How Much Juice Does The Pandora Use At Full Hog?


Phawx

Professional Derailer
Joined
Oct 1, 2008
Messages
1,345
with LCD brightness at full
all LEDs on the mobo on
all chips running at maximum usable frequency
data transfer of wifi and bluetooth

Has this been tested yet? Has anyone looked at the maximum power draw of each of these components and added them together yet?
 
MWeston posted some numbers of the CPU at 900MHz plus Bluetooth. It was about 700mA IIRC. Don't know about the rest.
Edit:
This is the quote I'm refering to:
QUOTE
Without a battery, the TWL4030 is designed to regulate the voltage with a current limit of 1A (the conversion is linear so it does generate heat). I don't know what happens if you try to pull more than an amp but I will try it once Wifi is up and running. BT doesn't draw enough and even overclocking to 900MHz doesn't pull more than about 710mA. This is just a limitation presented by TI in the datasheet so it could be a recommendation. We shall see.

With a battery, the system will regulate at up to 1.7A (measured to be 1.78A on real hardware). Again, the process in linear and the nominal input is 5.2V. The board regulates this down to 4V-4.2V so that extra volt turns to heat. I believe that is why TI imposes these current restrictions.
 
mali said:
MWeston posted some numbers of the CPU at 900MHz plus Bluetooth. It was about 700mA IIRC. Don't know about the rest.
Man that is so sweet. I guess we will find out more when everyone starts getting there Pandora's in. Also, couldn't you program the battery indicator to be more accurate if you knew approximately how much juice everything takes at different loads?

Or is it even worth it? I've read a few articles on Battery charge levels and how inaccurate they are and possibly a work-around if we knew exactly what each component "ate" and when the battery stops delivering the voltage and amperage needed for the Pandora is function.

It would just be neat if you looked at the battery and it said, "Based upon the past 2 minutes of usage, if you continue this trend you will have X:XX amount of time left."
 
Last edited by a moderator:
Phawx said:
It would just be neat if you looked at the battery and it said, "Based upon the past 2 minutes of usage, if you continue this trend you will have X:XX amount of time left."
Already possible.
 
Last edited by a moderator:
I am also wondering... the usb is fully powered so this can draw quite a bit of current right
(imagine you connect a 2.5' powered usb drive...)

If you also have the device overclocked a bit and full screen brightness, wifi and bluetooth working, would the device really be ok ? :blink:
That might seem a bit extreme but I am sure some ppl here are going to try this...
 
marshal said:
If you also have the device overclocked a bit and full screen brightness, wifi and bluetooth working, would the device really be ok ?
It should be fine. As quoted above, the system will handle up to 1.7A with a battery installed - which you will always have in real-world situations. With the Pandora using a max of 710mA and the USB port maxed out at 500mA, you're still well under the max.
 
Last edited by a moderator:
I think MWeston stated in his quote that the 710mA is JUST from "overclocking" to 900mhz. So not everything is accounted for. Well lets see here

710mA - Cortex pushed to full speed
40mA - Bluetooth worse case (Assuming this is correct)
500mA - USB 2.0 standard max power load
XmA - Wifi at max power sink
YmA - DSP at max power sink
ZmA - SGX at max power sink
+___________________________
1250mA + XmA + YmA + ZmA of total power sink

1250mA + XmA + YmA + ZmA = 1700mA (maximum potential power load with batery)
XmA + YmA + ZmA = 450mA of left over power

Now, assuming that MWestons initial figure of about 710mA for max OC did include a bit of power draw for the SGX and DSP (just low level running voltages) we can probably add *a little* bit more to our left over but not much. Our most power hungry module will most likely be the wifi in this case so it will probably get most. Lets just hope it leaves enough behind for the SGX and DSP.

Keep in mind these figures are mostly estimates at WORST case scenarios. Plus, will you be streaming HD video and downloading it to your usb hard drive that pulls 5 unit loads from your usb, while playing doom 3? If you answered yes then you probably should go buy something else :p
 
Soulkiller said:
I think MWeston stated in his quote that the 710mA is JUST from "overclocking" to 900mhz. So not everything is accounted for.
That's silly. I don't even know how you would go about measuring the current draw of just the CPU. Based on this post and the quote above, it is pretty clear that 710mA is the current draw of the entire system (minus wifi or any USB accessories).
 
Last edited by a moderator:
Thx for the quick answer. Seems like there is well enough margin :)

I guess I could have / should have figured that myself from the figure you provided... ><
 
while we are on the topic relating to the pandora battery performance, i was curios to know what will happen when the battery does actually run out. Will the pandora just sizzle out with a load of white noise like the gp2x or will the os enter a 'safe shut down' or hybermation mode (i know this will require more ram)

Please excuse me if this question may have already been answered. I couldnt find any posts which gave a clear answer.
 
djmaddis said:
while we are on the topic relating to the pandora battery performance, i was curios to know what will happen when the battery does actually run out. Will the pandora just sizzle out with a load of white noise like the gp2x or will the os enter a 'safe shut down' or hybermation mode (i know this will require more ram)

Please excuse me if this question may have already been answered. I couldnt find any posts which gave a clear answer.

All power / suspend functions are controlled through software, so you can theoretically set it up however you like. I don't know what the default will be though.
 
Last edited by a moderator:
Soulkiller said:
710mA - Cortex pushed to full speed
40mA - Bluetooth worse case (Assuming this is correct)
500mA - USB 2.0 standard max power load
XmA - Wifi at max power sink
YmA - DSP at max power sink
ZmA - SGX at max power sink
+___________________________
1250mA + XmA + YmA + ZmA of total power sink

1250mA + XmA + YmA + ZmA = 1700mA (maximum potential power load with batery)
XmA + YmA + ZmA = 450mA of left over power



There is a rather big power drain you forgot about: The LCD screen! Also, other possible power drains if used: Speakers.
 
Last edited by a moderator:
djmaddis said:
while we are on the topic relating to the pandora battery performance, i was curios to know what will happen when the battery does actually run out. Will the pandora just sizzle out with a load of white noise like the gp2x or will the os enter a 'safe shut down' or hybermation mode (i know this will require more ram)

It's possible that you can say "When battery reaches 15%, alert me, when it reaches 10%, hibernate". I doubt it'll be there from the beginning, but it will be possible to do.
 
Last edited by a moderator:
Benjiro said:
Soulkiller said:
710mA - Cortex pushed to full speed
40mA - Bluetooth worse case (Assuming this is correct)
500mA - USB 2.0 standard max power load
XmA - Wifi at max power sink
YmA - DSP at max power sink
ZmA - SGX at max power sink
+___________________________
1250mA + XmA + YmA + ZmA of total power sink

1250mA + XmA + YmA + ZmA = 1700mA (maximum potential power load with batery)
XmA + YmA + ZmA = 450mA of left over power



There is a rather big power drain you forgot about: The LCD screen! Also, other possible power drains if used: Speakers.


The quote mentioning 710mA is about forcing the system to take the most juice possible for testing the power chip. So not using the LCD in this test doesn't make sense. The 710mA should already include the LCD power draw.

Edit:
@Chip
Those links both don't work for non-Mods.
 
Last edited by a moderator:
mali said:
@Chip
Those links both don't work for non-Mods.
I guess so, as I always get on error page:
QUOTE
Sorry, an error occurred. If you are unsure on how to use a feature, or don't know why you got this error message, try looking through the help files for more information. The error returned was:
Sorry, the link that brought you to this page seems to be out of date or broken.
 
Last edited by a moderator:
I think it would be more interesting to know, if the SGX Unit needs much more Power for complex 3D stuff compared to simple 2D Stuff. I know from PC graphic Cards that modern 3D acceleration sucks more power than anything else in the PC. I know the SGX can't be compared with this but reduced a intense 3D game onto the Pandora the battery more than a simple 2D Game? Or is the difference low? (compared with the same CPU clock speed of course)
 
Check the BBcode. That second link is devoid of any special mod tags. I don't know why it won't work for anybody else. I'm not sure how else to link to it. Here's the quote:

MWeston said:
When Linux is idle at the console, regardless of clock speed, the current sits about about 420mA. Running FBA at 500MHz puts the current at 600-625mA and at 900MHz it is 695-710mA. Picodrive uses less power at about 650mA at 900MHz I think it was. FBA is the most current hungry of the emulators I have tested, but just by 20-30mA or so. Even the PSX emu uses less, but it's got a bunch of debug junk in it that really slows down screen updates so it might go up when I get to see those 150+fps Zod was getting. He has been too busy to make me a better binary so it's not a realistic version of something you would play on.

Yeah, it would be great to run some IR tests on the CPU. It's going to be pretty much the same as the EVM. The Beagle board I had feels the same at 500MHz as the Pandora board does at the same speed. If you are interested, you could do some EVM tests, or does the processor mount upside down in the socket? I forget.

Another of the many concerns to address was how the battery does at high current loads. So far, since we are operating in the nominal 0.2C range, it is cool as a cucumber. It's such a big battery that the internal resistance is very low. Obviously you can't test that one yet, but the fact that it still feels cold to the touch seems convincing to me.

Off topic: I did a battery test the other day with a new cell on its first full charge. Based on the current draw (about 625mA), I predicted the battery would last 6.5hours but it ran for about a minute over 8.5 hours. The current load was very consistent for the entire day so that calculated out to over 5000mAh! Who knows, we might be in for a welcome surprise with the final product when power management is applied.
 
Last edited by a moderator:
Back
Top