Yes, it started!

I finally got all of the parts I needed to convert from carbs to MS. My car has 96mm pistons and performance cam (Fat FC442 280 dur.) so going with a stock system would not work.

I bought a box of FI parts from Ebay, I thought they were all from a 2.0 liter Porsche but it ended up being a mix of just about everything. The throttle body and plenum are 914 2.0 parts, the injectors are 1.7l Porsche, the runners are 2.0l VW bus and the fuel pressure regulator was from someones trash can. It was quite a mix but it all seems to work for the most part.

The megasquirt computer I’m using is on a version 2.2 board using the MS-I chip (soon to be upgraded to a ver 3 board and MS-II) with an LC-1 WB O2 sensor. I’m running the extra code to get 12x12 tables. I have not hooked my ignition into MS at this time since I thought it would be too much to figure out at one time. I’m also using the MS relay board.

I modify the stock throttle body for a TPS. I was able to install the TPS inside of the stock throttle switch assembly and wire it to the original connector. It looks like a factory job. I rebuilt a stock wiring harness to include the extra wires that I need and am using the original relay board to run the auxiliary air valve and to supply the MS board with the switched positive.

I’m using a fuel pump from CB performance. I had a problem with the fuel pressure being too low. It turned out that the fuel pressure regulator was not adjusting properly. I had to cut it open to loosen the internal nut so the adjuster would work. It works now but looks terrible, I will probably replace it. I also found that the PVC had a hole drilled through it. I glued the hole shut but will probably need to find a new one.

Once I fixed the regulator and the PVC the car started just fine. The idle was ok but it would skip every now and then. I checked the logs and saw that the processor was resetting. The problem was that the tach signal circuit was being grounded internally in the MS box. I brought the ground out of the box (XG1 mod) and it cured the problem.

I had the wide band correction turned on and let it determine what the proper values for the VE table should be at idle using MSTweak. Judging from my idle values I think my VE table is going to be very different then what others have posted. I’ll post them once I get it tuned.

The only major problem I have left is with the CHT sensor. I’m using that for the MS Coolant sensor and it is just maxing out way to fast. I need to artificially scale it to typical coolant values. MS will only read the sensor up to 215 degrees which is too low. I think that if MS “saw” the value of the sensor as being 180 degrees when fully warmed up instead of the actual reading of 215+ it would work better and give a more appropriate warm up curve.

Anyway, now that the car idles it time to road tune. Wish me luck!

Dean

Congrats Dean

You're going to be very pleased with the performance advantage of the FI, and the improved drivability.

Lets see some pics of your install.

Enjoy

This post came at a great time. This weekend I'll be putting my order in for an MS kit. To hell with these carbs

Don't forget to use the air temperature sensor....MS is very sensitive to the temperature of the incomming air. I just wired mine into the ms box in the passenger compartment.

I do like the ms extra code. ms tweek also rocks. I still have a few very rich spots in mine. but the car hauls ass.

good luck

Please forgive the newbie MS questions. I am still reading up on how MS works in general, so...

If I understand things correctly you have a built in MAP sensor in the MS unit. That you would need to run a vacuum line from the manifold to this sensor. With a single intake manifold this would be straight forward. But has anyone tried to build a MS system that does not have a single intake manifold? Let say you try to replicate the look, feel and sound of dual carb setup that have a unique intake path and throttle body for each cylinder.

I could see that the throttle position sensor and everything else should be pretty much the same, but how would you obtain a manifold pressure if you effectively have four manifolds? Would you run a vacuum line to just one of them? If you did just one, would this pressure be pretty much the same or less than if you had a single manifold? Would you run a line from all four and join them together to build some type of single average manifold pressure?

Semi-related question regarding EGO sensor. If you have lets say a four into one exhausts, then I can see that you could have an EGO sensor right after the four into one. But if you don't do that, you would then need to have an EGO that is only looking at maybe a single cylinder. Is that a problem?

QUOTE (Richard Casto @ Aug 16 2005, 08:48 AM)

Please forgive the newbie MS questions. I am still reading up on how MS works in general, so... If I understand things correctly you have a built in MAP sensor in the MS unit. That you would need to run a vacuum line from the manifold to this sensor. With a single intake manifold this would be straight forward. But has anyone tried to build a MS system that does not have a single intake manifold? Let say you try to replicate the look, feel and sound of dual carb setup that have a unique intake path and throttle body for each cylinder. I could see that the throttle position sensor and everything else should be pretty much the same, but how would you obtain a manifold pressure if you effectively have four manifolds? Would you run a vacuum line to just one of them? If you did just one, would this pressure be pretty much the same or less than if you had a single manifold? Would you run a line from all four and join them together to build some type of single average manifold pressure? Semi-related question regarding EGO sensor. If you have lets say a four into one exhausts, then I can see that you could have an EGO sensor right after the four into one. But if you don't do that, you would then need to have an EGO that is only looking at maybe a single cylinder. Is that a problem?

It's been done, quite a few times. One way is to make a small plenum (a large fuel filter works for this), and plumb all four runners into this plenum, then read the other side. It works, as long as the cam you use isn't too wild. Sometimes it's difficult to get a good MAP reading near idle, as the intake pulses are the strongest and spaced out the most at low engine speeds. Reading only one manifold increases the intake pulsing problems. Plumbing in all of them tends to smooth the signal a bit.

Commonly, when using an IRTB (individual runners throttle bodies) setup like this, you'll see people use Alpha-N (throttle angle and engine speed only), not speed-density (manifold pressure, air temp, and engine speed). This eliminates the intake pulsing problems, and allows you to use really wild cams. However, it's very fiddley in terms of tuning, and generally isn't the best setup in terms of efficiency (fuel mileage tends to suck). An alternative is to to both, and blend the results together. Do Alpha-N at low engine speeds were you have a hard time getting a good MAP signal, and then use the MAP signal once it stabilizes. The current code for MS does not do this (it will do one or the other), but it's expected that code to do this will appear eventually, probably as part of the MS-II project.

As far as the O2 sensor goes, you can just read one cylinder if you like. When using a wide-band sensor (which, IMHO, everyone setting up a new system should be doing, now that the hardware is so cheap), you can place the sensor in the muffler, or even in the tailpipe. You only have to ensure it's far enough from the open end of the tailpipe that you're not diluting the exhaust gases with fresh air. Unlike most narrow-band sensors, wide-band sensors are accurately heated on their own, and so aren't dependent on using the heat of the exhaust itself to get up to operating temperature.

QUOTE (Richard Casto @ Aug 16 2005, 09:48 AM)

Please forgive the newbie MS questions.  I am still reading up on how MS works in general, so...

first off .... congrats !!!!!

I had the very 1st* megasquirted 914 about 3 years ago or so.....car ran great for the time I had it in there (broke an exhaust stud, dropped the motor, sold the motor, bought a new motor. sold the car....etc..etc... )

*that I knew of on any of the megasquirt websites

The on-line support for the MS is awesome....


Richard,

Running individual throttle bodies is done all the time, normally you just have a line from each T/B run into a common manifold.

If you go to this thread here:

http://www.914world.com/bbs2/index.php?act=...5&hl=motorcycle

you might be able to see the vac. lines for each T/B which all join together.

As for the EGO, yes, it's a comprimise right now for most systems that have 4 go into 1.....

A few very, very high end systems can read EGO for each cylinder....Innovate Motorsports is coming out with one that'll datalog up to 12 cylinders

Engine datalog system, up to 12 cylinders

My set up is 4 individual throttle bodies....

I ran vacuum hose from each bore to a tee then to a fuel filter then to the map sensor..

No problems.

I am running a pretty lumpy cam in the 2.666. this thing pulls like stink until the rev limiter kicks in about 6500 rpm....I need to raise it to 8k..

QUOTE (Mueller @ Aug 16 2005, 09:30 AM)

If you go to this thread here: http://www.914world.com/bbs2/index.php?act=...5&hl=motorcycle

Very cool! I was also thinking throttle bodies from something like a Motorcycle or some other engine would be a good idea. Nice to see that someone is already blazing the trail on this.

The car I just bought has a running 1.7, but is not EFI. I have 2.0 engine from my college days car, but the Dellortos on it have been un-loved for about 10+ years and I am thinking that it would be more fun to do a Megasquirt setup with individual throttle bodies than to try to get the Dellortos rebuilt.

Anyhow, I am such a geek and the idea of doing the Megasquirt just sounds like fun!