BD125-8 ZS190 swap and EFI conversion

[bogieboy]

From the sounds of things, could your timing be 180° out? I forget what all you are using for timing, i know most megasquirt builds in the car world use a crank and cam signal to verify engine timing... are you using just the crank trigger or? Im not too familiar with the microsquirt and what all it wants to see... its been a few years too since i was investigating MS for my truck...LOL

[bogieboy]

Also, i dont see why a cdi WOULDNT work with MS... how many people run aftermarket ignitions in thier cars with MS... most of them run accel or msd ignitions... however, i wonder if the built in timing curve of the cdi for our motors is throwing things out... i wonder if you need to just run a coil driver off the microsquirt, straight to the coil... or like previously mentioned run the ms for fuel only, and use the cdi somehow isolating the signal between the 2...

[glavey]

A few things happened; good and bad. Problems; good and bad, were mostly sorted out.

1. I realized that the injector and/or the throttle body were poorly made (mostly the part where the injector in seated) and the spray from the injector (if it is even spraying and not dribbling) is hitting the walls of the TB and running past the throttle vane and right out to the air filter. After seeing that for the first time, I took the whole intake/TB assy. off and felt the bottom of the intake - bone dry. If there is fuel getting into the engine (there must be to get a BANG out of it), it is dribbling off of the walls into the intake port in no controlled manner. To remedy this, I will need to buy a new throttle body without a injector seat glued to it as well as have a hole in the intake manifold precision drilled and honed for the injector to sit in, so it can spray fuel directly on to the back of the intake valve. Both of those will take a considerable amount of time and some money as well.

2. I under-estimated the NEED to go out and ride in warm weather. This is my first motorcycle, I bought it last year around July. This was my first winter project and this will be my first spring where I have something to look forward to when the weather gets nice, instead of doing yard work and getting bitten by mosquitoes.

3. This is my first attempt at an EFI conversion and attempt at non-oem EFI in general. I should have started with something that had FI to begin with, then I would hopefully be able to focus on tuning (which is what attracted me to EFI in the first place - tuning, not circuit diagnostics and parts compatibility) instead of troubleshooting.

4. You all have seen how many things I have had to order for this project. I am very near the limit of what I want to spend on this project, this year

5. I am working on something that I love to use. When I get frustrated with something in the EFI system, I want to be able to go out and have a good ride to clear my head; I can't do that if the EFI system is ON by only bike. Having the bike be unrideable for large portions of time while waiting for parts to arrive is no good.

6. (a bit emotional) This thread, while I have thoroughly enjoyed sharing my findings, discoveries, and insights, has caused me some anxiety; I feel like I have to do something so I have something to post about and I have to post about what I have been doing. I know I don't really have to, no one has a gun to my head, but I started this thread and if I stop posting to it, I'll feel like I let people who browse this forum down. I realize this is all in my head and isn't correct, but I am mentally unable to let go of those feelings.

It is for these reasons that I am suspending, NOT CANCELLING the EFI portion of this project. To get the EFI system into good working order as described in #1 above, I need to either buy all of the things that will allow me to drill a precise hole in the intake manifold for the injector AND buy another TB (or remove the injector seat and patch the massive hole that is left in its place) without a injector seat AND wait around 1-2 weeks for everything to arrive AND then spend time installing everything. That is no fun when there are roads out there waiting to be torn up and sunshine waiting to make me feel nice and warm while giving me skin cancer.

Luckily, I made the EFI wiring harness in such a way that I can and have removed it from the bike along with most of the supporting parts in an evening (~4 hours) and installed the carb and CDI back on. Since it only took about 4 hours to take off, it should only take 8 hours to re-install, now that I know where everything needs to go and all the connectors are already there in the correct location.

I will still probably keep an eye on this thread (for things specific to this bike) and on the forum in general (for everything else) and answer questions and post my usual lengthy review/walk-through/write-up on things that I am doing or have done to my bike. I still have a few more somewhat inexpensive things I want to do like re-pack the exhaust with a quality packing material, not the wafer-thin sheet of fiberglass that they have in them from the factory, the $15 grom shock mod, and replace the rear-shock-shaped-object with an oem honda grom one.

bogieboy; I am using wasted spark on this engine. There is only a single tooth on the flywheel/magneto that triggers once per 360 degrees, so the ECU and CDI can only fire the spark plug once per 360 degrees (at the end of the compression stroke and the end of the exhaust stroke), not once per 720 degrees, as would be needed for non-wasted spark. Also, I have had the engine spinning using the starter motor and have synced the timing that the ECU says it should be and what it is on the engine. 10 degrees BTDC on the engine matches the ECU's 10 degree base timing.

Those aftermarket CDI boxes for automotive engines probably have a tachometer output wire that can give a signal to the ECU and they probably don't mess with potentially shared signals like crank speed by grounding one of the leads. The 190's CDI could be used with the microsquirt if I could build a small circuit that could duplicate the crank VR sensor's signal into two separate, isolated outputs. But, I don't have the skills to do that; I already have difficulty with arduinos, making something that is used to assist in the control of an engine is way beyond me.

Also, those ignition boxes will probably accept a 5v square wave input from an ECU or sensor; our motorcycle CDIs need a signal that goes high, crosses 0 volts, and goes negative (again something that I could build, but my lack of skills are stopping me). The microsquirt only has logic-level outputs for sparks; you cannot directly drive a coil from it without hardware modifications.

Sorry to stop progress on this so abruptly, but I must ride, be it with injectors or carbs.

If anyone has any questions that they wanted answered with regards to anything I did or said I was going to do with my bike, please go ahead and ask.

EDIT: Just had the bike started up for the first time. Loud even with the silencer in. Vibrates like a harley at idle. Big-ass smile on my face. I'm gonna miss the counterbalancing the 125 had. I'm letting it cool down so I can do an oil change and then take the new engine for its first inaugural ride. I'm keeping the wideband controller installed on the bike; once I get/make a gauge that will display the AFR, I'll add the sensor back into the exhaust to aid my tuning. I'll try (no promises) to take a short video showing how the bike sounds.

[bogieboy]

hope all goes well with the bike Glavey, totally understand wanting to ride it.... on the coil driver side of things, i wonder... if you went to a junkyard and found an automotive coil, say an LS1 coil-near-plug setup, that would get its own 12v supply, and would only need a 5v trigger from the microsquirt, and just use the MS for everything, bypassing the CDI all together....

just bouncing ideas right now...LOL

[glavey]

Actually, that is very nearly the exact ignition system I used with the ECU; a GM "truck" coil with a built-in ignitor that used logic level signals for dwell and spark. The reason I even considered the CDI instead of having the microsquirt deal with the ignition was because the CDI has a known-good and safe spark advance curve for the 190 engine. I wanted to leave the spark timing to the CDI and just run the ECU in fuel only mode because I didn't trust my ignition timing guesswork.

[bogieboy]

hmm.... start conservative and creep up on it, when you hear any detonation, back it off a few degrees....LOL its not too terribly hard... not trying to be an ass or anything, its just really simple (at least to me, but i have been researching timing maps for car stuff since my first truck, back in 05)

[glavey]

It does make sense that it would be simple like that, but still I don't trust myself. I had contacted support @ tboltusa about the max advance in the 190 CDIs. They came back and said it was around 30 degrees at maximum, so I put a 27 degree max advance in the ignition table in the microsquirt and tapered it down from 27 at about 6krpm to 10 at 1krpm. That is what I was using when I got the BANGS from the attempted engine start.

I'm not out for maximum power on this engine; I don't want to tune the ignition table for max. timing before torque drop off, I just want a good amount of power without worrying about knocking, pinging, pre-ignition, or detonation.

Also, I was and still will be struggling with low-rpm erroneous spark firings. Anything below about 500rpm and I get lots of extra sparks not apparently correlating to any specific position on the engine. I'm guessing that is due in part to the crappy signal VR sensors can give when measuring low speeds, not to mention there was only one "point" of information per revolution on crank positioning.

I have been thinking about getting a toothed wheel cut to size and epoxying it to the flywheel/magneto where it won't be interfering with anything that is already there and using a hall effect sensor to detect crank speed. That would give up to 36 (35 really, the wheel has to have a missing tooth) points of timing information per 360 degrees of crank rotation. Even if that still resulted in fewer, bit still some erroneous sparks, I could then run the microsquirt in fuel only mode if I choose, as I would have a separate crank speed input for the ECU, independent from the one used by the CDI.

Don't worry you aren't sounding mean at all, just throwing ideas around.

[bogieboy]

i think the 36-1 tooth wheel would be a great idea, when you have free time, since thats what the majority of MS users run. even a lot of OEMs use a similar setup, with a 32-1 or a 36-1 wheel on the crank... if you could somehow couple a cam sensor in the mix as well, that would be even better, but a royal pain in the butt to get working....a dual signal=full sync, and can be traced to a VERY small percentage of the rotation of the engine, I.E. more accurate timing.....LOL

[RedCrowRides]

None of my business but i'd put a carb back on it and ride hell out of it this spring n summer and revisit this EFI deal when it turns too cool to ride again. You are NOT having fun at this point, versus riding it, which WOULD be fun. Stop looking at it and RIDE IT.

[glavey]

That's what I've done; yesterday I hooked up the carb and the CDI and went out for an inaugural ride. Other than the clutch friction zone being right at the beginning of the lever travel, causing more then a few embarrassing stalls at intersections, everything went swimmingly. I even met someone on a modded ruckus. I really want to quiet down the exhaust with a combination of new packing material, baffles, and modifying the exhaust silencer.



I really needed that ride and after coming back from it I didn't regret my decision to pause the EFI and put the carb on one bit.

[glavey]

I think I've discovered the reason guys who name bikes, cars, trucks, etc use female names... I was going to name by bike something really unusual, like... Bob. Then I realized if I was ever asked what I was doing all weekend long, I'll have to answer, "Oh man, I was riding Bob all weekend. He's a real screamer. I couldn't walk right for a while after getting off of him."

On a less sarcastic note, I took Bob out for a brisk ride today. About 50F, but with clear skies and sunlight to warm you up. I adjusted the clutch so I can actually start from a stop and not stall 5 times. I had read that the gearbox in this engine is quite smooth, but damn it really is. This bike is an excellent candidate for a quick shifter. I'm not sure if this is the correct terminology, but the gears are close ratio, meaning... for example lets say you are at 5000rpm in third gear, you shift into fourth and the revs only go down to... maybe 3700-4000rpm. Not a lot of drive line shock will happen going between those revs.

I have been looking for a wideband gauge that has the numerical AFR readout as well as a row/bar/graph/ring of LEDs that at a quick glance can give you a rough idea of what your mixture is. I could only find gauges with integrated sensor controllers for >$125. Nope, I already bought a controller and I'm not buying another just for a gauge. So I set out to make my own.

2 or 3 years ago, I made a dashboard for my scooter (pictured). Everything on it worked as it should (needles moved, numbers were displayed), but I had a problem with getting a reliable, clean signal for the engine rpm and the speedometer. I stopped working on it because I had one of those moments where you work hard on something, and in the end it was all for naught, combined with me not really riding it enough to warrant caring about it.

I dissected dashboard and preformed an display-ectome. I took out the 16x2 character display (blue rectangular display) and the two 7-segment, 4 digit displays. Organ harvesting done, I started trying to imagine what the gauge could look like with what I had to work with. The 7 segment display fit in the center of a gauge face I had (in a previous life, the gauge was for oil temperature on my scooter) with enough room around it to line the gauge face with LEDs.

I like the wideband gauges that have, from left to right, yellow LEDs for rich, green for stoichiometric, and red for lean. I wanted to copy that in my gauge, so I gathered some LEDs out of my electronics project box and glued them in a ring to the inside of the gauge face along with one of the 7 segment displays (pictured). I wanted to have two 7 segment displays and the LEDs, but the chip I am using to drive them can only do 2x 7 segment 4 digit displays (maximum of 8 digits) OR one 7 segment 4 digit display and the LED ring that I just glued in. I went with a 7 segment 4 digit display with the LEDs. I did have another driver chip in the scooter dashboard, but when I tried to de-solder it, I killed it.

After much soldering, gluing, and back-ache from being hunched over a desk, I got all of the LEDs and the display connected to an arduino nano and used a example program to test out all of the individual sections of the display and LEDs.



Success! Now, on to actually programming the arduino to make the display and the LEDs show meaningful data. After about 2 days, on and off (I'm not very good or fast at programming) I had working code that interpreted a 0-5 volt signal into 10.00 to 20.00 AFR on the display, and individual LEDs would light to indicate AFR within a given range (attached to this post, if you have the arduino IDE, you can just double click the afr.ino file to open it in the arduino IDE. If you don't have the arduino IDE installed, you can open the afr.txt file. Both filesa re exactly the same, just with different file extensions). Now that I had working gauge electronics, I had to cram and stuff them into the gauge housing. In the end I got everything in; the gauge face and everything glued to it, the arduino, all of the wiring, and a 5 volt step-down voltage converter. I plugged in the arduino using a USB cable and nothing happened and then I saw a small whiff of magic electronic smoke escape from the voltage converter. It appears that the converter was not able to have 5v applied to the output when there wasn't any voltage on the input side. I tried using a diode on the 5v output, but it was too late, the patient had died. Now I can either make a voltage regulator (it would get real hot inside a gauge case) or buy some different, better made voltage converters(like the one I used in the bluetooth case).

Thats about it for this post.

[loganlexow]

Dude, this is effing sweeeeeeeet! You have gathered so much info and shared it, and all the rad trial and error is just inspiring! I get discouraged too when trying to make something but fail. Please know you aren't alone there or with anxiety. Keep updating when you can please!

[glavey]

Thanks, dude. I really appreciate the kind words. It's posts like yours and others here who have encouraged me to keep building and posting.


<3

[glavey]

I salvaged the voltage converter from my DIY bluetooth module and wired it up to the AFR gauge. 90% success! The arduino did what it was supposed to and the voltage converter didn't go up in smoke, however for some power-related reason, the arduino couldn't communicate with my computer. It would work when the voltage converter was disconnected. Just a small thing I'll have to be mindful of when updating the arduino.

The different voltage converter was bigger than the last one I used; too big to fit in the gauge housing, so I needed to elongate the gauge to fit all of the electric modules. I decided to cut the gauge housing in half and glue/tape/epoxy something between the two halves to make more room. I cut apart the plastic cover for some old blank DVDs and made a rectangular-ish shaped piece that I could use as the "filler" piece. After I dremeled the gauge in half, I taped the gauge pieces and the filler piece in place with electrical tape. I tried to use super glue, but the filler plastic was too shiny and slippery.

The gauge, as it is now, is about 90mm/3 1/2 in. deep and about 55mm/2 3/16 in. diameter. It will update the display and the LEDs 10 times/second. I found that any faster than that, I can't make out the numbers in the display. I'll probably temporarily mount it somewhere along the top of the speedo cluster so I'll be able to quickly glance at it while riding. I haven't decided exactly how I am going to attach the gauge to the speedo, but preliminary thoughts are either more electrical tape or zip tie to speedo case.

Sorry there isn't much in this update but my brain is kinda foggy right now and not that much really happened. Lots of thinking and planning about how to do certain things, but not much doing things. If for some reason the AFR gauge doesn't work in the form factor that it is in now, I have plans to make it much smaller, simpler, and blend in much better with the styling of the bike.

Except for tomorrow, the next 5 or 7 days look like good riding weather. Stay tuned...

[glavey]

Apologies for the poor-quality post previously, I was really out of it but I felt like I needed to post something.

I think the few rides I have taken on the bike have given me enough information to give a bit of feedback/observations on the 190 engine.

Like most of the pit-bike-sized engines around, the 190 doesn't really idle smoothly... It doesn't idle roughly, just not smooth. By that I mean the engine is going putt-putt-puttputt-pu-pu-putt-putt-putt instead of a "normal" idle of putt-putt-putt-putt. I don't know for sure, but it think the combination of little-to-no intake piping, low speed, and no load make the engine kinda miss once every... 3-7 complete engine cycles.

The carb that came with the 190 is a Chinese copy of a keihin PE28 carb and (I am 90% sure of this, but check your jets to be sure) uses these slow jets and these main jets. I think, but am not positive that these keihin main jets will also thread in to the main jet tube and work MAYBE. My carb came with a #40 slow jet and a #115 main jet. I didn't record the needle# or the clip position.

I don't have any information on the AFR ratio for these jets; I still have yet to install the AFR gauge on the bike. I've either been too eager to ride it or busy with something else. I can say that the engine responds very well in all running situations I have put it under except one. There is a miss around 4-6k rpm regardless of throttle position. I don't know if this is an ignition or fuel problem, but I have read of other people having the same miss at around the same engine speeds. Other than that, there has never been a fueling issue.

The engine starts up with the choke on within 2 revolutions every time. With the choke on and the air temp ~67F the bike idles at about 4-5k rpm. after about a minute of running, the engine will start to sound a little boggy and rich. That's when I turn off the choke. I have the idle set around 1500-2000 rpm. A little high, I know, but the engine likes it there and responds very quickly to all but the most violent of throttle snap-openings. Since, as described above, the engine doesn't idle smoothly, using an AFR gauge to tune the idle would probably be futile. I've read about mechanics using a vacuum gauge to set the idle mixture, so I'll give that a try. Also something to note about the keihin PE28 carbs - the low speed mixture screw, NOT the idle screw, meters how much air goes into the engine at and around idle speeds. Adjusting an air-mixture screw, in my experience, is a bit more finicky than adjusting a fuel-mixture screw. When I was using a VM22 carb that also has an air-adjust screw, I was running about 1/4 to 3/4 turn from fully closed to get the idle just right.

At in-town speeds of <35 mph, up to 6k rpm the engine, as geared (I think 16/34?), feels much like my scooter felt after I put a big block kit on it; feeling the same amount of torque, just add 3-5 mph to whatever speed you felt you were doing with the smaller engine. I don't know if it is my lack of experience with bikes or perhaps a poor fuel mixture, but I have to rev the engine up to about 6k rpm when taking off and SLOWLY release the clutch or else it will stall. It could also be the new clutch is being broken in as well. I'll choose me and my lack of experience as the main cause of feeling lack of torque on take-off. In situations like this that have many possible causes of a certain symptom, If I am anywhere in the "might be the cause" list, I just automatically choose myself as the cause of the symptom. I'm usually right, as well.

In contrast, once the engine is above 6k rpm, the torque and the pull from the engine increases precipitously all the way to 10k rpm. Seriously, with how much the 190 pulls at wot, I am thinking that the 125 was very held back by the factory quiet/restrictive exhaust. It pulled as hard at 8k rpm as it did at 4k rpm. I might temporarily put the factory 125 exhaust on the 190 and see just how bad it is.

When the engine is up at those higher rpms, the entire bike and engine buzzes, vibrates, and shakes and I can't hear the intake or exhaust over the mechanical engine sounds. Now, I don't mean the engine is vibrating the mounting bolts loose, I mean the lack of a counter-balance shaft makes this engine, the bike, and everything touching or connected to either of those vibrate. Nothing on my bike has come loose except for a cell phone holder. Don't worry, I didn't have a phone in there at the time.

The transmission shifts quickly, smoothly, and effortlessly. Only twice have I accidentally shifted into neutral instead of 2nd. It is still difficult to get into neutral with the engine running and the bike stationary. I suspect that will improve with continued use.

I don't know if, again, this is either my fault with lack of experience or just the way things are, but it has been very difficult to keep the engine at a steady operating state in lower gears at lower speeds. What I mean by that is below 35mph in any gear but mostly in the lower ones, the engine wants to accelerate or decelerate (as well as jerk the bike when going from decel to accel), NOT run at a steady speed. I really don't know if that is inherent to this engine or just engines above a certain size. It has certainly given me cramps in my throttle hand, having to try and precisely control input.

The exhaust system I am using is too loud for my tastes. I have only run it with the silencer in the muffler, I don't dare take it out. I'm going to try and modify the silencer to reduce the volume a bit. The silencer that is in the muffler now only has a narrower diameter pipe for the exhaust to exit through (like this but without the holes drilled in the sides of the tube.) I want to drill holes, like in the linked piece, to force the exhaust gases to flow through smaller, but many more small holes. I also want to try cutting slots in the pipe instead of holes and see what happens with that. If none of that cures my sore ear's woes, then a muffler re-pack and possible baffles are the answer. At the extreme end, I could buy and install a different muffler. Now that I have an exhaust system that comes in pieces instead of all welded together, I can change out mufflers with the same pipe size, but that is only if everything else fails.

I have ordered some filament for my 3d printer so I can attempt to make a "backpack" for the speedo/dashboard. The big round taped-up AFR gauge, while perfectly working, is a bit too... it's just not what I like to see when looking at my bike. I am going to try and make something that will ziptie/glue/something to the front of the speedo (the part that all the bugs hit) and have the numeric AFR readout, some number of colored LEDs to indicate AFR at a glance, numeric battery voltage, and maybe either a numeric temperature display or just a temperature warning light.