I have 2 of these with power switch physically shorted closed.
ShortMethod1- 1cm2 of copper leaf across on-off switch contacts, replacing the dished metal contact.
ShortMethod2 - wires soldered to switch contacts & run external to cam -> jumper'ed together
For these cams,
Normal power on is - close a switch and put ~4.7V DC to batt eliminator device
Normal power off is - open the switch
Normal control of cams is a chain of chdkptp commands. 'connect; rec; remoteshoot; play'
Normal cycle is exec'ed - one cam, every 3 min; one cam every 5 min round the clock
After 7-8K cycles, both cams now prompt for date/time at every power up. Setting the date/time is possible only when CHDK sdcard is removed from cam. (the CHDK logo pops up over top of the cam's prompt for date/time, and the FUNC/SET button is unresponsive. chdkptp commands sent to the cam while in this state continue to work correctly. Remove CHDK sdcard, setting date/time succeeds, SET button responds, but the value does not seem to be retained by the cam for very many cycles (tens, hundreds, zero) before it starts prompting again.
No complaint about CHDK/chdkptp, rather, I wonder if shorting the switch 'permanently' the way I did caused a drain on the battery responsible for retaining date/time.. Another possibility is I'm supplying too high Volts, and continuous over-voltage, over time, fries the "not-set" value into the cam.?.
Not sure if the tiny battery at end of plastic wire-ribbon (behind flash) is responsible for date/time clock. On my external-wires camera, it measured 3.04V. Foil-shorted camera measured 2.97V.
Tested for a drain by measuring V across the wires I'd run outside the cam, with cam Off. 0.3 V. Not sure I am seeing this correctly.. I don't recall measuring any V at the on-off switch copper contacts before
butchering performing the soldering job.. It may be that I got it too hot and melted to a neighbor wire inside the plastic ribbon; not sure how else to explain measuring V across a switch that (in my mind) should only be 'continuity' and never 'current'.
I may have ruined something in the cam regarding date/time values, but both cams are still normal otherwise. And I must have at least soldered to on-off switch wires 'okay' because the cam's on-off switch action works correctly; touch those wires together (provide continuity, not current) by closing switch for 200ms, then opening switch causes the cam turn on & off. Lens retracts correctly for the off cycle. (Cam still prompting for date/time; even tho date/time is correct value, I can see it behind CHDK logo.) chdkptp commands all work good, they blow right past the date/time prompt.
Will have another rig soon to test with - one that will use Method2 only, and maybe I'll achieve better soldering on the on-off switch. If my suspicion above is correct, by using cam's on-off switch function instead of brute forcing it to power on, I'll avoid the problem. It may be drain on the date/time batt, over-voltage or something else but it's ok because using cam's on-off switch action gives me a more reliable lens retract than just giving chdkptp:play command followed by a power cut.