Probably similar to this thread, the T530 I recently received has fan noise issues. I've invested considerable time in solving this, but found that a solution can be achieved with reasonable effort only by Lenovo:
Introduction:
Most ThinkPad models do not provide the operating system (or the user) the ability to set the fan speed directly (e.g. in rounds per minute) or in a fine-grained way (e.g. as a percentage of maximum speed). Instead, there are "fan speed levels", usually 9 levels (level 0 through level 7 as well as the level "disengaged"). For each level, a particular fan speed is defined and the Embedded Controller tries to achieve this particular fan speed by measuring the fan speed and adjusting the electrical properties of the fan (e.g. voltage or pulse length or something like this).
The fan noise problem:
Every time the internal temperature sensor reaches 50 °C or so, the fan starts with about 1900 rounds/minute and then quickly goes to about 2737 rounds/minute, where it is loud. Then it takes about 20 seconds, the temperature drops to 45 °C, then the fan is switched off. Then, after maybe other 30 seconds, the cycle repeats and the fan is switched on again, ...
There are major problems with this behaviour, because:
- The fan is not running continously, but pulsating.
- The fan is running at level 1 at a much higher speed than needed, creating unnecessary noise.
I have measured the individual fan speeds for individual fan speed levels:
The fan speeds at different levels are as follows:
- Level 0: 0 rounds/minute
- Level 1: 2737 rounds/minute
- Level 2: 2944 rounds/minute
- Level 3: 3033 rounds/minute
- Level 4: 3074 rounds/minute
- Level 5: 3597 rounds/minute
- Level 6: 3773 rounds/minute
- Level 7: 3773 rounds/minute
- Level disengaged: 4754 rounds/minute
Solution draft:
A minimum solution to the problem above would be introducing another fan speed level between 0 rounds/minute and 2737 rounds/minute, with about 1900 rounds/minute. The reason this makes a difference is that there is a fundamental difference between laminar flow and turbulent flow, where the laminar flow makes nearly no noise and the turbulent flow makes a lot of noise. (Whether an airflow is laminar or turbulent is easiest influencable by affecting its speed.)
As you can see on the observed mapping from fan speed level to fan speed, level 6 and level 7 map to the same speed.
Thus, no particular fan speed would be lost if the mapping would be changed to the following:
- Level 0: 0 rounds/minute
- Level 1: 1900 rounds/minute
- Level 2: 2737 rounds/minute
- Level 3: 2944 rounds/minute
- Level 4: 3033 rounds/minute
- Level 5: 3074 rounds/minute
- Level 6: 3597 rounds/minute
- Level 7: 3773 rounds/minute
- Level disengaged: 4754 rounds/minute
Solution implementation by user is impossible:
The problem, however, is that, while a user can decide which fan speed level should be active at a particular time, the user cannot decide which fan speed should actually be achieved for a particular level. Thus, the user can either decide to have 0 rounds/minute (which is not particularly sustainable, because the equipment becomes hotter and hotter) or to have 2737 rounds/minute (which is not particularly sustainable either, because both the user and other people arund the user's Lenovo computer are distracted from work).
The mapping from fan speed level to fan speed is actually stored in the Embedded Controller Programm ("ECP"). This Embedded Controller Programm is freely changeable or configurable by the user. Only Lenovo has the source code to this Embedded Controller Programm, and actually only Lenovo may know how to disassemble such a binary Embedded Controller Program in case the user wanted to change it. Thus, even advanced users cannot fix this problem by fixing it where it should be fixed.
It should be noted that, because fan speed is a long lasting issue at Lenovo, a third party has been even developed an after-market hardware solution: TPFC. However, this solution probably voids the warranty, and it work arounds the problem at the wrong place, and it costs time and money and effort to actually install it.
Only Lenovo can fix this problem:
How Lenovo should fix the problem
Thus, this is my call to Lenovo to fix this problem (at least for the T530 series, but probably W530, T430 are also affected). Ideally, the solution should be as follows:
Change the Embdedded Controller Program such that the mapping from fan speed level to fan speed is as follows:
- Level 0: 0 rounds/minute (no noise level)
- Level 1: 1500 rounds/minute (low noise level, sweet spot)
- Level 2: 1700 rounds/minute (low noise level, sweet spot)
- Level 3: 1900 rounds/minute (low noise level, sweet spot)
- Level 4: 2100 rounds/minute (some noise level)
- Level 5: 2600 rounds/minute (high noise level)
- Level 6: 3100 rounds/minute (high noise level)
- Level 7: 3773 rounds/minute (high noise level)
- Level disengaged: 4754 rounds/minute (high noise level)
The reasons for this fan speed design are as follows:
- The "sweet spot" where both the fan will run most of the time and where the user would not complain is between 1500 rounds/minute and 2100 rounds/minute. If the fan needs to be faster to keep the system cooled, then the user is likely to run the system with considerable load to CPU or GPU. In this case, there is no option to reduce the fan speed, but also in this case, the user probably accepts the fan noise for the computing power delivered.
- The "sweet spot" in terms of CPU load is where the users do not accept fan noise while actually generating only minute CPU load (e.g. during web browsing or writing documents, where the CPU and the GPU is mostly idle). For many users, this is their normal working state (e.g. working in a quiet office, a very quiet library, a quiet bedroom, a quiet boardroom, ...). These scenarios are where Lenovo ThinkPads currently fail to deliver, unnecessarily
- The "sweet spot" needs higher resolution (more fan levels allocated to these fan speeds), because slightly different load patterns may make it necessary to go from one level to the next higher level, but then the next level still should be bearable. Conversely, the higher levels (e.g. from 2600 rounds/minute on) do not need such a high resolution, because for most of the users, these levels will be only attained rarely anyway, and then they are accepting the noise.
Currently, there is no fan speed level at the sweet spot. With the proposed fix, there would be 3 fan speed levels at the sweet spot, with a neighbouring fan speed level (2100 rounds/minute) still being bearable when considering the performance delivered. - There is no fan speed level below 1500 rounds/minute. This is ok, because the fan is nearly inaudible at this speed anyway, and below a certain speed, there is a risk (depending on the fan) for the fan to actually not rotate at all (e.g. due to friction).
How much time and effort would Lenovo need for fixing the problem:
The Lenovo engineer responsible for the Embedded Controller Program of the T530 should need less than one day to change the source code of the Embedded Controller Program and to recompile it. The engineer may need one additional day to test this program (e.g. for different types of fans), and there may be a need for an additional day to release the update to the BIOS and ECP. All in all, Lenovo should not need more than 3 engineer-days (maybe worth 1800 USD) to fix the problem.
Why Lenovo should fix the problem:
Probably the whole current T series (T530,T520,T430,T420) as well as the current W series (W530,W520) fails the users in any of these scenarios (among others):
- User working in a quiet office.
- User working in a quiet library.
- User working in a quiet bedroom.
- User working in a quiet boardroom.
Thus, the Lenovo T series is currently not usable for many business users and almost all students. This problem should cost millions of USD of revenue, while fixing it would cost it maybe 1800 USD.
Incidental remark: Lenovo subsidiary can do better:
Lenovo has bought the majority of the Germany consumer electronics company Medion. Medion currently offers a consumer level notebook "E7222" for 499€ including German VAT. Despite of this low price, this notebook delivers very good fan control (that is: the notebook is nearly silent when the CPU is near-idle). (However, being consumer-grade, it does not offer a docking-station or a FullHD screen or WWAN.)
What will I personally do:
Some weeks ago, I had a powerful W530. I returned it due to fan noise for cash refund. Now I have the T530 (Serial number: R9RRF26) which has exactly the same problem. The deadline until I can legally request a cash refund runs until 2012-09-07. I hereby ask Lenovo:
- To fix the problem and release an ECP or BIOS update until 2012-09-06 (it is doable, see above) or
- if the problem cannot be fixed in this time (e.g. due to bureaucracy): provide me with a written statement (e.g. e-mail or forum post is sufficient) that my legal cash refund deadline (German law: §355 BGB) will be voluntarily extended by Lenovo (to some date Lenovo thinks it needs to fix the problem).
If none of the above happens, I will have to return the T530 as well. This is the 2nd chance Lenovo has gotten from me, there will be no 3rd chance. I would have to turn to Fujitsu (they do have business-class machines with big batteries, FullHD, WWAN) and they probably have better fan control.
spidey101
4
Jack_Lenovo1
3