Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Brian H]

You may not see many empty Cxxx places on the CM15As main board.

Most engineers would add a .1uf 50 volt cap across the main power supply caps and ICs for some high frequency bypassing. Not there and I guess never thought of as there are no positions for them.

One user here added four .1uf 50 volt caps by soldering them directly across C11,C12 and U1,U2 VCC to VSS. He did say his over all performance was improved.

You can find many of the X10 devices in the FCC database. Some even have schematics available.
http://kbase.x10.com/wiki/Schematics
Note here. My RF Receiver Daughter Card from the CM15A does not match the schematic on the FCC site.

I suspect the clock maybe getting its timing from the Zero Crossing circuit. You can see how it is done in the main board schematic. Not too sophisticated. Small power supply derived from a tap on the power transformers primary. Switching transistor and an opto isolator to the controller circuit.

[JeffVolp]

I suspect the clock maybe getting its timing from the Zero Crossing circuit. You can see how it is done in the main board schematic. Not too sophisticated. Small power supply derived from a tap on the power transformers primary. Switching transistor and an opto isolator to the controller circuit.

It isn't so much the simplicity of the circuit, it is what is done with that ZCD input in firmware.  If it merely counts transitions, then there is the potential for a glitch near the zero crossing to cause double counting.  If it ignores that input except when a transition is expected, then it would not be sensitive to noise except for it possibly causing the X10 transmission window to shift slightly.

Jeff

[Brian H]

Jeff; Thanks for the clarifications.

[Noam]

I'm not an engineer, but did anyone look into the possibility of adding a real-time clock chip to the CM15A, so that it doesn't have to rely on the accuracy of the 60Hz to run the clock?

[JeffVolp]

I'm not an engineer, but did anyone look into the possibility of adding a real-time clock chip to the CM15A, so that it doesn't have to rely on the accuracy of the 60Hz to run the clock?

An accurate 60Hz reference is needed to position X10 transmissions, so that must come from the powerline itself.

Jeff

[Noam]

I'm not an engineer, but did anyone look into the possibility of adding a real-time clock chip to the CM15A, so that it doesn't have to rely on the accuracy of the 60Hz to run the clock?

An accurate 60Hz reference is needed to position X10 transmissions, so that must come from the powerline itself.

Jeff

Thanks for clarifying. However, could they have designed it so that it uses the 60Hz reference to time the signals correctly, BUT uses the clock chip to keep the actual time to run timers properly?

They use a clock chip to keep the time when the device is unplugged, why not use that all the time?

[JeffVolp]

They use a clock chip to keep the time when the device is unplugged, why not use that all the time?

There is no "clock chip" in the CM15A.  If it does keep time when the power is off, it must be using the microcontroller's internal oscillator.  That does not have the long term accuracy of the 60Hz power grid, which is normally good to seconds a month.  The internal clock in the CY7C63723 is accurate to 1.5% after it adjusts itself from USB communication.  (That's about 20 minutes a day.)

Interesting...   That is similar to the drift quoted at the beginning of this article.  Could it be using the internal clock rather than 60Hz as a reference?

Jeff

[Brian H]

Gee Jeff; That could be a possibility.

[bvw]

It has been very educational following this conversational thread, but it seems that determining the root cause of my CM15A clock problem may be on another tack.
Four days ago I installed AHP (v 3.271) and built a new .ahx on a laptop that had never had a previous AHP install. The purpose was to eliminate any possibility of this being a software issue.
All timers have since triggered as programmed. (The connected laptop has been shut down all this time).
I really didn't expect this result because I had already performed an AHP upgrade a few weeks ago on my primary connected PC, and it had no effect on the CM15A clock issue. Perhaps there is a corrupted remnant of the old install on my primary PC, perhaps there is a conflict with some other app. I will continue to monitor and build out the .ahx on the laptop and try to determine what to do next on the primary PC.
Thank you for all your input so far, and please share any thoughts you may have regarding this latest result.
(I wouldn't declare this issue resolved quite yet).

[JeffVolp]

The interesting thing is that you did this on another computer.  The Cypress Semiconductor data sheet on the microcontroller used in the CM15A talks about it adjusting its internal clock based upon the USB connection.  Perhaps the USB port frequency is a factor in the CM15A clock error problems?

Jeff

[Noam]

The interesting thing is that you did this on another computer...
Jeff

Did you verify that the clock on the old PC is keeping correct time (WITHOUT having to connect to the Internet to sync)?

If the clock on the PC is off, perhaps it is passing that inaccuracy over to the CM15A somehow?

[JeffVolp]

I don't thing the RTC in the PC has any relationship to the USB bus speed.  The RTC is normally a separate chip powered by a small battery.  The USB port would either run off the bus clock or some internal clock in the USB controller itself.

Jeff

[Noam]

Sorry, I guess my post was confusing.
I was under the impression that when connected, the CM15A syncs its clock to the PC clock. (There are a lot of people who complained their timers were off by one hour, and it turns out their PC time zone was incorrect). Perhaps if the PC's clock is off, and if the CM15A is syncing time to the PC, then it is syncing to a bad time reference, and starts to get off track. I'm not an engineer, so I'm really not qualified to do anything other than throw out my own hair-brained theories. ;-)

[Brian H]

Yes I believe the CM15A does set its time if connected to the computer and AHP is run.

[Dan Lawrence]

I believe you are correct.  My CM15A is plugged into my PC 24/7/365 and has been since 2005 so that's been 5 years of service with no interruptions.  I just updated to 3.185 as a clean install (did a total uninstall of AHP (3.271) first before installing 3.185.  Timers and my keypress Macro work as created.