Idea for EFI mod

[FastWoman]

Hi all,

In the same spirit that we replace thermal flashers with electronic flashers, do relay modifications on our headlight circuits, and replace incandescent bulbs with LEDs, I have an idea that would help to preserve the integrity of the ECU. Furthermore, it would substitute standard, off-the-shelf parts with known parameters for the power circuitry inside our black boxes.

In short, we should be able to remove the drop resistors in series with the injectors, substitute a power transistor assembly to regulate current, and then wire the injectors into two banks of three series-wired injectors. One end of each series would be grounded, and the other end would be fed by a PNP power transistor -- emitter to +12, collector to the string of 3 injectors, and base to a resistor and then to the ECU output (which drags voltage to ground to fire the injectors).

The injectors are 2.3-3 ohms, and the drop resistors are 6 ohms, so it is not unreasonable to have 3 injectors in series. In other words, the needed current could be supported. In fact one could probably ditch the drop resistors, wiring 3 injectors in series directly to the ECU output, but there was probably some reason that wasn't done.

Depending on the value of the resistor on the base and the current gain (hFE) of the transistor, the current through the injector series would be current-limited. That would be appropriate, as it is the current that determines the magnetic flux.

Benefits of this design:

* eliminates the big drop-resistor assembly

* offloads current demands from the ECU

* creates a new circuit with known and replaceable components

* eliminates 8 wires in the injector harness

* less power hungry = less draw on the electrical system

Modifications to design:

* If needed, a common injector control output could be used to power both banks (e.g. if one were to have failed)

* If needed, the design could be used to create three banks of two injectors or to wire each injector with its own power transistor.

Other factors:

* It's necessary to substitute a power transistor assembly with heat sinks for the drop resistor array. I presume the best place to mount the assembly would be where the drop resistors are removed. Heat dissipation should be minimal with the 2-bank configuration -- probably less than 2-3W total. It shouldn't require much heat sink.

* An interesting mod would be to attach the power transistors to a billet aluminum fuel rail, so that they would be liquid cooled. The supply to the rail would then be +12, GND, and the gate signal (i.e. only 3 wires).

I obviously haven't done this yet, but it will probably be on my list of things to do the next time I'm into my fuel injection system -- probably long into the future. Until then, I'll be sticking a postit note in my FSM with this circuit diagrammed.

Any thoughts, anyone, without telling me this is unnecessary nonsense?

Edited December 1, 2010 by FastWoman

[cozye]

I'm not sure how that mod, assuming it would work to preserve the ECU pulse signal, would accomplish anything? The dropping resistors have not drifted, and not only limit the current, but also take 2/3 of the voltage on the signal.

The drift is in the ECU, which is modifying the width of the injection pulse. It's a square wave. To correct our aging ECU, you would have to find out what in our ECU is causing the value to drift that is shortening the pulse width. I don't think it's a current issue. We don't need more current. We need a longer square or "on" state.

If you have access to a oscope, that would be the place to start. Unfortunately I sold mine years ago as I figured with the world going all digital I wouldn't need it anymore.

Edited December 1, 2010 by cozye

[Mikes Z car]

I wonder if ethanol needing a richer mixture could have an effect? This site has prominent headache making info for me anyway:

http://members.rennlist.com/pbanders/ecu.htm#TL

If ECUs tend to blow the power output devices this mod could greatly prolong their life I would think.

[FastWoman]

Cozye, I'm not proposing this as a means of altering the functioning of the ECU, but rather a way to offload current from the ECU's power circuitry. In other words, it's a way to make the ECU last longer.

[cozye]

gotcha. Still not sure it would work. In my case I think it's been age, and not current that is ruining the ecu. low miles, car sat a lot. There wasn't any current flowing through it.

I've got a spare ecu. and I may still go megasquirt. If I don't I've still got a spare. If it ages and drifts anymore I'll just crank up the resistance

[Z train]

Fastchick,why the idea might have merit,the money would be better spent switching to Megasquirt.

[FastWoman]

Yeah, probably. I'm just waiting for someone else to do it first!

[cozye]

Yeah, probably. I'm just waiting for someone else to do it first!

I'm assuming that I'm that "someone else"..

My plan is to take my car up and get the AFR checked with a gas analyzer before proceeding further. If it's close now with my band aid, I'll probably live with it for a while. Currently she's up on jack stands getting all new springs, struts, bushings, steering, and the MSA premium exhaust

[conedodger]

Better yet... I am in the process of switching from carbs on my 914 to twin individual throttle bodies with Ford Red Top injectors and Simple Digital System injection. The ECU will sort out and process input from TPS, CHT, O2 sensor, and MAP. I am replacing the SDS harness with a custom weather proof harness I am having built by an electrical engineer friend.

Once this is done, I will consider doing the same on my 73 triple Weber Z. The cost on the Z will increase by one throttle body (twin throat) and two injectors.

As some of you may know, I played around with modifications to the stock EFI and outlined them on my website PlanetZ (No longer exists) back in the 90's. I was disappointed with the results. An analog system just doesn't seem to be flexible enough to tolerate modification.

I suspect I have a couple thousand dollars in the 914 system and will have another 500 in it before I am done. Other than cold start up which is a bitch to set up on the 914 it should be as drivable as most modern EFI cars. For cold start I am setting up a mechanical throttle called a hand throttle that increases the idle until you move it back to the normal position. This is the same hand throttle used on the six cylinder 914 from the factory. I have just reverse - engineered it.

Edited December 2, 2010 by conedodger

[cozye]

As some of you may know, I played around with modifications to the stock EFI and outlined them on my website PlanetZ (No longer exists) back in the 90's. I was disappointed with the results. An analog system just doesn't seem to be flexible enough to tolerate modification.

I would be interested to see your data from your 90's experiments. Do you no longer have the info?

Also, I know in mine and Fastwoman's case, we aren't trying to gain more performance from the stock EFI, really just trying to tune around the drift in the ECU and make it run as it did before. Any thoughts on that ?

[conedodger]

I would be interested to see your data from your 90's experiments. Do you no longer have the info?

Also, I know in mine and Fastwoman's case, we aren't trying to gain more performance from the stock EFI, really just trying to tune around the drift in the ECU and make it run as it did before. Any thoughts on that ?

No thoughts on that. My data are long gone but I did have Brett Industries do the work on the Air Flow Meter and then build a variable tunable resister from the Water Temperature Sensor. John at Brett used to like projects like this but he has sold the company. Maybe someone there will remember the project. I couldn't tune out the lean bog at tip in and at that time, I didn't have an Air/fuel ratio gauge. These are readily available and cheap now. As a side note, I have since then discovered that A/F data without load is not a tuning tool! Duh!!! I would get things perfect in the driveway and then wonder why it sucked on the road. You need either a temporary system you can drive with or a permanent installation. (or a chassis dyno).

My point is that I think I was and you are chasing your tail trying to modify the LJet system. It isn't that you aren't smarter than the LJet, it is that the LJet is dumb-as-a-rock.

I went with the SDS because it doesn't require laptop tuning. Megasquirt is often preferred by geeks and nerds because of the rich data it provides. This is fine if you want to take the time to learn what all of that means and tweek the system to get a better response. I just want to drive...

If you want to stick with the LJet and I think this is certainly a worthy goal on many levels, I think you can find a company that cleans and reconditions the stock ECU. I know the DJet (914 and many others) have this done.

[FastWoman]

Cozye, yes, you're that special "someone!"

Rob, I agree that the L-Jetronic system is pretty dumb, but I have a strange affinity for primitive technologies. I want a system that utilizes the stock, vane-type AFM, just because it's old and different. I have no special affinity for the analog electronics in the ECU. I suppose my ideal system would be a MegaSquirt digital ECU that would utilize the signal from the vane AFM and perhaps also fine-tune via an O2 sensor.

Another possibility, since it's just a dumb analog ECU, would be to build a fully adjustable analog ECU with modern linear circuitry design. This would require that I reverse-engineer the original ECU, of course.

FAIW, I made a special extension cord for my AFR adjustment potentiometer, so that I could have my wonderful assistant drive my car while I made AFR adjustments from the passenger seat -- e.g. seeking out the fastest cruising speed at constant throttle. I found that the ideal settings under load were pretty similar to the ideal settings in the driveway. (I was rather surprised by that.)

BTW, I can buy a "rebuilt" ECU through AutoZone and am certain I can find one elsewhere too. However, all the electronics are proprietary, mysterious, and obsolete. I suspect the "rebuilding" process amounts to little more than cleaning the thing up, making it look pretty, fixing anything obvious (e.g. a bad solder joint), putting it in a nice box, selling it, and hoping it doesn't get returned. Short of having Hitachi's documentation and a stock of NOS replacement parts, I doubt there's anything else that can be done.