Dave Carlson, fellow 914 owner, has contacted me to let me know he is developing an engine management system (EMS) for the 914 and other air cooled cars. He apparently has a working prototype on a T3 engine.

I know he would appreciate our input, guidance and collective wisdom in developing his EMS system to meet out needs.

Whaddya think boyz and girlz...?

Post comments here or email him thru his site:

http://home.mindspring.com/~dave.c/kitcarlson/

That sounds interesting and EASY!

wow, that looks promising. i've thought before that such a venture would be possible, glad to see that it's happening.

Not much description there yet. It'll be interesting to see some more info about it. E.g., how much $$? Is the system user-programmable? Is it a build-it-yourself affair? And so on.

I had another person talk to me about the feasibility of doing something very similar for 914s, Bugs, and 356es. Nothing ever came of it that I am aware of.

--DD

He is an electrical engineer so it could be a reasonable product but I'd like to hear other people's experiences with it and see some of these before I'd consider buying it.

Charles

Thank's for all the comment's!

I have been dreaming, researching, developing in an endless loop since about 1997 on this EFI stuff. The commitment to actually do something has come recently. I have spent my career developing sucessful products for others, and neglecting to follow my dreams. Working for myself is a bit of a shift.

There are many more functions in the hardware that need to be implemented in software. The guts of the system is there, the dwell and timing control, the fuel calculation and timing, and the mmi. The system is interrupt based, there are sufficient timers and resources to keep things simple in the code. The calculations are done every 180 crankdegrees. The cold start, throttle enrichment, unloading, and other drivability things are roughed out, but not tested. The code is written in C. The code size is currently 12K bytes, less than 10% full (room for 128K). The code can be bootloaded over the communications connection.

The operator interface is all done on the EMS, only a VT100 terminal program needs to run on the Palm® or PC. Many pararameters are displayed at the moment. Parameters are easily changed by moving to the parameter, (inverted-video) then + or - to change the variable real time. Values are stored in eeprom if desired. The "VT100 Online" by Mark-Space ® enables Palm "keys" to be assigned to chars, the "Keys" are then easily used to make adjustments.

Having the system start up and run was bit of a surprise! I expected countless multiple bugs. Turning the distributor in a drill press and monitoring the ignition and fuel pulses with respect to the timing reference signals worked well to sort things out. Most all development was in my lab, (garage with A/C) off the engine. I seem to loose my head around a spinning engine on a stand.

By Monday I will have a 5 gas exhaust analyzer. This will help safely check out the fuel mixture. It is a portable unit. I hope to get the wife involved. It will be hard to drive, do the mmi, and monitor the Ferret14.

My website is new and many changes have been made recently. It will be a challenge keeping up with the e-mails, website, design, procurement, fabrcation, assembly, and check-out. The marketing thing will be new too. The good thing is I love what I am doing and I'm willing to spend countless hours. I have added a "Please Help!" link on the website, many you work at varied professions, please help me out in my many weak places. Comments and suggestions are appreciated.

Dave

Welcome Dave,

We have a section for member products for sale to the 914 community. When you are ready to sell them, post your info there.

what part of tn? can you make/design/modify the dizzy pick up to use an existing 'rotor' already made like a unilite 'fan'?

Did everybody read what he said:



Leave him alone. Unless your want to invest in his product (I dont mean buy one).. leave him alone.

You would be amazed at how little you get done on projects when your phone is ringing off the hook and people are buggin the crap out of you in email.

Thanks for joining the club.


B

Dave,

Will you keep the stock air-box and intake runners for your system? (That would be my preference.) I also like the fact that you are using a more modern (and hopefully, less expensive) MAP sensor.

Charles

All,

I live near Murfreesboro.

The dizzy is special because the design is for waste fire. Two sensor are required. If the IR (optical vane switches) are a problem a switch to Hall sensors would be fairly easy. I have had 3 vehicles over the years with optical distributors and zero problems, never cleaned them either. The Borg-Warner on my 66 Barracuda dates to the late 70's, still runs great! An Allison was on a VW T1 KitCar and the other was an OEM on a 88 (150K+?)Dodge minivan.
Since the sparkplug wires are not distributed by a rotor, high voltage is not present at the sensors. The sensor supply is only 5.0V. It is a great advantage to have the plug wires go directly from the coil to the plugs.

It should be possible to go to crank fire, one sensor on distributor(or fuel pump lobe, for cycle determination) and one on the crank (still 2) with 2 targets on the pulley. This will eliminate any gear slop, but will require the engine to sync at start (may need to turn a couple revolutions). It will not alway start with a touch of the key and var-room. It can also be driven waste fire.

The first method is easier to install. It will also work on about any Bosch 4-cylinder distributor (Type1-IV,Volvo,Saab)

I have been experimenting with more un-conventional ignitions, CDI derivatives, however more parts (high voltage inverter) and $. From a control stand point they are easier (no dwell prediction required). Too much dwell saturates the modern coils (saturation not limited, is a bad thing) in a few mS. On the other hand short dwell => high rpm potential, since the charge time is small.

The stock manifolds and throttle body are used if desired. On the L-Jet, the air flapper sensor is not required. A tube should be fabricated to replace it. By the way the MAP sensor is on board. It is made by Motorola, has built in calibration and gain. The raw cost ~$20. An auxillary, input is provided for MAF development. I have some used MAF's, just need the time. Other sensor inputs are TPS (variable resistance type), charge-air-temp, and head-temp.

Please keep the comments coming. As you all know, I need to work on my writing and spelling skills. I will write when I can. Today is a good day, resting up from a 1200 mile road trip.

Dave

Could you separate the ignition and fuel management functions and let me buy just the fuel management components? (I'm not unhappy with the stock dizzy and a compufire module.) Possible dumb question - is it a closed loop system where the ECU gets feedback about the exhaust CO and can adjust the mixture?

Thanks,
Charles

Can it play MP3's too?

It's been a few years since I was seriously into C, but if you need some fresh eyes to pick out any bugs, I volunteer. I need to warm up my coding skills anyway. I'm resonably sure there are some other rather accomplished computer geeks on the board too.

I was just down near Murfreesboro a few weeks ago....Henry Horton State Park. Real nice place.

Again, welcome to the board!

Hi Dave,
This is really cool. I'm very interested in digital engine management, and currently have a perfect power MIC3 on my 914 daily driver.

Regarding the dwell issue with the modern coils: If you haven't already, you might want to look at the Bosch 008 series of ignition amplifiers. They have internal circuitry to measure saturation time and automatically trim dwell time to be optimal for the coil. The only downside is cost.

Regarding the crank trigger idea: Another thing you could do is keep all the senors together on the crank. I have a dual hall sensor with both sensors in the same plane, but one only sees north-side-up magents and the other only sees south-side-up magnets. This lets you put two possition magnets and an (upsidedown) sync magnet all in the same plane on the crank with the sync magent leading the TDC magnet by a few degrees. SDS sells a dual-sensor pickup like this for about $35 I think. (They would also be cheap/easy to make as you know)

Regarding Charles' comment about not using the ignition features: Charles, trust me. You WANT the ignition control integrated. It's SO NICE, and if you do digital EFI you've already done 90% of the work anyway. It would be a shame not to have the benefits of digital ignition control.

-Ben M.

The EMS system does incorporate an input for O2 sensing. I built and idle speed controller about two months ago to test an electronic auxiliary air valve. I incorporated O2 sensing for test reasons. The modern solenoid valve is actuated with PWM (pulse width modulation). I will put some pictures on the website. This feature is also built in to the EMS.

Back to the O2 (lambda), since the AFR may be near 12:1, a normal O2 may be at the edge of its slope. I have some ideas how to make things work, but had to get the injection going first. The 5 gas analyzer is also needed.

Back to the fuel only system, the EMS need the timing sensor inputs to measure RPM for the fuel part. Air Six is correct, the 90% of the work is in the ignition, the fuel control falls in place after that. I could make a ignition input, and use a conventional ignition for RPM signal, this was done on the idle speed controller. Yes, it can be done. The fuel control may not be sequential.

On the sensors, I wanted to be different, not to copy any known system. It ends up being very simple. The Bosch idea is good but, coil pack charge time is much less than a conventional coil. At sub idle speeds (starting), timing is controlled by the sensors (locked timing), I found many OEM's do this. The charge current is limited in this system to protect the coil. It incorporates compensation based on derivatives, for positive dwell, avoiding misfire for fast rpm changes.

On a side note, the system incorporates stall detection to cut fuel and spark in the event the engine stalls. It comes back to life if the engine spins or restarted. I have also included fuel pump and am master relay for power to injectors and ignition. Any comments for safety related control?

The ignition system was first tested about two weeks ago on a carb engine. It took about a three days to track down 3 small bugs (1 mechanical, 1 hardware, and 1 software). There were problems with the bug motor that sat outside for 5+ years. The carb worked at first, then the crud messed up the idle, needed to rebuild the carb.

I switched between point ignition and the electronic a couple of times. The engine ran very smooth (almost un-bug like, that iron block engine feeling) with the electronic ignition, the difference is the cylinder to cylinder timing is more accurate. The little cam, bearing run-out and roughness in the advance system, makes a conventional ignition sloppy. The ability to adjust the timing based on RPM, manifold pressure, temperature, and throttle input is a good thing. Right now, just RPM and MAP.
I will need help with a dyno, and the experience of a tuner to learn the rules of timing limits of an engine under load.
Fuel also fits into the equation. My biggest concern is with all this adjustability, "what is correct? ".

I have thought of selling the ignition only.

No Mp3's at this time, a good sounding engine is music to me.

You all, have great suggestions and questions.

Dave