Weird humidity data

[tjaliwalpa]

Checked my graphs when I got home today and noticed some strange data had been recorded for outdoor humidity displayed on the trends. Consequently the apparent temperature was a mess as well.

I copied a backup of the data from this morning over todays data and reopened Cumulus and the graphed information was rectified.

[steve]

Some sort of problem with the values of humidity that the station was supplying, so Cumulus ignored it, hence the flat line.

Could you let me see the relevant diag log from the time, please? It may not show much as most of the logging that would help is to the debug log.

[tjaliwalpa]

Steve,

These diags should cover the period.

Bob

[steve]

Thanks; I didn't think it would show anything, all of the messages about discarding bad data go to the debug log. I'm looking at revamping the diagnostic logging, as the latest version of Delphi that I'm using comes with a much better system.

But the only reason that I can think of for the humidity to flat-line while all other data was OK would be for it to be giving a value above 100, which Cumulus would ignore. The only flaw in this theory is that the values left in the logger were clearly OK. The humidity was falling very quickly at the time, you haven't set a 'spike removal' value, have you?

[tjaliwalpa]

But the only reason that I can think of for the humidity to flat-line while all other data was OK would be for it to be giving a value above 100, which Cumulus would ignore. The only flaw in this theory is that the values left in the logger were clearly OK. The humidity was falling very quickly at the time, you haven't set a 'spike removal' value, have you?

Yes, I've set all my spike removal values to 5. I assumed that would be fine but perhaps it wasn't. What would you suggest they be set at?

Bob

[steve]

Yes, I've set all my spike removal values to 5. I assumed that would be fine but perhaps it wasn't. What would you suggest they be set at?

I can't really say, as I've never had any spikes from my station. I had assumed that people who had suffered from spikes would have an idea of what values to use. I would start with high values, and if that lets spikes through, then lower them a bit. If you've had spikes before, then start with values just below the amount that the value 'jumped'.

[tjaliwalpa]

Steve,

Thanks for that. I was working the other way by starting with a small value and planning to increase it if problems arose. I guess I found that they have for humidity. I increased the value for humidity before leaving for work this morning (to 15) and may well decide to change them again tonight to reflect your approach.

Bob