I think this hits the nub of the problem – the conditions are dictating the nature of the response. I’ll look forward to how your analysis gives any insight to deployment strategies.

Aside from the top segment which has a new reference number, I believe the order in which they are reporting might have altered. When copying data from the website it may be worth checking the column references to make sure you’re plotting like with like.

The torrential rain, aside from refilling our reservoirs (or is it the wrong type of rain ) should provide interesting data over the next couple of weeks.

Vincent van Walt

POSITIVE: There haven’t been any of the periodic skips of a few hours now for 2 months since I slowed the system scan rate down, so it looks like Campbell’s advice has worked. I’m optimistic that we’ve seen the end of that particular problem. Campbell have tested the table fill time change and concluded that it does indeed change with the varying scan rate. This seems to be because the table fill time estimate is calculated using the system scan rate rather than the amount of data being stored (doesn’t seem like the best system to me but at least now it makes sense). In some ways this explains our unease about the card fill times and conflated data earlier in the project as the real fill time is years! I don’t think I’ll be leaving it that long in case there’s any more problems but its nice to know for the future. As an aside, the battery life of the stations is impressive and they seem to be holding up, even as we enter winter.

NEGATIVE: The logger which the watchdog error is still causing problems. It had collected only two hours of data then stopped running. Connecting using an RS232 and my laptop was impossible, either using the connect program on loggernet or the device configuration manager. The logger seemed to answer the initial handshake but was unable to load the logger settings. This left me with very little choice but to de-wire the logger and bring it back with me to Birmingham. Once in the lab it took another half a dozen attempts to get the logger to communicate and once contact was made, it seemed to have reset. It seems almost certain that this particular logger has a hardware error so I’ve sent it back to Campbell to be bench tested. Fortunately, the logger I ordered for the lab turned up on Friday, and it was used to replace the missing logger at the start of the week. Hopefully this is just a temporary measure until my other one is fixed or replaced, but it should stop us losing any more data.