DIY Fuel injection thread.

[yoeddynz]

Just make one- I made one from 2mm steel disc. Marked out the teeth, cut it with hacksaw and filed smooth. did not take as long as expected and was a nice simple srarngely satisfying wee job to do with music cranked up.

[Yowzer]

Is there an edge on the crank pulley that you can cut the teeth directly into? That's what I'm doing on mine since I don't want the tone wheel way out the front

[ajg193]

I can imagine that will end up just destroying the belt.

[Yowzer]

Ah yeah V belt. Probably not then.

[NickJ]

Do as Alex suggested, you'll be able to file back 2mm more than accurate enough, the backlash in the distributor gearing will give more grief than careful filing.

[yoeddynz]

I think there's a buffer built into the code anyway. 

Or am I just talking lots of food. 

[fletch]

Snap off every 2nd tooth?

[kpr]

can you fit a hall sensor in there to read the orginal 20 tooth wheel?   rather than vr.   

[ajg193]

My next attempt will be to just turn the teeth off and drill holes for a 24-1 or 28-1 pattern (I think I will get more voltage on 24-1 though) and just weld in some 1.6 or 2.4 mm diameter tig filler rods as teeth.

 

In the mean time I put the car back to fuel only and bodged a few settings to make it not overfuel on starting with the ignition noise (ignore first 8 ignition pulses and no afterstart enrichment above 65 Celsius). Seems to have helped a lot, car now starts after the first compression. Not sue how the 8 skipped pulses will affect the cold starting though, but if the noise is the same when the engine is hot or cold then it should be fine.

[Yowzer]

Why so many teeth? Just run 12 and you'll be mint.

[ajg193]

With cam speed you get half the resolution compared to crank speed, so it would -probably?- be a good idea to have a few more teeth.

[Yowzer]

It really makes no difference. You could run four teeth and still be sweet.

If you're after all the kilowatts and screaming it out to 12k, running sequential injection and CoP then yeah, higher resolution could give you an advantage (depending on the ECU), but otherwise there's no need to overcomplicate things.

[yoeddynz]

Still better then dizzycrabs 

[Evan]

I'm just going to leave this here for future reference/anyone else who's interested

https://www.aliexpress.com/af/weber-throttle-bodies.html?SearchText=weber+throttle+bodies&d=y&initiative_id=SB_20190323183021&origin=n&catId=0&isViewCP=y&jump=afs 

 

[ajg193]

1 hour ago, Yowzer said:

It really makes no difference. You could run four teeth and still be sweet.

If you're after all the kilowatts and screaming it out to 12k, running sequential injection and CoP then yeah, higher resolution could give you an advantage (depending on the ECU), but otherwise there's no need to overcomplicate things.

4 teeth doesn't really give enough accuracy from what I have tried. But then again, I was using an inadequate igniter for that experiment.

[kpr]

 1 tooth per combustion event is all you need for dizzy.   ecu sees trigger pulse and fires coil.  dizzy sorts the rest.   will need to setup the teeth so you have one of them pointing at sensor when engine is at tdc

do you have any external vr conditioners.  or vr straight into ms?

 

[ajg193]

46 minutes ago, kpr said:

 1 tooth per combustion event is all you need for dizzy.   ecu sees trigger pulse and fires coil.  dizzy sorts the rest.   will need to setup the teeth so you have one of them pointing at sensor when engine is at tdc

do you have any external vr conditioners.  or vr straight into ms?

 

VR conditioner is built in to ecu.

 

Also the tooth needs to point at sensor quite an angle before from TDC (like 60 deg or so) with megasquirt, that point gives a hard limit on the maximum timing advance possible.

 

The problem with only one tooth is that the engine speed can change significantly over the space of 180 degrees of crank rotation, so it can really end up being a guessing game for correct timing. The original 7K-E distributor was just 4 teeth and timing controlled by the ecu, but Toyota seems to use a better vr conditioning system than MS, giving cleaner signals during cranking.

 

If 4 teeth was fine for giving accurate timing then why does Toyota use 24 teeth + a single tooth wheel in shitty old 4A-FE engines?

[Yowzer]

Because they can. 

Just put a tone wheel on your crankshaft already. The freeplay in the drive between crank and dizzy is going to be greater than any inaccuracy caused by low tooth count anyway.

[BlownCorona]

can you snap off the opposite tooth to the missing tooth and tell the ecu the sensor is as crank speed with a 15-1 trigger. 

thats what ive done with my dizzy mounted sensor (12 tooth hall effect wheel with two oposing missing teeth). unsure if 15 teeth is any better for the maths though. 

 

HOWEVER. 

i also have a 36-1 crank mounted trigger wheel waiting to go on when i also change the superchrager pulley for more boost (is doest fit with the current pully) because im 99% most of my shitty idle and roughish cruise is down to all the backlash in the dizzy setup and all the tricks made to get it to run thinking its got a crank trigger. 

 

[ajg193]

45 minutes ago, BlownCorona said:

can you snap off the opposite tooth to the missing tooth and tell the ecu the sensor is as crank speed with a 15-1 trigger. 

thats what ive done with my dizzy mounted sensor (12 tooth hall effect wheel with two oposing missing teeth). unsure if 15 teeth is any better for the maths though. 

 

HOWEVER. 

i also have a 36-1 crank mounted trigger wheel waiting to go on when i also change the superchrager pulley for more boost (is doest fit with the current pully) because im 99% most of my shitty idle and roughish cruise is down to all the backlash in the dizzy setup and all the tricks made to get it to run thinking its got a crank trigger. 

 

Tooth (prior to removal of one) count needs to be a multiple of the number of cylinders, as with a megasquirt the ECU links a physical tooth with a "tach" signal (one tach tooth per cylinder) and then counts a set time after that to send the ignition pulse. This lets it resynchronise timing every 720/#cyl, rather than once per 720 degrees, supposedly improving accuracy.

Since 30 doesn't divide by 4, the system can't accept it. And removing any other teeth to try and make it work won't work either because then the tooth spacing won't be correct.

It's a bit of a shit limitation but that's just how they made it.