Board reboots and strange behavior

[MobileMe]

I just got a new board so I could send the older one back to cai to be fixed and the firmware upgraded. After installing the new board yesterday, it seems to reboot every couple of hours. I woke up this morning and my living room lights were on and the garage door was up. I setup a program on my computer to monitor if the board is running or not and it stopped running a few times last night.

- 3 temperature sensors with no more than 15 feet of wire to each sensor.
- 1 humidity sensor on the boad itself
- 1 relay board that has its own power source
- 1 contact sensor for the garage door
- 1 moisture sensor that has its own power source
- 1 gas MQ7 gas sensor

I can't tell you how frustrating this is. I made sure I'm not over driving the board to prevent problems. It seems that when I add more than one temperature sensors, the board does some very bad things. I've tried different sensors for all of them.

Any help would be great.

Thanks,
Mike

 

[rossw]

Things I'd be asking.

1. How long is the wire to the door sensor? I would be very reluctant to run a long wire straight to the board because it will certainly pick up lots of spurious. I would use an optoisolator or a reed relay back near the board so you can use totally seperate power supplies and only short input leads.

2. Have you tried adding a decent capacitor across the DC supply to the board? (What voltage is it? Too low and it will certainly cause problems). I find 9V is about perfect, and adding a 1000uF electrolytic cap across the input is certainly beneficial if power is a bit ordinary.

 

[fwd03]

It must be a bug of the new firmware. Which version is that?

 

[CAI_Support]

I got a notice that this is a bug in 3.02.16 and 3.02.16a firmware. If you don't use WEBSET feature, you can save notify configuration first, then save the temperature sensor configuraiton after that, then they should work fine.

We have made a fix for this bug, the firmware version with fix to this bug is 3.02.16b. Sorry for the bug, Anyone with 3.02.16 and 3.02.16a firmware will need to send the board to us for firmware updates to be able to use WEBSET and temperature sensors together.

 

[MobileMe]

I'm disappointed that I have to send the board back to CAI just for a firmware upgrade. I just got this new board, but lucky I spent sometime setting it up so I can change out the board with minimal hassle. In this situation, it would be nice to be able to upgrade the firmware myself. Even if I had to make a jtag cable, it would be worth the effort to reduce downtime. I've had to make a jtag cable before so I could unbrick a router.

 

[CAI_Support]

We are so sorry for the problem. We have worked whole New Year holiday to locate and fix the problem. One of the big issue was the C compiler we used from the Microchip having a bug in it, causing code not operate correctly. We rolled back to the older version compiler that seemed working correctly. For any customer running 3.02.16/16a firmware, sufferred from this hassle, we will provide $5 credit toward your next new board purchase.

 

[MobileMe]

We are so sorry for the problem. We have worked whole New Year holiday to locate and fix the problem. One of the big issue was the C compiler we used from the Microchip having a bug in it, causing code not operate correctly. We rolled back to the older version compiler that seemed working correctly. For any customer running 3.02.16/16a firmware, sufferred from this hassle, we will provide $5 credit toward your next new board purchase.

Instead of a $5 credit, I would like a paid return shipping label, so I can ship the board back to you at no cost to me.

Thank you

 

[CAI_Support]

Just send that board back to us in first class, we will give you a check cover the postage in your return package. Is that okay?
Sorry for the problem.

 

[MobileMe]

Yes, that would be fine. I sent an email to notify you that the board is being mailed out tomorrow.

Thanks,
Mike

 

[CAI_Support]

Thanks Mike.
We apologize this problem caused you trouble. We appreciate your reported the problem to us.

 

[todster]

Would I be correct in guessing that this is what the bug looks like?????

 

[CAI_Support]

Yes.  compiler generated embrassing code that corrupted memory from one function to another. We end up on all the computers to uninstall and reinstall the compiler, that somehow fixed the problem.  The fix is firmware version 3.02.16b

 

[todster]

2 boards are on their way to be updated. 1 is the .16 bug and the other is a .14. Not sure of the fee for the update on the .14 but I stuffed $7  in there with the boards. If it's more, PM with the amount.