Idle Fueling and ITB's

[cheftrd]

I recently had a real fight with a set of TWM DCOE ITB's on an L, that had me going in circles for months. Car ran fantastic. Turn-key start and idle from dead cold with only ignition timing manipulation (No idle air control), awesome transient response, good power, and excellent drivability... But it would foul plugs at idle, no matter what I did. Tuned loads of engines with ITB's, but always with the injector at the port.

Finally got it figured out as an injection timing issue. I'm posting the write up here in case someone else ever comes across it, and hopefully won't have to struggle with it like this dummy did   You can also find it on my Facebook blog. Motorworx

 

Warning: This is going to be long, and go to nerd level-7. Eject before it’s too late.

I’ve been having a problem with a high compression, long duration cam engine running ITB’s that is fouling plugs on idle. Each fuel injector is placed just behind each throttle plate, and a significant distance from the intake port and valve. There are also two slight turns in the manifold before the fuel gets to the port. I was told the engine would idle without fouling the plugs on the Mikuni DCOE carburetors, and that left me with the question: "What's the difference between carburetor idle fueling and port injection (PI) idle fueling?" The most significant thing I came up with was the timing of the fueling.

I didn’t have the car, so I decided to dig a little. I bounced the problem off the good people on the FB group “Guild of EFI Tuners,” and some of the people with experience with this kind of set-up said “injection timing,” and “higher fuel pressure,” for better atomization. I also bought SAE Technical Paper 2000-01-2834 “The Effects of Port Fuel Injection Timing and Targeting on Fuel Preparation Relative to a Pre-Vaporized System.”

I have quite a bit of experience calibrating ITB engines in pure alpha-n, alpha-n in VE, etc., but it’s always been in PI engines with the injector at the port. In that case, injection timing has always been best with intake-valve-closed (IVC) injection, end of injection in the 400° to 450° BTDCC (before top dead center compression) range. The above SAE paper, as well as SAE Technical Paper 2003-32-0023 “The Effect of Fuel Manifold Injection and Injection Timing on Performance and Emissions for a Spark Ignited Engine” both back this up. So does every stock engine I’ve logged this on. This is because injecting fuel IVC, onto the back of the hot intake valve and port floor just before the valve, gives the fuel time to vaporize, for better fuel preparation before the valve opens. Injecting IVO gives poorer fuel preparation, and a decrease in port fuel film (lower X-Tau). The better fuel preparation of IVC injection is what gives sequential injection its low-speed advantage over batch-fired injection. Vaporized fuel burns more efficiently, and cleaner than atomized fuel. It’s also why DI engines get choked with carbon; they inject directly into the combustion chamber, preparation is poor, and particulates abound.

So a carburetor in an isolated runner circuit will obviously deliver the highest idle fuel when the port flow is highest. Flow and vacuum are not necessarily exactly linked in time, but for the sake of what I’m doing here, they happen close enough together. This should be when the intake valve has opened, port pressure is dropping, and peaks when the intake valve closes, but air is still flowing to fill the port. Fuel should flow the least when the pressure is highest, which is just before the intake valve opens (unless you have a lot of valve overlap and reversion). This is the time when fuel is being injected in the standard PI model, when intake port pressure is climbing, and will peak just before the valve opens. With the injector at the port, aiming at the valve, it sprays onto the hot valve, so port flow is irrelevant to the fuel injection.

Turns out that injecting fuel into an air stream that’s slowing down, possibly reversing, far from the valve, puts a lot of it on the port wall, meaning a lot of it that’s getting to the cylinder is probably doing so in liquid form from the port wall.

The first picture just shows the two laptops on the car; one running the Haltech ECU Manager software, the other running the Pico DSO. On the scope, I’m capturing #1 injector voltage signal on channel 1, and pressure in the #1 intake port with WPS500 pressure transducer on channel 2. This way I can see where the injector is opening vs. pressure in the intake runner.

The second picture is a marked-up scope capture of the injection happening at 440° BTDCC. Port pressure at the time of injection is about -20kPa and climbing. As the pressure climbs, the intake valve opens and a reversion spike from cylinder pressure (big cam, big valve overlap) drive the port pressure to zero, and over zero, very quickly. Then the vacuum very quickly increases, and I assume the port flow with it.

The third picture is a scope capture of the injection timing changed to 270° BTDCC. Port pressure at the time of injection is about -27kPa and falling quickly. At this time, I hadn’t changed anything but the injection timing, so it ran kind of rough, which was one of the hang-ups I had earlier... Everything in the book was tried, including changing injection timing, batch-firing the injectors, etc. When these changes were made, it ran worse, so in the spirit of "That didn't work" and "It's not the way we've always done it," it was abandoned quickly...dummy. What I hadn't keyed in on, and what one of the technical papers turned the lights on to, was the effect on AFR. Not the real AFR, as that doesn't change with injection timing, but as it's read by a sensor. The same amount of fuel is being injected for the same amount of air that's entering, but suddenly we're showing really rich because we're burning more of the O2. That's obviously a step in the right direction, but the excess fuel needed to be taken away to make it happy again.

The effect of changing the injection to mid IVO was a 7% decrease in the VE table (a lot less fuel) to achieve the engine’s “happy point.” The idle screw had to be reduced, as the engine was much “lighter,” (more torque) and sounded much smoother. Timing is zero-throttle-table @ 20° BTDC in both captures. 30 minutes of idling did not foul the plugs, where 10 minutes would do it before. A very interesting thing that happened was the effect on the fuel film. Where the engine had a crisp transient response on a snap throttle from idle at approximately 1.8ms of fuel in the zero-1000 transient fuel cell, it now required 7.2ms for the same crisp response. I assume this is because there is much less fuel on the manifold and port wall when the throttle is closed, and it has a lot of dry area to condense on when the throttle is snapped open. This is in line with the findings in the SAE paper on IVO vs. IVC injection timing. It also means it’ll go rich on closed throttle, and need some transient disenrichment tuning. Next round I will also bump the fuel pressure, and see how that affects it. Feel free to comment.

 

Matt

[Leon]

Great detective work! Ah the fun of engine calibration... Many think it's just dialing in fuel and spark tables but that's just the tip of the iceberg. It's work like this that really refines the experience. Well done.

[turbogrill]

On 10/30/2017 at 8:31 PM, cheftrd said:

The effect of changing the injection to mid IVO was a 7% decrease in the VE table (a lot less fuel) to achieve the engine’s “happy point.”

So with IVO you got the same AFR for some operating point with 7% less fuel? Does this translate to 7% better BSFC?

Very cool stuff! 

 

[TimZ]

On 11/1/2017 at 10:24 AM, turbogrill said:

So with IVO you got the same AFR for some operating point with 7% less fuel? Does this translate to 7% better BSFC?

 

This is at/near idle, so most likely no.  

At full power with properly sized injectors the injector is generally open much longer than the intake valve and the dwell time is generally much shorter due to higher rpms, so timing is less important there.  Probably some effect but it should be much less dramatic.  At idle/low power the effect is more pronounced due to the short injector open times still being able to "fit" within the intake valve open period.

[NewZed]

This seems to explain the inherently never-quite-smooth idle of the factory EFI system.  Different cylinders see different injection timing.  It's always bugged me.  

Maybe you could fabricate a small venturi in the ports and inject in to it so that the ITBs act more like a carburetor.  JK.

[Sam280Z]

Matt,

Check out this video put out by Andy Whittle.

 

The solution I came up with 5 years ago on my turbo ITB engine was a little more involved.

https://photos.app.goo.gl/hpt3FkUcJnnSSQ9s2

I cut a FI manifold flange off and welded it to replace the flange on a canon triple carb manifold. I have 19lb/hr mustang injectors at the ports and much larger injectors in the ITBs.

I got the idea from Joel's monster turbo stroker.

Sam

[cheftrd]

I saw that, and asked him about it. His timing is start of injection, and he found it best starting about 350°, which is 10°BTDC, or just when the valve is opening, so it's IVO injection. All I did above was changed it to somewhere I knew was going to be IVO, and its far from optimized. Once I play with it some more, I'm guessing that I will find the same results Andy did. The injectors on in the TWM DCOE throttles are also a little further away. It runs good now, but I'm also wondering how much better it might run with some Denso UC injectors with a 50μm SMD atomization vs. dirty DW modified 440 EV14's...LOL

 

Edited November 7, 2017 by cheftrd

[cheftrd]

On ‎11‎/‎1‎/‎2017 at 11:24 PM, turbogrill said:

So with IVO you got the same AFR for some operating point with 7% less fuel? Does this translate to 7% better BSFC?

Very cool stuff! 

 

Sort of, but not really...LOL.

 

It's only at idle, and because there was so much fuel going in dirty, it took a lot more fuel to get it to burn correctly. At 7k rpm, there's 17ms to get all the fuel in, so pretty much once you're past the mid range, there's no real benefit to be gained from tweaking injection timing, other than keeping fuel stand-off to a minimum, if it's even a problem.