MGExp

MG Midget Forum

DIY EFI for 1275 Spridgets

Moss Motors
AutoShrine Sponsor
AutoShrine Sponsor
AutoShrine Sponsor
AutoShrine Sponsor

Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
I finally began bolting on a DIY EFI setup I had configured for my '67 Midget 1275, but landed in the hospital for reconstructive surgery on the dominant shoulder. Unable to turn wrenches, I began playing with a DIY supercharger solution for the longitudinally mounted 1275, so for Spridgets and (I assume) for Minors: http://www.mgexp.com/phorum/read.php?3,3418377,page=3

That blow-through conversion will work with either an HIF44, after a handful of simple mods to seal it for forced-induction, or throttle body EFI--the latter being the superior choice for a number of reasons. However, since lots of folks seem interested in DIY EFI per se, I thought I'd at least put out the nuts and bolts of the setup until I can finish the install and testing.

My idea was to use existing parts to assemble a DIY EFI setup that requires no hacking, or tuning of the ECU fuel and ignition maps. Crazy, you say? Absolutely, but so what? The underlying theory is actually fairly simple.

Step 1 - Identify a fuel-injected car with a volumetric efficiency per cubic centimeter of displacement equivalent to the 1275cc A-series engine.

I settled on GM's 2.0L OHV 4-cylinder, which powered the 1985 Chevrolet Cavalier and its evil twin, the Buick Skyhawk. The manual transmission Cavalier and Skyhawk use GM's OBD-I ECU, service number 1226867. This particular ECU was used in a variety of GM vehicles, and can be purchased used for less than $30 at wrecking yards and on the internet, or new/remanufactured for around $100 at your local parts store or the internet. This mounts nicely under the package shelf on simple aluminum angle iron. (see photo)

This engine is superficially comparable to the 1275cc A-series. Both engines feature a pushrod-operated valve train, and they have similar valve lift and duration, ignition curves, and operating temperatures. Most importantly, the 88-h.p. 2.0L Cavalier/Skyhawk's power output per cubic centimeter of displacement is very close to the 65-h.p. 1275cc A-series, which is slightly more efficient.

Interestingly, the initial ignition timing and the curve appear to be just slightly retarded from a 1275, perhaps because the Cavalier used regular fuel (I think). That is less than ideal, but I believe is what one does when supercharging an engine. Typical starting point is around 2 degrees retarded, or so I've read. So, the thing may be a bit sluggish when running a naturally aspirated 1275, but could be spot on for a blown engine. Anyway, we'll see...


Step 2 - Find a mass-produced, 'all-in-one' throttle body that combines several desirable features.

After eliminating numerous options, I chose an Hitachi throttle body from the 1800cc 4-cylinder Subaru Loyale. (pictured) This little jewel fits nicely under the sloping Spridget bonnet, has a cable-operated throttle like the Spridget's SU carburetors, and features a 2.75" flange that connects to aftermarket silicone snorkels and performance air cleaners, such as the K&N model pictured.

Also, the Hitachi throttle body has a large bore, built-in circuit for an idle air control valve, a throttle position sensor, and low impedance injector compatible with the GM OBD-I ECU's injector driver or the Megasquirt II (in case the no-tune gambit doesn't work), and which is designed to satisfy the 90-h.p. 1800cc Subaru Loyale. The Hitachi's injector is sufficient to fuel a stock or mildly hot-rodded 1275cc A-series Spridget, Morris Minor, or Mini Cooper.

Finally, the Hitachi throttle body has a well-tapered throat that accelerates the intake charge as it approaches the 42mm throttle butterfly, which is almost identical to the bore of the Minispares plenum manifold designed for the 1275cc Mini Cooper. This should produce improved low-end torque and throttle response from idle to redline. (see photo) A simple aluminum adapter plate mounts the TB to the Minispares or cheaper MiniSport manifold and is easily Dremeled to match the bores.


Step 3 - Trick the Cavalier/Skyhawk's ECU into metering the right amount of fuel for the 1275.

The final innovation is to 'think outside the box' to exploit certain assumptions of the OBD-I ECU's algorithms when connected to the Hitachi throttle body, in order to feed the 1275cc A-series. Here's the simplified version:

The GM OBD-I ECU uses an oxygen sensor to tweak the basic fuel-air ratio contained in the mapping under certain running conditions, but it can do so only within narrow parameters. The OBD-I-era ECU's algorithms assume and do not monitor fuel pressure, but they do measure and act on changes in atmospheric pressure in a variety of ways. Most engineers would view these and other factors as limiting the applicability of the Cavalier/Skyhawk e-prom to the engine it was designed to service. According to this logic, using this ECU-and-e-prom combination on another car would require 'hacking' the ECU to adjust the fuel mapping.

Fortunately, I was not issued GM-spec pocket protectors and slide rules. moon

To use GM's assumptions to the advantage of 1275 owners, I calculated the reduction in fuel pressure necessary to put the Hitachi injector's output in the ballpark to service the less powerful but very efficient 1275cc A-series engine. The only limitation was avoiding the minimum pressure below which one might compromise the excellent fuel atomization that is a major benefit of running EFI.

Once the system is installed, all that remains is to fine-tune the fuel pressure to find the 'Goldilocks zone,' where the oxygen sensor produces inputs that the OBD-I ECU's algorithm can stomach...

...and voila! We'll have a 'no tune' DIY EFI conversion that yields a pronounced difference 1275cc A-series owners will enjoy every time they tun the key.

In theory, anyway. spinning smiley sticking its tongue out If it doesn't work, I'll either learn to hack the maps sufficiently and give people a flashed chip for a sweet deal, or swap the ECU for a Megasquirt II, which is compatible with 100% of the setup, save for the ECU plugs. But that's easy-peasy to swap with a soldering iron and shrink wrap. Also, if I'm right, the Megasquirt II will work with an HEI-converted Lucas distributor or a crank-fired ignition, which can be enabled by the bespoke crank pulley I've designed for the DIY supercharger solution.

Joel


Member Services:
Innovative DIY performance and reliability upgrades.

Attachments:
ECU brackets 1.jpg    62.1 KB
ECU brackets 1.jpg

IMG_0361.jpg    48.1 KB
Sign In or Register to view this photo
lower TB WOT.jpg    35.2 KB
Sign In or Register to view this photo
Ballybob Avatar
Ballybob Rob McKellar
Wollongong, New South Wales, Australia   AUS
1958 MG MGA 1500
1968 MG MGB "NED"
1969 MG Midget
Looks really interesting Joel. I hope it works out for you.

I would be interested in doing this to my 1800 MGB.

Keep us updated on how it goes.

Rob

. Hide banner ads & support this website by becoming a > Gold Supporting Member <
trevorwj Avatar
trevorwj Trevor Jessie
Louisville, KY, USA   USA
As I understand, the problem is that an injector will spray the same amount regardless which cylinder is firing. Whereas a carb atomizes fuel based on the pulse of air that exact moment. Heads with Siamese intake ports do not have even pulses on every intake stroke. This leads to charge robbing. For some reason this causes a lean condition on 1 and 4. Firing order is 3421, so three's intake is not totally closed when 4's intake is opening. It seems like this would only effect the volumetric efficiency of 1 and 4, but it actually introduces a mixture problem. I have a theory that it has less to do with actual charge robbing and more about how wet the intake wall are during the different strokes.

Guess what? Forced Induction minimizes this problem ... but it does not eliminate it.

So, once you get it running you might want to check the AF ratio in the center exhaust port and compare it to one of the outer ports.



Midget 50th - A Spridget Event
It was a Blast!
http://sprite-midgetclub.org/

Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
In reply to # 3440301 by Ballybob Looks really interesting Joel. I hope it works out for you.

I would be interested in doing this to my 1800 MGB.

Keep us updated on how it goes.

Rob

Rob,

You bet. In the interim, you may want to look at Rick Patton's nifty GM-based system (Patton Machining online). I'm rolling my own for a number of reasons, saving money being one of them, but messing about with MGs being the primary draw. Also, now that I'm doing the DIY supercharger thing, having an actual, high-flow, all-in-one throttle body appeals to me more than adapted SU carbs.

Joel


Member Services:
Innovative DIY performance and reliability upgrades.
Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
In reply to # 3440324 by trevorwj As I understand, the problem is that an injector will spray the same amount regardless which cylinder is firing. Whereas a carb atomizes fuel based on the pulse of air that exact moment. Heads with Siamese intake ports do not have even pulses on every intake stroke. This leads to charge robbing. For some reason this causes a lean condition on 1 and 4. Firing order is 3421, so three's intake is not totally closed when 4's intake is opening. It seems like this would only effect the volumetric efficiency of 1 and 4, but it actually introduces a mixture problem. I have a theory that it has less to do with actual charge robbing and more about how wet the intake wall are during the different strokes.

Guess what? Forced Induction minimizes this problem ... but it does not eliminate it.

So, once you get it running you might want to check the AF ratio in the center exhaust port and compare it to one of the outer ports.

Trevor,

Aboslutely, this is a concern. I actually did not include the discussion from my related research to keep the opening post shorter. That said, I have never heard anyone suggest what you do about mixture condensing in certain places within the intake tract during the strokes. Fascinating!

In a nutshell, charge-robbing is combatted in my setup--in theory at this point--in the following ways:
- Use of the plenum-style MiniSport / Minispares HIF44 manifold & Hitachi throttle body which (i) forcefully introduces a highly atomized fuel charge versus one produced by manifold vacuum, and (ii) accelerates the air charge via the tapered throat. In my bench study using smoke, this makes the air charge swirl around the longitudinal axis as it enters the plenum;
- Reliance on a GM OBD-I ECU, which pulses the injector multiple times based on the same firing order as the A-series--resulting in a 'fogged' intake tract;
- If you have an LCB header: placing the O2 sensor in the branch which serves the leaner, outer cylinders, 1 & 4. That way, the ECU is 'sniffing' those cylinders and effectively combatting overly lean mixture for you.

All that said, I will definitely use a broadband O2 sensor to check AFR--but it'll have to be at the tail pipe, since the bung in the leaner cylinders' branch will be dedicated to informing the GM ECU. I think (though I'm not positive) that the Megasquirt II can use a broadband sensor for dual-duty--reading the mixture to inform the algorithm and also providing readout for the dashboard of the software. But I might be wrong. At any rate, that option will not be open to me when beta-testing the GM-run experiment, so I'll have to test AFR at the tailpipe, which will produce aggregate data.

Joel


Member Services:
Innovative DIY performance and reliability upgrades.
1974MGMidget Avatar
1974MGMidget Silver Member Jack Orkin
Grayson, Georgia, USA   USA
So, you could adapt SUs for your throttle bodies, to keep the stock look? That would be pretty neat. And, I would think that anyone that could do this conversion to EFI could program a Megasquirt unit, maybe with the help of the 12 year old next door! That might be preferable to get the unit working best rather than trying to make a non-adjustable unit work. Also, doesn't FI require a high pressure fuel pump? BTW, all I know about FI is what I've learned on this forum, so please excuse my ignorance!!!

. Hide banner ads & support this website by becoming a > Gold Supporting Member <
Homerr Avatar
Homerr Jeff Kogut
Seattle, WA, USA   USA
1966 MG Midget MkII "For Sale?"
1970 MG Midget MkIII "Midget On A Spit"
1971 MG Midget MkIII "The Donor"
1975 MG Midget MkIV "General MG"
Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
In reply to # 3440687 by Homerr Have you seen this?

http://www.pattonmachine.com/TBI-Components.htm

Jeff - yes, look up at post #4.

Jack - adapting SUs for EFI is Rick Patton's method, mentioned above in post #4. I have no desire to compete with his cool idea.

I respectfully disagree that a 12-year-old can handle successfully tuning Megasquirt II maps, but if you've got an adolescent who can, you might have him or her start making maps available to DIYers on this forum. I'd take one...

Yes, high-pressure pumps and either a swirl tank or other options Trevor has mentioned elsewhere are required. I'll post some pictures of what I've done in that regard when I get a moment.

Joel



Edited 1 time(s). Last edit at 2017-01-30 06:21 PM by Yankeedriver.


Member Services:
Innovative DIY performance and reliability upgrades.
1974MGMidget Avatar
1974MGMidget Silver Member Jack Orkin
Grayson, Georgia, USA   USA
Well, didn't really mean to insinuate that a 12 yr. old would know about EFI mapping, but that the computer skills needed to do such a thing might be better handled by that 12 yr old better than me!! smiling smiley I don't know much about it, but am enjoying this discussion and your enthusiasm for doing it.

Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
In reply to # 3440904 by 1974MGMidget Well, didn't really mean to insinuate that a 12 yr. old would know about EFI mapping, but that the computer skills needed to do such a thing might be better handled by that 12 yr old better than me!! smiling smiley I don't know much about it, but am enjoying this discussion and your enthusiasm for doing it.

I'm sorry Jack. I missed your joke self-deprecating humor altogether. My apologies! I feel humbled by the welders and fabricators on this forum all the time. It's all relative, I suppose. smileys with beer

Joel


Member Services:
Innovative DIY performance and reliability upgrades.
Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
So, I've installed an inexpensive swirl tank in the boot, with the high-pressure pump, a Walbro, nearby. Then, I got some AN fittings and 3/8" stainless line to run forward to the Hitachi throttle body. That, in turn, goes to the pressure regulator, which has an option for boost reference. The pressure regulator burps out excess pressure into the 1/4" OEM fuel line.

I need to run another test to be sure, but it doesn't look like a problem for the 1/4" line to relieve sufficient pressure back to the swirl tank, which itself has an overflow leading to the OEM tank's filler neck, which has a return elbow screwed into it. The system will use somewhat reduced pressure from the OEM Subaru specs per the opening post in this string, and of course the injector output accounts for some pressure loss. So, the 1/4" return should be adequate. We'll see!

Joel


Member Services:
Innovative DIY performance and reliability upgrades.

Attachments:
swirl tank overflow routing.JPG    38.2 KB
swirl tank overflow routing.JPG

fuel pump mounting1.jpg    42 KB
Sign In or Register to view this photo
pressure regulator.jpg    44.8 KB
Sign In or Register to view this photo
1974MGMidget Avatar
1974MGMidget Silver Member Jack Orkin
Grayson, Georgia, USA   USA
Joel, does the stock fuel pump fill the swirl tank?

Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
In reply to # 3443160 by 1974MGMidget Joel, does the stock fuel pump fill the swirl tank?

Yessireebob. I mean, Jack. tongue sticking out smiley The circuit is: OEM pump > swirl tank > high pressure Walbro pump > throttle body > regulator > swirl tank. So, no matter how hard you corner, etc., there's zero starvation. Trevor mentioned another option I've got to move over from the supercharger string...

The added amps draw of two pumps is not a problem, for when fuel injecting, you need an alternator, and one that'll provide enough juice for the ECU, injector(s), sensors, IAC valve step motor, etc. I'll be swapping the late model Spridget alternator for the Saturn 96 amp version shortly, and going to a serpentine belt, too.

Joel



Edited 1 time(s). Last edit at 2017-02-03 06:47 PM by Yankeedriver.


Member Services:
Innovative DIY performance and reliability upgrades.
Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
So, at Cliff's request, I'm reviving this string, which I hope to keep more or less 'live' until I turn the key on this project!

Since the doctor approved turning wrenches . . . the 'day job' blew up, resulting in me accomplishing very little else! But, as the song goes, "Always look on the bright side of life..."

Basically, I've decided that I might as well get the EFI operational before mounting the supercharger (there's another dormant string), which in my case means an engine rebuild as I believe my rings are a bit worn. Here's my recently resumed activity:

1) I ran a test and discovered that the OEM 1/4" fuel delivery line is sufficient to handle return flow to the swirl tank in the boot while maintaining what I've determined is adequate pressure to atomize the fuel through the Hitachi injector, and not overwhelm the 1275, per the above discussion. Since the engine was not running during the test, consumption should render the experiment's findings conservative, unless I'm missing something;

2) finished wiring up the ECU and diagnostic port to the handmade loom - which only takes an hour or so to make. Devoted a ton of time to taking excellent photos of how to do this. It's easy once you get the hang of crimping the pin connectors, so long as you have the right tool. I've attached a photo of the loom (untrimmed when the pic was taken).

3) determined the 5/16" stainless fuel feed line routing. That line will move more fuel than the 1275 can ever consume. The Walbro high pressure pump really churns it out. The routing apes the OEM line's path more or less, to stay clear of the exhaust but will rise a little aft of the OEM line near the throttle body (TB), attaching to the pictured inlet barb. This Hitachi TB has multiple inlet ports, so one has some engine compartment design flexibility, but I've used one that aims the right direction with the throttle cam situated to work well with the stock throttle cable. See picture, which is of the upper TB only.

The picture also shows the injector, which sits in a cool, streamlined pod, looking a little like a rocket engine in an old Buck Rogers comic book. The intake flange, which connects directly to the aftermarket snorkel pictured in an earlier post, comes off with four screws to reveal the injector and wiring pigtail that routes through one of the streamlined stays.

The brass barb is aftermarket, and you have to slightly enlarge and retap the threads for NPT - which takes about 10 minutes. The other is one of three barbs that came with the TB from the used auto parts emporium. My TB cost $35 with injector inside (for core; after exchange the new one from Geek Auto was the cost of a CHEESEBURGER).

Oh: and yes, this TB has a passage with nipple to connect to the inlet pressure-compensating feature on the fuel pressure regulator. So, I believe when I mount the blower, it's at least possible that all I'll have to do is connect that vacuum line and possibly re-tweak the regulator slightly... but we'll see.

4) determined mounting positions for: idle air control valve (I'm using Patton's excellent stand-alone jewel, though I've found another alternative made in NM) and MAP sensor. I've fewer options than many folks for mounting the fuel pressure regulator due to my Volvo overflow tank, which slips onto an aluminum tab (pictured - this was before OEM fan removed). But best seems to be near to where the regulator's bottom output nipple will aim almost directly into the OEM feed line-turned-return line.

5) --and this was purely for fun--order a 1957 Corvette 'fuel injection' deck lid chrome script. I saw a guy with a fuel-injected TR4 who had mounted the script, called him to find out where it came from, and think it'll match the stylized 'MG' logo on the boot lid quite well.

The next step is, frankly, dictated by laziness on my part - or some would call it 'efficiency.' Do I pull the motor to finish the alternator mount design, which I will offer as a kit with a bespoke crank pulley so people can convert to a Saturn alternator (out of the box, no 'clocking'), and run a serpentine belt? Then put the engine back in to finish the EFI setup? Or do I just finish the EFI using the late model Lucas alternator I've been running for a few years, and which *might* supply sufficient current at idle to avoid gremlins when running the GM OBD-I ECU?

I haven't seen much interest in a serpentine conversion kit--though the benefits are four-fold: (i) out-of-the-bock mounting for a ~100 amp Saturn alternator; (ii) single auto-tensioner serpentine belt; (iii) crank pulley with electronic ignition trigger wheel; and (iv) enables mounting an Eaton M45 blower in minutes - though obviously the inlet-outlet adapter won't be ready until later this year.

So, for that reason and my sequencing discussed above re: the supercharger, right now my predilection is to get the EFI running, then once successful pull the motor, configure the mounts, commission the crank pulleys (short run - one of them is mine!), and rebuild the motor in anticipation of blowing her.

Joel



Edited 2 time(s). Last edit at 2017-06-04 08:22 PM by Yankeedriver.


Member Services:
Innovative DIY performance and reliability upgrades.

Attachments:
wiring diagnostic port-2.jpg    51.2 KB
wiring diagnostic port-2.jpg

upper TB with hose barbs.jpg    38.9 KB
Sign In or Register to view this photo
Radiator plumbing 1.jpg    40.2 KB
Sign In or Register to view this photo
Yankeedriver Avatar
Yankeedriver Platinum AdvertiserAdvertiser Joel Young
Albuquerque, New Mexico, USA   USA
A revision to the above: I think best interim step is to 'clock' the Saturn alternator, mount the V-belt pulley, and use it to run the EFI system. That will let me avoid pulling the engine until the EFI is working, and there's no harm to clocking the alternator again (to original configuration) to put the serpentine pulley back on. I'm not confident that the Lucas alternator will have sufficient stable current to run the ECU at idle, whereas the Saturn alternator is designed to do that.

I've been summoned to prep the swamp cooler, but here's the EFI script, which I've held up to various locations, trying to decide where it looks best. The '57 Vette and Bel Air it was originally used on had them in both locations: fenders and deck lid. The guy with the TR4 just mounted it on the boot lid, which on that car is kind of vertical and has a bit more room than the Spridget...

Joel


Member Services:
Innovative DIY performance and reliability upgrades.

Attachments:
EFI script, fender.jpg    32.5 KB
EFI script, fender.jpg

EFI script, boot lid.jpg    26.4 KB
Sign In or Register to view this photo
EFI script, rear panel.jpg    26.6 KB
Sign In or Register to view this photo
tr4fuelinjection2.jpg    41.8 KB
Sign In or Register to view this photo
. Hide banner ads & support this website by becoming a > Gold Supporting Member <

To add your reply, or post your own questions

Members Sign In   or   Create an Account

Registration is FREE and takes less than a minute!


Having trouble posting or changing forum settings?
Read the Forum Help (FAQ) or contact the webmaster