DIY Fuel injection thread.

[mjrstar]

Normally ends up ~5-8 degrees timing maybe, as for more air yeah just whatever you can idle bypass screw if it has one, crack the throttle if TPS is still happy(sime mitsi have idle stop feedback).

Probably depends on stuff like intake volume size of throttle compression ratio and a billion other things.

You should be able to set idle targets (pwm output)at a range of different temps. 

 

I had major dramas with hot idle stability on my evo with even some pretty mild cams until i started using this trick. It does take a bit of getting right but is satisfying when you do.

I don't think there is any magic formula it is some trial and error but seems to work well on both map and afm cars i have.

As for afr's at idle they can be fairly inaccurate as a discussed above i go for feel rather than a number. 

 

[Yowzer]

Factory ECUs generally won't run any more than 10° at idle, as mjrstar said 5-8 is pretty standard, but ramps up to 15° or more as soon as you touch the throttle

[yoeddynz]

I shall have a play.

But for now I shall continue fucking about with our NZ new Sentra with a GA16DS that has a Carb which only a few countries got. It has three solenoids. It keeps over fueling on light throttle and then bogs down massively until it stops and wont start again. Fuck I hate carbs that try to be all smart and shit. I want to burn this car!

Going to fit a Stromberg and make it good...like a an English car

Rage!

[Testament]

4 hours ago, mjrstar said:

Normally ends up ~5-8 degrees timing maybe, as for more air yeah just whatever you can idle bypass screw if it has one, crack the throttle if TPS is still happy(sime mitsi have idle stop feedback).

Probably depends on stuff like intake volume size of throttle compression ratio and a billion other things.

 

 

yeah and non bypass non clutched non intercooled supercharged and injected means other things haha

going from whats said here I need to drop the timing, wind my throttle stop out a bit more. complete opposite to what I had been doing - currently have about 20 deg since I tuning for highest idle and winding the stop in to bring the rpm back down. I other thing I think I need to play with is my air temp corrections since the ability to heat soak and go mega lean off throttle is horrendous. have the water injection there but really I want to tune it so it runs fine without it (in case you run out of water of have a pump failure etc.) and use the air temp correction to make finer adjustments to fuel or timing if the air is denser.

[kws]

Shit, this is all so relevant to me too. I'm sitting at about 20 degrees timing at idle, which I always though was far too much. Increasing air and dropping advance sounds like something good to try. 

Im not fancy enough (yet) for an idle controller. All bypass screw and tuning here 

[kpr]

watch out things don't start getting hot.  if go too low on the timing,  you'll end up  with half the burn in the exhaust manifold.   i run about 15deg on 4age stuff,  much  less and will start pumping a lot of heat into the exhaust

[scooters]

On 9/8/2017 at 09:16, yoeddynz said:

Are you adding air with the idle air bleed screw or by using the idle control valve?  I'm asking because I only use my idle control valve, a PWM unit, to control the warm up circuit. After its warmed up I have set my idle using just the fuel table and spark settings. I'm idling around 13 13.5 and pretty smooth at that, but I get that not so nice rich smell about the car.  If I go much leaner, say to 13.5 - 14,  then it starts hunting and going a little rough.

And how much timing are you pulling?  My engine is somewhere around 15 degrees at idle so there is room to pull some and still be close to what I think would be a 'factory' type idle tune?

Edit; I must add that I'll never achieve a factory perfect idle, at low rpm, with an inherently unbalanced V6 running a super light flywheel. But its fun to try...

hows injector dead time ? was that set up with air fuels ? I can dig you up some how too stuff if your into it.. 

[scooters]

@yoeddynz or get me your injectors and i'll make a latency table up and characterise your injectors. Then it will be bang on. 

[ajg193]

So I discovered that converting a car to EFI doesn't actually stop it from needing petrol.

 

I spent an hour trying to prime the new fuel system yesterday, but it turned out that the tank was empty.

 

A quick walk of shame down to the local gas station to get some fuel and the car started right up!

 

It then died after about 10 seconds, it was no longer detecting RPM.

 

It turns out it is a common problem with the microsquirt boards for the R20 resistor to fail, rendering the opto-in ports useless. I soldered on a 0.5 W 1k resistor over the tiny surface mount R20 resistor and the car fired right back up. Now I've just got to tune it.

 

I've also got to replace the temperature sensor for the in-dash temperature gauge as the one that came with the EFI thermostat housing just reads zero. Hopefully I can get it out without breaking the housing.

 

The EGO just reads 12.1, I'm not sure what the story is with this. I might hit the calibrate AFR table button in tunerstudio and see if that gets it going. I calibrated the LC-2 sensor as per the manual that it came with.

 

EDIT: I hit the calibrate buttons it is now reading fine.

 

Car was idling at 22:1 though, explains the colour of the plugs..

 

Gotta get the idle sorted, currently it goes all over the pace and the RPM signal occasionally spikes up to 7000+ rpm

 

Edit: Idle and stuff all sorted, the key was to enable and tweak the noise filtering option for the RPM signal. Car runs better now than it ever did with the carb.

[ajg193]

Well tonight's tuning run didn't end too well. We managed to get the car tuned pretty much spot on, no hesitations or anything and then all of a sudden it went full rich and crapped out.

 

I'm thinking either the injector drivers died or the pressure regulator (35 years old) is stuck shut.

 

ECU still gets all of the signals correctly, so reads MAP, TPS and RPM correctly.

 

Any ideas?

[kyteler]

Whack a carb on it mate.  Problem solved.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pewhuehue

[yoeddynz]

Yep...

[Testament]

go home your both drunk

[ajg193]

Turns out in the excitement of getting the engine running I forgot to secure the loom out of the way of things. It contacted exhaust manifold and melted through an injector ground wire, locking injectors 1 and 2 full open. I've now fixed up the loom and tidied some stuff up, securing the loom properly. I pulled the head off the engine yesterday to make sure the flooding and fuel diluted oil didn't damage any cylinders (long story behind that choice). It all looks good in there so I'll throw a new gasket/oil/filter in on Tuesday and should hopefully be away laughing. 

 

I'm also going to install a fuel pressure gauge, even if only temporary 

 

I also put on a more durable MAP sensor. I calibrated the upper end of the readings to give atmospheric, but still need to see about calibrating the low end. It's a denso unit and I haven't found any data sheet for it. 

 

 

 

 

[Roman]

Hey are you using a normal NA Toyota map sensor? 

I've got some calibration data for them. if you can take a photo of it or post a part number. 

[Roman]

One thing I've not played with on my car until this weekend is trying to optimise the VVTI cam timing (inlet) under part throttle conditions.

Obviously at full throttle you're trying to stuff as much air into the engine as possible but what about at cruising when you are trying to get best economy?

The documentation from Toyota says that at cruising it advances the intake cam "about half way" to give some overlap to cause an EGR effect to improve emissions/economy.

I did some tests where I was driving a stretch of road with cruise control set to 3000rpm and closed loop lambda turned on. Starting and stopping a log file at same points.

I set cam timing to 0 degrees, then repeated at 10,20,30,40.

Closed loop lambda would keep trimming the fuel to reach goal AFR, and then I could monitor how much fuel in total each setting was using. 

Interestingly enough it seemed to make sweet fuck all difference, even when introducing a bunch of overlap. 

Might not apply similarly at a lower or higher rpm range but it certainly didnt have the obvious impact that this same test had when I was adjusting ignition timing the same way.

Or otherwise, I'm thinking it might impact EMISSIONS but not really economy.

[ajg193]

89420-97201

 

One of these, everything is the same except 4840 instead of 4841

 

I think I set something along the lines of 180 kPa @ 5.00 V and 10 kPa at 0.00 V (standard lower limit) to get 100 kPa at sea level on a 100 kPa day.

 

It's off a little Charade or something, I didn't actually pay enough attention to the car I pulled it off.

[Roman]

Ahh ok mine is 1.4v = 34kpa and 3.58 volts = 100kpa. 

So similar slope but offset by quite a bit

What voltage do you get when the car is not running? (~100kpa)

 

[Roman]

In other news, I just paid for a copy of Megalogviewer HD. 

As you can import .CSV files from most other ECU types into it.

The main feature that I'm interested in, is the ability to do maths on measured fields to calculate new ones. 

So for example, my ECU currently outputs CC per min of fuel consumption but you really need to factor the vehicle speed into this to get a true measure of economy. 

So you can do exactly that and then see a log of litres per 100km or km per litre or whatever. Which is cool.

But then secondary to this, you can use the histogram function to generate a table which shows you what economy values you are getting in each cell of a table. 

So you can make changes across a map (say globally add or remove 2 deg timing) and then see how that changes economy really easily. 

The UI is just generally speaking disgraceful but it goes hard for what it is.
 

But the table generator thingy and a few of the other calculated values will be awesome to use, I can see there are already some pre-existing calcs for working out coast down losses and power estimations and stuff like that. 

[yoeddynz]

I just like all the pretty colours on those squiggly lines that go up and down when my engine makes more noise.