Sunday, February 10, 2013

Items noted from 1st live run

I have been looking over the log files from yesterdays live run and making some corrections to the design.  Overall the biggest area of change is around the RPM sampling, trying to improve it as well as dealing with the transitions as it 1st comes on-line so to speak.   Here is what I have changed and will be testing today:

  • Removed D-17 Phantom Power path.  Going forward the choices for this one and only Low Drive are: place this Diode at the starter relay, leave it out all together and let the FET run 'naked', or assure that the remote starter relay AND the controller will always have (and more important not have) power to them at the same time.
  • Adjustment to Stator RPM sampling hardware:  Am going to remove the voltage divider, this will help someone on stability.  The Series capacitor will also be removed, seeing as I am IRQing on the Rising Edge it does not help with its falling edge acceleration function..  And will add in an appropriate filter cap to form an RC trying to cut down some on  higher frequency noise.  
  • Will bypass fault-checking the engine RPMs during Alternator RAMPing phase.  I noted form the run data it takes time for the alternator RPMs to stabilize   The hardware change above should help, but even so when the PWMs are very low I suspect getting a good lock on the stator with just the Amtels input ckts will be difficult.
  • Added calibration factor for the Amps shunt, mine was under reporting by 7-11%
  • And fine tune several of the throttle control timing values - As the design has no throttle position feedback (it uses measured RPMs to close the loop) this is to be expected.  I would think these will need to be customized for each engine / throttle actuator combination.

Will see how today's runs work.  And on the overall stator derived RPMs, if these do not settle down more early on i am going to remove the RPM sensing method option of determined if the engine has started.  Watching the Starter AMPs draw is working spot on for this, so if the RPMs are unstable it goes.

And on my generator- I need to tighten down the throttle cable clamp a bit more, using a wrench as opposed to a screw driver.  The data looked like the throttle cable was slipping and when I inspected it sure enough, about 1/4" had slipped out.   Glad I did not have to chase down that programming bug around throttle management during alternator ramping and bulk charge phase!


  1. Interesting approaches to determining when the engine has started. With my auto-start project, I decided to measure RPM using a Hall Effect sensor as there was a large gap between starter cranking RPM and minimum engine running RPM.

  2. Hello, and thank you for taking the time to actually read this and your comment.

    Yes, using RPMs is a traditional approach for deciding if an engine has started or not. But given I already had the amp shunt in there (a foundation of the overall project - measuring Amps as well as Volts) it seemed like a great way to decide the engine has picked up. And, to be honest, it is working spot on!

    Also for this project I wanted to sense the stator to synchronize voltage sampling. This gave me a free way to measure RPMs as long as the alternator was charging. Which, for a "generator" is most the time :-) For this application it works well when managing the throttle speed. Though I have noticed a loss of RPM sensing when the charging state 1st enters Float mode. Will work on that a little, perhaps add a little field until the throttle control gets the RPMs down to idle speed. Or, just let it stay where it is for a few minuted until the battery settles down as it transitions out of Acceptance and into Float stage.

    I do have an 'option' in the code of using RPMs to determine if the engine had started and during that time some field current is added when needed. But for my install I will disable that option as the Amp Shunt works so well.

    If I could only depend on RPMs for detecting starting, then yes - using a Hall Effect sensor is a much more stable way to determine RPMs. And I would put it directly on the engine flywheel instead of the alternator. (many diesels have a way to position the sender over the flywheels starter ring teeth - that would be rock solid)

    Thank you again for your comment, do you have you have a page for your project as well?