Camless Engine Capstone project WVU

[stealhty1]

Mitsubishi was trying to came with this idea, as well as an electric water pump

and further electric brake system, but it never happen

[wshelto6]

How did you wire the encoder, I got the same one but I have no idea whats up with all the different cables they sent.

[MrSirosis1]

Very interesting. How are you syncing the valve timing with the crank rotation? Some kind of pickup on the flywheel?

[bbaldwin1987]

@MrSirosis1 - There is an absolute rotary encoder measuring the angular poisiton of the crankshaft.

[naturalyshocked]

i think the engine rmp is to slow for the usage of a carburator.

try sourcing a fuel injector from a aprilla SR 50 scooter or any other fuel injector from the Piago imperium from itally.

with a slight mod in the algorythm to inject fuel.

is that a demo board from microchip?

you don't need a rotary encoder, just use math and a hall sensor.

you require 2 pulses from the hall sensor for 1 full cycle.

just devide 2 rotations into 4 labor strokes.

[bbaldwin1987]

@naturalyshocked

You right. It is going to slow. The solenoids could not keep up at any higher speed unless we bumped up the voltage. When we bumped the voltage up it snapped our coupler

we wanted exact angular measurements. Not just rotational speed. So we went with a absolute rotary encoder

[naturalyshocked]

but as far as i know, you only use 'rotary encoders' in industrial applications.

i don't know how long they work with lots of rpm.

cars in real live have 1 sensor on the crank, flywheel, enough to serve 4, 6, 8, 12, etc, etc, cyl.

i think your rotary encoder will wear out soon.

how you store the crank position after you powered off?

how do you detect 'upper death point'?

if your math is good, there should be no problem.

i think your looking the wrong direction

[bbaldwin1987]

@naturalyshocked

Our rotary encoder is a "abosulte MAGNETIC rotary encoder" it has no moving parts.

It will last as long as the magnet will stay magnetized. I think you might be missing the point to camless engine. We could completely control the valve train digitally. SO it doesn't matter what position the engines in when it stops we can start the engine on any of the 4 cycles when we restart the engine. Retaining the crank position after power down is not necessary.

[naturalyshocked]

'Retaining the crank position after power down is not necessary.'

why not?

how you know, you reached the upper death point?

i have worked as service mechanic in the industry, the machines i saw and worked on, need recalibration after a unexpected los of power happend.

'my machines' calibrate automaticly and have sensors on each end on the all axis.

ofcourse i'm not aware of all models out there.

: ) but yet i'm curious.

[naturalyshocked]

if your microchip = 40khz and your pulse from the upper death point sensor takes 1 second for a full rotation, your mcu has done 土 40 million cycles.

make a main(){...}

that measures 1 revulution and than devides into 2.

that's the time you have for 1 valve to open en close just in time.

if you use assembler you could countdown (decreasing the number or cycle required). 

[bbaldwin1987]

@naturalyshocked

I wish our microcontroller was 40khz. We had six months to design and construct. With limited resources. I'm sure if we had more time and money we could have gone a lot further with the project. 1 second per rotation of the engine is terribly slow. If an engine is running at 3000 RPMs thats 50 revolutions per second thus the intake or exhaust valve will be opening 25 times per second. An absolute encoder proved to be the easiest to implement, and code.

[naturalyshocked]

i know 1 rpm = 1 sec is slow, but was just an example.



[naturalyshocked]

main(){ wait for pulse loop IF pulse port goto 'openintakevalve' ELSE 'wait for pulse loop' goto 'openintakevalve' countdown goto 'closeintakevalve' countdown wait for pulse loop goto 'spark' countdown goto 'openexhaustvalve' countdown gotoc 'closeexhaustvalve' countdown goto 'openintakevalve' etc, etc. } you 'countdown ' loop is dynamic here. but you have lots of tools in your MCU, watchdog timer, interrupts. this is a RAW idea of code, requires some more logic. been a while, last pic probes

[naturalyshocked]

use bitshift to keep track of the 'strokes' in the cycle.

pulse ... bitshift ... countdown ... bitshift .... countdown ... pulse ... bitshift

a countdown is made by storing a number (i.e. 11111111 (255)) in a register is decrease it with 1 bit and compare with 0

don't forgot some intructions require 2 cycles on the MCU

[naturalyshocked]

it's easy to make a algorythm that measures / devides 2 turns from crankshaft.

than make various bit configuracions for various rpm's, amount of fuel, amount of time to open / close valves.

i gues you need to buy the entire fuel system from a aprilla SR 50 or design your own MAS flow sensor with heated wire.

Voila drive by wire!

the simplify the design even more, use springs to retract valves (like classic engines) and use only proper made selonioids or the downwards movement.

[bbaldwin1987]

@naturalyshocked

Our original design incorporated fuel injection, with a fuel management system, but they don't give our groups a ton of money at our school. We maxed our budgets out after the solenoid purchase and construction.

[alphabetica]

Hi. I am working on a project similar to this for my own engineering degree and I was just wondering what kind of solenoids you used. They both look to be different kinds?

[bbaldwin1987]

@alphabetica - A company called "Ledex" just google them. We used two low profile, and two tubular.