MX data/chart resolution

[TRB]

Since upgrading to MX I have been fairly happy, however...

I observed on my graphs that rain minimum increments are mostly displayed as 0.4mm yet my bucket-tipper corresponds to 0.3mm

On old Cumulus the rain graph increments/steps were always 0.3mm so that I would see 0.3, 0.6, 0.9, 1.2 etc. With MX I can usually see 0.4, 0.8, 1.2, 1.6 but every now and then a 0.3 increment occurs.

I figured this was a graphing resolution matter until I noticed these odd values also on the main page i.e. 'Now'.

Further to the above I also noticed on the external temperature graph where it resolves to 0.1 degrees that sometimes it steps as much as 0.4 in 60 seconds and that is during steady weather conditions i.e. not the sun in/out behind clouds.

As I am typing this I realise I have not examined the data files and shall do so later in the day - but I'll post anyway in case others have had similar issues.

[mcrossley]

Interesting - the graph data also has 0.4 as the first increment, so it is coming from Cumulus.

You don't have a rain multiplier set in the configuration do you?

[TRB]

Thanks Mark,

When I posted I was convinced that this issue arose with MX as I can’t recall seeing it with older Cumulus graphs and I usually study rain graphs carefully.

Data files however have proven me wrong. Looking at July 20 data (pre MX) I did see rain increments of 0.4 (rather than 0.3) occurring at times with the odd 0.3 thrown in. I assume the odd 0.3 is there as a cumulative correction?

Yes, I have always used a rain multiplier but it has never been a factor.... until it seems when I re-calibrated around April taking it up to x 1.2 so that must have been the Camel’s straw.

So the multiplier is the answer thanks Mark


Whilst I have you attention, is there an easy way I can have a lost sensor alarm displayed on my uploaded webpage?

[mcrossley]

Ah, 0.3 x 1.2 = 0.36 which rounds to 0.4.

[mcrossley]

You need to use the web tag for the alarm - <#SensorContactLost> and embed that in either in the name of an image file (it has a value 1 or 0), or use it in some JScript to set some text. If you search the forum for examples of using the record set web tag you should find something suitable.

[TRB]

Further to the above I also noticed on the external temperature graph where it resolves to 0.1 degrees that sometimes it steps as much as 0.4 in 60 seconds and that is during steady weather conditions i.e. not the sun in/out behind clouds.

Mark,

I mentioned the above in my first post and now I have a good example, attached.

Note the larger steps around time 05:30 and again around 07:30. They are 0.3 degrees and on most parts of the graph the smallest resolution of 0.1 degrees appears often.

Also the 0.3 steps are abrupt occurring over only 1 minute whereas my graph data is only uploaded every 5 minutes.

[mcrossley]

Well 0.3 degree change in a minute isn't unheard of, but normally you would see it as part a larger rapid rise or fall in temperature.

Also the 0.3 steps are abrupt occurring over only 1 minute whereas my graph data is only uploaded every 5 minutes.

I'm not sure what you mean, the graphs plot a data point every minute.

I'd say the base station is missing some transmissions from your remote sensor. Iirc it takes some time before the sensor lost flag gets set, so it could be missing some messages, then picking up again and the temp has changed a bit.

[cliftonweather]

I was just thinking that it looks like missing data from the sensors due to the long flat period before each jump

[jon_iz]

Perhaps a failing sensor? My console pressure sensor started exhibiting that kind of symptom, and that was the trigger for replacement of the component.

[TRB]

Thank you all and upon carefully examining wind speed and external humidity graphs at the same time interval as temp they seem to confirm a loss of signal from the remote unit.

If I ever suspected a loss of signal I could check that on the old Cumulus with a permanently set fail light.

[mcrossley]

I know its not the same, but you can enable the Data Stopped and Sensor Contact lost alarms, and leave the Dashboard running. It will alarm when the signal is lost.
The log file should contain a record of the events too.

[TRB]

Mark,

I enabled those alarms a few days ago but they are not flags, they are dynamic and off/on/off.

I tested with a temperature alarm and unless I stay glued to the screen I don't know if it ever briefly goes red.

Is there some logging where alarm activity is recorded?

[mcrossley]

The alarms are not logged - except for Data Stopped, which appears in the MXdiags log file.

The alarms may be be getting a bit of love in the next release