So, there's this forum I know that's broken - no, wait, it's fixed now...
(Didn't want to gratuiously want to welcome its return, as of yesterday, but I've got something else to query.)A bit off "computing", but perhaps electronics...
Situation: A wall-mounted "Radio-controlled clock" (synchronised with so-called 'Rugby' time-signal, these days the UK broadcasting mast being based in Cumbria, but that's closer and the transfer was a whilecago now) went wrong. Showed odd times, whirred round as it adjusted (analogue, mechanical, standard clock design is to never adjust 'backwards' so it might need to sweep the hands round the best part of a full hour's revolution). Eventually, died and wouldn't work at all, no matter how much battery fiddling and button-pushing was tried. Replacement bought. Very similar, both 'powered' by a 'standard' MSF-style "quartz alarm clock" unit (single AAA battery type) set behind the 12"-or-so-diameter face.
New wall clock
also has started to go wrong. Sometimes seen an hour-and-a-bit fast, independently moving second-hand seems to be correct (against LCD clocks with their own 'Rugby' connection, 'pips' on analogue radio, probably also TIM/speaking-clock and websites like time.is, etc), at least when checked, but gear-tied hours and minutes seem lost/adrift on frequent occasions. Or you hear a whir and they're readjusting back to correct (not having noticed precisely what up-to-11h:59m of 'correction' was required, but obviously the clock thinks it needs it). Not yet had the hands stop entirely, unlike when the original broke.
The clock hangs (as did the one it replaced) on a wall near a windowsill upon which an LCD RC-clock keeps (seemingly) perfect time. Obviously glass vs brick (insulation-filled cavity wall) is slightly different, but given the line-of-sight-ish direction to the Cumbria mast, probably not a great factor. And this was not a known problem prior to the predecessor starting to go off-kilter.
The only factor that seems to coincide with the errors starting to crop up, might be the installation of a smart-meter. Within the same stretch of wall is some of the electricity-feed (and, for some time before, bits of the FIT-equipment that the roof solar panels also connect to, though their inverter is elsewhere). Since the smart-meter, a battery-storage was installed (last autumn, not yet reached the time of year when it is useful, and this is after the replacement clock), also located near the inverter (though includes changed cabling to/from the meter-box). But, though it is hard to pin down, the most logical (local) change was the smart-meter installation itself. Thus intimating that either it's the (SIM-based?) call-home signal doing something to corrupt 'Rugby' reception (unlikely?), or some novel electrical flux from whatever rearrangement of supply (external and solar) got fed through the new meter and other bits and pieces inserted betwixt incoming supply and onwards towards the separate (different housing) fusebox unit.
As an experiment, have asked the owner to moun the clock on a different (internal) wall. Too early to tell, but not so much obvious periods of inaccuracy. Second-hand retains customary precision (against other clocks), and I
once caught it whirring maybe from "two-ish minutes slow" to become correct again, once, but never noticed to be "running wrong" for any time at all, despite greater awareness and thus comparing with other clocks/etc, nor adjusting forward for a majority of the 12 hour sweep.
Had, at one point looked into the actual time-signal encoding. The time difference (in minutes) seemed to be 01010101
2 minutes 85m, 1h25m) fast, prompting ideas of a beat interference getting entangled with the broadcast 'data', but this didn't seem to hold up with either the way the signal is encoded nor the other phases of error and correctness observed. (This, admittedly, is probably the closest this issue comes to being a 'computer' problem.)
I've had both clocks 'opened'. More so with the old one, once "broken", but I've had the face-'glass' off the new one. No apparent mechanical seizure. Could not quite free the quartz-unit's casing (on the proven 'dead' old one) to poke away at the dead circuitry/internal gearing (and not even gone that far on the new 'working' one), very much designed as "no user-servicable parts", at that level, and screws and casing 'tabs' seem to have been designed as much more impregnible than the '80s-era same-style integrated travel alarm clocks (just quartz/mechanical, not RC) that I regularly dismantled and remantled
just for fun way back then.
(I was going to actually break open the old-one's casing, but the owner went and put it into IEEE bin at the refuse site, shortly after its replacement was installed.)
Anyway, for now I'm opening up this question to anyone who might have a better technical insight than myself, as far as any more remote disgnosis might be made. If the clock's owner can't handle turning to glance at the different wall, to get an idea of the time, it may get moved back and we'll see if it goes crazy again on a(n, ironically,) irregular but very frequent basis. If the new clock actually breaks (like the old one), then
its replacement might be permanently moved across. If it just runs its battery down quicker than expected (all that 12-hour forwarding?), as we think it's also been doing, then it'll probably just be lived with.
Author
Topic: The Generic Computer Advice Thread (Read 574386 times)