The PC running my weather station is in my observatory. When the temperature exceeds 32C the computer temp rises above 45C and shuts down because of excessive heat. When the temperature drops and the computer restarts the WD program begins logging new weather data immediately following the last data logged. If there was an 8 hour outage then the data that actually happened 9 hours previous is reported as happening an hour in the past. The history record has been mangled is no longer valid.
This application is very feature rich but seems limited with respect to its history recording. In simpler words, WD is not as reliable as it initially appears.
Could I suggest WD places a timestamp at the time the last set of weather data was recorded. If the next time WD attempts to record a value and an interval longer than <typical_measurement_interval> (maybe 1 min) has elapsed then WD enters <no_value> for all one minute intervals from the last recorded and the current time and begins recording new data from this point on. It isn’t necessary to record a timestamp with every record. A timestamp is only required when the last record was recorded. Each time a new record is added the timestamp is updated.
This ensures that missing data is recorded as <no_value> and not an average of previous values or a repeat of previous (which screws up analysis). If the WD is shutdown for any reason the history data remains absolutely valid and reliable.
Is anyone else impacted by this deficiency?
Can this be fixed?
hi
yes, i see your point, and it is a problem for looking at the trend comparisons, and also that data is not included in the averages/extreme data
you can get wd top put a gap in the default 12 hour graph, via graph setup, graph time delay, yes, but that only works best for shorter periods.
the other thing you can do is pad out the log file with the expected or last data record, and then convert the log file to graphs (under action, import log files, convert wd log files to grpahs), which will then pad out the grpahs and hence the trend analysis
use say wordpad or word to copy/paste and slot in a new section of logfile into the log file (which is say 72003lg.txt for this month)
Brian
I believe I’ve understood your repsonse. In short… the solution is for me to manipulate the data records. If this was an isolated instance then this approach could be acceptable. However, given the frequency I must shutdown the PC due to excessive heat…
I was hoping a coded solution could be provided. Once the coded solution is characterized then users know to what extent they can rely on the data and the recorded history.
I understand you’re busy… but please consider a coded solution.
My spouse believes the cost of an air conditioner for my observatory is an unnecesary cost. When the normal summer air temp goes to 30C or more, the internal observatory temp rises to +35C or higher. Despite my efforts to keep the PC temperature down, the internal observatory heat rises to such an extent the internal PC heat rises above 50C and the PC automatically shuts down.
All this means I have to shut the PC down for 8 to 10 hours on hot days. When WD starts up again it continues to graph data values as if no time was lost. The 24hr graph often shows the day’s peak temp appearing to have occured at midnight. This is why I’ve asked Brian if WD can some how recognize a break in the recording pattern and skip the power down interval.
I understand your request, there are users who can’t run the PC 24 hours and it will help them too. I run an (old) HP P2 24 by 7 for WD but it’s not where we live so this time of year it’s often running in 90F ambient. It’s specified to be OK up to 100F ambient. I’m surprised you are seeing such a rise in/can’t control internal temp, what type of PC is it?
To keep the temp down you could hack some holes in the side and add some more fans, they are low cost and easy to fit or if like me a bit tight you could try running with the case off, it can keep the temp down but tends to be a bit noisy.
Heres a handy programme that puts the current temp on the desktop and if you have the right kit alows you to adjust the fan speeds
Also the shutdown temp maybe settable in the bios, mine is set to around 75C for the processor
Thanks for the info on SpeedFan and on modifying the PC. I am quite handy at modifying or designing improvements for my astronomy hobby. I’ve installed the PC in rather small ventillated enclosure. I’ve also installed a 3 inch vent in the floor with a muffin fan to pull cool air from under the eleveated observatory. This is obviously still not good enough. (A picture of my “observatory” is at http://www.ccdguy.com/RASCv3_files/frame.htm under slide “Inside and Out”)
I’ve looked at installing a peltier thermo-electric cooler on the PC case. This unit would be scavenged from an electric cooler from a local supplier for $70. Again, the cost factor has to be weighed and the correct solution selected.
Wow, that’s cool! (Well, you know what I mean…) What a great project.
Looks like the shed walls are just plain metal as supplied? I’m sure you thought about insulation. I lined my well house with aluminum foil coated styrofoam sheets (Home Depot) and it made a huge difference to the internal temp. Being a solid sheet it is very easy/clean to install, on metal you could probably just use double sided tape.
This has strayed way off-topic but…
The shed is a Royal Outdoor Products double wall vinyl shed. It does remain rather cool inside despite the soaring heat we’ve sen in the past weeks. The shed is vented and I have installed fans but the heat still climbs. Short of installing AC I have no choice but to turn off my PC. Low cost efforts to keep things cooler have not been successful. Other means that require $$$ will have to wait for the next while…
Brian
Dowloaded and tested this out. This appears to be what I was asking for. I’ll monitor it for the next while and ensure I see no residual undesireable effects.
Your response and attentiveness is excellent. Thankyou for an excellent application and superb Best-In-Class support.
This is why I’ve asked Brian if WD can some how recognize a break in the recording pattern and skip the power down interval.