hours straight, the much more realistic battery life test should evaluate runtime as well as standby stamina.

What we do is run a device for 40 minutes straight, and then turn it off for 23hrs and 20 minutes. Repeat the same 40 minute ussage the next day. And so on....

This will give you a much more realistic battery life result - and I think you would see that the extra memory probably has a major affect.

I really wish that more reviewers would use this sort of "real world" method for measuring battery life. I think the results might be surprising.

- chris

[jojaboi]

I can't remember where but, I just read an article yesterday about the same thing. The author upgraded his xda to 128 mb RAM with ppc techs and tested battery life to find very similar results. He used windows media, playing through speeker (can't remember if he specified volume), with screen off. I believe the results were 8 hrs before upgrade and 7.5 after

[enemy2k2]

I really wish that more reviewers would use this sort of "real world" method for measuring battery life. I think the results might be surprising.


- chris

I like his method of reviewing much better. I for one plan on charging my unit every night any way, and most people synch theirs all the time so it's always full of juice. Not only that, power users use their devices quite a bit more than the average joe. I think this looks good for the future of these devices, 128 and 256 MB in the not too distant future, now that will be cool! Not just that, but replaceable battery packs also help quite a bit.

[Janak Parekh]

So, the manufacturers told us they were not being cheap with RAM, only trying to give us better battery life. And they LIED! What a surprise.
There's cost of the device too...

Also, power consumption is not linear with RAM size.

--bdj