[sysmon] System Monitoring

I have improved my logging method by including an MSF clock display in the video frame, and have discovered something disturbing about the HDR time reference (I shall start another topic).

At 2GB per video log, I don't particularly want them hanging around on my PC (though I could shove them off to the Hummy!).
 
It's looking like the Humaxtv program is changing the fan on 1 Min. intervals but there is no temp. change (that our sysmontools can see) at that time, however my latest guess is the fan comes on at 55C + Time, Half = 51C, Quarter = 48C and Off = 46C on my Humax at any rate

Code: 
Log Temp. Change each Second
Mon May 14 13:0X:XX BST 2012    054
Mon May 14 13:16:10 BST 2012    055
Fan  Full  13:21:10 Humax Time stamp
Mon May 14 13:22:10 BST 2012    054
Mon May 14 13:25:10 BST 2012    053
Fan  Half  13:27:10 Humax Time stamp
Mon May 14 13:28:09 BST 2012    050
Mon May 14 13:31:09 BST 2012    049
Fan  Qtr.  13:33:10 Humax Time stamp
Mon May 14 13:34:09 BST 2012    047
Fan  Off  13:35:09 Humax Time stamp
Mon May 14 13:37:09 BST 2012    046
Mon May 14 13:40:09 BST 2012    047
Mon May 14 13:46:10 BST 2012    048
Mon May 14 13:49:09 BST 2012    049
 
My last run had everything bang on 3 mins (fan on 3 mins after the logged temp reached 55, temp dropped to 54 3 mins after that...). Still trying to get hard evidence of the 1 minute subdivision - but the point is that the fan does not seem to turn on at a fixed delay after 55, so it must be responding to something we can't see. Ditto at the point where the fan turns off - no corresponding temperature transition.

Am I correct in saying that sysmon is interrogating the HDD SMART data directly? If so, and if we are seeing less data than the HDR firmware is seeing, there must be some kind of back door into the temp sensor - and it seems a distinct possibility that this back door might not be the same (or exist at all) on other models of drive. This calls into question the viability of installing alternative hard drives.
 
Black Hole said:
Am I correct in saying that sysmon is interrogating the HDD SMART data directly?
Click to expand...

The humax appears to be running its own sysmontools type package, there is the smartd (smart daemon) command smartd -i 60 that should set the update interval to 60 seconds, but having set that earlyer Today, I still got changes every 3 Mins. or more
 
All todays runs have hit 3 min quanta (just my luck). The timebase has drifted by 4 secs over 8 hours though.
 
Ha! Here's one (phew!):
FanSpeed2.png
The fan timings have been corrected to system time to plot in proper alignment, note that they are on a 4 minute quantum independent of the SMART steps.

I regard it as very odd that this is the second instance I have observed on 4 minute quanta, and all the others were on 3 minute quanta. If we hypothesis it is actually a 1 minute quantum, why have I never seen a mixture in one run? Strange.

I realise you all think this is obsessive in the extreme, but I'm trying to work out what the mechanism is. Where is the data coming from to run the fan? - it may well be the same source as is used for the SMART data, but it is being obtained independently of the SMART table.
 
Now af has the technology to actually read the fan setting instead of having to assume its trip points based on the temperature data, the sysmon package could be updated to add real fan data to the temperature plots.
 
Black Hole said:
Now af has the technology to actually read the fan setting instead of having to assume its trip points based on the temperature data, the sysmon package could be updated to add real fan data to the temperature plots.
Click to expand...
Sysmon 1.1.1 does now correctly display the 'Full' 'Half' 'Quarter' and 'Off' (or minimum setting) phases of the fan, though not the minimum value. Was you suggesting more / different info. ?. It makes more sense to display the effect the fan has on temperature than to display the on-off ratios sent to the fan
 
You know damn well I was talking about adding info not replacing existing info (the use of the word "add" signals that). AFAIK the existing fan annotations are deduced not measured live, so the only actual data is the HDD temperature. There could easily* be another variable plotted on the same axes representing the fan control signal.

* By "easily" I mean it would be straightforward in a spreadsheet graph - how easy it is using the WebIF graphics libraries is another matter.

This post adds little or nothing to my post 132, just spoon-feeds it.
 
While I'm looking at sysmon, I don't understand why the colour boundaries are ramped. There seems no logic in it - why not straight-forward vertical boundaries?
 
Black Hole said:
You know damn well I was talking about adding info not replacing existing info . . .
Click to expand...
You have overreacted here, af123 has stated, (In this link) that the fan turns on at 55, goes to a lower level at 54, a further lower level at 51 and off (or minimum) at 49 Deg. C and has placed this info. into sysmon. My question was simply what would you add / change
 
Black Hole said:
Now af has the technology to actually read the fan setting instead of having to assume its trip points based on the temperature data, the sysmon package could be updated to add real fan data to the temperature plots.
Click to expand...
It /could/ but I don't see any real benefit in doing that. The fan speeds shown on the graph do now correspond to the exact thresholds used by he Humax software (with the exception of minimum speed/fan off).
 
Black Hole said:
While I'm looking at sysmon, I don't understand why the colour boundaries are ramped. There seems no logic in it - why not straight-forward vertical boundaries?
Click to expand...
It's an artifact of the graphing package and the way I'm using it (the different colours are actually stacked, just that only one is ever non-zero now) - I made some quick changes to just get it working. if I were to reimplement it there might be a better option that would make the lines vertical.

On the other hand, the fans don't spin up down immediately so I quite like the slight lean : )
 
You must log in or register to reply here.
Back
Top