Fuel Injector Retrofit Project

Description of your first forum.
admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Thu Sep 06, 2018 7:00 pm

Connir_rider your questions.


1) You said you only have 1 pulse and your temporarily telling the system it is only a 2 cylinder.
A C-10 has 2 pick up coils so I assume you could use both?
Does this bike only have 1 pick up coil?

2) At the start of this project you planned a low budget version of what others have built.
With all the special parts you have been ordering, isn't your budget higher than expected?
NOTE: I realize the development process costs extra as you determine correct components.
So, I'm not referring to the development costs as you work out details..
I'm referring to the costs for the special boards etc that you finalized on.

3) My guess is; this has become more complicated that you originally though it would be?

4) Have you cranked it anymore?


1. The magna has 2 pickup coils on the crank but they are not 180 degrees apart. They are only about 85 degrees (as near as I can tell from the signals). I have tried combining them but end up with a distorted double pulse that are so close to each other I dont think the ECU can distinguish between them. The real solution is to bolt a toothed wheel to the side of the started clutch and add another sensor dedicated to the RPM signal for the ECU. I think it will fit. For now I will limp along as it until i can get a toothed wheel to the right specifications. As a side note I noticed that the stater on the generator has 18 poles which will evenly divide into 360. A bit of wire wrapped around each pole may yield a good enough signal to accurately determine the crank position (within 20 degrees) . I will leave that experiment for some future date.

2. Budget. yes I am way over budget as far as total cost. I probably have close to $650 in it so far. That includes stuff that I bought and ended up not using and test equipment needed and tools. Remember I built a lot of experimental manifolds and throttle bodies that ended up in the trash.

Here is a quick n dirty estimate to replicate what I have done. ( i intend to replicate this on my 99 Connie and on a 97 manga that is just sitting in the back yard taking up space. )

ECU $159
Wide band controller and sensor $100 Not completely necessary but makes fine tuning so much easier and faster.
Injectors and home made fuel rail. (used off ebay) $40
Manifold $50 in actual costs.
harnesses and connectors. $50
fuel pump $17
Misc stuff, $30
So I am at $450 more or less. Cost will vary based on where you get your parts and what sort of deal you get on ebay.

3. Complicated? Yes and no. Yes it was tedious tying up all the loose ends and verifying everything was done correctly. The harnesses and the connectors were a real pain to get the connectors crimped and cut to the right length. And NO because I tossed my arduino design aside and adopted a similar ECU design that had already been field tested and although still in development it is miles ahead of where I was. I'm sure it saved me 9 months of software development. I have learned so much from this project. I am still on the learning curve but so far ahead of where I was when I started.

4. Have I cranked it anymore? My lord that is all I did for a week is crank it , adjust it , crank it again. At this point it will start immediately every time and the throttle is fairly responsive. Yesterday I decided it was time for a road test. I got nothing better to do while I am waiting for my wide band controller. I need a battery for that which I ordered from Amazon last night. https://youtu.be/WY8xhAa_ZQA

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Thu Sep 06, 2018 7:21 pm

Thinking on how to do this on my C10.

I have a set of throttle bodies and injectors off another Kawasaki but I don't think i am going to use them. Instead I think instead I will use the manifold approach like I did on the Magna.

Looking at the C10, after you remove the carbs there is a lot of room in there for creative thinking. Just running a few ideas through my head I came up with the idea of plugging the holes in the back of the air box, mounting injectors on pipes running from the air-box to the head and mounting a throttle body on the front of the air-box directly below the ports.

Any thought?

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Thu Sep 06, 2018 7:35 pm

Some thoughts on why the crank signals on the Magna are so hard to deal with.
Here is a photo of the crank (starter clutch). Notice the one large tooth which is used by both pickup coils to determine when to ignite the sparks.

Rotating clockwise you see that when the tooth has just finished passing under pickup coil number 2 it has only ~45 degrees before it starts passing under pickup number 1. I assume the ECU would prefer these pulses to be 180 degrees apart and at worst 90 degrees but maybe confused by 45 degrees of travel from the end of one pulse to the beginning of the next.

These are the signals being generated by the two pickup coils.
One is low going pulse the other is high going pulse. If I try to invert either one it kills it. As near as I can determine they are wired with a common ground and to reverse one ties its signal to ground. I tried to use as is and or the outputs of the VR together but it is a open drain and any low output makes both low. I could try exclusive or of both outputs. That would produce two low going pulses . But still they would be very close together and may confuse the ECU.

Scope traces are the outputs of the VR board pins 8 and 7
Attachments
pickup1and2.JPG
pickup1and2.JPG (136.39 KiB) Viewed 2105 times
crank sensor.JPG
crank sensor.JPG (57.35 KiB) Viewed 2110 times

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Thu Sep 06, 2018 11:42 pm

took the two outputs from the VR board and xored them together to produce this signal. Seems the ECU is missing one of them. As the rpm is half what it should be. Hard to start and wont throttle up without difficulty.

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Thu Sep 06, 2018 11:47 pm

xored the two outputs of the VR board. ECU is missing one of the pulses as the RPM is half what it should be. Hard to start and slow throttle up. Runs with sputtering.
Attachments
CCI_000003.jpg
CCI_000003.jpg (279.23 KiB) Viewed 2097 times
xor2signal.JPG
xor2signal.JPG (79.15 KiB) Viewed 2099 times
trigger 4-2.JPG
trigger 4-2.JPG (58.26 KiB) Viewed 2099 times

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Fri Sep 07, 2018 1:16 am

A good place for a trigger wheel.
Attachments
361.JPG
361.JPG (17.67 KiB) Viewed 2096 times
241.JPG
241.JPG (16.69 KiB) Viewed 2096 times
magna crank.JPG
magna crank.JPG (69.94 KiB) Viewed 2096 times

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Wed Sep 12, 2018 10:09 pm

Finally arrived, my wide band controller and O2 sensor.
A little disappointing because it came as a kit, 2 circuit boards and a bag of parts to build the harness and solder the plugs on the board and no instructions.
Well half a day later its working and working quite well. Running with the narrow bad sensor the bikes take a while to smooth out the idle and even longer to be able to throttle up.

With the wide band sensor enabled the bikes starts much quicker and goes quickly into a smooth idle. Able to rev up when cold.

Tuning the fuel tables can start in earnest.
https://youtu.be/Nsd6Jsibokg
Attachments
20180911_090018.jpg
20180911_090018.jpg (7.88 MiB) Viewed 2078 times
20180912_144538.jpg
20180912_144538.jpg (7.43 MiB) Viewed 2078 times

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Thu Sep 13, 2018 1:12 am

The ECU is not running closed loop at the moment. I am using the feedback from the O2 sensor to play with the fuel table manually to get things in the ballpark. Tunerstudio also has a logging function that you can use to analyze how the engine is performing with the current table settings.

Tuner studio has an auto tune feature that will read the sensors and make the necessary adjustments to get the AFR correct under various conditions of load and throttle. It is really amazing to watch. You begin with a reasonable fuel map and if you can get it started and have some throttle control, tunerstudio will begin to dial it in for you.

That feature is disabled on my version because I have not yet paid the $59 upgrade. I do intend to get the upgrade I just was procrastinating on the purchase since I did not have the wide-band sensor yet. I will do that on my next day off work (next Tuesday) .

I need to shrink my setup. As of now I have a desktop computer on my bench running tuner studio and talking with the ECU over a USB port. I cant drag my PC out on the highway. I need to duplicate my setup on a laptop that I will strap to the back seat.

Once I am satisfied with the performance I will freeze the tables and run it open loop only and probably switch back to the narrow band sensor
(if it continues to run well with it)

A demonstration of auto tune with tuner studio. https://youtu.be/jVht8KxsPr4
https://youtu.be/zxYOUAL9R_o

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Thu Nov 01, 2018 10:25 pm

Been playing with the project a couple times and I am still having trouble tuning. Even with the auto tune. After I get it started and warmed it seems to run ok and throttle is somewhat responsive. But it is a bitch to start. I have set the crank and warm up enrichment to 200% and it is still taking a minute of cranking before you hear a popping in the exhaust. After warm up it idles around 1000 and the lambda runs about .97 < 1 < 1.1 so the sensor and the closed loop feedback is working.

Aside of the hard to start when I rev it up (WOT) and then close the throttle too quickly it will die. If i back off the throttle slowly it will purr right into idle at 1000 rpm.

At the moment it is using the spark signal on one coil to sense the rpm and there is no other input to the ECU to tell it where the crank is in the 360 degrees.. I don't think there is enough crank feedback to get the auto tune to work properly and the ECU to know exactly what is happening in the engine.. I may have to bite the bullet and install a multi-toothed sensor on the crank. I can order a toothed disk to attach to the attach to the starter clutch or.... I was thinking of winding a few turns of wire around each pole of the generator (there are 18 poles on the generator and 18 will divide into 360 evenly. Either one of these solutions involves removing one of the side covers, which I really wanted to avoid.

Its been too cold to work lately let alone take it for a ride. I have ported the software (tunerstudio) to my laptop which I intend to strap onto the passenger seat and take if for a spin around the neighborhood. If that works out I will take it for a longer ride on one of the back roads.
That's the update in a nutshell.

admin
Site Admin
Posts: 44
Joined: Mon Jul 02, 2018 1:02 am

Re: Fuel Injector Retrofit Project

Post by admin » Sat Nov 03, 2018 12:32 pm

" The lambda calculation determines the ratio between the amount of oxygen actually present in a combustion chamber vs. the amount that should have been present to obtain perfect combustion.

Lambda represents the ratio of the amount of oxygen actually present in a combustion chamber compared to the amount that should have been present in order to obtain "perfect" combustion. Thus, when a mixture contains exactly the amount of oxygen required to burn the amount of fuel present, the ratio will be one to one (Ll) and lambda will equal 1.00. If the mixture contains too much oxygen for the amount of fuel (a lean mixture), lambda will be greater than 1.00. If a mixture contains too little oxygen for the amount of fuel (a rich mixture), lambda will be less than 1.00.

The Wide-Band sensor generates a variable signal as opposed to the simple rich/lean signal of a standard oxygen sensor. Because the signal varies in strength and also in current flow direction (polarity), it's impossible to directly view the signal with anything except an oscilloscope. However with the right supporting equipment, the Wide-Band sensor can be used for adjusting air/fuel mixture on any engine.

Perfect combustion requires an air/fuel ratio of approximately 14.7:1 (by weight) under normal conditions. Thus a lean air/fuel ratio of, say, 16:1 would translate to a lambda value of 1.088. (To calculate, divide 16 by 14.7.) A lambda of .97 would indicate an air/fuel ratio of 14.259:1 (derived by multiplying .97 by 14.7). " more at this link http://www.austincc.edu/wkibbe/lambda.htm

In this project the air fuel mixture starts out lean when the bike is cold started but I think that is more to do with the O2 sensor has not reached operating temperature yet and the exhaust is still full of cold oxygen rich air. Soon after starting it goes rich while it is warming up (which is what it is supposed to be) . After it reaches operating temperature the air fuel mixture alternates between slightly rich to slightly lean.

The O2 sensor is still mounted on the top of the muffler. There is a spot under the bike between the arms of the center stand that would be out of the way. It would also be a mix of all 4 cylinders rather than mostly the left side. It seems to be working just fine where it is so I will leave it there unless I find a compelling reason to move it.

A wide-band O2 sensor with a digital display is a must have for a one-off project like this. It gives you visibility into what's going on. You can make changes to the engine setup and see the effect in air fuel ratio. I would be lost without it.

Changes in engine and air temperature must be factored into the air fuel ratio as the air density changes with temperature. The air temperature sensor is a known sensor with specifications. But nowhere online could I find the temperature/resistance chart for the Magna's sensor. I just guessed at the settings. That is probably why it is runs so ragged while warming up. I am changing it to a Toyota temperature sensor that has known specs. .

I have learned a lot from this project. First was that i was naive about how simple it was going to be. EVERYTHING is critical. Air temperature, coolant temperature, battery voltage, fuel injector delay times...everything makes a difference. Very small changes can effect how the engine performs.

https://youtu.be/L_JAWalFRXE

Post Reply