Need tips for detecting AC power failure

... Takes about 10 seconds for the Elk to broadcast the SS command, but yes it does send it without being polled. Plugged it back in and took a couple of seconds to indicate AC was restored.

Harleydude, many thanks for confirming how it operates. No polling required; that's very good news! I guess it is time to upgrade my M1's firmware to take advantage of this new feature.

AceCannon, the wall-wart idea is a viable solution. However, given that the M1 can report AC availability (via the SS command) it's a better fit for my operating environment (ELK M1 with Premise) and it requires less peripheral hardware (i.e. the wall-wart).
 
Wall wort works for everyone with a PLC not just Elk or HAI. Also some have their alarm on an UPS, wall wort works that way too.
 
I would like to address the shutting down of the PC only.

My first question is; are you sure the 20MB of Triplite software is really necessary? I actually use APC, and it comes with a CD that contains their monitoring software. I have the software installed on my work PC and it tells me all kinds of nice info about battery condition and warns me that the battery is more than 2 years old and that I should order a new one, etc... At home I have the same UPS but I did not install the software. Windows (XP) automatically detected the UPS and will shutdown my system when the battery is low. It doesn't have all the extra, and in some cases annoying, bells an whistles, but it works great for it's basic purpose.

I know that Windows directly supports APC, and a couple of "off brands". Triplite may be different, but I suspect they are more alike than their marketing departments would like to admit, especially for units with USB ports. Check it out.

Next I would NOT recommend using anything other that the UPS low battery signal to shut the system down. Why? It is the only thing that is reliable.

Experiences I have had in the past:

I generally try to size a UPS to give me 20 minutes of run time. Statistically most power outages are 10 minutes or less. Of course this varies by area...

- When the battery is new and fully charged I can usually count on the full 20 minutes, but this time can be affected by current system load. If the outage occurs while the system is idle, the UPS run time is significantly longer than if it occurs while I am rendering a DVD.

- When the battery is getting old run times also get shorter.

- Multiple outages can cause short run times. One day we were having power problems at work. I think they were working on the lines down the street. We had several outages that only lasted for few minutes each, but they occurred over the space of about an hour. Then we had another outage and about a minute into the outage the my system reports low battery and begins shutting down. The reason was that it may take 4 hours to fully recharge the battery after an outage. Basically the UPS just didn't have time to fully recover between the outages.

IMHO any system/mechanism that considers only time is not reliable. I generally feel the simplest solution is often the best. Use the built in and established mechanism.
 
You'll get no argument from me that a UPS driver can provide more detail and is a better approach. My goal is to be apprised of a power failure (programmatically) so that my HA application can initiate other activities like maintain a log and shutdown sensitive appliances to protect them from "dirty power" (spikes, over and under voltages) when the AC is restored. My HA software would need to be notified by the UPS's monitoring software.

There appear to be two versions of software available from Tripplite, one is a 10 Mb download whereas the other is a 20Mb. I haven't installed either version because I've found nothing on Tripplite's site that suggests their driver can be accessed programmatically. Perhaps the documentation that comes with the driver reveals more information ... I may have to install one or the other to learn more.

I just tried to use Win XP's Generic UPS driver with the Tripplite and the result wasn't good. Pulling the plug on the UPS caused it to shutdown immediately. I restarted everything, disconnected the serial cable between the PC and UPS, and repeated the test. This time the UPS did not shutdown. Clearly, there's an undesirable interaction, between the PC and the Tripplite, when using the Generic driver! I accepted all of the default Generic driver parameters and they probably need tweaking. But I don't feel any of this is worthwhile if I can't interact with the UPS driver on a programmatic level.
 
Back
Top