LampLinc 2000STW - mystery solved

Hey Dave.. I have a bunch of their lamplinc and appliancelinc modules and have several with the same problem. They work fine until the power goes off, when the power comes back on they are as if they were reset. They lose their programming and code setting..Very frustrating, some of them are behind heavy furniture..I emailed Smarthome and they said "sounds like your getting surges, we recommend a whole house protector"..Well..I have two..One in each panel... Now we have had several power failures and most of them last only 1-2 seconds..But seems the only thing not liking this is the smarthome modules...

Anyone else having similar troubles? John

"John" wrote in news:ddgc3m$hs$ snipped-for-privacy@geraldo.cc.utexas.edu:

Yes John,

I have about a dozen or so of the Smarthome modules/Switches. Out of the

12 I have 2 that will lose their address during brief power outages. I also have a whole house surge supressor. I don't know why only 2 out of the 12 have this problem, I have even switched the location of the modules with locations I have no problems with, and the problem follows the module.

Hi Kevin.. That stinks, if these things are that sensitive that a power outage will cause them to do this..I emailed Smarthome and they blame it on my power..They need to build them more robust apparently..Funny thing is, if I unplug them, and replug them they are fine..But after a power outage they lose memory..Can't figure that one out..Only thing I can figure out is our power usually only goes off for a couple of seconds until the substation can switch to a different circuit..It's very frustrating..All of mine are still under warrenty, I may start sending them back a couple at a time for exchange..Maybe I got a bad batch..I bought so many over a period of time I don't have any way of knowing which was what..But all within the two years.. Other than that I love them..They do work (when working) better than the X10 units...One thing I do like that if the power goes off at night and comes back on your not left in the dark...

John

"Dave Houston" wrote in

Did you think that problem's been engineered out or is it likely to affect the Insteon line as well? How would you simulate a brownout for testing?

There's nothing more annoying than having to reset stuff after a brownout!

-- Bobby G.

>

Very interesting. Switch mode power supplies cause all sorts of havoc, I'm a little disappointed it took you this long to figure it out ;)

You are probably experiencing a brownout (a lowering of voltage, perhaps just before a full outage) that is affecting the PIC. This is a design flaw

- it has nothing to do with your power being substandard as everyone experiences these at one time or another.

(snipped)

Hi Dave.. You could be right. We just moved in last November..New developement, all underground electric. We have had terrible occurances of very short power outages..Like one to two seconds..Also have noticed as you mentioned some brownouts..Once was very odd..I was standing in the kitchen (open floor plan) and it even appeared that different areas of the house were dimming at different levels.(this was non X10 lighting) I reported it to our electrical coop. They were unaware that our subdivision was having any problems so they came out and removed our meter and installed some sort of a "recorder" and they asked me to email them every time any outage or brownout happened. He said part of the problem is electrical storms cause the substation breakers to trip and switch to another route..I told him I think the fact it does it so fast is what causes problems. Electrical stuff doesn't even get a chance to fully turn off..He compared my readings with the readings at the substation and said they didn't match so was why they wanted to put the meter/recorder thing on..They just picked it up late last week so hope they find the problem. I have never in my life had such problems with power outages... Thanks for the heads up on the units..Strange thing tho, it's always the same five modules..The others seem to maintain their memories.....

John

[more stuff snipped]

I don't know. The recent event here was a few days before I started testing the Insteon gear. I did not open up the Insteon modules to see what MCU and EEPROM they use as they were on loan and I did not want to risk marring them. It's not easily simulated as there may be characteristics (e.g. slope, duration, magnitude) that are unknown.

There is one factor that may help with Insteon (or make this a bigger PITA). Each Insteon device gets an ID number at the factory - analogous to an Ethernet MAC number. You go through an enrollment procedure (see the Insteon white paper) to add new modules to your network. Individual devices will retain their hard coded ID number but the controller may lose its memory of them. If so, you'll have to re-enroll _every_ device which is a bigger problem than having a few anmesiac modules or switches.

However, SmartHome's ostrich-like behavior regarding an obvious design flaw that is still affecting new LampLinc & SwitchLinc devices has to make one nervous about Insteon. I _think_ I even saw reports here of recent (but pre-Insteon) PowerLinc controllers losing their memory.

The latest PICs will operate down to 2.5V. You see 2-3 second power glitches. Pressing and holding the programming switch for ~3 seconds puts the units into address mode. It may be that internal capacitors store enough energy to keep the PIC going but that the circuit associated with the programming switch drains off much faster and the PIC sees the button as "pressed". Differences in component tolerances or firmware could account for the fact that only some are affected.

3-4 years ago we had frequent outages but I cannot recall any for the past two years until the recent one. There was about a 1.5 hour outage. The power came back on just long enough for me to start resetting clocks, VCR, etc. before flickering off for a few seconds. The latter flicker got the LampLinc module.

Hey Brian, Long time no see!!!

Dan

When did you get out? ;)

I just looked through Google, searching on various forms of *linc, reset, power loss, etc and it's clear this issue goes back to at least 1999 with the same sorts of issues being reported repeatedly. Some, but not all, switches in a house will reset to A1 after a power loss.

I suppose we'll learn sooner or later whether the Insteon line has inherited its parent's flaw as more of them get into the field. I tend to think that they're going to use the same design as before on the X-10 side of the switch. Let's hope, for end user's sakes that they haven't made it worse instead of better.

In terms of designing a brownout simulator, it might be very difficult because in a real power "incident" large appliances will suddenly come back on line simultaneously and create both an enormous current flow and a great deal of back EMF. That might be very hard to simulate without flipping the main house breaker on and off rapidly. I'll bet it wasn't part of their testing procedure even though that's the real world these devices operate in.

-- Bobby G.

Heh... Busy in an information security startup, for some reason I just decided to pop in here and have a look around.

Yeah, I was actually kinda of wondering/hoping if this group would've gotten back to some core concepts via newbies being drained off to specific fourms/bbs's (no offense to newbies).

If the messages weren't date-stamped, it would be hard to tell most of today's messages from most of 1995's messages...

"Dave Houston" wrote in

That's what I was trying to say; if they tested the units with a whole house load flipped on and off at the main breaker a few times in a few seconds, the might have identified the problem in their circuit. Quickly flicking the main breaker only *sort* of simulates the powerline conditions found when the power company switches loads at the substation. It's not a real simulation because when a substation "chatters" every large load in the neighborhood is being rapidly cycled on and off at the same moment. Since the reports almost always say that only a few modules exhibit the problem, it might be load related as well. Only switches driving large loads or near large loads might be affected. Maybe someone with some duff *lincs lying around can run a few tests.

I certainly didn't mean to imply that the power would be off long enough to cause the EEPROM to lose its settings. In fact, the ECS K7S5A motherboards have exactly this problem. If you don't have them on a UPS, a brief series of power blips will cause the CMOS settings to default - which drives a lot of people crazy. They would go to sleep and the machine as fine and wake up and have to reprogram the CMOS. It was the fatal flaw in an otherwise very nice, very cheap motherboard.

Fortunately a $50 UPS was the cure. While it's practical and even a good idea to put a PC on a UPS, it's not so practical for a wall switch. I wonder if the Lincs and the ECS mainboard share any components in common. They certainly share the same behavior.

-- Bobby G.

Aye that. Between Ocelot forums, Homeseer forums, etc, I can barely keep up.

[snip]

That's why I added the caveat to my otherwise positive comments on Insteon. That the complaints have spanned six years raises questions about their dedication to quality.

I suspect it's simpler than that and related to the design of the address programming circuit. IIRC most of the reports (and my one incident) appear to be associated with momentary powerline glitches rather than long term outages.

EEPROMs are designed to retain data for years without power so its highly unlikely that the power glitches directly affect the EEPROM. It's more likely that a flaw in the hardware design or firmware logic is the culprit.