x10 - CM12U (CM11) Clock Accuracy

The CM11/CM12 clock is usually kept roughly in sync with the AC line frequency, which can fluctuate. In the USA the power companies tweak the frequency up or down so that AC clocks will be accurate on the average over the long term.

When disconnected from AC and running under battery power, the CM11A clocks I've tested tend to lose about 6-8 minutes per day.

When you disconnect the CM12U from the PC, you _must_ disconnect the cable at the CM12U end. The circuitry is very susceptible to stray pickup from a dangling cable which can screw up the works.

Note: If AC power to the CM12U is interrupted even momentarily, then once power is resumed it sends a request for a clock update to the PC. ActiveHome silently complies with this request.

Regards, Charles Sullivan

The powerline frequency may be more accurate than you think. As you say, long term is very precise. Short term can deviate slightly, but peak errors are less than 0.1%, and typically better than 0.025%. The deviation is caused by load variation. Generators have enormous rotating inertia, but sudden increases in load will begin to slow them down. Many power plants use the steam turbine cycle, and it takes a finite amount of time to respond to the increasing load. So there can be instantaneous variations in frequency as the control loops respond. The short-term frequency errors occur on both sides, and long term average is very accurate.

Years ago the MIT National Magnet lab had a huge flywheel generator brought in to provide the instantaneous energy necessary for their experiments. It was impossible to pull energy of that magnitude off the grid.

Beacon Power is a company using flywheel energy to help stabilize the electric grid. This is a more efficient way to handle instantaneous increases in electrical demand compared with having excess capacity available in a steam boiler.

Jeff

Thanks for the info Jeff.

A CM11A that's doing nothing else seems to keep time within a few seconds of NTP time, but will drift a little when given other tasks (like reading its clock?) so I've never trusted it to accurately reflect the daily frequency accuracy.

Regards, Charles Sullivan

Hi -

Thanks for this. Do I understand from your message that the CM11/12 uses the mains supply frequency to regulate its internal clock? If so I may have a problem as we are connected to a private generator supply and the frequency does vary somewhat. Even so the variation in frequency doesn't account for the large errors I'm experiencing with timed events. Other postings (c.f

talk about a modification to the internal clock to improve its accuracy but this doesn't seem to apply to the CM12U as it doesn't have an 80C48 chip.

I would appreciate hearing of any ideas for improving clock accuracy? Thanks.

A private generator (not connected to the grid) could certainly be a factor. One Onan manual for a RV generator says it should run at 63Hz no load, dropping to 60Hz under full load. Even if the generator is only a few Hz off, that can accumulate to a clock error exceeding an hour a day. However, that same error would also show up in any other electric clocks that take their reference from the same powerline.

As I recall from the CM11A beta test, it updated the clock whenever it was connected to the PC. So it would never register a clock error.

Jeff

The post cited refers to the CP290 (an antique), not the CM11A. They're two entirely different beasts, not just in the clock. The CP290 used a crystal oscillator which could be improved; the CM11A/CM12U uses the AC mains frequency, which you're generally stuck with.

Does your CM12U gain time or lose time? If it gains time (runs fast and timers execute earlier) it's possible that excessive noise or spikes on the generated AC waveform could cause it to do that, in addition to whatever changes are due to fluctuation in the mains frequency.

Jeff, I believe ActiveHome will only update the CM11A clock on user command -or- (automatically and silently) if AC power to the CM11A has been interrupted. The OP never stated whether or not he unplugged his CM12U from the AC socket when he disconnected from the computer.

That was a long time ago, but I don't recall the Beta version ever reporting the clock was wrong. Of course there were many cycles of unplugging and removing batteries, so it could have always queried for a time update when I reconnected.

Jeff

Jeff & Charles - many thanks for your responses and information. I'm now beginning to understand the situation a little more!

With reference to your comments ...

When the generator is running the frequency varies between 48-50Hz. When not running the load is supported by a full-sine wave invertor stabilised at

50Hz. The timed events are downloaded to the CM12U interface. The PC to which it is connected is not running the whole time.

The CM12U is left plugged into the AC socket all the time - even when the computer is switched off - with a small number of timed events stored in the interface. The CM12U is plugged into an AC socket that is part of an 'essential' supply supported by a battery-invertor system when the generator is not running, and the generator when it is running. The primary purpose of the CM12U is to ensure that services that may have been left switched on are switched off at fixed times to conserve battery power at times when the generator is not running.

Ok - understood about the CP290/CM11A; so the clock mod won't apply to the CM12U? Is there anyway of 'modding' a CM12U to use an 'internal' clock rather than using the mains frequency to regulate its timing?

I haven't kept detailed logs of timing errors. Timed events occur both earlier and later than planned so it is possible that errors may be due both to the noise spikes and poor frequency regulation. There is an x10 filter on the input side to the supply to the 'essential' service.

With thanks,

Martin...