Go Back   DIY Go Kart Forum > Building Plans And Advice > Engines & Clutches

Notices

Closed Thread
 
Thread Tools Display Modes
  #1  
Old 04-09-2016, 10:45 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default Honda G200 - EFI Conversion

Well , the plans to get a mikuni carb and finally get this engine running kinda went kaboom ..

So im thinking ... i have a Arduino kit coming my way .. why not convert this engine to EFI ?

After all .. EFI kits from the likes of ecotrons go for up to 600$ ...

For those who dont know what an Arduino is , its basically an open-source platform that revolves around inexpensive electronics , coupled with some #C language knowledge , you can make pretty much anything .

So here is the plan :
- I will use a ATMega328 microcontroller chip , which has 13 digital pins ( outputs or inputs ) , and 6 analog inputs .
- For ignition timing / RPM measurement and injection timing i will be using an NPN Hall Effect magnetic sensor
- For temperature indication i will use 4 K-Type thermocouples coupled with MAX6675 ADC modules , to convert the given analog signals to digital ones.
- Im yet to decide on how to get pressure readings ( intake ( MAP) , oil pressure for an eventual turbo to come , and fuel pressure )

The white / blue circles are points which are to be drilled and tapped to accept the K-type thermocouples . The blue sketch near the engine crankshaft is the way im going to mount the magnetic sensor inside the flywheel radius .

Since the sensor is threaded , i can adjust the pickup to make fine adjustments to the ignition timing .

Links to sensors / Arduino kit :
http://www.ebay.com/itm/New-Hall-Eff...0AAOSwrklVWiZF
http://www.ebay.com/itm/MAX6675-Modu...gAAOSw~1FUVKpW
http://www.ebay.com/itm/Sintron-Ligh...UAAOSwKrhVZSBv
Attached Thumbnails
20160409_160909.jpg   20160409_160916.jpg  

20160409_160921.jpg   20160409_160926.jpg  

  #2  
Old 04-10-2016, 08:57 AM
Toycrusher Toycrusher is offline
Senior Member
 
Join Date: Dec 2015
Location: Deep Deep South Texas
Posts: 168
Thanks: 4
Thanked 36 Times in 25 Posts
Default

That's awesome but do you have engine programming experience? Is there an available baseline map you can start with? Most efi engines start around 30 HP. Where will you find a small enough injector? I'm pretty sure you will need a pwm driver for your injector. I'll be watching this closely!
__________________
-Professional Bench Racer...
  #3  
Old 04-10-2016, 09:21 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Quote:
Originally Posted by Toycrusher View Post
That's awesome but do you have engine programming experience? Is there an available baseline map you can start with? Most efi engines start around 30 HP. Where will you find a small enough injector? I'm pretty sure you will need a pwm driver for your injector. I'll be watching this closely!
Yes i do ! We do #C in school , so if i do get stuck , programming support isnt far away

Injector wise im looking at an injector from a 250cc scooter , which i will also use the throttle body from .. and for controlling it i will use PWM , with help from a IRFZ44N N-Channel MOSFET . Basically , the injector is "hot" all the time , meaning it is connected to the positive 12V all the time , while the MOSFET is hooked up in series to the ground .

PWM signal can be taken from the ATMega328 chip via any of the 13 digital I/O pins so that is already taken care of .

I havent seen a baseline map before , so i will need to start from scratch . But , there will be 4 major parameters :

IGNITION ( TIMING / INJECTION ) - self explanatory
TEMPERATURE ( oil , intake air , exhaust and cylinder / cylinder head)
PRESSURE ( oil - if i decide to turbo this , manifold pressure and fuel pressure )
THROTTLE POSITION - self-explanatory

Now .. right there it becomes complicated .

Throttle position dictates ignition and injection parameters , while in the same time , temperature and pressure readings can influence the ignition parameter without the throttle actually moving .

For example - throttle is at 100% opening , but due to wrong calibration on my end , the engine is overheating ( temperature reading is out of scale ) the controller will cut the ignition to slow down the engine or if its severe enough shut it down all together .
  #4  
Old 04-10-2016, 09:58 AM
itsid's Avatar
itsid itsid is offline
Moderator
 
Join Date: Jul 2013
Location: Ruhrpott [Germany]
Posts: 9,714
Thanks: 1,326
Thanked 4,113 Times in 3,074 Posts
Default

A few hints:
try to drill and tap the engine case as little as possible!
the pwm signal can easily be taken from the stock magneto flywheel;
even easier, coil a length of wire around the sparkplug cable and take that as you ignition timer signal.
(you might need to clean the signal on it's way to the input, but it'll work nicely .. that's how many very expensive tachs read the rpms )

engine temp is highest at the head, and you can easily epoxy the thermocouple between two fins.
Oh btw.. I'd suggest an N-Type instead of a K Type, it's a tiny bit less accurate in the expected temp range but it costs only a fraction ..
frankly, I'd rather use a thermistor anyways, since it's even cheaper, easily covers the temp range needed, is smaller and since it'll read either 0 or infinity when it fails, can easily be monitored for internal failures

The engine already has a low oil sensor (with respective hole in the crankcase)
swap the existing oil sensor against your own (additional temp reading) and there's no need to drill another hole at all

good luck with the conversion...

'sid
__________________
Jokes about german sausage are the wurst.
  #5  
Old 04-10-2016, 11:57 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Quote:
Originally Posted by itsid View Post
A few hints:
try to drill and tap the engine case as little as possible!
the pwm signal can easily be taken from the stock magneto flywheel;
even easier, coil a length of wire around the sparkplug cable and take that as you ignition timer signal.
(you might need to clean the signal on it's way to the input, but it'll work nicely .. that's how many very expensive tachs read the rpms )

engine temp is highest at the head, and you can easily epoxy the thermocouple between two fins.
Oh btw.. I'd suggest an N-Type instead of a K Type, it's a tiny bit less accurate in the expected temp range but it costs only a fraction ..
frankly, I'd rather use a thermistor anyways, since it's even cheaper, easily covers the temp range needed, is smaller and since it'll read either 0 or infinity when it fails, can easily be monitored for internal failures

The engine already has a low oil sensor (with respective hole in the crankcase)
swap the existing oil sensor against your own (additional temp reading) and there's no need to drill another hole at all

good luck with the conversion...

'sid
Well , the PWM for the RPM / Ignition is going to be provided by the NPN Hall Effect magnetic sensor . It can support up to 320Khz switching frequency , so its all good there .

PWM for the injection is going to be done by calculating the time in ms , of how long is the intake valve open @ set RPM .That will need some clever mathematics , an easier solution is to use 2 more senors at intake opening and closing positions , and then set the injection period as time till it takes the flywheel to pass from opening to closing position .

Reason using sensors is that due to how they work , it enables me to play around with ignition timing ( retard and advance ) on the fly . If i position the sensor closer to the flywheel , it will register it earlier thus again using clever math to determine when is the piston at TDC and how to fire the spark in relation .

Il see about the thermocouples , but il be using ADC converters anyways ( MAX6675 ) to hook them up to the digital inputs .
The analog ones are going for the pressure sensors .

And actually , the engine doesnt have an oil sensor , the hole i pictured is filled ! Thats why i have to drill it . But thats the only place il be drilling the block , anyways .

But ty for the tips anyways !
  #6  
Old 04-10-2016, 04:50 PM
Toycrusher Toycrusher is offline
Senior Member
 
Join Date: Dec 2015
Location: Deep Deep South Texas
Posts: 168
Thanks: 4
Thanked 36 Times in 25 Posts
Default

You won't want to be spraying fuel the entire time the intake valve is open
__________________
-Professional Bench Racer...
  #7  
Old 04-10-2016, 05:52 PM
karl's Avatar
karl karl is offline
Senior Member
 
Join Date: Feb 2011
Location: North east Ohio
Posts: 1,266
Thanks: 86
Thanked 222 Times in 179 Posts
Default

This will be interesting. What about the throttle position sensor? And fuel injector? Anyway I Want to try this one day, with a turbo.
  #8  
Old 04-10-2016, 11:15 PM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Quote:
Originally Posted by Toycrusher View Post
You won't want to be spraying fuel the entire time the intake valve is open
Yes , what i mean was during that period , the injector is going to be "enabled" so to speak , so it can inject during that period .

The amount is in turn dictated by the throttle position , RPM , manifold pressure and intake air temp.

---------- Post added at 11:15 PM ---------- Previous post was at 11:09 PM ----------

Quote:
Originally Posted by karl View Post
This will be interesting. What about the throttle position sensor? And fuel injector? Anyway I Want to try this one day, with a turbo.
Easiest to do it ( i should be ordering the injector and throttle body with position potentiometer ) is with the two from a scooter .

I got both from a local site for 600 kuna ( 90ish $) Used but functional .

Il get some pictures up later today , and the design to the timing / ignition signal pickup
  #9  
Old 04-11-2016, 05:18 AM
Toycrusher Toycrusher is offline
Senior Member
 
Join Date: Dec 2015
Location: Deep Deep South Texas
Posts: 168
Thanks: 4
Thanked 36 Times in 25 Posts
Default

Most automotive engines actually inject fuel on each revolution, even though the intake valve is closed every other revolution. 1/2 of the fuel required is injected each time. This allows for more exact to-the-milisecond decision making by the PCM
__________________
-Professional Bench Racer...
  #10  
Old 04-11-2016, 05:39 AM
Poboy kartman Poboy kartman is offline
Banned
 
Join Date: Jul 2013
Location: White Settlement Texas
Posts: 12,463
Thanks: 3,769
Thanked 2,952 Times in 2,298 Posts
Default

Quote:
Originally Posted by Toycrusher View Post
Most automotive engines actually inject fuel on each revolution, even though the intake valve is closed every other revolution. 1/2 of the fuel required is injected each time. This allows for more exact to-the-milisecond decision making by the PCM
I find that very hard....no...make that impossible to believe!!!!! What you're saying (whether you realize it or not) is that the injectors are spitting fuel into the exhaust cycle ....ummmm....not very energy efficient!
  #11  
Old 04-11-2016, 06:15 AM
itsid's Avatar
itsid itsid is offline
Moderator
 
Join Date: Jul 2013
Location: Ruhrpott [Germany]
Posts: 9,714
Thanks: 1,326
Thanked 4,113 Times in 3,074 Posts
Default

And I agree with Doug..

My 1970's K Jettronic fuel injection didn't, and no fuel injection I've ever seen newer than that did that.
I don't say there aren't exceptions.. but I also find it impossible to believe the "Most"-part of the claim
especially since fuel efficiency is a big selling point these days.

'sid
__________________
Jokes about german sausage are the wurst.
  #12  
Old 04-11-2016, 06:25 AM
Poboy kartman Poboy kartman is offline
Banned
 
Join Date: Jul 2013
Location: White Settlement Texas
Posts: 12,463
Thanks: 3,769
Thanked 2,952 Times in 2,298 Posts
Default

Now....I might believe it would inject 1/2 at the top of the compression stroke...(as the valve is open) and the other half as the piston nears TDC and just before it fires (when the valve is closed) ....but then again...I might believe in Bigfoot and UFO's. (both of which exist....and I can prove it!) Don't bet me....you'll lose! !!
  #13  
Old 04-11-2016, 09:00 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Here is the current design i will be basing the timing / ignition / injection control on .

Of course , this is just a sketch for now , i havent taken these measurements on the engine yet , which i will do as soon as i get some time .

Sorry for the crappy photos / sketch , but i did this to help anyone else who is attempting the same , and to explain the procedure along the way .

S1 and S2 are the magnetic sensors which i will use to determine the intake valve opening and closing . As you can see on the graph which i sketched up , the time period in which it takes the flywheel magnet to pass from S1 to S2 is the injection period . The amount or duration of injection is controlled by other stuff , to which i will get later .

S3 is the ignition / timing sensor , which does the equivalent of your magneto . But , it does get a big improvement , or at least few improvements over the stock magneto . First , there is no need to fiddle around breaker points , or the distance of the magneto to the flywheel , and the sensor doesnt need maintenance , and as an additional bonus , it already has all the needed electronics inside of it !

Meaning , it outputs a perfect 5V square-wave , which can be directly connected to the ATMega328s digital ports.

Tho there is a catch , as some of you may know : This is a 4 stroke engine , meaning ignition happens every 2 turns of the crankshaft , and so does the intake cycle !

And this is easily solved by programming the chip , so it reads 1 turn , second turn ( injection -> igniton ) , then resets the counter and keeps going till the engine is running , so a loop program is needed.

Again , i will need to find the engines timing in the manual sheet , and measure the intake duration in degrees , the sketch is only a "model" .

Also , the mounting plate is going to be 2mm thick steel sheets , thick enough to provide adequate support , and thin enough so i dont need heavy machinery to work on .
Attached Thumbnails
20160411_164046.jpg   20160411_164017.jpg  

  #14  
Old 04-11-2016, 09:18 AM
Toycrusher Toycrusher is offline
Senior Member
 
Join Date: Dec 2015
Location: Deep Deep South Texas
Posts: 168
Thanks: 4
Thanked 36 Times in 25 Posts
Default

Quote:
Originally Posted by Poboy kartman View Post
I find that very hard....no...make that impossible to believe!!!!! What you're saying (whether you realize it or not) is that the injectors are spitting fuel into the exhaust cycle ....ummmm....not very energy efficient!
Yes, most multiport injection systems fire into the intake (not exhaust) each revolution. First spray sits in there with valve closed and second spray hits at appropriate time during intake cycle. Put test nodes on an injector and you'll see what I mean.
__________________
-Professional Bench Racer...
  #15  
Old 04-12-2016, 09:01 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Continuing on with the timing and ignition , a bit of explanation about how im going to use the magnet sensors , and why .

First off , the sensors themselves and why or how do they give advantages over the standard magneto system .

The sensors in question are NPN - Hall Effect sensors with built in electronics enabling them to be directly hooked up , either to a microcontroller or a tachometer . They have a maximum switching frequency of 320Khz ( 320 000 times in a second ! ) , so these will work at very high RPMs too without needing modifications !

They do not require any maintenance , just a power source , and something to use their output signal . Tho , they do require positional calibration , just like the magneto . Simply said , like the magneto they need to be a specific distance from the flywheel magnet , but its a bit trickier .

Before you view the pictures , my drawing skills arent best , so any confusion is on me .

See , the sensors have whats called a hotspot , and secondary pickup fields . Reason to this is , the sensor itself emits a small magnetic field , and the sensor measures minute changes in its shape , thus outputting a signal . Also , the top of the sensor is plastic , while the case is metallic .

Thats why it has the smaller secondary sensor fields , and the strongest one , the "hotspot" in the middle .

This presents an issue when you need 1 pulse signals to measure RPM or whatever , because if the sensor is too close to the magnet , it will read the same magnet , 3 times in a single pass !

That is not what you would want in a precision measuring environment where signal counts are critical .

So to fix this issue , experimentation is needed , and is usually straight forward . The sensor needs to be close enough so that the magnet can be sensed by the "hotspot" but not too close that it touches the smaller secondary fields .

This is easily solved by experimenting , and since the sensor case is threaded , it can be fine-tuned for the application , which i will demonstrate once the sensor arrives .
Attached Thumbnails
20160412_163854.jpg   20160412_163410.jpg  

20160412_163417.jpg   20160412_163422.jpg  

  #16  
Old 04-14-2016, 10:30 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

The arduino programming PCB has arrived !

As you can see all the connection points are clearly labeled , 13 digital I/O ports , 5 analog inputs , and power inputs / ground connections.
Attached Thumbnails
20160414_153344.jpg   20160414_153336.jpg  

20160413_224102.jpg  
The Following User Says Thank You to NewbCarter For This Useful Post:
Jrgunn5150 (04-15-2016)
  #17  
Old 04-16-2016, 01:25 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Quote:
Originally Posted by chancer View Post
What do your weeds have to do with EFI on a G200?
Buy you books send you to school you eat the teacher!
Im not sure i understood what you meant with that
The Following User Says Thank You to NewbCarter For This Useful Post:
chancer (04-16-2016)
  #18  
Old 04-16-2016, 04:19 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Here im making a plate to hold the electronics , which will be housed in a plastic waterproof box to help protect the microcontroller from the elements
Attached Thumbnails
20160416_111347.jpg   20160416_110619.jpg  

20160416_110354.jpg   20160416_110405.jpg  

IMG-20160416-WA0026.jpg  
  #19  
Old 04-16-2016, 04:58 AM
NewbCarter NewbCarter is offline
Senior Member
 
Join Date: Sep 2015
Location: Solin , Croatia
Posts: 352
Thanks: 71
Thanked 59 Times in 56 Posts
Default

Here is the finished plate , and a look thru the engines flywheel / fan shroud .

The hole thru whitch the sparkplug wire used to go will serve nice for the sensor wires
Attached Thumbnails
IMG-20160416-WA0032.jpg   20160416_125338.jpg  

20160416_125349.jpg  
  #20  
Old 04-16-2016, 05:07 AM
Desertduler Desertduler is offline
Senior Member
 
Join Date: Mar 2013
Posts: 1,735
Thanks: 469
Thanked 668 Times in 414 Posts
Default

Kudos to you for doing this conversion, may I make a suggestion? The abrupt 90 degree that you have on the exhaust really impedes flow, if indeed you need to have a 90 degree angle in the exhaust then for maximum exhaust flow it needs to be constructed with a radius bend in the 90 degree section of the exhaust pipe, that abrupt 90 degree section in your pipe is bad for performance.
Closed Thread

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -6. The time now is 10:21 AM.