Megasquirt W108 3.5 M116

Website · Post by **ADow** » Thu 25 Feb, 2010 7:37 pm

Over the last 12 months I have been assembling the hardware, software and the knowledge to install a Megasquirt FI control system on my d-jet M116 3.5. Today the car started :lol: :lol: . I'll start this post with a couple of mentions. The hardware came from http://www.diyautotune.com/index.html. DIYautotune is based in Georgia, USA. They have set the benchmark in sales and aftersales service. Matt Cramer from their customer service department has been incredibly helpful A+++. The other asset to anyone wanting to do this work is Barri (Gurunutkins) who has an excellent website http://www.mercdjetmegasquirt.britautor ... /index.htm that is full of helpful information. He is also willing to answer questions. Terrific bloke.

You can get the MS in two ways - as a lolly bag of components that you solder together, or as a preassembled unit. I chose the preassembly route, although they are more expensive. Having said that, with the strong A$ you could probably get everything you need for less than $1000. I purchased 1. the ECU, 2. the underbonnet relay (not essential but recommended), 3. the 37 pin cable connecting the relay and the ECU, 4. the nine pin cable connecting the ECU with the laptop, 5. the bundle of colour coded wires that go from the relay to the various components and 6. a thing called a Stim, which emulates the engine, and can be used to teach you some of the theory before you attach everything to the car (again, not essential but recommended).

With the existing d-jet system you use the original 1. injectors, 2. coolant sensor, 3. air intake sensor. These you connect to harnesses that you fabricate - not difficult if you can use a soldering iron. You disconnect the original 1. MAP sensor, 2. throttle position sensor, 3. ECU and 4. trigger points. I have left these and the entire original loom on the car at this stage. You can use/butcher the original loom (not recommended) but it's so easy to make another one and it means that you can mothball your original and fairly easily go back to the original system if you want to.

You add 1. a wide or narrow band O2 sensor in the exhaust (any exhaust mob will install one of these for about $100 (narrow band), and 2. a SAAB throttle position sensor from an 88 - 93 SAAB which you can pick up from a wrecker. Make sure you get the three wire connector that comes with it. The SAAB TPS needs a little spacer plate fabricated but it's not hard for anyone who can cut a piece of metal and drill holes in it.

Tools needed - something to fabricate the spacer plate, a soldering station with a 'third hand', a set of screwdrivers, solder, heat shrink, various eye connectors - nothing you can't get from Jaycar for a few bucks.

You also need the software to run the ECU and interface with your laptop and this is all available free on the web. There is also a lot of datalogging and tuning software - my next challenge.

Particular M116 quirks - I was not able to find anyone who had installed a MS on a W108 3.5 M116, and there is no good advice on the web about how to connect the fuel pump relay on this model car. I ended up splicing it into the power wire near the loom next to the fuse box. There should be a more elegant way of doing this but if there is, I was unable to work it out after many hours of trying. The Bosch yellow injectors are technically 'low impedence' injectors, but they won't work with the low impedence settings recommended by MS. I can assist with settings that do work, and if you send me a PM I can send you a file that will get your car started. You also need the right settings to 'flash' the ECU so it can read your existing sensors. Again I can provide these.
What's been the main challenge? Definitely learning fuel injection theory. There's a vast amount of information on DIYautotune and the Bowling and Gippo sites - I don't reckon that you can shortcut the 'input' phase.
Anyway, it's been an adventure in learning so far. I'm more than happy to pass on what I have learned, and will update this thread as I move into the tuning phase.
Alastair

Post by **AMG** » Fri 26 Feb, 2010 12:09 am

Nice work Alastair, when do we get pics???? :o)

Website · Post by **ADow** » Fri 26 Feb, 2010 9:17 am

Joe - will post some next week. The unplugged original loom is still in place so it's a bit of a snake pit. I have to decide whether to remove the original loom. At one point an enormous bundle of wires goes through the firewall. If this is actually a multipin plug it should not be to hard to undo. If it is hardwired though..... :?: :?: :?:

Post by **CraigB** » Fri 26 Feb, 2010 11:22 am

Fantastic Alastair - she's alive!

Post by **drew56cus** » Fri 26 Feb, 2010 5:08 pm

Well done Alastair, it will be interesting to see if it is worth it after it sounded like the Don was going quite well anyway. Fuel economy will be interesting too.
Ok, you are in luck with the engine harness - it is a separate loom joined to the main loom by two 4-pin plugs. See pictures.
So this one shows the D Jet brain board plug, so the 2 connections are close to that. Note that one plug of the main loom has been cut off already.

Pics 003.jpg

Zoomed out a bit - main loom on the right, engine loom on the left.

Pics 004.jpg

And this is what I have been having fun with, trying to use a W108 loom in a finny with a K-Jet 450 V8!! So if you need any plugs or wiring to do a proper job, let me know. I have been trying to use the factory colour coding throughout. I am finally getting the hang of soldering without picking up the hot end of the iron :shock:

Pics 005.jpg

Cheers,
Drew

Website · Post by **ADow** » Fri 26 Feb, 2010 6:05 pm

Thanks Drew - very helpful indeed. I enjoyed your photos of squashed and tangled snakes. What are you doing with your k-jet?
It sounds like I should start at the injector/sensor end of the loom and work back towards the ECU.
With regard to the performance of the Don, I am not expecting much more HP although another 10 - 15 percent would be nice. I understand that you need to go the full fuel AND ignition route to get the big improvements in HP. However, the aforementioned Barri reckons he has managed considerable improvements in fuel economy and some power increase. To be honest, there is nothing wrong with Robert's original system, but there is something tasty about tuning the car from the laptop, and I'm looking forward to learning about data logging etc.

Website · Post by **ADow** » Fri 26 Feb, 2010 10:19 pm

Here's some pix.
Apart from the installation of the O2 sensor which I had done professionally, the main hardware job was replacing the original TPS with a SAAB TPS. Here is the original TPS.

original TPS.jpg

This is the finely crafted mounting and spacer plate for the new TPS

New TPS plate.jpg

This is what the new TPS looks like in situ,

SAAB TPS in situ.jpg

I built up the injector harnesses and the sensor attachments, and the TPS and O2 connectors, and tested that they were all working (not the o2 sensor, but the others).

early testing.jpg

This is the relay - not essential but I reckon it is worth it. Otherwise you have to connect everything directly to a 37 pin plug that goes into the ECU. The relay is also heavily fused to protect your investment.

Relay.jpg

The relay is attached to the ECU with a 37 pin plug that is passed through a convenient hole in the firewall. I had to remove the metal casing around the end of the plug to get it through the hole, and then replace the casing. The hole also carries the MAP sensor vacuum hose and a wire that goes to the fuel pump.

ECU relay cable firewall.jpg

On the inside of the car is the ECU itself. This will eventually be fixed out of site. You can see the 9 pin cable that goes to the Dell Laptop.

ECU inside.jpg

There is a switched power point conveniently located on the battery cradle. There is also a fused line that goes to the +ive terminal on the battery (you can see the pink fuse), and a line that goes to the negative terminal of the coil. This drives the 'tach' function of the ECU.

Switched power point.jpg

This is what it looks like with the old loom and the new ECU co-existing.

With old loom.jpg

But inspired by Drew, I plucked out the old loom tonight. It was about as flexible as uncooked spaghetti. The old ECU itself was permanently attached to the loom but it slid out of its cradle easily. I'm intending to renovate the old loom.

Old loom out.jpg

When I have time to tidy things up, I'll post a pic of the completed engine bay - I will cover the relay and my aim is to make the new system 'disappear'.

Post by **benzfan** » Sat 27 Feb, 2010 12:33 am

Thanks Alastair, I have bookmarked this thread and am watching with great interest!

Post by **PC230SL** » Sat 27 Feb, 2010 11:30 am

Great work Alastair,

I am also following your work with interest, please keep us posted with your progress & impressions!

Regards

Paul

Website · Post by **ADow** » Thu 04 Mar, 2010 9:58 pm

Update - A couple of years ago I installed a Pertronix Ignitor and a Flamethrower Coil - IMHO a winning combination for the M116. However the tach signal from this proved too much for one of the resistors in the Megasquirt ECU and it lost tach signal, meaning that the computer thinks the engine is not turning over and therefore shuts down the fuel pump. Fortunately this happened in the driveway. The latest models of ECU have been uprated with a more robust tach signal resistor with a heat sink and I will be replacing the dead one with the upgraded one under warranty. I'm using the down time to rewire the injectors in a pattern that improves the vapour flow. Instead of wiring the injectors in two groups of neighbouring pairs(1 -2, 3 -4) and (5 -6, 7-8) it will be (1 - 6, 3 - 5) and (7 - 4, 2 -8). This will stagger the wires to the injectors to keep the cloud of gas in the plenum more evenly distributed across the cylinder banks, theoretically resulting in a smoother idle. The proof will be in the eating. It won't be quite as neat because the wires will cross the "V" from one bank to the other but the aircleaner covers it.

Website · Post by **ADow** » Sat 06 Mar, 2010 11:48 pm

Update - Saturday - replaced the ECU resistor and reworked the injector wiring pattern and the car started and idled nicely and more smoothly than it did with my first attempt at an injector wiring pattern. Now to get on with the tuning part of the story. :) :)
Update Sunday - checked everything for the 1000th time and went for a ten minute drive around the neighbourhood. The car ran well and felt normal. With the laptop on the passengers seat I was able to glance at the dials on Megatune - the software that programs the ECU. These are the dials.

Megatune dials.JPG

Megatune can also log the inputs from the engine as an Excel file. In turn this can be read by Megalog, which produces a scrolling graph like a hospital monitor.

datalog.JPG

This allows you to review the drive you have had and look at the various inputs from the sensors. Megalog will also analyse your VE table - which is the table in the ECU which controls the richness and leanness at different revs and MAP pressures.

At this point I have done no actual tuning apart from the idle settings. All the other ECU settings have either been calculated by Megatune based on my data (size of engine, number of cylinders, injector flow etc etc) or have come from people who have trodden this path before.

Post by **markb** » Sun 07 Mar, 2010 7:37 pm

A great thread Alastair, and a mod. that every thirsty 3.5 needs !!!!

Post by **CraigB** » Mon 08 Mar, 2010 10:42 am

Thanks Alastair, I'm watching and learning. You've always been a good documenter of your stuff and you know how hard it has been to find this sort of info. Well done and I just need to clear a bit of space and find some time and this would be a great basis for a fun car. Have you thought about coming out for a sprint? You only have to be as hard on your car as you want to be but I know you don't mind a good blast through the hills. All you need is a fire extinguisher, $50 helmet or any old one with the AS sticker on it - hasn't changed for yonks, extra return spring on the throttle and a blue triangle which can just be blue contact. And we would just about have enough now for a 6 hour team but of course that is a little harder on the car than the 4 lap sprints. Give me a call if you want to know more but I would love to see what a 3.5 would do around Mallala. And some will say about the swing axle but I have had my W113 around and it is just one bend where it can be a bit tricky but fine once you master it. Certainly Paul and Ad have done many laps in their 250.

Website · Post by **ADow** » Mon 08 Mar, 2010 11:38 am

you know how hard it has been to find this sort of info

Indeed! There's been a few M117s Megasquirted but I've not been able to find anyone who has done it with a D-jet M116. I keep hoping to be contradicted on this so I could share tuning files. Which reminds me, the tuning files are just like any other digital artifact - they can be sent as email attachments and then uploaded into the ECU. My original file came from the US from Gurunutkins who has MSed a couple M117s. The capacity to send and upload tuning files is probably a bit ho hum for the turbo riceburner generation but this little old baby boomer finds the concept quite amazing! :shock:

Have you thought about coming out for a sprint?

Appreciate the encouragement Craig and sprint is definitely on the agenda when I am comfortable with the tuning, especially mixture tuning through the RPM and MAP ranges. Too lean and the consequences don't bear thinking about

. I'm going to install a permanent air/fuel gauge in the next week or so that will allow monitoring of the mixture without having to carry the laptop. I have a feeling that this is going to get addictive.
ps These days a gauge is not just a gauge. It's full blown sound and light show, complete with opening sequence, changing colours, warning sounds and some programmable features. Anyone under 50 probably knew that but I am still impressed with a dimmer knob :roll:

Website · Post by **ADow** » Sun 14 Mar, 2010 7:28 pm

Update - Sunday 14th March. This weekend spent more time driving around and datalogging. Getting more confident in the reliability of the system, and the car is running at least as well as it was before the conversion. One of the immediate benefits has been hot starting. It is possible to tweak the hot start injector settings, and the flood clear override (foot flat down when cranking) seems to work better than the original system. My next step is to get a wideband O2 sensor which will allow for more automatic tuning - as I don't have a ready slave who wants to drive around fiddling with the laptop, and the idea of autotuning while driving the car seems somehow appropriate to a digital system. So, immediate impressions are positive, and it has been a relief that the preliminary tuning files have not resulted in any nastiness, at least while driving in the normal ranges for city driving.

Post by **markb** » Sun 14 Mar, 2010 8:10 pm

Alastair,

I think all 3.5 owners can relate to the 'hot start' issue...I even tried running my fuel through an additional 'cooler' but to no avail......it ended up being used as an additional trans. cooler for towing !!!

Website · Post by **ADow** » Sun 14 Mar, 2010 10:58 pm

Mark. It's time you got yourself another 3.5! As you know the hot start issue has many suggested cures - insulating the fuel rails, increasing fuel pressure, reducing fuel pressure, putting a cold bucket of water on the bonnet, maintaining a full fuel tank, jumpering the fuel pump cut off so that you can pump cooler fuel through the system. My most successful method pre megasquirt has been to open the bonnet, wait 15 minutes, pull the fuel pump fuse (fourth from the firewall) kick over the engine, replace the fuel pump fuse and turn the key. Of all of these I think that opening the bonnet is probably the most important as it helps reduce the heat soak. I have not yet tested the MS on a really hot day, so the apparent cure might not be a reality.
Graham from the US posted this interesting piece on the topic.

Vapor Locking.pdf

Post by **PC230SL** » Mon 15 Mar, 2010 10:52 pm

ADow wrote: - as I don't have a ready slave who wants to drive around fiddling with the laptop, and the idea of autotuning while driving the car seems somehow appropriate to a digital system.

Alastair, let me know if you need someone to help out with the laptop or to drive while you fiddle, I will be happy to help out!

Keep us updated, it all sounds pretty promising so far.

Paul

Post by **AMG** » Tue 16 Mar, 2010 3:29 pm

PC230SL wrote:
ADow wrote: - as I don't have a ready slave who wants to drive around fiddling with the laptop, and the idea of autotuning while driving the car seems somehow appropriate to a digital system.
Alastair, let me know if you need someone to help out with the laptop or to drive while you fiddle, I will be happy to help out!

Keep us updated, it all sounds pretty promising so far.

Paul

It's not a good thing to fiddle with oneself or one's laptop whilst on the move. Suggest a Dyno, and a number of 20 min sessions (allow to cool between sessions) to get really accurate closed loop function. but you need to get a wideband lambda. Don't waste your time with a cheap one, buy a really good one and leave it in there. the best place for it is in the balance pipe, in front of the cats (if you're going to use cats - these days there is no excuse not to other than price - they're pretty much free-flowing)

Alastair, an NTK probe is the best on the market, but also comes with an appropriately heavy cost. you can get good ones like the AFX which has afr as well (standalone) and these are about the best thing on the market in a combined package, outside callibration equipment (which I can sell if you really want one - with logging fucntion)
http://www.mmsport.com.au/products/955/ ... and-Sensor
grab it at that price.

Website · Post by **ADow** » Tue 16 Mar, 2010 10:07 pm

It's not a good thing to fiddle with oneself or one's laptop whilst on the move.

Especially while texting.
I have a wideband sensor winging its way as we speak and I am hoping to install it on the weekend. Paul thanks for your offer and I will take it up when I am a little more tuning savvy. I will bring the beast to the next MB meeting for inspection if I have not melted the pistons by that time :shock: and you are welcome to drop by before that. The WB will allow for more meaningful data logging, and also provides sufficient data for the autotuning capacity. The dyno idea is also a good one Joe, and again, when I have a few other issues sorted I will be having a dyno session or two. I've noticed since I installed the MS that the engine is pinging under load. This could be simply a coincidence (the timing chain is relatively new and is still stretching - retards valve timing and advances ignition timing) or it could be a result of the MS system processing the tach signals more efficiently than the original trigger points, or something else, such as a bad tank of premium unleaded. I can't imagine that AFR changes would cause it to ping. This weekend will check the timing.

Website · Post by **ADow** » Thu 18 Mar, 2010 6:47 pm

I have had a couple of requests for the injector harness wiring diagram. There are two. Both work. One is neater

Fuel injection wiring.doc

and one produces a smoother idle (courtesy Gurunutkins)

's system.doc

Smoother is better.
The blue coloured wires are the injector 'input' wires from the four 'outputs' from the MS relay - these are injector 1a and 1b, and injector 2a and 2b. These need to relate to pairs of injectors in a V8. In reality I have also used separate colours for 1a/b (blue) and 2a/b (geen). If you are tracing a fault at some time in the future a bundle of undifferentiated black wires will require a lot of work with the multimeter to sort out.
The red wires are the 'earth' or 'return' wires that can go back to a set of four terminals on the end of the relay in any order.
Injectors have no polarity so if you are using the original Bosch injectors it does not matter which you choose as 'earth'. However, for neatness and consistency I always chose the darkest coloured wire (usually brown) to connect with the red wires.

Post by **PC230SL** » Thu 18 Mar, 2010 7:16 pm

CraigB wrote:I would love to see what a 3.5 would do around Mallala. And some will say about the swing axle but I have had my W113 around and it is just one bend where it can be a bit tricky but fine once you master it. Certainly Paul and Ad have done many laps in their 250.

If anyone is interested, here is some footage of Craig in the 280SL around Mallala:

http://www.youtube.com/watch?v=XfcIbkc5KWI

and here is some footage of Ad driving our W108 250SE manual around Mallala also:

http://www.youtube.com/watch?v=usYNbykTQGQ

I can tell you it's not the fastest thing out there but it sure is a lot of fun!

One of my fondest memories is being on the start line next to Craig in the 280SL and trying to keep up with him, not easy as he is a pretty hard charger, boy that was a blast!

OK back to the topic, just let me know if you need me to help out, will be only to glad Alastair.

Regards to all

Paul

Post by **ggr** » Fri 19 Mar, 2010 1:03 am

Very interesting reading Alistair! I will surely be contacting you when I install my Megasquirted engine in my car, though mine is a 5.0 M117, not a 3.5. I have grafted a 4.5 Djet intake manifold on it (in replacement of the Kjet one) as it was designed as an EFI from the begning, and to be able to use all the Djet stuff I need without having to fabricate fuel rails etc.

For the hot start issue: moving the air temp sensor from the front of the air cleaner box onto the manifold plenum may help improving things: when you restart a hot engine that has been sitting for a while the temp reading of the sensor is relatively cool as the air it reads is being sucked next to the opening in between the hood and the top of the fender in the front. Instead, the air admitted in the cylinders is very hot has it has been sitting in the hot plenum. Also the new air admitted at cranking rev speed has all the time to get heated by the time it reaches the cylinders. So, the combination of a wrong air temp info and heated/vaporized fuel in the rails (though theyir effect may cancel eachother up to some extent) can render hot starts problematic.

I am planning to run the fuel pump permanently from the contact and move the air temp sensor in the back right of the plenum where I have a threaded port. For security purposes, I may fit one of these impact switches in the trunk to cut the fuel pump in case of impact. The plan is to let the pump replace the fuel in the rails by some fresh one and build up pressure by letting the pump run a few seconds before cranking. And then a proper air temp info to the ECU may send the right mixture into the cylinders.

Website · Post by **ADow** » Sat 20 Mar, 2010 8:11 pm

I'd be delighted to share my limited knowledge with you. Your 500 project sounds really interesting. This whole MS process is massive learning event, but that is what makes it interesting. Using autotune I have eliminated my 'pinging' problem and the car is going really well - at this stage still have not tried any WOT tuning as this requires a wide band 02 sensor (its in the post) but an narrow band sensor is fine for tuning the lower and middle ranges of the table. I'm also intending to install a vacuum controlled fuel pressure regulator (picked up one on Fleabay) which should ensure that the fuel pressure is more stable through the rev range, and thereby increase the likelihood that the ECU settings will be based on a real fuel pressure (and injector flow) reading. Theoretically this should also improve the accuracy and stability of the tuning.

It would be excellent to see a photo of the threaded port where you have installed sensor 1. What you suggest certainly makes sense as it would give the ECU a more accurate reading of the actual air temp when it is sucked into the gizzards. With Megasquirt the ECU converts the resistance from the sensors into an algorithm using a table that you create using software based on the resistance data about your sensors (I have these if you need them).

airdenfactor word.docx

This table is 'flashed' into the ECU as firmware where it resides permanently. The end result is that the ECU can read the messages from the sensors in your car - in effect it customises the messages to the ECU based on the variable resistance of your sensors. If you look at the upper end of the table attached (I use Fahrenheit with my MS because the temp gauges in the car are F) it suggests even when the air is getting really hot, the table is still generating changing messages to the ECU as it goes from really hot to bloody hot. If this is the case, it would support your theory. BTW, the MS regularly registers air temp as high as 140F, and this is with the sensor in the airbox horn, and a cold air intake that I have installed.

The fuel pump cut off is obviously for safety, in the event of a crash or fire the f/p won't continue to pump fuel if the engine has stalled. The existing MS relay has about a 3 second delay before it cuts off - I would be hesitant about permanently disabling this feature. There might be a way of achieving your aim with a switch that allows you to switch between continuous running and controlled running. In fact it's pretty easy to do if you have a four way switch and a second source of 12v. At this stage, I have had not trouble starting the car in the heat. The combination of reducing the cranking pulse at 170F, and occasionally giving it a bit of throttle has resulted in quick starts.

Post by **ggr** » Sat 20 Mar, 2010 9:31 pm

In fact I am using a manual gearbox so I will be using the port on the back right where the auto trans originally connected for MS MAP sensor and Teed FPR. I will then be using the port on the back left where the Djet MPS originally connected to fit the ATS. I'm still to check if the thread is the same. If not I may have to work out something.

Website · Post by **ADow** » Sat 10 Apr, 2010 4:19 pm

Update - Wide band sensor and gauge installed, and a vacuum controlled fuel pressure regulator.
The WB sensor provides a wider range (doh) of O2 readings and the gauge monitoring those in real time without the laptop - it also makes autotuning easier and allows for the use of AFR (mixture) targets that the ECU uses for mixture control. This WB sensor can be programmed on its own right, and when I eventually remove the Megasquirt and put the original ECU back, this will stay.

The Megasquirt ECU operates on the assumption that fuel pressure is constant, and the vacuum controlled regulator does two things - it reduces fuel pressures under increased vacuum (during idle or deceleration), and under stong acceleration, restricts the flow of fuel back to the fuel tank, thereby preventing momentary drop in fuel rail pressure that is caused by the sudden demand from the injectors.
This is a Sard fuel pressure regulator (ebay) that I have modified with an extra inlet opposite the original inlet - this allows the regulator to sit between the two fuel rails in the original position of the original fuel pressure regulator. It's beautifully made.

P1020379.jpg

Here it is installed. You can see the vacuum hose patched into the MAP vacuum hose.

P1020382.jpg

The next part of this project is using the Megasquirt for ignition control. There are various ways that you can do this. I am intending to use an original distributor and the existing Pertronix Hall Effect points system. It involves locking the mechanical advance on the distributor, but that's the extent of the hardware changes. I have a spare dizzy and today locked the advance mechanism.

Picture 082.jpg

Theoretically you only need to lock one of the advance weights but the second rivet is for balance and additional security. Like all the work on this project it is reversable. Oh and this dizzy won't need a vacuum advance either. Will post the wiring diagram later.

Picture 085.jpg

Post by **Al Zheimer** » Wed 14 Apr, 2010 9:24 pm

Thanks for the great thread, as `d' jet is the early predecessor of modern electronic fuel injection so to see a comparison between the two technologys especially any power & economy increases will be very interesting.

Thanks again for documenting this great project.

Website · Post by **ADow** » Thu 29 Apr, 2010 8:22 pm

any power & economy increases will be very interesting.

The project is on hold for a month. When I return to it I intend to do some serious air/fuel ratio mapping work, and leave any further hardware mods until I get a reasonably settled map. BTW I have removed the SARD fuel pressure gauge as I was getting fuel starvation with it. One of the learnings from this project is to change one variable at a time!
ps I expect significant improvements in economy as the d-jet runs pretty rich, but with a fuel only conversion, any improvements in power will be a bonus. Others advise that ignition control yields the power improvements, as you can control the ignition advance very accurately. Time will tell.

Website · Post by **ADow** » Sun 23 May, 2010 12:34 pm

Started the mapping process in earnest this weekend. I have found the wideband O2 sensor and gauge to be VERY helpful in giving an immediate readout about the air fuel ratio (AFR) - that is, how rich or lean the car is running. So after each mapping session I can see immediately whether the changes I have made to the map are making a difference the air fuel ratio (AFR) simply by looking at the gauge.

The AFR 'map' is basically a graph with 12 MAP sensor readings on one axis and 12 RPM readings on the other axis. There are 144 squares or 'bins' in the fuel map. The idea is to get optimal richness or leanness in each of these bins. In broad terms the MAP reading is a reading of how hard you have your foot on the accelerator, and you know what RPM means if not send me a PM :)

Tuning each of these while driving around with a laptop would be a long a slow process and would need a very patient assistant. Fortunately there is software to make this easier. The software that actually runs and tunes the ECU (Megatune) produces a constant stream of data from the ECU that can be saved into an Excel spreadsheet. There is another piece of software (Megalogviewer) that reads the Excel spreadsheet and creates a graphical representation of the data stream like a hospital heart monitor. Megalogviewer can also analyse your existing map in relation to a theoretical ideal and produce an alternative map, which you can then upload and use. I will post something to make more sense of this later.
In summary, my car was running very rich, and the process of logging and analysis retuning the ECU so that it runs more lean where it should run lean (cruising, low to mid revs) and more rich when it should run rich (idle, acceleration).

Post by **ben108** » Sun 23 May, 2010 12:54 pm

Once you get it sorted, I will be interested to see how the fuel consumption changes now its running a closed loop efi system, as the old one has no feedback from an EGO sensor obviously..

I have done a few efi retrofits in the past and if the 3.5 I just bought ever plays up, I'll be doing similar.

Website · Post by **ADow** » Mon 24 May, 2010 10:08 pm

Ben, the addition of the exhaust gas sensor adds a whole new dimension to tuning for sure.

I will add updates to this project as it unfolds. This is for people starting out, not experts who would be better off looking here http://www.bgsoflex.com/megasquirt.html

For those interested in looking at a map, one is attached below. The side axis is manifold pressure readings in Kpa, and the bottom axis is RPM. This 'VE map' is produced by the tuning software based on engine capacity, torque and HP and is a good starting point. The higher the number the richer the running. So in the top right hand corner of the map (high revs, wide open throttle) you'd expect to see a higher (richer) number than the bottom left hand corner (idle, no throttle). This map was my starting point and the car ran fine on it, but rich.

Extra VE table.docx

To make a new VE table you can change the numbers in the bins while driving around, or you can use software that reads the stream of data that comes from the ECU and analyses it. This is the easiest option and the tuning software is improving rapidly and getting more user friendly. The stream of data is saved as a datalog, and the Megalogviewer software can do two things. Firstly it turns the datalog into a streaming, rolling graphic like a hospital heart monitor, and secondly it can compare the datalog with a theoretical ideal fuel air mixture and then suggest a new VE map. You can then save the new map and the next time you drive the car upload the new map into the ECU. This seems to work well, as the process has resulted in leaner running with no decline in performance. In fact the car is going very sweetly indeed.

Megalog viewer.docx

Post by **ben108** » Tue 25 May, 2010 9:51 am

The MegaSquirt software looks quite nice. I'd like to have a play with it!

Your VE table looks odd though.. Normally the highest VE numbers would be where the engine makes the most torque, not at the peak RPM.
Also, why is the y-axis scaled to greater than atmospheric pressure? On a naturally aspirated engine there is no need to go above 100 kPaA.

Website · Post by **ADow** » Tue 22 Jun, 2010 11:15 pm

Your VE table looks odd though.. Normally the highest VE numbers would be where the engine makes the most torque, not at the peak RPM.
Also, why is the y-axis scaled to greater than atmospheric pressure? On a naturally aspirated engine there is no need to go above 100 kPaA.

Well spotted. You are right. The VE table above is produced by the software and is a starting point for tuning only. It also goes to 150 Kpa because it can be used by the forced induction people.

The current VE table pasted in below only goes to 100Kpa but still has 144 bins. The low rev, mid Kpa bins are now pretty settled and the tuning software is not changing them much because there have been thousands of data points that the software has been able to analyse. But if you look along the 45 Kpa line you can see some big jumps above 3000 RPM. This will smooth out when I do some data logging at higher speeds and with the foot harder down on the go pedal. Then the datalogs will record more 'cross over points' in those bins, make comparisons with the target air fuel ratio table, and alter them to make the car run either richer or leaner (probably leaner by the looks of things).

VE table clip Extra June 10.jpg

Overall the car is going really well. I've been able to tune the warm up enrichments so that it starts and idles nicely (there is a clever 'wizard' that allows you to do this in real time) and by making tiny adjustments to the hot start cranking enrichments it will start and run when hot without any need to touch the accelerator.

I've also hidden the ECU under the dash of the car so that the passengers don't have to nurse it.

The real time feedback from the Wideband AFR sensor and gauge is reassuring. (Innovate LC-1).

LC 1 wide band pic.jpg

It gives you constant feedback about the AFR ratio and you can see the effects of any changes to the VE table.

Post by **ben108** » Wed 23 Jun, 2010 12:26 pm

That table looks much more normal Alastair.

I'd be doing the same as I enjoy tinkering with EFI tuning, but mine runs too well as is to justify the effort. If the ECM ever dies though, I won't be looking for another :)

I have a couple of widebands, including a LC-1. It's a good unit.

Can we see your spark advance tabletoo ? The other big advantage in this case is you will have the ability to simulate vacuum advance, which the 3.5 doesn't have in standard trim, and will probably benefit fuel economy to a measurable degree. Combine that with closed loop fueling and it could be quite an efficent thing.

Ben.

Website · Post by **ADow** » Fri 25 Jun, 2010 9:47 am

Can we see your spark advance tabletoo ?

When I set up the ignition certainly. I'm still working out how to get link the Pertronix ignitor to the ECU.

Post by **ben108** » Fri 25 Jun, 2010 8:58 pm

ADow wrote:
Can we see your spark advance tabletoo ?
When I set up the ignition certainly. I'm still working out how to get link the Pertronix ignitor to the ECU.

Hmm.. I think you will end up using a GM ignition module and coil...

Post by **ggr** » Fri 25 Jun, 2010 10:15 pm

Ford EDIS-8 works very well with Megasquirt and has prouved to be a roughed and reliable system. It requires some work and fabrication to adapt it on an M116/M117 engine, but nothing out of reach. This is how I did it:

EDIS set up 1.jpg

EDIS set up 2.jpg

EDIS set up 3.jpg

EDIS set up 4.jpg

EDIS set up 5.jpg

Post by **ggr** » Fri 25 Jun, 2010 10:17 pm

EDIS set up 6.jpg

EDIS set up 7.jpg

Post by **ben108** » Sat 03 Jul, 2010 7:06 pm

Alastair,

I don't want to hijack your thread, but this is somewhat related.
A few weeks ago I worked through all my D-Jets components and tested everything and reset the fuel pressure etc..

Today I had a bit of spare time, so I welded a bung in the exhaust and installed my LC-1 WBo2, as I was curious to see what AFR's it was running.
I was very impressed :) Idle was a 14.5 - 14.8:1. Steady state cruise was a nice efficent 14.9 - 15.5's. As expected, WOT and part throttle acceleration was a bit rich, full throttle seeing low 11's. I knew it was a little rich as I have seen a bit of smoke behind me in other cars headlights at night when giving it a bit of a bootful :) And I have also read that the D-Jet is typically a bit rich at WOT anyhow.

I have read that many people disconnect the air temp sensor as it makes some cars run better. When I got back home and it was idling at 14.6ish, I disconnected the sensor and the AFR plummeted to low 11's! I didn't test, but am interested to know what effect it has when driving disconnected. If it makes it significantly richer as it does at idle, those who leave it disconnected are wasting a LOT of fuel.

Anyway, I thought that was worth sharing. It seems the D-Jet although primitive does a respectable job of fuel control and should be capable of returning good fuel consumption.. Now if it only had vacuum advance to help out too!!

Post by **Chai** » Sat 03 Jul, 2010 11:14 pm

An analogy of the disconnected air-temperature sensor is like driving with the choke pulled out partially. It fools the computer into thinking the air temperature is very cold.
The D-Jet manifold pressure sensor enrichens the air-fuel as the pressure drops, like when the car climbs up a high mountain.
The computer gets its acceleration and decelleration instructions via the throttle position switch.

Website · Post by **ADow** » Thu 08 Jul, 2010 1:58 pm

oday I had a bit of spare time, so I welded a bung in the exhaust and installed my LC-1 WBo2, as I was curious to see what AFR's it was running.
I was very impressed Idle was a 14.5 - 14.8:1. Steady state cruise was a nice efficent 14.9 - 15.5's. As expected, WOT and part throttle acceleration was a bit rich, full throttle seeing low 11's. I knew it was a little rich as I have seen a bit of smoke behind me in other cars headlights at night when giving it a bit of a bootful And I have also read that the D-Jet is typically a bit rich at WOT anyhow.

This is very interesting indeed. I've been wondering what the original d-jet would show up on the AFR gauge. Your experience with disconnecting the sensor 1 is logical - I've never been able to understand why some people recommend doing this as it makes the car run permanently rich as Chai has pointed out.
I wonder if you have considered tuning the existing ECU by adjusting the Map sensor.
http://members.rennlist.com/pbanders/ma ... sensor.htm

The cruise and idle AFR figures that you describe could be leaned out a little I suspect without hurting anything, and you might get a better result at WOT.

I'm currently refettling the original loom with new wires and connectors, as I intend to put it back on the car at some stage , and use the MS for another project.

Website · Post by **ADow** » Fri 23 Jul, 2010 8:01 pm

Today I returned the Don to original and replaced the MS with the original D-jet, having tidied up the original loom with some new connectors. The MS will be used in another project that I am incubating. I've left the AFR in the car and note that the original system runs richer - no surprises there. On a scale of 11 (very rich) to 16 (lean) the original system runs about 0.5 richer - that is, if the MS was running at 11.6 under WOT, the Bosch system runs at about 11.1
It did make me wonder why the gnomes set the system to run as rich as it does, especially as the emissions problems with the d-jet led to its replaceement with a mechanical system. I wonder whether it is linked to the ignition timing - a rich mixture allows you to advance the timing further before it starts to ping, giving sharper engine response.

Post by **Chai** » Fri 23 Jul, 2010 8:37 pm

Alistair,
If the original D-jet is providing a richer mixture than the MegaSquirt, adjusting the MPS may be your answer. This should provide a proportional AFR adjustment at any constant rev-value.
The MPS allows the mixture to enrichen as a car is driven from sea level and up a mountain. My understanding is the fuel-mix must be richer at low air-pressure (less oxygen).
If you were happy with the engine performance using the MS, then you should be able to 'tune' the MPS using your AFR meter. Basically, fooling the ECU into thinking you're running in a higher-pressure environment should lean the mixture.
However, after saying all this, I'm still not sure if this will change the anount of fuel squirted at WOT as it is possible the ECU's WOT mode may ignore the air-pressure setting.

Website · Post by **ADow** » Sun 25 Jul, 2010 10:37 am

you should be able to 'tune' the MPS using your AFR meter

This site explains how to do just that in some detail.
http://members.rennlist.com/pbanders/ma ... sensor.htm

Post by **Chai** » Sun 25 Jul, 2010 10:21 pm

A very informative site although the picture of the adjustment screw on the Porsche MPS is not the same as in my 1974 107.
I'll probably install a MS when the ECU starts to play up - one day. Not driving enough nowadays to justify spending money on the car.

Website · Post by **ADow** » Mon 26 Jul, 2010 10:54 am

I'll probably install a MS when the ECU starts to play up - one day. Not driving enough nowadays to justify spending money on the car.

You could be waiting a while. The main weak spots appear to be the diaphragm in the MAP sensor and the exposed wiring near the sensors and the injectors and the connecting plugs themselves. BTW I did have a sensor failure - coolant. It was intermittent but when it happened the car ran really badly. The MS made it easy to diagnose.

Post by **Somniferous** » Sat 24 Nov, 2012 8:56 am

Hi!

I am New to this forum and to Megasquirt. I've bought all the things necessary to convert the m116 in my 350 sl -71 from D-jet to Megasquirt.
What version of MS did you use? I have the Ms2 3,57 and the relay board.
I'm going to use and edis8 module and ignition coils, standard injectors and a 36-1 trigger wheel ond the flywheel.

I thought maybe you could help me out with some files and wiring diagrams. I could not PM you since I havn't been active enough in the forum yet.

This will be an ongoing process during the winter and I might need some help now and then.

I live in sweden btw, so the winter is loooong :)

Post by **CraigB** » Sun 25 Nov, 2012 5:37 pm

Welcome. I will email alastair to let him know about your message, I think he might go away a bit though but he is definitely still about so hold tight. I thought we had a swedish guy on here that did all sorts of performance stuff but can't remember his name.

Post by **Mercmad** » Sun 25 Nov, 2012 6:39 pm

Didn't he sell his mega squirt stuff to a guy in Scandinavia?
http://www.peachparts.com/shopforum/sl- ... 560sl.html

Website · Post by **ADow** » Sun 25 Nov, 2012 9:25 pm

Hi all
Yes I did sell all my MS stuff to a guy in Sweden. He is on this forum or at least has been in the past, and has successfully installed my system on his 3.5 - he seemed like a smart guy with excellent English and would be a good contact for anyone in Europe.
Most of what I know and did is already in the posts on this thread. I don't know anything useful about the ignition side. However I'm happy to share my fading memories of MS on this thread.
Cheers
Alastair