TBI vs Port injection

[Sledge]

Yes, I've done many searches, looking for the latest opinions as to whether port injection is worth the extra $$ vs. a good TBI setup. My funds are not of the unlimited kind.

Thanks

[bobcowan]

It depends on the application.

The air:fuel mixture needs to be moving along towards the cylinder fast enough to keep the fuel in suspension. If it slows down too much, the fuel will drop out of suspension and puddle up in the manifold, and the engine will run lean.

With a big cam, big heads, and big intake track, that happens at normal idle speeds. That's why you see high performance carb engines idling at 1200 rpm.

If you're building an engine like that, then I would recommend a port injection. It helps to avoid most of those problems, and it will idle pretty nicely at low rpms, and be very easy to drive.

Also, the port injection system can time the injectors to fire at exactly the right moment, just as the intake valve is opening. That makes more power with lower emissions, and be more efficient.

If it's a fairly mild engine - stock cam and heads - then the intake charge will be moving along pretty quickly at low speed. And then a TBI system will work fine.

[joyridin']

A lot depends on the look you are after also. You can pretty easily hide a TBI unit under and air cleaner with ease. A buddy has a 66 Corvette with a 427 and TBI and unless you remove the air cleaner, you can't see it. A port fuel set-up is pretty obvious unless you go with stack injection and can run the wiring under the intake or something.

If you do not care about looks, the port fuel is the way to go. In some cases, you can literally tune each cylinder throughout the whole RPM range. Can't get any better than that.

[Luce]

Actually, OEMs open the injector shortly after closing the intake valve so the fuel boils to a vapor on the back side of a closed valve. THIS gives 100% atomization and minimum emissions at low speed.

Also, near max HP high rpm, the injector if not grossly oversized is open 75% duty cycle. the intake valve is only open ~30% of the time.

Port will give a much better idle quality and drivability, but a TBI actually can produce a tiny bit more top end power because the wet fuel in the manifold evaporating gives a cooling effect to the intake charge, like in a carb. Carbed engines can make a few more horsest han port injected, all other things equal.

[cycleguy55]

Luce: Thanks for your comments. I concur.

FWIW, given the operating conditions of most vehicles (Cobras included), MPFI would seem to be the better choice. Improved idle quality and low RPM drivability would seem to be more important than a few extra high RPM horsepower, IMO. I know there are those that track their cars, so their decision may be different, but MPFI seems the better choice for a 'street' car.

Question: Given the improved low RPM behaviour of a MPFI system, can a 'slightly' more radical cam be used to provide more high RPM horsepower? What's the impact on midrange drivability?

[bobcowan]

Originally Posted by cycleguy55

Question: Given the improved low RPM behaviour of a MPFI system, can a 'slightly' more radical cam be used to provide more high RPM horsepower? What's the impact on midrange drivability?

Yes! This is where the MPFI really shines. I had the "biggest" cam I could get into my 427W. It idled smoothly (for a big cam) at 850 rpms. Low and mid range drivability was smooth as silk. But mash the throttle and it would scream like a banshee and take off like a rocket.

I don't think I'll ever go back to a carb.

[eschaider]

TBI was a stepping stone in Detroit's transition to a port injection model EFI. They don't use TBI anymore. There's a message there. Maybe you shouldn't either.

Ed

[Luce]

With a victor or supervictor manifold and 4bbl throttle body, other than fuel rails, it can look fairly traditional, and yes, you can tame a much bigger cam with port injection. (I'm running megasquirt2 in batch fire, so not full sequential injection, but much better than wet wall. And as Bob says, you stick it there is absolutely zero hesitation.

But due to reversion throwing some of the fuel back up the manifold below 600 RPM it still wont quite idle like the wifey's lexus.

A thought I had if you really wanted max power and best possible idle would be staged injection where you have port for low to mid power with smaller (more accurate) injectors at the port and some big ass foggers at the throttle to cool the charge along the manifold path. I seem to remember some race engines with trumpets, and the injectors 2" above them for this effect.

[Gaz64]

The reason we run at 75-90% max power duty cycle is so the injector is just large enough for the engine to produce the desired max power.
No point in running excessively large injectors to try and get all of the fuel in at 30-40% duty cycle (typical inlet valve timing), since the idle fueling would be near uncontrollable.
I do like the idea of staged injection per cylinder, to have a small injector for idle low rpm, and a second high mounted injector for high rpm.

[eschaider]

The low speed idle is not that difficult to manage with large injectors if the tuner is willing to do some work in terms of getting the injectors properly characterized.

I use dual 2200cc injectors on a 'little' 281 inch SVT Cobra race engine. We needed the dual injectors because at high boost, on methanol,l the engine's appetite for fuel substantially outstripped a single 2200cc injector's capabilities. After a day racing and before the car is put up each evening we drain the methanol, put in gas and start the car on its gas tune to flush out any remaining methanol to protect the fuel system.

Even with dual 2200cc injectors the car idles around 1000 rpm and is very drivable on gas with the big injectors. The drivability issue revolves around having properly characterized the injectors using SAE's J-1832 low pressure gas fuel injector standard and then properly using that data in the tune.

This J-1832 characterization gives the tuner the ability to know exactly what the injectors do, rather than assuming generic published data for the injectors. The differences especially in terms of the low slope or idle range of the injector can be substantial. Knowing the injector flow characteristics in that range goes hand in hand with good idle and off idle throttle response.

For a n/a engine I usually recommend sizing the injector to the owner's target power level and adding another 15% of head room for 'good air'. For a supercharged engine you want to determine the maximum useable boost the supercharger is capable of and size for that boost even if the owner is claiming he is not going to go that high — he eventually will.

If you don't build in headroom for increased boost, the owner will finding himself spending lots of money to take an incremental but necessary step up in injector size to accommodate his new found boost levels. Of course even at the high boost levels you want to leave another 15% of head room for those 'good air' days.

The biggest challenge with idling the engine on big injectors is the low slope flow characteristics of the injector. It is tedious and uses up a lot of dyno time to map out the fuel and ignition tables at engine idle and just off idle speeds for varying load and voltage situations. The tuners ought to be doing this as SOP no matter what injector is being used. Sadly some tuners will skip over this because of the drudgery and the 'lack of time' to get the injector characterized.

Most injector flow tests are done as a static max flow exercise. Dynamic flow tests are run across a wide range of pulsewidths and voltages to properly characterize an injectors’ behavior from its initial point of opening, through its low slope nonlinear range, up into linear and finally static flow.

The availability of dynamic flow test data in the hands of a tuner who knows how to put it to use brings with it a much better drivability with significantly improved low speed and mid range throttle response.


Ed

[Gaz64]

Originally Posted by eschaider

The low speed idle is not that difficult to manage with large injectors if the tuner is willing to do some work in terms of getting the injectors properly characterized.

I use dual 2200cc injectors on a 'little' 281 inch SVT Cobra race engine. We needed the dual injectors because at high boost, on methanol,l the engine's appetite for fuel substantially outstripped a single 2200cc injector's capabilities. After a day racing and before the car is put up each evening we drain the methanol, put in gas and start the car on its gas tune to flush out any remaining methanol to protect the fuel system.

Even with dual 2200cc injectors the car idles around 1000 rpm and is very drivable on gas with the big injectors. The drivability issue revolves around having properly characterized the injectors using SAE's J-1832 low pressure gas fuel injector standard and then properly using that data in the tune.

This J-1832 characterization gives the tuner the ability to know exactly what the injectors do, rather than assuming generic published data for the injectors. The differences especially in terms of the low slope or idle range of the injector can be substantial. Knowing the injector flow characteristics in that range goes hand in hand with good idle and off idle throttle response.

For a n/a engine I usually recommend sizing the injector to the owner's target power level and adding another 15% of head room for 'good air'. For a supercharged engine you want to determine the maximum useable boost the supercharger is capable of and size for that boost even if the owner is claiming he is not going to go that high — he eventually will.

If you don't build in headroom for increased boost, the owner will finding himself spending lots of money to take an incremental but necessary step up in injector size to accommodate his new found boost levels. Of course even at the high boost levels you want to leave another 15% of head room for those 'good air' days.

The biggest challenge with idling the engine on big injectors is the low slope flow characteristics of the injector. It is tedious and uses up a lot of dyno time to map out the fuel and ignition tables at engine idle and just off idle speeds for varying load and voltage situations. The tuners ought to be doing this as SOP no matter what injector is being used. Sadly some tuners will skip over this because of the drudgery and the 'lack of time' to get the injector characterized.

Most injector flow tests are done as a static max flow exercise. Dynamic flow tests are run across a wide range of pulsewidths and voltages to properly characterize an injectors’ behavior from its initial point of opening, through its low slope nonlinear range, up into linear and finally static flow.

The availability of dynamic flow test data in the hands of a tuner who knows how to put it to use brings with it a much better drivability with significantly improved low speed and mid range throttle response.


Ed

I like reading your replies Ed. I hope others are absorbing your writings as well.

Gary

[eschaider]

Thanks for the kind words, Gary but you give me too much credit.

Most of the stuff I post is basic blocking and tackling, to use a football analogy. Once someone gets involved with EFI, if they go to any professional training sessions the instructor will take you through the same knowledge very early in their training sessions.

There is an excellent knowledge source in Michigan called Calibrated Success, which is where I got a lot of my knowledge. It is owned and operated by a man named Greg Banish.

Greg is a professional fuel systems calibrator that works on a contract basis for GM, Ford and Chrysler sometimes simultaneously I think, but he speaks writes and trains in everyday plain English that we all can understand. He is stunningly good and provides both live and video training for EFI suystems. He usually does the live training in the US although I suspect you guys might be able to lure him downunder with the right size class.

This is his website CalibratedSuccess.com (<= clickable). You will find links there to his books and DVD training videos. The two videos he has that I found particularly good for me were the Advanced Training Episodes 1 thru 4 and the Ford Advanced Tuning Training DVD - DVD_3.

The Advanced Training, episodes 1-4 is what he calls the universal truths and is applicable to all brands and styles of EFI systems. The Ford Advanced Training DVD is specific to the Ford OEM ECU's and OEM fueling strategies. Both are excellent.

I think the biggest benefit of the DVD's is the ability to go back a week, a month or years later and refresh the training in your head — I like them a lot. If you follow his lead when it comes time to tune your car you will only buy dyno time. The actual tuning will be done by yourself and one or two extra guys and probably better than the dyno operator would have done for you.

For those interested in the stuff I shared in the earlier post, here is a link to a downloadable paper on some of the subject matter. It is in PDF format. Greg writes and talks the same plain English we all speak and use. If you are interested in EFI I believe you will find his paper on Are all Fuel Injectors Created Equal? (<=clickable) and his other work equally if not more interesting.


Ed

[Luce]

Good read Ed.

I have mine running well beyond acceptable, so didn't bother going this deep, but I remember one method of calculating "dead time" or all of the dynamic variable you mentioned, involved running the engine in different pulse patterns where you found the correct milliseconds for idle with 1 squirt per cycle and milliseconds with 2 or 4 squirts per cycle. Of course, the engine should need the same amount of fuel, so if the sum of the 4 squirts is different from the 1 squirt you can back into a theoretical dead time.

Dead time is the injector constant megasquirt uses to assume the time needed to open the injector minus time needed to close the injector plus the time needed for the fuel to actually start flowing. It ends up being pretty close to a constant number if the injections are not at a resonant frequency (spring/mass oscillation) of the fluctuations in fuel pressure.

Different ECU's allow different back EMF bleed voltages that effect pintle closing speeds, as might the size and geometry of the fuel rail effect just after open flow rates, more so with batch fire, so published data is of limited use. (If you want to get into the weeds about this, consider the injector electrically is a coil, and takes volts over time to build the field that opens it, and also has to dissipate this energy to collapse the field to allow the spring to close it. If you just disconnect the power without a controlled bleed, you'll have a spark, and that's not OK.)

Your second link is a great illustration of the "dead time" the non linear kink for the most part can be baked into the fuel table cake if you're willing to take the backyard hack rout and tune it in. But the closer your characteristics are modeled into the fueling equations the more accurate and predictable your adjustments are across the board.

I've seen quite a few useful megasquirt videos of pro tuners going through driveability issues like accel enrichment and low speed timing. It's really interesting how you can use data logging and see what causes things.

Tuning for WOT top end is easy. It's the idle, off idle and transitions that take understanding of what's going on.

My real world experience has been limited to the Cobra and a 383 SBC with 8 stack where a little bit of idle lope is accepted and maybe even expected, but stalling or stumbling certainly is not.

[eschaider]

You have a very good working knowledge of the various variables Luce. In general it is better not to attempt to characterize the injectors in car. The number of variables necessary to control consistently, makes it challenging to duplicate individual results test to test. Aggravating this, is the fact each injector needs to be characterized individually. By far the best (and least aggravating) way to have the injectors characterized is by a shop with proper flow bench equipment already set up to do SAE J-1832 characterizations.

The MS3Pro systems are very, very impressive — especially for their price. They will allow you to 'tune' each injector to provide proper fueling on a cylinder by cylinder basis if you want to split hairs. Once you have the J-1832 characterizations the job is not done yet. Even when injectors are done on a bench that is specifically designed to do the characterizations, you will still find data points that are 'flyers'.

Virtually (but not all) physical processes have a relatively predictable progression as you increase volumes, temperatures, voltages, pressures etc. As a result when you plot the points you should get a smooth curve. Invariably when you get the characterizations back, even from Ford, you will find flyers in the data that do not land on the predicted curve and in fact tug the curve from its natural position on a graph. Those flyers, along with how you set up your spark and fuel tables will, depending on their number and severity, affect part throttle drivability and throttle response.

To correct for this I use an Excel spreadsheet that takes the characterization data and plots it to look for the flyers. When I identify a flyer I move it, inside it's data set, to fit by using Excel's curve fitting tools to render the curve and then reposition the flyer, reflecting a real world physical behavior. When you do this, Excel will provide you with an "R" factor that essentially tells you how closely the curve fits your data or vice versa depending on your point of view. You want an "R" value of 0.99 or better. A 0.99 value means a 99% accurate fit. This extra effort takes additional time but in the end is worth it.

Below is a pic of what the sheet looks like. Each of the graphics is carefully massaged to be certain any 'flyers' are properly accounted for and then correctly rendered. This process is repetitive and duplicated for each injector. The process is also time consuming.


Ed

[Luce]

I know better than to get lost in the weeds trying to read a SAE-J spec (or spend the $80 they want to download it from SAE), and the allowed flyback voltage may be an industry standard (that can easily be held constant by the injector driver electronics), but why I was suggesting the backyard hack method of using the various injector spray modes was because it would bake in that variable as well. I also had a fuel pressure gauge on my regulator, and at just above idle, the needle would violently bang on the pegs on both sides until it hammered the needle clean off of the stem. That may have been more spring mass resonance of just the gauge excited by the injectors, but it did make me think there was possibly resonances in the rails as well.

This is coming strictly from an electronics geek direction, but as you can see by the voltage offsets, the voltage is a very critical number in calculating opening time. Unless held constant, allowed flyback voltage is at least as big of a variable. If I don't recall the exact number but I think it was around -20V to -30V in the MS drivers. For example, if it was -24V flyback volts, the magnectic feild that opens the injector could collapse twice as fast as it opens. This doesn't necessarily mean the fuel will stop twice as fast. That would be determined for closing by the spring/mass of the pintle as well possibly fuel flow, and opening, by the force of the feild/mass and maybe fighting fuel pressure.
But something must be done to clamp the flybackvoltage or the resulting spark would wreck the transistor switching it. Since it's not mentioned, I'll have to assume the flyback voltage is constant or high enough to become irrelevant.

But wow. There is a bit more minutia in the specs than I was aware. I know MS has some rudimentary voltage compensation, but I think it's more limited in usefulness to starting conditions where battery voltage can sag to 6-9V. I think they rely more on a stout and well regulated alternator. And also, I haven't really felt a need with my applications to progress beyond MS2. There's a lot more just going to MS3 that was released 6 or 8 years back.

And I've been limited to relying on vendors claims of "matched sets" and with MS 2, you only have 2 driver circuits, so no individual cylinder trims. But even so, it's still an order of magnitude more accurate than a carb and able to compensate for weather and altitude.

If anyone is interested in what all of my yammerings about flyback is, check this out. I didn't find anything about a controlled flyback circuit, but the MS driver circuit does something similar to the frewheeling diode, but more complex so it can allow 20-30V instead of the .7-1.5 of just the diode in the article.

https://www.circuitbread.com/ee-faq/...ack-diode-work

Add- Here's the MS schematic. It is set at -36V by the zeneer diode and TIP-42 transistor. it's about hald way down, title block says output injector Megasquirt 1 - V3.0 board schematics MS 2 and 3 use the same MS1 board but add newer and more powerful CPUs. The board just has driver circuits and basic sensor reading circuits.

[eschaider]

Luce, my EE schooling was minimal and almost 60 years ago. Even worse I have not used it for almost as long. As luck would have it (usually) I can sneak around the need for having to use any of it. All the points you are making are valid but if you stick within the SAE spec, virtually all the EFI systems will provide the tables and screens to use the data of the SAE spec characterized injectors.

Some do better jobs than others. Megasquirt is impressive in the extreme for their thorough approach to the characterization issue. They also allow you to ignore it if your supplier did not provide the characterization data — very nice save there.. Max power WOT tuning is not particularly challenging. Part throttle, off idle, coasting, decelerating and then accelerating, and proper accelerator pump enrichment schemes are much more challenging.

The electronic redesign of a system is not always the shortest way home when you are attempting to tune an engine. If the basic design falls short of the mark for specific purposes that are important to your build, it is usually less expensive to buy a different system that handles the particular issue more to your liking than re-engineering the one you already have.

Stay inside the sand box the manufacturer supplied and pick a manufacturer that provides good tuning tools inside the SAE fenced in playground. Everything becomes easier when you do this.

On the other hand the Megasquirt systems provide the code logic so you can get creative if you want to. It is just a matter of deciding what's most important and how big a chunk of work you want to bite off.

My EE skills are old, rusty and dull so I just buy the box and stay inside the sand box to play with tuning. For those with the prerequisite skills there is absolutely nothing wrong with going about re-engineering portions of the system you want to improve.

My old hands are full just doing the tuning


Ed