Very laggy X10 performance

[bkenobi]

My system has been working fine for quite some time...until last weekend.  When out of town, we apparently had some sort of power issue due to a wind storm.  When I returned, I found that the entire system is working very poorly.  To be specific, if I try to turn a light (e.g. WS467) on locally, it works perfectly.  If I send a command from any source (HR12A, MS16A, CM15A, XPT, SC503, etc), it takes 5+ seconds for the command to go through.  I pulled out my XTBM and found that I do not get a response from the XTB-IIR with a signal strength so it appears I have an issue with noise/bad module/something else.

So, I took the XTBM to the basement and turned EVERYTHING off.  I then turned on the breaker for the XTB-IIR and the closest outlet.  Perfect 9.99 with 0.0 noise.  I went through and found that one circuit seems to be problematic intermittently.  I took the XTBM up to that room and unplugged everything and still had a problem.  But, then I realized that that circuit also feeds my attic fan on that end of the house.  I confirmed that SOMETHING in that setup is the issue, but what?

My fan is not simple, but I think there are 3 components that are possible culprits.  I installed an Arduino Uno with temperature and humidity sensors to control when the fan is needed.  The Arduino is powered by a wall wart and commands the fan on via a PSC04 one-way transmitter.  The fan is hooked up to 3 switches:

• WS13A
• temperature switch (http://www.homedepot.com/p/Master-Flow-Manually-Adjustable-Thermostat-for-Power-Vent-PT6/100098472)
• humidity switch (http://www.homedepot.com/p/Master-Flow-Manually-Adjustable-Humidistat-for-Power-Vents-H1/100023416)

It's a long story why I have 3 and how they are wired, so let's just say that the 2 mechanical switches don't exist cause I know they aren't the issue.  I believe that the problem is related to either the Arduino PSU, the PSC04, or the WS13A.

To test, I unplugged the PSC04 and PSU and pulled the load wire from the WS13A.  I think after plugging/unplugging/combining these modules, I can say that the PSU is likely not the problem.  But, I can't really say that either the transmitter or switch are the problem.  I suppose if the transmitter were at fault, it could be related to the Arduino commanding things, but I haven't investigated this yet.

I'm curious if anyone can offer any advice/insight into either of these.  When I see the issue, the XTBM does not get a response from the XTB-IIR with signal strength and I tend to see "HIGH NOISE LEVEL" and "BSC..." messages.  I assume that the delayed response when turning on lights is the XTB-IIR waiting for the flood of noise to subside before transmitting the command.

[Tuicemen]

I would suspect the fan it's self or the Arduino power supply
I have seen power supplies go noisy and fans also.

[bkenobi]

My first thought when I realized it was related to the attic was the PSU, but then I realized that the environment up there gets hotter and colder as well as more humid than inside where these modules are designed to reside.  That said, I assume people in HI might use X10 and they don't even have doors on some structures!  But, it looks like the switch functions just fine when tested without load and even with load it clicks away as expected.  I realized that I never tested with everything plugged in with the RJ11 cable unplugged.  If the Arduino is the culprit, it could be sending garbled commands, right?  But then if that were the case, I should be seeing the PSC04 LED flash on transmit.

[Noam]

Could it be possible that the PSU is quiet initially, but starts to get noisy the longer it is on? How long did you wait on your PSU test before determining that wasn't the likely culprit?
What did the noise levels on that circuit look like when you removed each component?

[bkenobi]

My noise was never really an issue when I tested each circuit (breaker).  It appeared that I was at near zero noise when I was at the basement outlet.  When things stopped working, I didn't see all that much noise either.  It seemed to fluctuate between 0.2 and 0.3 (I think it's 1 digit on the XTBM, if not, those are the last digit).

When I was testing in the attic, I didn't see any noise with any configuration, but I was just plugging the module in, waiting a second, then plugging in the XTBM.  Perhaps you are right that it takes a few minutes to kick in when it becomes a problem.  I didn't notice a problem this morning with everything hooked up/plugged in.  Next time I see the issue, I'll try going back to the attic and trying different durations with the PSU et al.

[bkenobi]

When I got home, everything was working as normal.  I went about my chores and this evening found that the lights didn't work.  When I checked the attic, I found that the PSC04 had it's LED constantly illuminated. I unplugged the PSU and nothing changed.  I put it back and unplugged the PSC04 and the XTBM saw the XTB-IIR again.  I plugged the PSC04 back in and the LED lit up again.  with the PSC04 plugged in, I see 0.27-0.43 noise!  I suspect strongly that I have found the culprit but I don't know why it failed.

[Brian H]

You may want to disconnect the equipment using the PSC04 to send X10 power line signals.
It could be holding the transmit keying signal enabled.

There is a schematic of the PSC04/PL513 and PSC05/TW523 near the end of the Technical Manual.
http://www.authinx.com/manuals/X10/technicalnote.pdf

[bkenobi]

The manual didn't help all that much, but I did a quick search on PL513 to see how the LED should act.  It appears that it should be off when not doing anything and turn on ONLY when transmitting (blink on when it transmits).  At this point with nothing connected to it, the PSC04/PL513 turns the led on and it remains solid when I plug it in.  I suspect that it may be stuck in a transmit mode of some kind.  If so, is there any more debugging that could be done to possibly bring it back to life?  I'll go ahead and pop the cover off to see if I see burned traces or an exploded cap or something.  Other thoughts?

If I can't get it working again, my options are to:

• replace the module and move on
• buy an Arduino network adapter (wifi or wired) and use MQTT to transmit temp/humidity to HomeGenie and let it trigger the fans
• add wireless temp/humidity sensor to each attic and eliminate the Arduino in favor of HG controlling the fan

It would be quicker, easier, and cheaper to go with a new PSC04 but if it's likely to fail down the road I might as well go with a different solution now.  This setup lasted around 3 years, so it's not all that big a deal to replace at that rate, but I like robust solutions which this may not be.  I like having the Arduino controlling the fan since it's separate from my HA setup and having the fan X10 controlled allows for override situations.

[dhouston]

If you look at the schematic, the 120kHz oscillator and the LED are essentially operating in parallel so I'd suspect the optocoupler which is common to both.

That said, I do not recall seeing many complaints about PL513/PSC04 failures over the years so I suggest replacing it.

[bkenobi]

I have 2 on order from ebay so I'll have a backup.  I didn't get a chance to open the faulty one up, but I'll post pictures if there's noticeable damage.  If it's a single faulty component (such as the optocoupler) is that something that can be replaced to put it in the box as a 2nd backup?  Most X10 modules are through-hole so I would think I could replace just about anything except for the logic chip(s).

[dhouston]

If it's a single faulty component (such as the optocoupler) is that something that can be replaced to put it in the box as a 2nd backup? 

Probably - but it's been awhile since I've seen its innards so the optocoupler might be SMD.

[bkenobi]

I don't have the module with me.  Does anyone know the FCC ID on it?  Perhaps they would have published photos that I could take a look at.  If not, I'll look this evening at the actual module.

[Brian H]

The PSC04 didn't have any RF transmitter in it.
Most likely will not have an FCC number or file in the FCC database.

[bkenobi]

Ok, wasn't sure.  I thought any time it transmitted and could have interference it could have an ID.  I'll just look in a couple hours.

[bkenobi]

Nothing obviously damaged.