05-27-2010 04:59 AM
I really appreciate your volunteer spirit! Thanks!
I'll follow up with you later today to work out the logistics and collect a few more details on time until this occurs and some of what you do on the system - just it happen if it just sits, or after you stress it - computational, graphics, general use, wifi use, etc...
05-27-2010 05:50 AM
Not to be a party pooper, but Steve's machine is a U450, not U350 - and these machines are not the same. At least that's what we found out when trying to apply a U350-modified NHC ACPI-module on the U450.
Mark: Is it that hard to adjust the temperature thresholds in BIOS? I would think it would be a piece of cake for developers to just up it from 45 to 50-55. The best would be if us users could set these values ourselves in BIOS.
05-27-2010 06:48 AM
I also have an issue with my U350,
I am actually glad i saw this issue, since I just checked my temp with CPUID.
From this image I am just happy nothing bad has happened while having the laptop in my lap :/
From the Microsoft proof of license on the bottom:
Hoping for a reply.
05-28-2010 05:14 AM
Today i unboxed my u350, very nice maschine if there was not the fan....
From Minute one the fan started blasting, while cpu temp is around 41°C.
No idea what todo, the noise is unbearable, is there a fix incoming, otherwise i have to return it and buy an acer or dell.
too bad, i liked that u350 series.
05-28-2010 07:25 AM
Thanks for returning to this thread. Not sure if you saw my previous post, but do you think it would be possible to at least provide a BIOS that allowed the fan speed and trigger temps to be controlled by software?
Right now there is no interface that software can use to control the fans. I believe this is most likely because the BIOS does not pass this information through to the OS.
05-29-2010 07:22 AM
I think that we are all wanting more or less the same fix, but asking for it in different ways. I made a personal list with priorities:
*** Top of list, really needed to make the U350 bearable ***
1 - Adjust the fan set point to something more forgiving. (example: fan blowing fully only after 47~50ºC)
2 - Make the fan set points adjustable in BIOS. (or at least some "fan profiles" options)
3 - Release a driver or program to configure the fan speed on windows. (some Thinkpad's have drivers like this)
4 - Release a more adjustable BIOS, with processor and bus options for example. The existing one is too much crippled. If Lenovo is worried about the average Joe missing with the settings, make a shortcut to access the advanced bios options.
5 - Change the bios so the clock generator of the U350, SLG8SP513V, can be adjusted in Windows. (using for example the program setFSB)
*** Bottom of list, features that I would like but can be fixed latter ***
I understand if other member priorities are different from mine. In that case, maybe the thread creator could make a poll so we vote on things that need to be addressed faster. Anyway, the fan problem will surely be at the top of the list.
On a side note, I didn't see yet a member from USA wanting to help us all letting Mark have a look at its U350.
Come on guys, lets make this happen!
ps.: If Lenovo Brazil can also investigate the problem, I volunteer my U350. I can also help if there exists any remote diagnostics or reports that can be done on my U350.
05-29-2010 01:42 PM - edited 05-29-2010 06:52 PM
... a few more details on time until this occurs and some of what you do on the system - just it happen if it just sits, or after you stress it - computational, graphics, general use, wifi use, etc...
One more, complaining or surprised by this noise.
I do currently nothing. Processor use is at 0 to 5%, under Linux FC13. I'm using Firefox and only reading this forum since half an hour (no flash, nothing special). System temps seem to be most the time at 44°C, fan at high speed, sometime at 42, fan at lower speed, for a room temp at approx 22 or 23°C.
I've left the keyboard for several minutes, it remains at 42°C, fan being on, CPU load being at 0% (I'm typing now, with the cores most or all the time at 800MHz). Fan is still audible, at low speed.
I now watched a 5 min video, over Wifi, from dailymotion. Temp quickly raised to 52 and remained stable, for a 70% proc load. After this, temp slowly decreases, 47... 46... 46... 46... remained at 46, with the fan blowing plus CPU at 12%. I had to close the dailymotion Firefox tab to reach again 0% CPU; with that tab closed, the fan speed slowed down and the temp remains at 44°C for 0% CPU load (i'm again just typing here).
The temp raises to 54°C for a full screen dailmotion video and a 85% CPU usage. Now the PC is like stucked, at 43°C, 0% CPU load and with the fan on.
I brought the laptop to sleep for 3 minutes. At restart, temp was lowered to 41°C and the fan was off. Immediatly after, due to linux activities at restart, the temp raised again to 45, quickly, and the fan was on again...
I have other laptops that do not behave like this. I think my R61running linux remains fully quiet for a low load - 20/30% CPU - and the same room conditions.
I'll follow this discussion - and may return this dual core su7300 U350 during the comming week. I think I won't be able to use this PC, it would disturb in places I intend to use it.
05-29-2010 05:15 PM - edited 05-29-2010 06:29 PM
I had a first look at its internals... and may keep it. Is this a space for a 3G mini PCI board, antenas being included?
After PC restart, its temp raised to 44°C, almost stable, fan being off (or really quiet), CPU at almost 0%. And what I guessed happened: temp once raised to 45 and the fan was definitively on, already loud.
Temp then lowered to 41°C and the fan remained on, with no noticable speed or fan noise level change compared to the 45°C noise level. From there, it seems that the fan is too slow to reach lower temp than 41°C. CPU is at 0%, temp is stabe at 41°C.
I ran a burn test... One CPU being fully loaded, room still at approx 23°C. Fan speed/noise raised again at 50°C. Then the temp reached 52°C, stable.
Loading also the second CPU, the fan speed quickly raised more, at 53°C, and again at 55°C. I reached 100% load and 58°C. 8 cell battery was at 93%, discharging, 5h30 remaining... and not more than 62°C at full CPU load, since approx 30 minutes, in a 22-23°C room.
To load the CPU, I used this tool (adding "-m32" gcc option inside the makefile, for my 64 bit system):
I would agree with other people, CPU and system could get a little bit hotter, with the fan being noticable only after 50°C or more. And the fan should switch off (or be at minimal speed) as soon as temp goes lower than approx 45°C, for a quiet PC, at almost no activity. Can this be achived by bios update?
Assuming the PC could survive it (I am thinking about summer time, 45°C room or outdoor temperature), with such an hysteresis, this laptop could be much quieter.
PS1: I stopped the two burn tests. Fan slowed down at 48°C, at 46 again... at 44 again... the case under my left hand, close to Ctrl key, is still warm.
After 10 minutes, nothing more happens, the PC seems stucked at 41-42°C, with a 0% CPU load, fan remaining active, loud.
PS2: I did another test, without the RAM/disk trap cover, PC being flat, on a desk. At 100% CPU load, temp raises only to 55°C. Those are 7°C less compared to the previous test. 56°C after 5 minutes CPU stress, in the same room.
I killed the burn tests. 48, a speed change. 46, a change again. 44, one more change. Load is at 0%. It takesa long time untill temp reaches 42°C. 41... I open the window, to lower the room temp. 42°C! I went outdoor, for 15°C "room temp". The PC reached 39°C, and it is now quiet.
PS3: Once again indoor, PC reached quickly 45°C and the fan is now loud again. So, one basic "hardware" option would be to drill more holes into this RAM/disk cover?
PS4: for my own basic usage, the load being at approx 10-15%, the temp is at 45°C, fan being too loud.
05-30-2010 01:35 AM - edited 05-30-2010 01:38 AM
I'm using it again. If possible, another really basic bios option would be to increase the last threshold - at which the fan is in fact turned off - to 42°C. This could preserve the hardware. The cooling system can reach 40-41°C in a 23°C room temperature, this would then switch off the fan in common conditions.
Currently, this threshold for fan cut off seems to be at 39 or 40°C. But even at almost 0% CPU load, the PC never or rarely reaches this temp level after a heat peak.
PS: I'm using lm sensors package for temp measurements.
Adapter: Virtual device
05-30-2010 07:53 AM
If the threshold for fan off is set to 42, then it would alot more on/off-cycling since the threshold for on seems to be @45 degrees. It could be done, but then the threshold for fan on must be set alot higher (~55 degrees). If this was done, my fan would almost never turn on due to passive cooling (I seldom use my machine for extensive tasks, and when watching movies the noise from the fan is eaten by the sound from the movie).