I was thinking today about gathering data for our N.M.E. project. In order to accurately control our Hi-Speed Needle valve we need to data log temperature to know when the motor is effectively warmed up to start correction, when the head is too hot, too cold, and the exhaust is too hot or too cold. All of these temperatures will remain fairly relative from motor to motor. A learning algorithm would have to be used to implement our controller in many different engines and vehicles. Perhaps a difference function could be used, based off of the gap between head and exhaust temps. For our experiments we will be data logging the temperatures during deliberately rich and lean conditions, also logging our best efforts to properly tune the carburetor. In a controlled environment we will need to simulate race conditions, which would be 4 to 5 minutes at at least 85% of that time at full throttle. We will also need to learn the minimal amount of movement our servo can produce that will also change our air/fuel ratio.
In theory the best performance should happen at High EGT and High Head temps, this would mean that the Air/Fuel is near stoich. This should give us efficiency, but it is important to keep the cylinder head temps under control. It is important to note that 2 cycle engines use fuel mixed with oil to lubricate the engines. This means that as 2 stroke engine goes lean it also is low on oil, which causes premature wear on the bearings within the motor.
| Tuning Tactics | LOW EGT | HIGH EGT |
| HIGH CHT | Too Lean | OK |
| LOW CHT | Too Rich | N/A |
No comments:
Post a Comment