Wednesday, February 29, 2012

NEX Overheating: More Surprises Revealed

Since my last post (Nex-5N Thermal  Shutdown Issue) I've been doing a series of tests on the effects of various heat sinking arrangements to see if I could achieve a goal of 1 hour recording time without any notification of thermal overheating (NOTE: the camera has an intentional, 30 minute, maximum continuous record time for apparently unrelated reasons).

All my efforts thus far have failed.  And I've tried many things, including removing the top of the camera (4 impossibly small screws) to see if it would help rid the camera of excess heat.



Finally, I became convinced that the problem was that heat accumulated in the top of the camera, and due to the LCD hinge arrangement, warm air could not escape quickly enough.  I determined to test this by inverting the camera.



But I had conflicting results.  I tried it at night and seemed to have initial success, but the battery depleted before I could finish the test.  I recharged the battery overnight and in the morning re-ran the test.  Immediate failure.  Perplexed, I added the heat sink back in.  What follows are the very surprising results.



TEST CONDITIONS  ("D" = Different from first round of tests)

The temperature probe is placed in the threaded coupling for the flash at the top of the camera. (D)

The label is OFF (see earlier posting above)

A heat sink of 1/8" tapered aluminium bonded with 3M 8810 thermal PSA tape is present in all cases (D)

The LCD is open approximately 15 deg. and not articulated (D)


RESULTS

CLICK ON THE CHART FOR LARGER VIEW
THINGS TO NOTE ABOUT THE GRAPH


  1. "30 minutes" indicates the camera successfully recorded for 30 minutes, stored the data and returned to idle status with no issues.
  2. There might be a modest effect of increased ventilation of having the camera inverted, but this is not the main issue.   Remember, the surface temperature sensor is, for these tests, located on the top of the camera.  Logically, one would expect that to be the hottest in the upright position and coolest in the upside down position.  It may be that there is almost no effect on the cooling of the camera electronics.
  3. The rate of temperature gain at the exterior diminishes with back-to-back recording.
  4. The initial recording displays the high temperature warning for some period of time and then shuts down.
  5. The warning light is NOT displayed during the second and third recordings which complete successfully even at substantially higher temperatures.
  6. Shutdown occurs almost immediately after the 3rd recording.  The warning my or may not have appeared.
  7. After a 5 minute cool-down the camera may be able to record for an additional extended period.  (Test aborted in the upside right case only because results were already clear).
  8. In this (and in all other tests I've done without exception), the camera asymptotically approaches some stasis temperature, but never flat lines.  That is, (according to my tests and under the test conditions) even after extended operation, and with significant addition of heat sinking material, the camera is still generating more and higher heat than these measures are able to shed.
NOTE: The delay in restart after the second recording is purely a failure of attention  on my part.  The  same delay was added to the second test.



DISCUSSION

Well, where to begin.

How to explain the successive decrease in the rate of measured temperature gain and the successful recording of two segments after the initial failure?  Some ideas already mentioned by others:

  • Does the battery generate more heat on initial discharge than when somewhat depleted?
  • Conversely, does the battery itself act as a sink when depleted?
  • Do the thermal materials perform better once heated?
And some of my own:
  • Does the thermal algorithm look at rate of change as well as maximum temperature?
  • Is cooling enhanced at higher temperatures by modest convection effect?
  • Does the LCD, slightly opened (and importantly, not tilted (articulated) , maximize its function as a heat sink and/or enhance any venturi effect?

Certainly it seems clear that the camera's initial shutdown is premature in light of its successful operation at higher temperatures.  The lack of a warning light at high temperature shutdowns with low slopes supports the idea that the warning light may only be coupled to rate of change.

If this is true, then it seems this could be fixed in software.  Given the rapid rate of cooling after shutdown or even in idle state, sensing the rate of change seems unnecessary.

It seems very strange that the camera doesn't appear to reach a maximum temperature much earlier.  Totally speculation on my part, but it seems similar to behavior of an open feedback loop or a process spinning out of control.  


NEXT STEPS

I've ordered the AC charger to test the effect of removing the battery and some conventional heat sinks to test the original suggestion of richg101  to add DRAM heat sinks in this post.

I've even order a very tiny cooling fan.




13 comments:

  1. have you managed to reduce the heating towards your goal of 1hr issues free?

    just bought a 5n... Love this baby even with the heating. If only there was a way to cool it down...

    ReplyDelete
  2. a very simple answer to the flattenning curve:
    The sensor puts out the same finite amount of heat every 30s recording period, but the surrounding casing temp. keeps increasing.
    Heat transfer between bodies of widely dissimilar temperatures is much faster then when their temperatures are closer together. Just basic physics.

    ReplyDelete

  3. NOTE: the camera has an intentional, 30 minute, maximum continuous record time for apparently unrelated reasons).

    First of all, thanks for the awesome blog. Thanks for sharing. Unfortunately, there is a restriction that limits still cameras from recording more than 30 minutes. Here are some links:

    http://www.dpreview.com/articles/0794343949/wto-looking-at-moves-to-remove-30-minute-limit-from-digital-cameras

    http://photo.stackexchange.com/questions/14563/why-is-there-a-limit-restriction-to-the-1080p-film-video-recording-time-duration

    -Quang

    ReplyDelete

Your comments welcome.