X10 Community Forum
🖥️ActiveHome Pro => ActiveHome Pro General => Software Problems & Bugs => Topic started by: david k on April 10, 2005, 11:42:21 AM
-
X10 Pro
Since unplugging the baby CM15A from its
mommy PC last night it gained 1.3 hours in
just less than 12 hours. I guess you want
us to buy a PC for these things and keep
them running 24/7 so baby will work. I'm
not going to do that, as other options cost
considerably less.
If the upcoming release does not fix the
clock issue I will be returning your baby
to you for my money back as "it has never
worked like it is supposed to do" (RMA
time).
I really want to keep the device so I
sincerely hope you fix this now or I will
be putting my money on something that
works. I am tired of waiting any longer.
Best of luck, and make it soon!
-
David: Sorry, software will not fix this
problem. It is a known hardware bug where
powerline noise can cause the clock to run
fast. When connected to the computer, the
computer will keep resetting the clock.
-
David K: I can swap your interface for a
diagnostic unit and see if that helps, if
you're willing.
-
X10Pro
If you could acknowlege or denigh Roger H.
post below about the clock it would be
enough for me to make up my mind....
I have even tried filtering the unit and it
makes no difference. Hours off after just
24 hours on its own. Can't live with it
like this at all. Can the clock be fixed
or not, just let us all know.
Thanks
-
David: There isn't a hardware bug in the
clock, though there are (fairly rare)
circumstances where powerline noise can
affect the clock. I believe you have an
interface I sent to you in the beginning of
February, right? Can you tell me what the
sticker(s) on the back of it say?
-
X10Pro
As you stated in an email to me concerning
this posting, I may have (without knowing
it) one of the first diagnostic units with
an error that makes the clock worse than
normal. I will await the (new repaired)
diagnostic unit to see if it is even close
to acceptable. I surely understand that I
had a bad unit. You know it would have
been nice if someone who knew I had a
defective unit had told me earlier.
We'll see how it goes this round and
hopefully the new release and unit will
clear up all my problems.
Thanks
-
X10 Pro: I guess you are right. It isn't
a bug. It was intentionally designed not
to filter out noise to prevent clock
errors. I was wrong in assuming that the
designers wanted it to have an accurate
clock.
-
BTW. The schematic for the CM15A is
published on the FCC's website (https://gullfoss2.fcc.gov/cgi-bin/ws.exe/prod/oet/forms/reports/Search_Form.hts?mode=Edit&form=Exhibits&application_id=476706&fcc_id=B4SCM15A).
TR10 (in the top left corner) is the
transistor that is used to convert the 60Hz
power into a 60Hz pulse. You will notice
that the base is connected to the neutral
wire through a 100K resistor and the
emitter is connected to the hot wire. The
transistor is biased as a switch, meaning
that for most of the time it is in either
saturation or cut-off (creating a high or a
low). If there is noise during the
operating region of the transistor, a
glitch could occur, causing the clock to
count twice.
In my opinion this is a design flaw, but
what do I know. I didn't write the design
specs.
-
X10 Pro
I wonder how much you know your hardware?
Roger H, is absolutely correct in stating
flaws in the clock circuit, else this
flawed circuit is identical to the TW523
interface and not on the CM11a clock, in
the TW523 is used only as a “zero crossing
detector” this circuit it’s not good enough
for precision clock.
Just to keep you informed the
Microcontroller used on the CM15a CY7C63723
has a faulty GPIO interrupt(there are false
interrupt triggered), there is NO SILICON
FIX PLANNED by Cypress, probably because
they are migrating the encore parts to the
PSOC line.
*************************
• SCOPE OF IMPACT
The chip enters the GPIO Interrupt Service
Routine (ISR) in error.
• WORKAROUND
Workarounds will need to be tailored to
individual applications based on the
flexibility of changing the GPIO
usage, the timing of the GPIO interrupt
sources and firmware interrupt latencies.
• FIX STATUS
No silicon fix is planned.
References
1. 38-08022 CY7C63722/23/43 enCoRe™ USB
Combination Low-speed USB & PS/2 Peripheral
Controller
-
OK, I've run into this problem also. Someone mentioned that there was more of a problem in earlier units? (I actually have a newer one that I'm not using. What can be done to correct this? I think someone said that a filter didn't seem to correct the problem (and wouldn't the filter also filter the signals?)
I could keep the unit connected to my computer most of the time. I'm a little worried about power surges getting to my computer (I live in Florida with lots of lightning)? If I plug the unit into an APC UPS, would it filter all of my signals? Also, does the activehome software have to be running in order for the clock to be kept accurate, or does it just have to be plugged in through the USB?
One problem with keeping my unit plugged into the computer is that it wouldn't be centrally located in my home, and some remote RF signals wouldn't be picked up by it. Since I have another unit, can I use one (with no programming?) just to relay RF signals? What should I do to have that happen?
P.S.: I can't believe that they didn't use a better clock....In my opinion, this is a major design flaw.
-
About that zero crossing input, it can be very effectively filtered in software. When an edge comes in you can internally set a timer to lockout more input until it's very close to the zero-cross time, then open it up again. This works very well from experience :). BTW, 60hz is very,very accurate, it is regulated by law so the power companies don't cheat us. Unfortunately the link for the schematic doesn't work so I can't look at the circuit :(. Do we know if the GPIO interrupt functionality is even used? It may just be turned off.
-
logical1;
The APC UPS will filter out all the X10 signals if the interface is on one if it's filtered AC outputs.
My APC BX1000 sucked up over 50% of my X10 signals when its AC Input [interface on outlet not ups] was in the same outlet as the interface. I added an X10 type filter to the UPS AC Input to stop it from sucking up all the X10 signals
Also I would NOT put any X10 interface on an UPS's backup outlet that runs on the inverter when the AC is out. The cm15a maybe OK as it has a small power transformer; but things like a TM751 with a powerline derived power supply; will overheat and maybe emit smoke; on battery.
-
emil;
Is the FCC site down? That is where I found my set of cm15a schematics.
Some have said that adding some bypass caps [there are none] to the microcontroller, serial ram and two power supply filter caps has fixed many problems. Not sure if clock glitches are one of them.
-
Brian:
Which X-10 filter are you using to keep the UPS from filtering out your X-10 signals? What do you use to measure your signal strength? I've got 4 UPS's in my Den, so that might have something to do with the poor performance that I've noticed when I'm operating my CM15A in here.
Thanks,
logical1 (Scott)
-
I am using a Smarthome 1626-10. It is a 10 Amp type as my BX1000's AC requirements where close for most of the 5 amp ones.
ACT makes a nice 15 amp one AF120. If your needs are not that high. Many X10 vendors have the X10Pro XPPF 5 Amp one.
I use the ELK ESM-1 for most of my tests. Also have a Smarthome TesterLinc that has a 2 row LCD readout and connects through a 1132B Serial Interface. It can store; I believe the last 40 X10 messages and gives things like Quality of signal and can give noise pulses in the X10 range per second.
-
Unfortunately the link for the schematic doesn't work so I can't look at the circuit :(.
I fixed the link. It was left over from the old forum interface which didn't like long lines.
-
So, knowing that the clock circuit has potential problems,
and also knowing that the clock will presumably not register
during a power failure, wouldn't it be easier if someone came
up with a low-power crystal controlled clock that could be
piped into the circuit, in place of the on-board kludge?
That way, it would at least keep time when the power was out. . . .
T
-
The CM15A will keep time if the power is out as long as it has batteries in it.
I want to emphasize that while some people see potential for problems in the design, the clock works accurately in practice. We have a number of employees, including my boss, who use the CM15A -- not connected to the PC -- to run their homes. If it didn't work right, I would have people on my case. Of course, our customers are plenty vocal as well. If it weren't running well, it certainly would come up here all the time.
I'm not going to say the CM15A always works perfectly, but I feel very confident that you can use it to reliably run your house.
-
So, knowing that the clock circuit has potential problems,
and also knowing that the clock will presumably not register
during a power failure, wouldn't it be easier if someone came
up with a low-power crystal controlled clock that could be
piped into the circuit, in place of the on-board kludge?
Some have suggested that a dedicated (low power) clock chip be used instead of the existing circuit. This would also have the advantage of making the batteries last much longer, however it would increase the construction cost slightly. Also, now that the CM15A is in production, it would be more difficult (expensive) to make the changes.
-
...there are (fairly rare) circumstances where powerline noise can affect the clock.
I want to emphasize that while some people see potential for problems in the design, the clock works accurately in practice.
X10 Pro: You have admitted yourself that it is more that a "potential" problem. It might not be a common problem. It might even be quite rare, but if it has happened even once it is not a "potential" problem. I had problems with the clock running fast for a short while (about a week) last year and I know of others (through this forum) who have experienced the same thing. Although I don't know conclusively what was causing the problem in my case, I suspect it was noise from an HP DeskJet power supply (it wasn't plugged into any type of noise filter during that time).
-
:DI just found this forum, and boy am I glad...I'm on my 3rd CM15a, and so far, none of the units have had a stable clock! It appears that when left on it's own, the CM15a's clock is running too fast (sound familiar?) so my modules start coming on at strange times.
I'm going to start keeping an eye out for any fixes that come along. Am I correct in understanding that this is a hardware issue on the unit, and not a function of a software bug?
Steve
-
:DI just found this forum, and boy am I glad...I'm on my 3rd CM15a, and so far, none of the units have had a stable clock! It appears that when left on it's own, the CM15a's clock is running too fast (sound familiar?) so my modules start coming on at strange times.
I'm going to start keeping an eye out for any fixes that come along. Am I correct in understanding that this is a hardware issue on the unit, and not a function of a software bug?
My best guess is that there is some noise on your powerline that is causing the clock to run fast. I had this for a short while and I think it was power supply for my HP DeskJet that was causing the problem (although I never confirmed this). I now have it plugged into a noise filter and haven't had a problem with the clock since (although it may be a co-incidence). Swapping out the CM15a won't solve the problem as it is a problem with the design (although X10 won't admit it).
-
Very Interesting...After 3 CM15s having the same 'problem', I took a look at commonalities. Same Power, and same softwaer. I'd tried running one of the CM15s on a surge protector, and it didn't help, so I finally got an update (2.03 I think). For some reason, my original software would telll me that an update was available, but it'd never get it! I finally found a link here in these forums for an update.
After installing the update, lo and behold, the timers are working perfectly now! I don't remember which version of sfotware I was running, but this one is MUCH better ;D.
I've seen that there appear to be some issues with time delays in macros running too fast in this version, but I don't have any macros right now, so it isn't an issue for me.
Steve