1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Project CBR250RR (MC22) EFI Project. Now with Turbo

Discussion in 'Your 250cc Projects' started by akeizm, Sep 26, 2020.

  1. DannoXYZ

    DannoXYZ Well-Known Member

    Messages:
    354
    Likes Received:
    216
    Trophy Points:
    248
    Joined:
    Sep 7, 2017
    Occupation:
    IT Manager
    Location:
    Mesa, AZ
    My Bike:
    Honda CB125TT, Kawi EX250F, EX250J race, Honda CBR250RR-MC19, NSR350R-MC21 VF500F, CBR600RR, VFR750F
    Looks like Koso RX2N ?
     
  2. ruckusman

    ruckusman White Mans Magic Master Premium Member Dirty Wheel Club

    Messages:
    3,225
    Likes Received:
    1,415
    Trophy Points:
    918
    Joined:
    Oct 1, 2013
    Gender:
    Male
    Location:
    Narrabeen, NSW Australia
    My Bike:
    Yamaha FZR250 3LN1
    You buggers have got me seriously considering EFI - previously the cost of the ECU which is more than these bikes are seriously worth made me dismiss it out of hand - however just found RusEFI with all the necessary inputs and outputs to make it viable
     
  3. ruckusman

    ruckusman White Mans Magic Master Premium Member Dirty Wheel Club

    Messages:
    3,225
    Likes Received:
    1,415
    Trophy Points:
    918
    Joined:
    Oct 1, 2013
    Gender:
    Male
    Location:
    Narrabeen, NSW Australia
    My Bike:
    Yamaha FZR250 3LN1
    • Like Like x 2
    • Funny Funny x 1
  4. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    Koso r2xn, got it second hand.
     
  5. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    Was working on the tuning software today and the initial map I was given was rpm vs tps based. I want to run off a map sensor and Im going to have to redo the map.

    Does anyone have any idea how much vacuum these motors make at max power or max rpm?
     
  6. ruckusman

    ruckusman White Mans Magic Master Premium Member Dirty Wheel Club

    Messages:
    3,225
    Likes Received:
    1,415
    Trophy Points:
    918
    Joined:
    Oct 1, 2013
    Gender:
    Male
    Location:
    Narrabeen, NSW Australia
    My Bike:
    Yamaha FZR250 3LN1
    Kind of a hint - I've made a thread on the rusefi forum as your work inspired me

    https://rusefi.com/forum/viewtopic.php?f=2&t=1842&p=38004

    from a post part way down
    https://rusefi.com/forum/viewtopic.php?p=37992#p37992

    "A TPS actually isn't strictly necessary, but in your case it probably is. I'd recommend running alpha-n at least to start, since MAP can be finicky on small engines with ITBs like yours."

    I've just reread Thomas' blog - probably need to reread to see if there's a post on MAP measurements as it does sound like it's potentially difficult

    hope this help
     
  7. crshbndct

    crshbndct Active Member

    Messages:
    46
    Likes Received:
    35
    Trophy Points:
    188
    Joined:
    May 15, 2019
    Gender:
    Male
    Location:
    New Zealand
    My Bike:
    1991 ZXR250R
    I imagine none? Usually MAP shoots to 0(ambient) when you open the throttle wide
     
  8. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    Interesting, thank you. Ive been speaking to Thomas and he was going to go back and revisit Map but thats probably a bit a way for him yet (ecu repair needed for him).

    Im going to redo the VE table using the tune for the tps as a base and see how I go from there.

    I dont think my wideband is working properly too. I can see it do a test mode on start where it sets voltages but when its done it just sits on 1.36 lambda all the time
     
    • Like Like x 1
  9. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    Yea youre probably right. Haven't looked at map sensors for a long while. Mine is a 2bar gm sensor so 100kpa is 2.5v and 0kpa will be 0v. It allows me to go to 200kpa if things go to plan...
     
    • Like Like x 1
  10. ruckusman

    ruckusman White Mans Magic Master Premium Member Dirty Wheel Club

    Messages:
    3,225
    Likes Received:
    1,415
    Trophy Points:
    918
    Joined:
    Oct 1, 2013
    Gender:
    Male
    Location:
    Narrabeen, NSW Australia
    My Bike:
    Yamaha FZR250 3LN1
    Interesting - how have you got yours plumbed in - I see some map sensors inside ECU's with just a hose going to them - so they don't look to be overly sensitive to location
     
  11. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    I've got 2, one built into the megasquirt which I want to use for barometric correction and the other I've mounted near the airbox (only with the cable tie, can't think of a good mount as of yet). I connected all 4 of the balancing nipples under the intake ports together and then ran a line back to the MAP sensor.

    20201004_110131.jpg 20201004_110149.jpg
     
    • Like Like x 1
    • Informative Informative x 1
  12. jmw76

    jmw76 Well-Known Member

    Messages:
    1,151
    Likes Received:
    585
    Trophy Points:
    498
    Joined:
    Mar 12, 2017
    Gender:
    Male
    Location:
    Melbourne, Australia
    My Bike:
    Honda cbr250rr mc22 road + 1 track, Honda RC 162 tribute bike, Honda Spada Vt250, Honda CBR900RR
    Sounds like great little project.

    Back in the early 90's I built a fuel injection system for a little car from scratch (all built on Vero board and mounted in a little box).
    It was built around a Motorola 6805R3 cpu, as I had easy access to them at the time.
    Very unconventional. The 6805 is an 8 bit cpu with a grand total of 3700 bytes of on board EPROM and just over one hundred bytes of RAM. It was clocked at an astronomic 1MHz.

    I hand code everything in assembler to make it as efficient as possible.

    The system used MAP and speed inputs (connected directly to old fashioned ignition points) as well as a simple engine cycle sync input (induction pickup off number 1 plug).

    The system was designed as a direct carb replacement for a Triumph Herald engine (used in a little clubman special). Initially the original ketering ignition system was retained.
    The challenge with this system was that the Herald engine has siamese intakes. One intake feeds 2 cylinders.
    The system had to inject sequentially to ensure fuel was delivered to the correct cylinder at the correct time.
    Each injector had to fire twice during each engine cycle, with a max inject time limited to just over 3ms. As a result, I had to use big injectors, which didn't work that well at low throttle openings.
    I ran 2 x seperate throttle bodies with a vacuum balance pipe between them. The map sensor hooked onto this.
    I run a fuel rail pressure of 7- 8 bar. This was needed for rapid injection with good atomisation.

    I used a 1 bar MPX style board mounted pressure sensor (similar to used on Megasquirt). I used a differential pressure sensor which sort of worked as an ambient pressure correction mechanism. I did not do any other correction (pressure or temperature).

    The fuel map was hand coded into EPROM. So a reprogram was needed for every map change.
    It took a little trial and error to debug but it ended up working quite well and I used the vehicle successfully on the road and track for about 7 years.

    The code was very efficient. I think it worked out at about approximately 200 bytes with another 2 x 256 bytes for the fuel maps. Each map was 256 bytes with a separate map used for cold start (simply selected via a switch).

    The system operation was modeled off the old mechanical Lucas fuel injection systems used in the 60s/70s which I had on other vehicles at the time.

    The system would never meet emissions requirements. During the entire time, the vehicle never saw anytime on a dyno. Just seat of the pants tuning.

    I have since used Megasquirt kits on Toyota 20 valve 4AGE engines and plan to do a GSXR1100 installation soon.

    Enjoy.


    Peter.
     
    • Like Like x 3
    • Winner Winner x 1
    • Informative Informative x 1
    • Nice Work Nice Work x 1
    Last edited: Oct 4, 2020
  13. maelstrom

    maelstrom LiteTek Staff Member Premium Member 250cc Vendor Contributing Member

    Messages:
    5,102
    Likes Received:
    3,472
    Trophy Points:
    1,148
    Joined:
    Dec 23, 2012
    Gender:
    Male
    Location:
    Thailand
    Home Page:
    My Bike:
    Yamaha FZR400 3TJ1, Honda MC22
    In the early 90's in assembly language. :bowdown:
     
    • Agree Agree x 1
  14. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    Peter,

    I bow to thee, I wouldn't even know where to start doing that! I've had experience with MS before, I had a Datsun 260z that I put a L28E in (injected 6cyl from the 280ZX) then turbo charged it. Made so many mistakes on that project and wish I still had it.

    But this one has been fun and is working well so far. Well I haven't actually given it the beans yet as I'm still sorting some small issues out (only thing left at this point is trying to make sure the wideband o2 is working) then I can do some small datalog runs and check my afr's and ensure my vr sensor for crank angle is working throughout the whole range (I may need to filter it yet).

    Marc.
     
    • Like Like x 1
  15. jmw76

    jmw76 Well-Known Member

    Messages:
    1,151
    Likes Received:
    585
    Trophy Points:
    498
    Joined:
    Mar 12, 2017
    Gender:
    Male
    Location:
    Melbourne, Australia
    My Bike:
    Honda cbr250rr mc22 road + 1 track, Honda RC 162 tribute bike, Honda Spada Vt250, Honda CBR900RR
    On memory constrained micros, assembly language still has its place. What is more, I didn't have an assembler for the 6805 so I just used macros on an old Z80 compiler that I had. Then programmed the UV ROM in the micro with a little programmer that I built. No flash or EEROM on these old school micros. Erase cycle under UV lamp took many minutes. No debug tools to speak of either. Debug was all done using tiny incremental code programming steps and lots of testing. Even with <200 bytes of machine code it still took quite a while.
    Extremely satisfying when I got it going. Even more satisfying when I would blow past other competitors on the race track.

    Some pics attached.

    Note, hand would induction pick up on spark plug lead, old school dizzy, cut and shut manifolds of Triumph 2.5 PI, big air bleed screw for idle speed adjustment.

    Hell, it worked.


    DSCF0183.JPG DSCF0220.JPG DSCF0218.JPG DSCF0195.JPG
     
    • Like Like x 2
    • Love it! Love it! x 1
    Last edited: Oct 4, 2020
  16. jmw76

    jmw76 Well-Known Member

    Messages:
    1,151
    Likes Received:
    585
    Trophy Points:
    498
    Joined:
    Mar 12, 2017
    Gender:
    Male
    Location:
    Melbourne, Australia
    My Bike:
    Honda cbr250rr mc22 road + 1 track, Honda RC 162 tribute bike, Honda Spada Vt250, Honda CBR900RR
    This is the megasquirt 20 valve conversion we did on my son's little AE82 Corolla.
    We tried to make it look as stock as possible. Of course, they were never delivered that way.
    It was a fun little machine, but never properly sorted. It really needed some dyno time, which was beyond our budget at the time.
    We used a very early megsquirt kit with lots of mods and hacks to the main board to make it work on the car.

    No more car pics now.
    Sorry for the diversion, but I thought the EFI background might be of interest.

    Peter.
    engine2.jpg
     
    • Like Like x 2
  17. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    Aww, went to do some tuning today and something wasnt right. Sounded like it was down a cylinder. Well turns out 2 cylinders were down, 1&4. Traced it and somehow a wire broke between the coil pack and my coil driver, which I havent touched in the last few weeks.....now I have to strip the wiring harness out of the bike and trace it back.....sigh.
     
    • Bummer Bummer x 3
  18. jmw76

    jmw76 Well-Known Member

    Messages:
    1,151
    Likes Received:
    585
    Trophy Points:
    498
    Joined:
    Mar 12, 2017
    Gender:
    Male
    Location:
    Melbourne, Australia
    My Bike:
    Honda cbr250rr mc22 road + 1 track, Honda RC 162 tribute bike, Honda Spada Vt250, Honda CBR900RR
    @akeizm are you using a wide band sensor for tuning? How are you connecting it into the Microsquirt?
    Peter.
     
  19. akeizm

    akeizm Active Member

    Messages:
    20
    Likes Received:
    22
    Trophy Points:
    73
    Joined:
    Aug 21, 2020
    Gender:
    Male
    Occupation:
    Control Systems Engineer
    Location:
    Brisbane, Australia
    My Bike:
    CBR250RR (MC22) EFI Turbo
    Hi,

    I will be using a wideband. I bought a 14point7 spartan 2 wideband with controller as theyre meant to integrate well with megasquirt. Its a 6 wire controller which has three wires for the reading and general power, one for a warmup led and two for the heating circuit. Only two of the wires lead back to the ecu.

    I found my broken wire, the wire had snapped at the original 3 pin plug for the coils. Wire is getting brittle so I replaced the plug. Then my idle was high again, so I hooked up my vacuum gauges and balanced the buttlerflies again. Now I can get the idle back down.

    Ive also installed a new relay, which is controlled by the megasquirt to run the cooling fan. Dont trust the original switches fitted to the radiator.

    Now Im ready to go for a little ride and try to get some data. Need to see how the map sensor is working so I can build a new table for speed density control instead of alpha-n.
     
    • Like Like x 2
  20. Mike Green

    Mike Green Well-Known Member Premium Member

    Messages:
    177
    Likes Received:
    246
    Trophy Points:
    248
    Joined:
    May 9, 2020
    Gender:
    Male
    Occupation:
    Taking it easy
    Location:
    Auckland
    My Bike:
    Yamaha FZR250
    Very cool project. With ITBs I suggest you plumb all 4 vacuum lines to a common chamber and then to your MAP sensor. This smooths out the pulses plus with all 4 cylinders providing MAP you have a more responsive signal to use. Re the home signal or whatever MS call it, there are people who have the syncing done using the vacuum signal on cylinder 1. Ecotrons do on their Ninja EFI kit. I know Link had a look at it and may have it working now. With a bit of effort you may even be able to come up with something on the MS system. Without it your injectors are firing EVERY crank rotation and will quickly run out of time to cycle as they are firing twice as often as required. I use way smaller injectors and even there the minimum pulsewidth still results in a slightly rich mixture at idle. With the Lambda meter reading you need to make sure the output of the lambda meter to your ECU matches to what the ECU is expecting. Typically a bit of fine tuning is required so the expected voltage matches the actual voltage.
     
    • Like Like x 1

Share This Page