Fuel injected 250/400s?

**Monsterbishi** · 15th February 2006, 09:14

Flyingdutchman, depending on when you get the project underway, I could have some small injectors towards the end of the year that should suit your application, Denso 200cc Top-feed, high impedance injectors, so should max out at about 27hp each.

They'll be from our Mustang, 44,000k's on them. It's a shame I didn't read this sooner, I had a alloy fuel rail that would have been perfect too that I binned.

**FlyingDutchMan** · 15th February 2006, 12:19

Originally Posted by imdying

Sounds like you're going to have fun

The sort of memory that you want isn't particularly cheap, I'm not sure why you want to use 2GB of it. I would've thought 128k of storage for maps would've been sufficient? What rev/vac/tps/granularity of values are you planning on sampling at?

There's a small number of cheap readily available microprocessors you could use, might just be easier to buy one (you can generally get a developers kit from their websites for working with the instruction sets) than to make one yourself.

I think if might've been the speed of the older microprocessors (as often used in cheaper car diy efi systems) that was insufficient for high revving bikes. Also worth considering which injectors... you'll want a smallish unit to keep the duty cycle up, otherwise you'll have bugger all granularity there too. Can't suggest good units to use sorry.

I'd be suprised if as turbo 250 would be faster than a late RGV/RS anywhere but in straight line, maybe not even then considering the extra weight of a four stroke motor with a lump of cast iron bolted to it.

Not so sure on the $250-$300... inlet manifold, custom air metering thingy, injectors, ecu, wiring, fuel pump. A decent enviro resistant 20pin connector isn't cheap by itself.

Sounds like you're going to have fun

Well I've got a 2GB SD flash card from my dad. The PCB in it was cracked, so he let me keep the flash chips if I could get the photos out of it for him. Damn cheap way of getting nv memory imo. Also will have plenty of space to have an imdependant map for each cylinder too - this isn't the initial plan, but will be considered if the variation between cylinders is too great. The timing resolution of the injectors/spark plugs will around the 100ns mark. One of the harder things will be accurately determining eactly where the crank is and when. More than enough processing power for 18,000rpm machine. (its only 300 rotations per second).

As for the microprocessor, I'll be doing it inside an FPGA, with my own custom instruction set (native handling of switch statements for one). Have some useful instructions unlike a lots of the damn things out there at the moment. Also since its an FPGA, do independant timing for each injector/spark plug etc, is a breeze since it lots of things can run in parallel on it (though a seperate microprocessor core for each cylinder is going a little too far me thinks).

Not really interested in doing a turbo though.

**FlyingDutchMan** · 15th February 2006, 12:23

Originally Posted by Monsterbishi

Flyingdutchman, depending on when you get the project underway, I could have some small injectors towards the end of the year that should suit your application, Denso 200cc Top-feed, high impedance injectors, so should max out at about 27hp each.

They'll be from our Mustang, 44,000k's on them. It's a shame I didn't read this sooner, I had a alloy fuel rail that would have been perfect too that I binned.

It sound awesome... I was dreaming about getting low impedance ones (they can turn on faster don't they?), but anything that is the right capacity will work fine. A basic fuel rail is pretty easy to knock up - some nice thick pipe with holes drilled at the appropriate spots. Pump one end, pressure regulator at the other.

**R6_kid:Me** · 15th February 2006, 12:41

as the aussie guys found out, the only way to get injectors that work properly at 18,000rpm is to talk the F1 people... and they arent gonna give it away cheap.

From memory that aussie guy managed to get it working, and it did give him more power but it was a shitter to start and sometimes would randomly cut out for no reason.

Good luck with it, i dont see why you dont just buy a bigger bike - no one picks up chicks on 250's trust me.

**imdying** · 15th February 2006, 12:56

Originally Posted by FlyingDutchMan

(though a seperate microprocessor core for each cylinder is going a little too far me thinks).

Not really interested in doing a turbo though.

Hahaha, for sure. Don't ignore turbos though, you'll want one eventually

I wonder just how bad single point would be... Lowers the tech of pretty much every part, might be worth investigating. Should be easy to setup a crank trigger... I've some ABS sensor rings you could have it you like

**stanko** · 15th February 2006, 19:50

I think the big problem will be with the injectors at 18,000 rpm that gives 3.3milliseconds per rev , but you have 2 revs per cycle so 6.6mS to do an injection cycle. I think thats about how long it takes for an injector to open and start flowing so while the electronics may be fast enough for a quick wank between injections the mechanicals will still be rubbing the sleep out of their eyes. If you have a really short time to inject the fuel increasing the pressure (normally 2 or 3 bar over manifold pressure) to about 10 bar might help.

**FlyingDutchMan** · 15th February 2006, 22:05

Originally Posted by stanko

I think the big problem will be with the injectors at 18,000 rpm that gives 3.3milliseconds per rev , but you have 2 revs per cycle so 6.6mS to do an injection cycle. I think thats about how long it takes for an injector to open and start flowing so while the electronics may be fast enough for a quick wank between injections the mechanicals will still be rubbing the sleep out of their eyes. If you have a really short time to inject the fuel increasing the pressure (normally 2 or 3 bar over manifold pressure) to about 10 bar might help.

Thats why I was interested in low impedance injectors - they can turn on quicker than high impedance ones. Involve a bit more electronics, but I handle that. Can do the same with the high impedance ones to by driving them with an larger voltage (more complicated than low impedance electronics)... 100 volts should kick those things open pretty quick.

But even if they stay partially on, by applying PWM you can adjust how open they are and still get reasonable (though not ideal) fueling.

**imdying** · 16th February 2006, 10:14

Originally Posted by FlyingDutchMan

still get reasonable (though not ideal) fueling.

Which sounds like what your (free ) carbs give you :/

**DEATH_INC.** · 18th February 2006, 03:56

At high rpm your injectors are open most of the time anyway....unless your talking F1 stuff the operation is pretty basic, there's not a lot to be gained from firing the injectors sequentially either, most are just batch fired.
The abs ring idea is good,but you'll need to either fill in between two of the points or cut one out to give ya a reference....but why not just use the ignition trigger that's already there?
Another idea for high rpm, use two injectors per cylinder, firing them alternately....not that hard to set up.