Out with the Megasquirt!, and in with.... Options

[JamesM]

and in with.... well... MORE MEGASQUIRT!!!!

Little back story here and then a bunch of pictures.

I have several 914s but the one I refer to as my "autocross car" (which started as my daily driver in the late 90s) has AFAIK been running Megasquirt possibly longer than any other 914 on the planet. For the last 17 or so years it has been the platform I have experimented on and in that time has seen at least 5 different Megasquirt configurations across 2 different motors. This ongoing project is one of the reasons I am so intimately familiar with both Megasquirt and D-jet as to date all the configurations I have ran in this particular car have done so using an unmodified d-jet wiring harness which required a good deal of study/reverse engineering.

At one point my focus was on making a completely stealth setup (looks exactly like a d-jet system) which I did, however as the Megasquirt feature set improved over the last 17 years and my focus moved more towards the overall performance of the system I kept hastily hacking in additional features (mainly having to do with the addition of full ignition control and O2 feedback) and an ancillary rats nest started to grow around my once stock looking system. Remembering what all extra wires plugged in where started to become tedious, especially if I would go extended periods without doing anything to them. With the desire to change/add additional elements to the system I decided I needed to just bite the bullet and build a whole new system from the ground up. Technically i have been planning (sitting on the parts) to do this for years but as the current system ran just fine it was lower on my priority list. The whole COVID situation has allowed me lots of time to get pretty much every other project done on all my other cars, so it was time to move forward with this one.

Details on the system coming out:
ECU: Completely custom (see homemade) ECU board based on the MS1 2.2 circuit design however I modified the injector circuits to supply a switched current rather than a switched ground in order to work with the D-jet harness, and then later retrofitted an MS2 CPU to expand the feature set.

Injectors: Stock 914 2.0 off a stock pressure regulator

Ignition: Full timing control currently triggered via Pertronix pickup modified with a pullup resistor in a stock distributor with a locked out mechanical advance. Output is then from the MS system to a Mallory 6AL CD box (this was a really easy way to trigger a coil from Megasquirt back in the day as it can take a logic level input)


The Goals of this project
1. Cleaning up the engine bay wiring and simplifying the connections (single harness plug to the stock relay board and thats it!)
2. Move to a 36-1 crank trigger
3. Add a 2nd MAP sensor for real time barometric correction (we have some high mountains around these parts)
4. Eliminate the distributor and CD ignition box
5. Swap the LC1 wideband controller for the newer(and smaller) LC2
6. Use a modern Fuel Injector with known voltage offset values (will get more into this later)
6.Stop using the d-jet harness (Will elaborate on this as well)


WHY???!?!
Back at the end of the 90s I was a d-jet purist, carbs to me have always been a step back and never an option for me, so when my d-jet system started acting up one of the things I did at the time was to purchase a new wiring harness made by a company called R.E.S. Systems. This was the only NEW replacement harness available at the time and is still an absolute work of art as far as i am concerned, overbuilt like no other d-jet harness you have ever seen. Needless to say it was not a piece I wanted to take off the car and replace with whatever homemade harness my limited experience could coble together at the time, so i invested time in figuring out how I could retrofit an MS system (also very new at the time ~2003) into an otherwise stock d-jet setup. Now, while doing this does result in a system that works as well as a good d-jet system I am here to tell everyone once and for all that D-jet does have some inherent faults with its design and that those faults become very apparent when you can see them in data logs and also when you realize that a lot of the known d-jet operational "quirks" follow you when running a modern ECU on the d-jet harness. The nicest d-jet harness in the world does not eliminate these issues, they are inherent in the design.
So in the interest of having the best running car possible a new design harness is needed, which is fine by me as i want to integrate everything into a single harness anyways.
Other things to come with the new build, modern 43 PSI injectors should atomize the fuel better, in addition having known voltage offset tables is mandatory to achieve the most consistant tune across all operating conditions. The change to the 36-1 crank trigger is for increased accuracy but also to improve ease of cold starts over the 4 tooth cam trigger.
Getting rid of the distributor was not mandatory but I had an idea for a new setup I wanted to try out, so its getting worked in the changes as well.

Had a week of vacation time to burn after Christmas that seemed like the perfect time to get this all done so...

on with the show...

[JamesM]

Will start with the "before" pictures, this is the rats nest of years of hacks im looking to clean up. The ECU there is not d-jet but contains my custom MS board internally (note the blue vacuum line running into the bottom of it.





Other than that the system is very "stock-ish looking" though on the latest motor I changed things up a bit. On my previous 2.0L motor I was running with the stock 2.0 intake step, the current built 2056 has modified 1.8 heads so im using a 74 1.8 plenum modified to accept a 50mm throttle body from a 2.1L Vanagon... almost looks like it was supposed to be there!






Weeks prior I had "borrowed" the 2.0 injectors off this car to put on a 74 bumblebee im bringing back to life so figured the injectors and pressure regulator were a good place to start in order to have the fuel system buttoned up before messing with the wiring.

Out with the old....





and in with the new...

New high impedance injectors (to eliminate the need for resistors) also have a 3 bar minimum pressure so the regulator has to go. New piece from Mario at TheDubshop that accepts modern bosch regulators going in its place with minor modification to the engine tin.

[JamesM]

Factor contributing to the rats nest, coming out!




And the ridiculous amount of wiring connection the LC1 wideband needed

[JamesM]

and now the part that makes me sad... pulling the beautiful RES Systems D-jet harness. Almost 20 years later this thing still looks practically brand new. just a tank of a wiring harness with every connector and junction properly sealed to water proof it and every connector labeled.

If i don't wind up putting this in another car it will probably go up on my garage wall as art!

[JamesM]

New ECU is just going to be an off the shelf Microsquirt. No more 12 hours of labor custom building circuit boards for me!

I had a gutted d-jet ECU case from another one of my previous MS builds, so i re-purposed that as a bracket for the ECU, both MAP sensors and the new (much smaller) LC2 wideband controller.





My old MS 2.2 custom board d-jet harness compatible ECU next to the new Microsquirt setup, not quite as trick looking but I may revisit that later.

[JamesM]

And now the REALLY fun part... putting this in there... and trying to make something out of this new mess!




Laying it out as an initial test to get a rough idea of spacing from the ECU bracket to the relay board.
Not looking complicated at all at this point (IMG:style_emoticons/default/biggrin.gif)





Making the runs for the injectors and MAP sensors
Not using pigtails for this build, everything is color coded un-spliced wires out to crimped pin connectors. Minimizing the the junctions minimizes the potential problems. NO SOLDER!

[JamesM]

After i got the spacing relative to the engine I pulled it back out to mock it up on a spare engine on the floor.






Decided to leave all the unused wires in the Microsquirt harness intact at full length in case I wanted to add something else later (I tend to do that) so they got sealed up with liquid electrical tape and cover in heat shrink. A couple wires im not currently using i still ran out to where i may use them, as I already have some addition upgrade in mind (IMG:style_emoticons/default/biggrin.gif)




Pulled back on the bench to get everything sheathed and connectors installed.


Just about ready to go, just need to add fuses inline on the power circuits at the relay board and finish sizing the wire lengths to my coil experiment. I was going to go with a familiar setup there but i just cant help myself and had an idea for something new!!
Not nearly as pretty as the RES harness, but i think i did ok. May revisit it again later as im not super happy with the braided sleeve.

[JamesM]

And now for the next surprise idea(s).

I want to keep the plug wires all running through their stock locations so I took an idea (and a spare distributor) down to Ye local house of 914 art (AKA PMB) and had them Frankenstein this piece up for me. They did this way prettier than I ever would have been able to.




Why you ask....



I had a very specific coil pack in mind. Looks almost like it was designed to go exactly there.

[JamesM]

The coil pack is the biggest unknown/experiment of this current build as I was not able to find a lot of documentation on it or reports of people using it.
I have previously used VW 2 channel logic level wasted spark coils on build i have assisted people with and good results however wasted spark coils have a short coming in that 2 of the plugs fire with reverse polarity resulting in increased and uneven spark plug wear. As i am going for maximum precision and reliability here (and i don't enjoy regular plug changes on Type 4s) Newer cars run dual platinum plugs to offset this wear, but i wanted a solution where i can run the stock copper plugs. I figured i would experiment with a different potential solution.

This experiment comes by way of another VAG coil pack, only this one is a 4 channel (CPP) logic level pack. First issue there though is that without additional harness modifications the Microsquirt only controls 2 coil outputs however being a logic level coil in theory I can tie the input pins together in pairs and batch fire them, essentially firing the same as a wasted spark but with all plugs using the proper polarity. Only concerns I have are is the logic built into this coil going to allow me to fire off 2 coils in the pack simultaneously without issue, and as this motor spins to 7k RPM and ill be firing the coils at essentially twice the rate they would in a factory application, can they take it or is there protection built into the logic circuitry that wont allow this???

TBD

It sure does look cool though!

[JamesM]

The entirety of everything that came off the car in this endeavor

glad to be rid of the mess

[JamesM]

Pre flight checks on the new 36-1 crank trigger looking good!
This should help a lot.





Static tests from Tuner studio confirmed injectors are firing properly and
I can indeed batch fire the coil pack as hoped.

Only issue during pre-flight was a fuel leak at the new regulator during the fuel pump relay and pressure test. This was at the plugs where you can install pressure sensors. Had to take some extra measures to seal them up, but all is well now.


Always nervous as hell to fire a new MS build, but happy to report she fired up on the first attempt.

Now to wait for warmer weather so I can get a proper tune on her.

[BeatNavy]

James, that's excellent. Thanks for posting this. I've got a small rat's nest in my setup as well, and I'm looking to clean it up with better wiring connections and routing. But I'm in the same sort of boat -- there's a tension between wanting to test something and improve it versus cleaning it all up and declaring it done and final!

One question: what model injectors are you using?

Again, looks great and thanks. Hopefully you do get some warm weather soon (IMG:style_emoticons/default/beer.gif)

[Porschef]

(IMG:style_emoticons/default/popcorn[1].gif) (IMG:style_emoticons/default/beerchug.gif)

[rhodyguy]

Incredible! Your skills are impressive.

[GregAmy]

Sounds like we're brothers from another mother...

I can't say enough how happy I am with the Microsquirt system I did last winter (with "bespoke" harness).

However, I'm now building an AW11 Toyota MR2 race car, and I am even more impressed at DIYAutotune's PnP system for it. I mean, it is seriously one well-thought-out bit of kit; I'm shocked that not all AW11 cars are running these ECUs. If we could only design and fab a PnP system for the 914 DJet harness... On the other hand, I can't leave well enough alone either: I've already ditched the AFM, added wideband, and I have COPs and a 36-1 coming in for it...

Subscribing! - GA


Edit: for future consideration, I am now using the 14Point7 Spartan 2 Lambda controller for my installs. Tighter install and has a much heartier lifecycle reputation than the LC2 (I've already replaced the O2 sensor in the LC2 in the street car). I get mine from a buddy, Conover Motorsports: https://www.raceconover.com/shop/spartan2

[GregAmy]

I. LOVE. THIS.

Care to make another?

@Montreal914

[jesse7flying]

Wow! I'm looking for the Google translator to put this into dumba** English for my consumption. Impressive. Thanks for sharing.

[brant]

Nice !!! work James !

[914werke]

(IMG:style_emoticons/default/aktion035.gif)

[JamesM]

I. LOVE. THIS.

Care to make another?

@Montreal914

You would have to ask @Eric_Shea about that one. His guys at PMB did the fabrication work for me as I didn't have the necessary tools to do it myself. I brought them a core distributor and an idea and they returned that beautiful piece. In the end it is just a hacked up cheap distributor (i wouldn't waste a d-jet one) with a plate welded on it but the result is exactly what I was looking for. Really nice for a one-off part but im not sure it would make sense to do it this way in a production environment given the labor and need to sacrifice a distributor. If someone wanted to produce these in any volume I suspect it might make more sense to just machine the whole thing.