View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0000276 | SpeedFan | Fan control | public | 2005-05-25 23:13 | 2005-05-25 23:35 |
| Reporter | Single | Assigned To | alfredo | ||
| Priority | normal | Severity | feature | Reproducibility | N/A |
| Status | assigned | Resolution | open | ||
| Summary | 0000276: Better fan control | ||||
| Description | There is some suggestion to better fan control method. I have some words to say: There are must be an integrating circuit (virtual, of course) instead of /in addiction to preposed lowpass filter to accumulate an error. Also it's almost impossible to set right timings and coefficients to integration and lowpass circuits, also fan drivers have discrete control, so there are must be a small hysteresis curve to prevent oscillations. Integrating circuit is the only right way to keeping set temperature. This method widely used in electrotechnology. We are already have primitive integrating circuit in speedfan with "delta value" as a abstract timing value. Also i think there are must be an alternative method of driving fan speeds, not to keep temperature, but functional dependence of temperature. And THE BEST WAY will be hybrid of two methods. The sum of integrating circuit and functional dependence of temperature. While integration circuit report "right" emperature difference it report zero value and the sum represent only value of functional dependence of temperature and we have desired speed (and noise) and it is ALWAYS SAFE speed. When temperature exceed desired value than integration circuit begins to accumulate temperature error ang give incremental positive temperature error value that affect fan speed and lowering temperature. It will be better to have integration circuit timing in seconds. | ||||
| Tags | No tags attached. | ||||
| Motherboard Model | |||||
| Video Card Model | |||||
