Anatomy of the Leica M8's Power Consumption

[scott kirkpatrick]

Advertisement (gone after registration)

Mark, that is really nice data. It fills out something that we were both interested in back when the M8 was new. Remember the battery run-down curves that were plotted back in 2007? I've uploaded one that I found in my files. With the camera doing nothing, but not permitted to shut down, the battery runs from fully charged (almost 4.2 volts) to a point at around 3.5 volts where the camera shuts down in about five hours. You show that the minimum cost of operating the camera is 0.8 Watts (4.2 V times 200 mA). That's

without running the exposure meter, controlling the shutter, or firing up the circuits that extract data from the imager and write the dng file out to the SD card. So now we know that the battery has a capacity of about 4 Watt-hours. I don 't think that Leica has made that spec public, and it may have improved with later batteries.

 

This sets some bounds on how many pictures you can take with one charge of the M8. With the shutter half-pressed so that the camera can take a picture we see 2.4 Watts power consumption, slightly more when the image is being written out. The minimum time to produce one image, not using continuous mode, seems to be about 5 secs, or 12.5 Watt-sec in power. Adding in the M8c's rewind energy brings that up to 14 Watt-sec. You haven't shown the power-on and power-off transients in enough detail to integrate them and get their total energy cost, but let's say the most efficient picture takes no more than 15-20 Watt-sec.

 

The least efficient picture would be one in which we leave the camera active for two minutes after each exposure. The background activity for that costs 96 Watt-sec, so one image plus two minutes of waiting for shutdown could be as much as 110-120 Watt-sec, or more than five times the energy cost of the most efficient picture.

 

Four Watt-hours is about 14 thousand Watt-seconds, so one could take about 140-160 exposures on one battery charge in this least efficient mode. Turning the camera off between shots could increase that figure to as many as 600-700 exposures. That's about the range which has been reported.

 

Of course we want to know more. Is the background power consumption the same then the battery only supplies 3.5 volts? Does using more smarts (aperture control of exposure and auto ISO) cost additional energy when the shutter is half pressed? When you get time to check, of course...

 

thanks,

 

scott

 

PS As you point out, running in discreet mode adds nothing to the total power consumption if you lift your finger within 2-3 sec after taking the exposure. If you hold the shutter down for minutes, that's a different story.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[lars_bergquist]

Well, this brings me back to one of my pet obsessions---a battery pack. Ideally, it should be able to take both AA rechargeables and disposable lithium cells, optionally. It should be about the size of a Leicavit and like it, just replace the baseplate. I think quite an number of serious shooters would buy it.

 

The old man from the Age of Thumb Wind

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[graeme_clarke]

Mark,

Sincere thanks for this impressive piece of clearly presented really useful research. It seems to me that when wandering around taking pictures frequently, perhaps in batches every few minutes, a 2 minute setting for auto power off is likely to give longest battery life (together with dng, full auto rewind and no review). Does this concur with your advice?

 

Now, I have a DMR as well, and I wonder if the same principles will apply. Its battery life is never as long as the M8s and the battery is physically bigger so I assume the DMR draws a greater current. I've now got it set to a 2 minute auto-off and no review as well. If necessary it can be recocked manually to prolong active life (sounds like thethe old dog-food advert for PAL).

 

Thanks again, Graeme

[marknorton]

Scott

 

Thanks for your comments and analysis which I agree with. The battery is nominally (and optimistically) rated at 1900 mAH, which at 3.7V would mean 7 watt-hours. That is theoretical, a bit like gas-mileage numbers you never seem to achieve, and not all the battery capacity is usable before the battery voltage falls away, so something like 4-5 seems a real world value.

 

I didn't measure the current at lower voltages, I'll do another test when I am back in the office.

 

What I can say is that the battery meter works completely independently of supply voltage - a simple voltmeter it is not - and I had the camera shutdown because of an "empty battery" even when I was supplying it from the lab supply. Remove the dummy pack, cycle the power to it, put it back, full battery indication.

 

I didn't show the wake up current drain, as I recall, it's a few seconds of "meter on".

 

What's interesting in all of this is that the battery capacity is so strongly influenced by the camera idle time - my calculations showed 4.8 pictures lost in a 2 minute timeout, your's showed similarly so I think there would be a case for an additional shutdown option, say, 10 seconds to eek out maximum capacity from the battery. On the D3(x), Nikon go down as low as 4s and that camera (D3) will take thousands of images on a single charge, if you are careful. Its battery is labelled 28Wh.

Edited March 14, 2009 by marknorton

[marknorton]

It seems to me that when wandering around taking pictures frequently, perhaps in batches every few minutes, a 2 minute setting for auto power off is likely to give longest battery life (together with dng, full auto rewind and no review). Does this concur with your advice?

 

Graeme, we posted at the same time. The answer is that it will definitely pay to use the 2 minute auto-power off option for general walk-around photography and I can see no reason really to use the longer times other than if you need to react quickly to get that shot. The M8 get-ready time is not bad...

 

As in my previous post, I think there's a case for shorter times than the 2 minutes to improve battery life. I'd add 10 seconds, 30 seconds, 1 minute to the list of options.

 

As regards the DMR, I'm still amazed that Leica managed to integrate the back to the camera without changing the camera whose design was finalised long before the DMR was thought of. It might be that there are some compromises in what can be switched on and off; the M8 has a number of ways of switching parts of the camera on and off and I remember before it was launched this was one of the things they said they were working on.

[marknorton]

Well, this brings me back to one of my pet obsessions---a battery pack. Ideally, it should be able to take both AA rechargeables and disposable lithium cells, optionally. It should be about the size of a Leicavit and like it, just replace the baseplate. I think quite an number of serious shooters would buy it.

 

The old man from the Age of Thumb Wind

 

I agree, a clip on battery pack like the Leicavit would alllow a higher capacity Li-Ion battery as well as alternatives but their energy density is not great so I think it would be for "emergency use" only. In particular, NiMH batteries will have trouble meeting the peak current requirements.

 

This would also be one way of making more space available in the camera to support a FF sensor. Currently, the battery housing intrudes into the lens throat which does not affect the cropped sensor but would a FF sensor.

[Theodor Heinrichsohn]

Advertisement (gone after registration)

Mark,

Thank you very much for your analysis. As a result I have set both the M8 and M8u to DNG only, 2 min. shut off, no discrete mode. When I need a quieter shutter I'll set the M8u to discrete mode.

This will be most useful for me in April (Shanghai, Tokyo and Kyoto). As a back up I'll take a M5 or M7 body. All kinds film are readily available there.

Teddy

[marknorton]

Teddy, discreet mode is fine, it's just best not to delay it too long... the benefits of discreet mode far out-weigh any impact on battery life.

[TEBnewyork]

Thanks Mark,

Very interesting with practical application. So this goes into the classic thread column next to the anatomy course.

 

terry

[scott kirkpatrick]

This thread is already producing very reasonable firmware suggestions for 2.006 or whatever comes next -- a shutdown time of 30 sec to 1 minute, for example, which could double the number of pictures that you get from a battery charge.

 

If we see that the LCD is a pure power adder, then leaving auto review off, and pushing the display button twice when you want to chimp, is a potential power savings, but Mark's data will show whether this is a big deal (factor of two), or just a small gain (if the autoreview display adds, say 20% to the cost of taking a picture).

 

In traditional DSLR power management, when the display is turned on, other things that prepare for the exposure, like the autofocus servos, can be turned off, so the extra power cost is slight. The M8 doesn't seem to present such tradeoffs.

 

scott

[dpattinson]

Teddy, discreet mode is fine, it's just best not to delay it too long... the benefits of discreet mode far out-weigh any impact on battery life.

 

Hmmm, I think the benefits of getting the shot far out-weigh any issues of battery life. I generally leave my M8 on because I need it to be ready when I am. I can carry 4 batteries in my pocket if need be, certainly less bulky than the equivalent amount of film.

 

What I would really love, is a more reliable power indicator so that I can get more out of a battery and still choose a good time to change batteries. The current one sucks to be frank, it needs at least one more bar of accuracy. That way I could just change when it gets to one bar - If I do that now, I think I'm wasting 50% of the battery.

[scott kirkpatrick]

Hmmm, I think the benefits of getting the shot far out-weigh any issues of battery life. I generally leave my M8 on because I need it to be ready when I am.

 

When we leave it "on", it actually turns itself off after two minutes, and needs a brief wakeup to resume. (Unless you have set the auto-off time to never.) There's dispute on this, but personally I find the wakeup time is less than my time to raise the camera from my side to my eye.

 

scott

[marknorton]

Hmmm, I think the benefits of getting the shot far out-weigh any issues of battery life. I generally leave my M8 on because I need it to be ready when I am. I can carry 4 batteries in my pocket if need be, certainly less bulky than the equivalent amount of film.

 

If you need the camera to be instantly ready all the time, you leave auto power-off disabled and pay the price in battery life. If not, you gain from using auto power-off and there are benefits to be had from reducing the minimum time from 2 minutes.

[fefe]

800mW when idle sounds like quite a bit compared to most portable devices (cellphones?) today.

[farnz]

800mW when idle sounds like quite a bit compared to most portable devices (cellphones?) today.

In Post #1 Mark wrote:

 

"You can see that the current consumption is about 200mA while the camera is idling, about 600mA when the shutter release is half pressed (presumably, sensor and DSP are now running) with a sharp rise to kick the shutter off and a peak current of 1.8A to start the motor wind, decreasing to about 1.35A during the wind, and back down to around 0.75A during the image processing."

 

Pete.

[marknorton]

Pete, fefe is multiplying by a nominal 4v battery voltage to give 800mW.

 

I don't have an issue with such a figure. The M8 is built from largely standard parts which will never compete with custom silicon used in mobile phones. The key points of all of this are:

 

- there's scope for extending the battery life by providing faster auto power-off

- the motor wind hardly affects battery life

- the same is true for delayed wind if it completes by the time the red LED stops flashing.

[fefe]

True, but the Blackfin is supposed to be under 100mW in sleep mode (http://www.analog.com/static/imported-files/application_notes/EE_293_Rev_2_06_07.pdf can be run at 0.8V, 100mA at 50C), the M16C draws pretty much nothing when inactive according to their datasheets. For the Spartan 3 it's hard to say though but considering the set of components they used 800mW still seems quite high if they took advantage of all the power saving features available.

[farnz]

Ah thanks, Mark, I missed that.

 

Pete.

[marknorton]

True, but the Blackfin is supposed to be under 100mW in sleep mode (http://www.analog.com/static/imported-files/application_notes/EE_293_Rev_2_06_07.pdf can be run at 0.8V, 100mA at 50C), the M16C draws pretty much nothing when inactive according to their datasheets. For the Spartan 3 it's hard to say though but considering the set of components they used 800mW still seems quite high if they took advantage of all the power saving features available.

 

In the mode we're talking about - where the camera is ready to take pictures/display a menu/enter play mode/listen for a USB connection - not all the components are asleep. The Blackfin and sensor may be, but the Intel PXA270C and M16C will not be, you only have to remove the baseplate to see that's the case. The lens code sensor is running as well, if not continuously, then regularly. Attach a WATE in this state and the menu pops up without any other user intervention. That requires both processors and the I2C bus to be active.

 

When the camera is truly asleep - after auto power-off or the mode switch is moved to "off" - the current drops to 130 uA, about 0.5 mW.

Edited March 16, 2009 by marknorton

[Advertisement]

Hallo,

Look here: Leica M

[fefe]

Looking through the datasheets to try to make sense of this in my head .

 

M16C 0.75mA /MHz typical active current draw operates between 3 and 5V and 20-30MHz. That's at 4V another 90mW if it is fully active and not in power saving mode.

 

The Intel Xscale is another ~100mW when in standby mode (as for the Blackfin, transitions to full active mode from these sleep modes are in the microsecond range so I doubt they can't use them).

 

We are starting to get there, but it still looks to me that some component is not put to sleep as it should be when the camera is inactive. It might just be as you are saying the 6bit coding sensor or something else.

 

Anyway interesting thread, and with the integration on a SoC for the S2 they are likely to get this in much better control.