Xmega 256 a3u power consumption

Log in or register to post comments
Go To Last Post
12 posts / 0 new
Author
Message
riccardomanfrin
riccardomanfrin's picture
Level: Hangaround
Joined: Tue. Apr 27, 2010
Posts: 103 View posts
#1
Posted by riccardomanfrin: Thu. Oct 11, 2012 - 03:34 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi all,
I'm messing with these micros and using the same code on all of the below described cases.

I'm experiencing a power consumption of about 60uA in sleep mode with an xmega 256 A3U QPFN mounted on an STK600 board.When I send the micro to sleep I can observe a power reduction from 20mA (@32MHz-IRC) to 60uA, so I assume that the power state transition does occur.

On a custom board with the same uC model, but in the surface mount flavor, with all the other components unsoldered I experience 149uA in sleep.

Finally, the same code, on another custom board with a xmega 256A3 model (without USB), goes down to 4 or 5 uA [which is reasonable as I also have a linear stepdown)...

Can you provide your experience on these micro's consumption?

Regards,
Riccardo

Tags:

AVR Microcontrollers, AVR XMEGA
Top
clawson
clawson's picture
Level: Moderator
Joined: Mon. Jul 18, 2005
Posts: 98819 View posts
Location: (using avr-gcc in) Finchingfield, Essex, England
Posted by clawson: Thu. Oct 11, 2012 - 04:03 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Are you following all Atmel's guidance about things like floating inputs and so on?

Top
mojo-chan
mojo-chan's picture
Level: Raving Lunatic
Joined: Thu. Jan 24, 2008
Posts: 2203 View posts
Posted by mojo-chan: Fri. Oct 12, 2012 - 07:13 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Arrrg, forum keeps eating my posts.

Make sure all I/O lines are set correctly. On your unpopulated board try setting them all to output low. Having floating lines can dramatically increase current consumption.

Also make sure you have all the power reduction registers set up.

What sleep mode are you using? I get 30uA with the RTC running in Power Save mode on a 128D3, and that is for the whole board which has other stuff on it.

Top
riccardomanfrin
riccardomanfrin's picture
Level: Hangaround
Joined: Tue. Apr 27, 2010
Posts: 103 View posts
Posted by riccardomanfrin: Fri. Oct 12, 2012 - 09:54 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Power reduction registers are said to be effective only in active and idle power states. In sleep mode they are overridden by the power state configuration (AVR1010 section 2.7).

As for the IO lines, I'm currently not touching them.
AVR1010 section 2.9 says that their level is, by default, floating, and section 12.2 of Xmega UM (doc8386.pdf - B rev.) specifies that they are tri-stated after a reset, which I assume to be synonym for totem-pole push-pull setup (doc8386.pdf section 15.3.1). The docs state that this floating configuration could trigger digital input buffer switching.Currently all port pins aren't connected to anything, so I could try to hold them in totem-pole with pulldown and see if something changes... (?)

Just a note:in section 37.1.4 of the UM they report power consumption in sleep power mode and assert a consumption lower than 1uA (~0.80uA). To tell the truth, they report this "normal" and "low-power" mode without ever mentioning the meaning and how those two setups are obtained. In any case, my 149uA is almost a 150 times greater value then the expected *_*...

I don't know what is draining all that current..

R

Top
clawson
clawson's picture
Level: Moderator
Joined: Mon. Jul 18, 2005
Posts: 98819 View posts
Location: (using avr-gcc in) Finchingfield, Essex, England
Posted by clawson: Fri. Oct 12, 2012 - 10:34 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Suggest you read AVR040 and AVR042

Top
mojo-chan
mojo-chan's picture
Level: Raving Lunatic
Joined: Thu. Jan 24, 2008
Posts: 2203 View posts
Posted by mojo-chan: Fri. Oct 12, 2012 - 03:45 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

riccardomanfrin wrote:
Power reduction registers are said to be effective only in active and idle power states. In sleep mode they are overridden by the power state configuration (AVR1010 section 2.7).

Right, so I'll ask again, what sleep mode are you using?

Quote:
As for the IO lines, I'm currently not touching them.

Then that is almost certainly your problem. Floating lines will mess your power consumption up. As I already said try setting them all to output low. Pull ups will also work but do consume very slightly more power, which all adds up if you have a lot of them.

You should never leave lines floating if they are not connected to anything. They default that way so that the AVR resetting does not mess up any buses it is attached to or try to force lines against some peripheral.

Quote:
Just a note:in section 37.1.4 of the UM they report power consumption in sleep power mode and assert a consumption lower than 1uA (~0.80uA). To tell the truth, they report this "normal" and "low-power" mode without ever mentioning the meaning and how those two setups are obtained. In any case, my 149uA is almost a 150 times greater value then the expected *_*...

1uA is presumably something like this:

- All I/O lines set to output low
- All possible power management enabled
- Power Down sleep mode
- 1.6V supply voltage
- No external circuitry, not even pull-ups on reset lines etc.

I would hope that they allow at least one async interrupt to wake the AVR up in that budget, but they could also be relying on resetting it to come out of Power Down mode.

Quote:
I don't know what is draining all that current..

I already told you.

Top
riccardomanfrin
riccardomanfrin's picture
Level: Hangaround
Joined: Tue. Apr 27, 2010
Posts: 103 View posts
Posted by riccardomanfrin: Mon. Oct 15, 2012 - 12:11 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

clawson wrote:
Suggest you read AVR040 and AVR042

Consumptions are being tested on STK600.. if they did wrong design it would surprise me..
R

Top
clawson
clawson's picture
Level: Moderator
Joined: Mon. Jul 18, 2005
Posts: 98819 View posts
Location: (using avr-gcc in) Finchingfield, Essex, England
Posted by clawson: Mon. Oct 15, 2012 - 12:24 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:

if they did wrong design it would surprise me..

Indeed but the STK600 has no control over what you are doing inside the AVR like setting unused IO to outputs etc.

Top
riccardomanfrin
riccardomanfrin's picture
Level: Hangaround
Joined: Tue. Apr 27, 2010
Posts: 103 View posts
Posted by riccardomanfrin: Mon. Oct 15, 2012 - 12:25 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mojo-chan wrote:

Right, so I'll ask again, what sleep mode are you using?

Power Save mode

mojo-chan wrote:

Quote:
As for the IO lines, I'm currently not touching them.

Then that is almost certainly your problem.

We also tried with output low config but no luck.

mojo-chan wrote:

1uA is presumably something like this:

- All I/O lines set to output low
- All possible power management enabled
- Power Down sleep mode
- 1.6V supply voltage
- No external circuitry, not even pull-ups on reset lines etc.

We use "Power Save" mode and the graph attached here by shows that the consumption @3V3 is 0.80uA. We don't enable RTC and we are testing on a STK600 whose schematics are not provided but whichi is supposed to be used exactly for these kind of purposes. I don't know if there's a pullup on reset line but all the remaining I/O pins are left open disconnected.

I already told you.

Attachment(s): 

Image icon Power Save mode consumption for XMEGA 256A3U
Top
mojo-chan
mojo-chan's picture
Level: Raving Lunatic
Joined: Thu. Jan 24, 2008
Posts: 2203 View posts
Posted by mojo-chan: Mon. Oct 15, 2012 - 12:43 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

riccardomanfrin wrote:
We also tried with output low config but no luck.

On the STK600 or on a bare PCB with nothing else connected? Are you sure the PCB is not loading it in any way?

I have the 128A3U here and with nothing connected at all except a 32.768KHz xtal which is not enabled I read 1.2uA.

Quote:
We use "Power Save" mode and the graph attached here by shows that the consumption @3V3 is 0.80uA.

The datasheet says 1uA to 3uA at 3.0V. Section 36.3. It assumes that the RTC is running. I would assume power saving mode for the 32.768KHz xtal is enabled.

Quote:
We don't enable RTC and we are testing on a STK600 whose schematics are not provided but whichi is supposed to be used exactly for these kind of purposes. I don't know if there's a pullup on reset line but all the remaining I/O pins are left open disconnected.

I'd try it on a bare board. Alternatively you could look in the ASF for examples of power management on the STK600 and see what they do.

Top
riccardomanfrin
riccardomanfrin's picture
Level: Hangaround
Joined: Tue. Apr 27, 2010
Posts: 103 View posts
Posted by riccardomanfrin: Wed. Oct 17, 2012 - 07:55 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You were absolutely right. The troubles were not related to the uC but rather to the STK 600 which I must say, sucks as a testing board if it doesn't even provide separate power lines for micro and external circuitry.

We took one of our boards and unsoldered every component, and tested and it came out that the consumption is 0.1uA with our testing code...

Thanks in advance to all of you for the support.

I must also thank atmel support which I wrote to in parallel. Fast and professional

Top
theusch
theusch's picture
Level: 10k+ Postman
Joined: Mon. Feb 19, 2001
Posts: 39693 View posts
Location: Wisconsin USA
Posted by theusch: Fri. Oct 19, 2012 - 06:32 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

October 11:

Quote:

On a custom board with the same uC model, but in the surface mount flavor, with all the other components unsoldered I experience 149uA in sleep.

October 15:
Quote:

Consumptions are being tested on STK600.

:(

BTW, How does one separate the power for the microcontroller and "external circuitry"--if that external circuitry is attached to the microcontroller and shares grounds and sends signals and the like?

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

Top