dc-12 i2c error

[rockhead]

I just got the chance to shovel 1.3.9 onto a 12-250 DC I have , oddly it continues to claim that an AF24 runs on ~25 watts and the log has this in it :
Dec 31 16:00:33 UI: i2c error setting 0x47 12 110
<snip STP messages >
Dec 31 16:00:34 UI: i2c error setting 0x47 14 122

[sirhc]

You may need to power cycle the unit. Power it off for 20 seconds and then back on.

This happens every once and awhile with a DC switch firmware upgrade. It is an event that is infrequent and hard to nail down as to why. We can do 200 upgrades in the LAB and it never happens but when you have thousands of units in the field someone is going to see it every now and then. All I can say is we are working to improve that, sorry.

As far as the current reading being wrong and if that condition remains after the reboot that is often from a current sensor being damaged from poor grounding. I have explained this many times on the forums and point people to these 2 post about grounding.
viewtopic.php?f=30&t=188
viewtopic.php?f=30&t=1429


I also discuss blown current sensors on the RMA post:
viewtopic.php?f=6&t=1259

There are several threads on this forums discussing blow current sensors and how they happen and how to prevent them.

If the tower and service ground are not properly bonded and there is not a dedicated earth ground wire for surges and static charges built up around the antennas or ESD to travel instead of the Ethernet cable and is of less resistance and length than the Ethernet cable it can expose the current sensor to greater than 80V then the current sensor can be damaged.

We are working on a design change now to try and combat this but the we are just masking the issue. If you are getting blown current sensors then your grounding system is insufficient. Its a 2 edge sword, if we mask the issue then you may take greater damage down the road but if do not people will lose current sensors. Keep in mind this 80V limit of the component is positive or negative current limit.