Dim efi

-
KitCarlson

KitCarlson

Well-Known Member
Joined
Jan 17, 2008
Messages
2,624
Reaction score
462
Location
Middle Tennessee
My 1966 Barracuda has ran well for the past 34 years. Recently I have found it hard to tolerate the gas smell after driving when parked in the garage. It is too early of a car to have evaporative controls. It also requires a few seconds of cranking to fill the carb after sitting for a couple days. It seems that others share this problem. Gasoline has been revised over the years. I am considering converting to a "do it myself" ignition and fuel injection.

I have done this work for other old vehicles with success. I have the electronic advance ignition already completed and starting with the the EFI. Since the car is very original, I wish to do the conversion in a way that it can easily be converted back to stock. I plan to use TBI, while not the best, it will be a significant improvement, solving the evaporative control and starting delay.

I have purchased a surplus 2-barrel TBI, I think it is found on 1989 Dodge trucks. I am disappointed that the TBI will not be a direct bolt-up, so I will have to make an adapter.

I should be able to put this together in a few weeks depending on my other chores that steal my time. I have electronic hardware to do about 90% of the tasks. I will need to add a circuit to interface the idle speed motor. Also it needs a few days of software revisions.
 

Attachments

  • TBI_R3361B.jpg
    76.3 KB · Views: 477
  • TBI_gasket.jpg
    69.8 KB · Views: 471
looks like the unit on my 89 Dakota.... rather than an adapter, what about grabbing a junkyard intake off a late 80's TBI 318? If that TBI is for a 360, you can have the 318 TBI intake milled to accept it. I'm interested in the set up and how to make it work, as it would be a nice economical upgrade to Jake's 65 when we get that far..... I'll be watching.
 
The intake swap may be the best. I am hesitant about the adapter offset and reduction messing with flow. Will an 80's 318 ports match to narrow ports of 66', 273 heads? If it does that is what I will try to find.

I may have a few questions about the use of the vacuum ports on the truck. I think I can use one that seems joined to the barrels for the MAP sensor connection, a large base one for PCV,and perhaps cap the rest.

I also may need to add a #1 sync sensor for the TBI injectors to fire in the correct sequence. I think this will provide more optimal control of fuel timing. The manifold design may impact this requirement.
 
66 273 will have different bolt angles than an LA right?

There's already a guy doing this on this board.

Got the adaptors for 2bbl intakes or 4 bbl intakes, custom harness and everything.

He's looking for beta testers, and I'm thinking about doing it.
 
The 66' 273 has the standard intake bolt holes, the 64-65' have a different angle and hole size. The 273 ports are smaller that the 340, and 360 heads. A large port manifold on small ported heads is undesirable, however small manifold on large ported heads will work.

I am doing this as a hobby, so I will use my own ECU and software. I have complete tuning control ability with it. For me it is easier than fixing or re-programming an OEM ECU.
 
Interesting.

You and that other guy need to get together.

The biggest gripe in that thread was lack of tuning ability.

I think he was using 89-92 truck 318 TBI intakes.

I have a couple Eddy performers with 318 size runners. LD4B also has that size. Weiand action plus as well IIRC.

Wonder if the runner design (not just size) difference between stock 2 squirter TBI and "performance" 4 bbl would be as noticeable with TBI?
 
Is the guy 66plyValiant? I think he is in Canada, a long way to visit or exchange parts.

I think the injectors are driven in alternating manner in normal running, the order of the runners in conjunction with firing order needs to be correct. Most likely to get a get a carb to work correctly, the same considerations.

I looked at some pictures of an SP-2P, and it would work. It would be helpful to know other manifold runner patterns. What I need is a picture with arrows showing cylinder positions and the barrel that feeds them.
 
KitCarlson said:
Is the guy 66plyValiant? I think he is in Canada, a long way to visit or exchange parts.

I think the injectors are driven in alternating manner in normal running, the order of the runners in conjunction with firing order needs to be correct. Most likely to get a get a carb to work correctly, the same considerations.

I looked at some pictures of an SP-2P, and it would work. It would be helpful to know other manifold runner patterns. What I need is a picture with arrows showing cylinder positions and the barrel that feeds them.
Click to expand...

I could be wrong, but I think most TBI's are batch fire. Although mostly GM related, there is a huge amount of info over on Binderplanet.com about TBI conversions. It may not specifically apply to what you are doing, the basic info is still relevent.
 
Oh, and 66plyValiant is the guy. He has a pretty good handle on the Mopar setup and it would be worth talking to him. Parts are cheap, information is gold.
 
For one.if you want tunability go for the G.M style tbi. The tbi are plentifull and easy to install,all that is needed is the computer and wiring harness from your donor vehicle.Tuning is easier than all get up,there are downloadable tunes here ( http://www.moates.net/ ) and cheap chip burners there too.

Trying to stay brand lotal and get the needed component's from a mopar is admirable but near impossible. You will get stonewalled by Chrysler's untunable system.
 
With 2 injectors on a dual plane manifold you need to batch fire the injectors.

assuming a 5500 rpm redline, your max pulse width per engine revolution would be 9 milliseconds. With a minimum pulse width of 1 millisecond for idle operation.

You can just fire them together, no need to stagger them. One driver.

Assuming 85 lb/hr injectors, you can support 280 horsepower at 9 milliseconds.

I would predict about 1.5 milliseconds for idle @ 850 rpm.

Avoid a full fuel cut under deceleration. Are you planning a feedback loop?

B.
 
Thanks for the timing information, that will helps.

The information I have found suggests the the chevy TBI alternates in normal running, however batch fires for startup and high load or WOT conditions. I can do the same. On low end the alternation may improve idle quality and reduce emissions.
Oops, you are correct a dual plane will require batch fire. I guess it might be best to have two modes of operation.

My ECU has provisions for O2 feedback. I have used wideband O2 with mixture table on my 2.2L turbo application. I also monitor a few conditions before entering closed loop operation.

When I say my ECU, it is not Mopar it is one I designed. I have been playing with it on different applications since 2003.
 
This is starting to sound like a win/win/win!

66plyValiant has the TBI aadaptor in several different variations, and you have a tunable ECM.
 
The problem is: I find it is mostly tunable by me. Not many want to read, understand, think and follow directions. My experience with others has mostly been a disaster. I don't have infinite people skills :). I try. Settings for starting, just started, temperature calibrations, acceleration enrichment, are some of the settings beyond the basic 3D fuel, and timing maps. There are many things to consider outside the O2 closed loop operation. It is about making the correct change in the right place. Engine theory is important. Many would be satisfied with less, I am a perfectionist, that is a problem. The system data logs 10 variables at a rate up to 10 mS. This is helpful. Tuning is done real time, however the amount of information is intimidating. I never drive and tune, the wife drives, I tune.

Most of my time has been spent with computers. They do a great job with instructions. For me it is a never ending hobby.
 
Check the Megasquirt MegaManual for much info on EFI and spark control. Batch fire is fine even on MPI setups, so certainly OK for TBI. I have no idea why GM would try to synch TBI to valve position. Maybe they just alternated firings to smooth the load on the supply wires. The TBI on my Holley Pro-jection appears to fire both injectors simultaneously, and there is no sync to the cam, it just uses a tach signal from coil-.

The TBI you have has a unique 4-bolt pattern, so you will need a custom adapter plate or the cast-iron Mopar TBI intake. Most TBI's are 3 bolt, like my Pro-jection, which I understand was a factory GM model. Your Mopar one also used the older injectors (round-top connector). The later Delphi injectors (Metripak type slide connector) give more options.

If you have a 2 barrel intake, mounting options are limited. A 4-barrel alum would save a lot of weight and allow you to use a Holley adapter plate, which comes with the Pro-jection (available separate). I got a complete Projection 2DI for $180 off ebay a few years ago. I have seen the TBI's alone for $25 used, and expect many Holley GM TBI's would work. You could use a Pro-jection controller w/ O2 feedback until you get your ECU ready, and have theirs as backup. Holley's ECU doesn't work great (not enough sensors), but does work.

Most people can't do the roll-your-own ECU as you plan. Megasquirt is easier, but still a very high threshold. Easier to use and not much more is the Holley Pro-jection 4DI ECU or Commander 950. They use more sensors than the original Pro-jection, so work better. I got a Commander 950 ECU (only) off ebay for $120. You occasionally see ads where people bought another ECU hoping to fix a problem, and that wasn't the problem, so they sell the ECU alone.

My hope is to go one step better on my 65 Dart and install a Magnum beer-keg MPI intake (a few posts), but not even working on the Dart now, since a backlog of other projects.
 
Just use the ford 2 barrel TBI.

They made it fit the existing 2V manifold, and it takes a standard air cleaner just like the 2V carbs do. Simple. Uses the more conventional Bosch style injector too.

B.
 
I appreciate all the comments, they help me think.

I am quite close on the code changes in the ECU. I somehow write the code in modules that pertain to each basic function. I merely change code in a few places and all is well. I have the ignition component finished and well tested. It also serves to sync the fuel delivery, one per pulse combustion event. This is a speed density system MAP, not MAF. In a MAF system the fuel delivery can be asynchronous and controlled with duty cycle.

If I choose to alternate injections at light load, it will serve to add resolution in fuel control. Going to batch will be similar to staging injectors, the PW timing will need to change along with other details to insure smooth transitions. I like to start simple so batch injection may be the first try. However the ability to drive in alternating sequence already exists.

My prior injection experience has been multi-point on 4 cylinder engines. I used two injector drives on 4 injectors. By observing the engine order pattern, two drives work well. I learned that sync significantly helped with emissions and idle quality.

I have been at this about as long as the MegaSquirt guys. I lurked on the EFI332 forum, then stopped, and started working on my system. They are brilliant and great engineers. I am not sure I am smart enough to completely understand and incorporate the MS system. I have worked independently, it is more fun for me to be original. I spent about 7 years researching and dreaming prior to cutting my first hardware. I made my first unit in 2003 and had it working in 3 months of spare time. I got the hardware correct, and perhaps robust in features to last my lifetime. The software is in my head, so I see what to do, and mentally work out changes. This ability helps me complete tasks in place without a pc or hardware. I have about 50 years of electronics experience 35 as a EE, 20 years in embedded systems development. It is great being at my peak, using my skills for my hobby, just like I did when I first started.
 
I never had any luck with firing an injector less than once per revolution with TBI.

The split plane manifold feeds two cylinders per revolution from each plane. Four cylinders fire per revolution.

If you skip a revolution the part throttle lean misfire crops up after a few revolutions. The intake tract was designed to be a "wet runner" and with a carb there is always some fuel & air flowing in both sides. When you keep flowing air but send the fuel in gulps the mixture is all over the place. Better to have an injection cycle every revolution.

If you have the processing power, I always wanted to try using an injector in the fuel return line. By PWM driving the injector as a relief valve you can have control over the fuel pressure. That would give you even more range from a smaller injector. You run out of pulsewidth, just jack the pressure...

B
 
The injector events will be 4 per revolution. One per cylinder event. With the correct manifold and injector sync the injector will be fired on the side to fill the correct cylinder. It will help at low load conditions. As the injector durations increase to provide more fuel, the window of opportunity, is replace by batch firing to provide the necessary fuel. The fuel hangs in the wet manifold. This is why MPI is better, DFI better than that.

The Holley TBI incorporates a fuel pressure regulator that is referenced to manifold pressure. It cuts pressure for high vacuum low load conditions. This is common to most injection systems. The differential pressure is constant, so the fuel delivery is more linear. Otherwise it is necessary to deal with flow changes, with square root of the pressure difference.

I am familiar with pwm closed loop control for pressure and volume control. I create special equipment for diesel flow benches.
 
I am very familiar with the existing OEM fuel pressure regulation - and it's shortcomings.

two events per revolution? Not really necessary, but ok if you want to. Trying to time events on a TBI takes a lot of processing. OEM's never bothered and still passed emissions just fine.

Later Ford MPI tunes off of the individual cylinder exhaust strokes and just moving the O2 sensor a foot will shake it up. The timing of the intake valve openings at various speeds with different runners will mean a big table...

B.
 
It will be 2 events per injector/rev, and the fuel is delivered in the correct place. The timer interrupts will still be executing for other activities for either, so insignificant time change either way. There is more timer granularity in short delivery times plus more error related to injector open/close delays at short on intervals for the batch delivery. The separate injection will have nearly twice the pulse width, however it will happen at low RPM, so it should work fine. I verify everything, so as I get into it in the running system I will know what works.

I like your idea of the Ford CFI TBI. I think is is used between 1982-87' on some vehicles. The direct bolt-on, or less adapter work would be great. Now to find one from a 5.0L H.O.
 
Hi,

I have shelved this project for for a few months. At the present time I am finishing a few details on the DIM Electronic Advance distributor project.

I hope to return after the ignition project. For the time being I replaced the Carted BBD, with a Holley 2280 and added evaporative emissions control. This solved the stink in garage issue. I am also getting fuel mileage in the low to mid 20's with 273 and A/C on 100%.
 
I had a few hours yesterday so I started on a 3D fuel table control for my phone over Bluetooth. The table scrolls, so is much larger. I need a bunch of work on vertical plot.
 

Attachments

  • Screenshot_2013-09-05-11-46-22.jpg
    19.6 KB · Views: 254
You must log in or register to reply here.
-

Back
Top