1. Welcome to the new SAOCA website. Already a member? Simply click Log In/Sign Up up and to the right and use your same username and password from the old site. If you've forgotten your password, please send an email to membership@sunbeamalpine.org for assistance.

    If you're new here, click Log In/Sign Up and enter your information. We'll approve your account as quickly as possible, typically in about 24 hours. If it takes longer, you were probably caught in our spam/scam filter.


    Dismiss Notice

MegaSquirt (MS3) project

Discussion in 'Modified Alpine' started by jumpinjan, Jan 22, 2011.

  1. jumpinjan

    jumpinjan Bronze Level Sponsor

    This isn't Sunbeam related, but more of an adult science experiment, that would be fun to do and to post updates about it. Don't encourage me too much, because I don't want to spend all my time writing tutorials in detail about what I have done here. I'll just post some highlights of where I am in the experiment from time to time. I'll give you some pointers too of where I learned to avoid the pitfalls. I have been building microcomputers since 1978, starting with my Wintek 6800 microcomputer and I always wanted to build a "micro-controller" to do controlling of machines. I finally have my controller now. I can't wait to hear my engine project running for the first time.
    Here's the MS3 on the bench running (talking to the laptop) after about 8 hours of stuffing parts on a bare board. This was back on 12/8/10

    There is the "JimStim" MS stimulator board kit that provides the tach and other sensor signals to MS to bench test only. I made the aluminum sheetmetal box to support the board.
  2. jumpinjan

    jumpinjan Bronze Level Sponsor

    Here's the engine project (2.5L I6 GM-Daewoo). I bought it back in June 2010, and have been rebuilding it for about 5 months. it had 50k miles on it but was sludged up really bad. There wasn't much wrong with it except the salvage yard stored it outside.
  3. Bill Blue

    Bill Blue Platinum Level Sponsor

    So how are you going to load the engine?

  4. jumpinjan

    jumpinjan Bronze Level Sponsor

    Geez, I haven't even fired one coil yet, although I'm very close to doing that on the bench.
    When I get to the point when the engine is running on the test stand, and then I need to load down the engine to rough in the fuel & sprark tables, I'll devise something. Right now it's all hardware building and learning TunerStudio

  5. jumpinjan

    jumpinjan Bronze Level Sponsor

    I have a coil firing now, using the single output on the v3.0 board. Yes, these COPs are just a plain ignition coil, with NO ignitor. The MS3X (MS3 Extension board) coil drivers are low, 5V levels, so a custom "spark box" is needed to house the 6ea, Bosch BIP373 high level switches (switches the 12v coil voltage to ground or the 4-5ms dwell needed)
  6. RootesRacer

    RootesRacer Donation Time

    Very few COP coils will take 4 to 5ms for any length of time.
    Most run in the 2.5 to 3ms range.

    You should dial back the dwell time and give the coil only what the engine needs. The nice blue spark you get is all for naught if it kills the coil in the process.

    I think you will find those coils to be very weak and unforgiving as far as overdwell is concerned.
  7. jumpinjan

    jumpinjan Bronze Level Sponsor

    Those dwell numbers that you mentioned are right on. I check my MS3 ignition parameters, and it's set to a max of 3ms (cranking is 6ms). I reset the dwell to 2.5ms just to see any difference in spark amount/color.
    The 4-5ms number stuck in my mind from observing the GM HEI's dwell on my O-scope some days ago. (5.6ms start, 4.3ms running)
    I would assume its the heat that breaks down the secondary wire insulation?
  8. RootesRacer

    RootesRacer Donation Time

    Sounds like MS still doesnt have a 2D or 3D dwell tables.

    Dwell interval should be varied with battery voltage and sometimes manifold pressure.
    Not just cranking or running conditions.

    Old school distributed coils (particularly ones with ballast) have higher inductance and if the coil impedance is ballast limited, more dwell time is needed to achieve adequate spark, hence the higher dwell interval as compared to dwell controlled individual coils.

    Heat can be the culprit, but what usually kills the coils is that once coil magnetics is saturated, the magnetic field collapse during spark is so hard and fast that both primary and secondary insulation cannot handle the DV/DT of the voltage rise during the back-emf kick.
    This degrades the insulation and makes coil weaker though still functional for a while.

    In the field, the coils tend to misfire under load once the insulation dielectric is compromised, with greater severity over time until they are replaced.
  9. jumpinjan

    jumpinjan Bronze Level Sponsor

    TunerStudio indicated a 2D battery dwell (6-point) correction curve that I can adjust. Also available, under dwell selections are:
    1) Standard Dwell (this is what I'm using, and can select dwell in tenths mS)
    2) Fixed Duty
    3) Time after spark
    4) Charge at trigger

    Oh, by the way I can't see any difference in 2.5 vs 3ms, a variable spark gap device would be something to try.
  10. RootesRacer

    RootesRacer Donation Time

    The standard dwell and the dwell table are the only valid parameters to change.

    Fixed duty would be if you wanted to take two steps back and run a "points" style dwell percent control.

    Time fater spark insures that at high coil duty cycles, there is sufficient time for the primary to avalanche prior to starting the next dwell cycle.

    Charge at trigger would probably be to trigger an external CD box or the like.

    If the coil energy is sufficient to ionize the gap, you wont see big changes in spark intensity between various dwell times.

    The place you will see if the dwell is sufficient is on a real vehicle under load.
    The dwell needed to idle a car is lower than that needed to fire at wide open throttle.

    I have spent a great deal of time evaluating ignition technologies and various coils. The only way to see the dwell impact on actual energy is to employ a pressurized chamber with a looking site.
    I use CO2 as the gas, and go to over 500PSI, which mimics the conditions of a highly boosted engine. Most coils wont light off a mixture beyond 100PSI CO2, which correlates to about 15 lbs of boost.
    The looking sight shows a yellowing of the arc just before the ignition starts to fail to arc. At this point, you need a pressure based dwell table, or better yet, a 3D dwell table with battery voltage in one axis and manifold pressure in the other.
  11. afssanders

    afssanders Donation Time

    Good stuff guys. Makes me want to get one of these MS and plan around with it.
    That engine will probaly fit into a Sunbeam without to much trouble. Length isn't that bad. Width is another story.
  12. jumpinjan

    jumpinjan Bronze Level Sponsor

    Here's my StimX on my MS3X on the bench. This is a good way to learn the software and try different parameters and seeing the result.
    I made a 90deg plug (from a wirewrap & solder tab Sub-D). I wanted the board level so its much, much easier to see the symbology on the PCB and to see the LEDs.

  13. jumpinjan

    jumpinjan Bronze Level Sponsor

    I also have a set of LS2 Corvette Coil-near-plug coils that I'm going to use on an Alpine engine (using the MS3x)
  14. jumpinjan

    jumpinjan Bronze Level Sponsor

Share This Page