[HarshReality]

ok, question is.... when will it show up? each time i bring up the chnagelog it shows last alteration as 7 days ago (most recent). unless the devs arent bothering with the changelog anymore or is a new sf issue

[elupus]

sf won't reenable sync to anoncvs untill tomorrow friday since their major hardware breakage.

[pablot]

this is now in cvs.

sweet, then we can get some real testing done on this

if there is any other similar stuff you want me to look in too just tell me.

[Serg]

no, guys, i don't believe it works correctly. sat away my 1.6 box for a day, 22c room temterature, then turn it on and boot directly into xbmc. info shows something like 50-55c. i really do not believe cpu temp could increase by 30 degrees in 15 secs. yes, use arctic cold plus 2 additional fans, cvs_2006-05-13

[pablot]

the absolute first value of the temp sensor will be wrong, i know that. and the calibration of the temperatures interpreted from the temperature sensor might not be correct either, but it is about right. we use the right sensor just that its not calibrated. to calibrate it we would have to find some way to get the real value and compare it to the value we read out.

at least this is better than nothing at all like it was before.

[Serg]

yes, looks like it just not calibrated, that i want to tell
ok, let me try to borrow some kind of electronic thermometer or
ir one, then at least it will be possible to measure heatsink temperature

will be posted if i could find the thermometer

[pablot]

awsome! thanks.

[Serg]

hey pablot,

all i can do is just take my hat off for you. u r awesome :) got ir therm. from radioshack today
(http://www.radioshack.com/product/index....age=search).
it's not super-pooper, but we have to make some measurements here :)
first of all, all readings i took with all fans off, dvd and hdd placed aside. i've tried to measure temp w/o cpu heatsink, but it went up too fast to 85c (about 5-7secs) i could'nt boot into xbmc to check readings :( then box turns off by himself :) good thing by the way. ok, here is the results (but do remember, readings was taken in hottest spot on heatsink, that means you were right and i was wrong. ok, here it is - separate gpu and cpu:
(xbmc - xbmc reading, ir - reading from radioshack's ir thermometer)

cpu first:

xbmc ir
54.52 49.8
58.14 52.0
57.91 53.6
62.70 59.2
64.18 60.0
64.68 61.0
70.70 66.4
73.30 69.8
74.04 71.6
76.01 74.2
77.20 72.4
77.85 74.4
79.21 75.4
80.00 76.8

now, gpu:

xbmc ir
52.00 58.0
53.16 58.04
58.04 60.6
59.20 60.8
60.80 62.4
62.40 62.4
63.20 63.2
67.20 66.4
68.80 63.8
69.60 65.8

as you can see, cpu temp. diff. is about 5 degrees, i would assume this is the temp.diff. betwen real cpu temp. and heatsink "floor", should it be like that :?

now, take a look at gpu readings. the diff. tends to be zero at about 63c,
then something weird happened :) i do believe that is somehow connected to the "needle" design of gpu heatsink (imho again)

but as i can say, u were absolutely right not only on speculations but in calculations as well, man!!! great job!!!

thank u very much, man! apologizze if i was pain in *** for you :)

best wishes,
serg

p.s. btw, more temp. reading on 1.6 motherboard:
nvidia mcpx x3 chip - 58c
xcalibur - 48c
memory chip - 45c
hdd temp. reading are almost equal to xbmc reading

and one more - let me know if you nee any temp. reading on 1.6 motherboard, i have to return ir back in 3 days :( i wan't loose $50 :)

[pablot]

okay, awsome. thanks!

question is if the heatsink temperature is the same as the "real" temperature. i have no idea about that one since i don't have much experience about it.

the reason that the gpu tempearture is strange is because someone found the temperature to be to high on a 1.6, so then he lowered the temperature by multiplying it with 0.8
i just did some calculations my self based on the temperatures i got from your read out and according to them the read value from xbmc minus the "real" value differs with a k value of 0.8 so that calibration is "correct"

so, should any of the temperatures be adjusted?
i have no idea. you guys decide.

you can safely return that ir-sensor now. thanks for the help!

/pablot

[Serg]

cpu temperature should be higher then heatsink, imho, but have no ideas
what the difference could be. it should depend on heatsink design, material, and so on. i'm not a specialist here as well :) one thing i got - it's not a big deal if cpu runs at 50-55c, because as you can see, other chips runs hot about this temp., but w/o heatsink (like mcpx chip). in my tests (all fans off) cpu temp went up to 80c and box still running, well, of cource, it's too much for "normal" env.

i would like not to change your temp. numbers, just keep in mind heatsink 5c hotter :)

best wishes, guys.