Megasquirt on MPI
Turbo Mini Forums

Hi all, I am new to this forum, I normally reside over at The Mini Forum.

I understand there has been a lot of researching and developing of fuel injection systems for the A-series engine going on over here at Turbo Minis

Simply put, Is it currently possible for me convert a currently standard 2000 Mpi Mini over to run on MegaSquirt??

If so what parts do i need?

Thanks, just really looking for a yes or no at this point.

Regards
Mark

Try doing some research yourself. Use the SEARCH button, all will be revealed. Put a few hours aside though and don't expect a simple answer.

http://www.twin-turbo.co.uk
http://www.hillclimbandsprint.co.uk/default.asp

A man without a project is like a like a woman without a shopping list.

what I want is a simple answer though! Is it possible to fit megasquirt to a standard mpi, I understand that if the answer is yes it will not be a simple process to actually achieve it but many of the cars I have read about it being fitted to do not use the mpi injection parts

Ok, if it makes you do a SEARCH then, yes it is possible.

http://www.twin-turbo.co.uk
http://www.hillclimbandsprint.co.uk/default.asp

A man without a project is like a like a woman without a shopping list.

What a nice, welcoming forum you have here!!

And a huge asset of combined knowledge, once you know how to use it.

And what a welcomed new member we have met here!!

as a quick overview,

there is a megasquirt code that allows the injector time to be played with,
I presume you already know this, but you need to move the injection points to get the fuel to go in the right cylinders, this means you end up with less time to get the fuel in, this is what they did on the MPI originally hence the large injectors.

as most people on here are looking for more power they end up with more injectors/larger injectors due to this shorter injection slot.

Hope that helps!

I'm not being funny, but you have asked a question that can't be answered the way you want it answered. Once you have done a search you will see this and realise where I am coming from.

Normally on forums it is good manners to introduce yourself and let people know who you are and what you are about. I have no idea if you are 17 and can't tell a spanner from a screwdriver of 50+ with a wealth of knowledge about minis. You can't put up your first post, that is almost the same question that has been asked hundreds of times before and expect to be treated like someone special, sorry.

However, stick with it the answers you want are here and the search really does work very well. Once you have had a brows then ask a more informed question.

http://www.twin-turbo.co.uk
http://www.hillclimbandsprint.co.uk/default.asp

A man without a project is like a like a woman without a shopping list.

lol i asked the same about spi and got the same replay but ben h is right you need to do your home work as for every 1 answer to you problems there will be 2 more problems to work out i have been at it for all most a year off and on and only just got to the point of worknig out what kit and add on in need so good luck

ms2 is beyond me come back carbs all is forgiven why did I decide to modify a spi mini to late to turn back now so please put up with my stupid question and humour me

This has been asked before in various guises but the answers below are how to relicate an MPI setup - WHICH IS A SEQUENTIAL PORT INJECTION SYSTEM - with a Megasquirt MS2 programmable ECU using the existing MPI components.

Engine speed/position sensors:

The MPI (or SPI) flywheel has the crankshaft position "wheel" built onto the back of it to run its own VR sensor.

As a result there is no need for a seperate 36:1 wheel on the pulley and an externally mounted sensor like most aftermarket systems use.

However, the patterns on the back of the MPI/SPI flywheels are not 36:1 but are unique Rover patterns. The MS2-Extra code has these two patterns available and the JimStim (the simulator used for setting up and testing an advanced Megasquirt unit) can generate them so they have been "bench tested" but I'm not sure if anyone has actually used that part of the code yet on a real engine. The tests show it will work, and if it didn't I'm sure it would quickly be fixed on the MS2-Extra forum.

The MPI has its own phase (cam) sensor installed on the back of the block reading direct off the camshaft. The phase sensor is required for sequential injection. However, on the MPI, it is another VR sensor.
For the majority who don't have this MPI cam sensor but want to run sequential injection it is a case of making our own, usually built into a dismantled dizzy, and we use Hall switches or Opto switches. This is well documented and photographed. The standard MS2 will take one VR and one Hall/Opto input so is perfect for the non-MPI setups.

So to use the standard MPI inputs which are both VR you need to add one small extra board for the second VR input. Jean on this forum (JBperf.com) sells them, as do other specialists, or you can make your own.

Temperature sensors:

If you want to use the Rover CTS and IAT sensors will need to use their calibration tables in the MS2 as it defaults to the american GM sensor tables. Entering new sensor values into the MS2 is straightforward, but the figures are hard to come by for the Rover sensors.
There are other sensors like Bosch, Weber Marelli (Cossie), etc where the information is more widely published but if you can find the calibration tables for the Rover sensors, they will work fine, all these temperature sensors are basic NTC devices.

Ignition:

The MPI uses a coilpack (wasted spark) and no dizzy. The usual MS2 build has a single ignitor chip to run a single coil and dizzy but it is a VERY common modification to install two ignitor chips and set the code for wasted spark. Usually we do this and use a Ford coilpack (no EDIS required, MS2 runs it direct) so using the Rover coilpack instead should be very straightforward.

Others:

The MPI has a MAP sensor mounted on the manifold whereas the Megasquirt has an internal one. Discard or ignore the Rover one and run a suitable sensing line to the MS2 box.
The Rover TPS and IACV should be compatible but may take a bit of figuring out which wire is which, especially to the IACV but the JimStim and tuning software provide a test routine to confirm it. Again, I'm not sure if anyone has actually confirmed the suitability of the Rover IACV on a real engine but, if not, there are plenty of alternatives that are proven to work.

O2 sensor - Megasquirt does not include a wideband controller. I'm not sure if the single MPI one is narrow or wideband but, in any event, it will need to be replaced with two (or more) wideband sensors and controllers. See below.

MS-Extra code:

The hardware answers are easy but it should be understood that the MPI setup was Rover's attempt to overcome the siamese port charge robbing issues enough to meet '95 on emissions regulations. It is sequential port injection with asymetric injection timing. The MS-Extra code allows for sequential injection and (the MS2 version only) has settings available within it (usually refered to as the siamese code) to provide all the options for variation of injection timing. Most of the development is discused openly on this forum but there is no "out of the box" or "one size fits all" answer. To make it work will take a lot of patience and will require more than one wideband setup for adjusting the AFRs between the inner and outer cylinders, ie, one in the centre exhaust and one in an outer exhaust.

There are other simpler ways of dealing with fuel injection (throttle body, wet manifold etc.) but it is unlikely any of these will meet '95 onwards emission requirements and will probably provide worse fuel distribution than an SU.



EDITED - to add further information and make it more suitable as a reference for future questions.

Edited by Rod S on 28th Jan, 2011.

Schrödinger's cat - so which one am I ???

Rod:
I appreciate the information that you have shared. Without the work you and many others have done I would never have attempted this type of project. I hope I can contribute at some point.

I would like permission to copy your response on my web site with proper attribution. It is a very good summary that many could use as a reference.
Terry

Terry,

You are more than welcome to pass this on, via your own website or otherwise, just refer it's origin back to the TM forum.

But please make it clear there are a few MPI bits that haven't actually been tested yet (as I point out in the thread) as most of us here are using pre-MPI parts or parts we have fabricated ourselves.

Rod.

Schrödinger's cat - so which one am I ???

I've got nothing to add to this, but I see that on the other forum, it has been suggested that it will take two years to implement.

It may have taken us several years to get this far, but now all the initial problems are sorted, there is no more to implementing port injection on the 5 port using the MS2/Extra code than any other form of fuel injection except an injection timing table and an extra wideband.

I believe Graham got his working in just a few months and uses it as a daily drive.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Yes I took about 3 months from recieving the bare Megasuirt to having a reliable running daily drive.
I started the project in Nov 2009, so its been on the road around a year now.
There were alot of teething problems, mainly due to dodgy wiring on my part and a hugh learning curve to understand the fundamentals of fuel injection during that time.
Although the car was running fine once it was one the road, it took a fair while after to properly get to grips with the MS extra code and so get what I consider now a very good tune.

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

Graham,

I don't know if you have a reference point for your current setup but how would you compare your current setup with a carb (or OEM injection) with respect to power, fuel economy and general engine behaviour? Any annoying points for a daily driver?

Jean

http://www.jbperf.com/

I never had a chance to take any logs when the car was running carb, so I have not got anything to visually compare. Also, I never had a reason to put the car on rollers before the conversion, so again no data to compare power against.
However, I now believe it would be simple enough to swap back to carb to get some data. I understand that fuel can be drawn through the MPI fuel pump if it is not powered (?), so I can use a small low pressure electric pump to supply the carb. I've previously tried screwing down the pressure from the MPI pump to get back to a low pressure to run carb, but my FPR won’t go low enough. If I can do this, I'm going to run with carb for a short while, just so that I do have logs to compare carb with EFI.

Equally I have also been thinking about taking the car to the rollers. Although I have the fuelling just about where I want it, I have done nothing with the ignition settings, so I think it’s worth giving it a go. Perhaps in the near future I will have some graphical data to compare.

The Engine is an unmodified 998. The only modifications are: an LCB and RC40 and of course what was needed to be done for the fuel injection, so I was never really expecting power gains. That said, my perception is that there is an increase in performance. Generally it feels like there is much more torque low down. It seems to pull better up hill and accelerate marginally faster . It revs out at around 4500rpm, but it used to with the carb as well.
This perception of increased performance could of course be attributed to the fact that the carb was never played with. It ran and that was good enough... or that I want to believe that it goes better.
Fuel economy wise, I don't do a lot of long runs, generally, the engine has only just reached temperature by the time I have got where I am going. So I cannot really give any firm opinion on this. If I had to guess, I would say there is not a lot of difference, purely because of how I have it set up. Although for general cruising I have the AFR’s at around 14.2 – 14.7, anything above 70Kpa and the fuel mixture is richer at around 12.7 – 13.4 AFR. And because it’s only a stock 998 NA, My foot is nailed to the floor for most of the time I’m driving it!
Generally the engine is a lot smoother running and much more responsive to the throttle since changing to injection.
I've not quite got the IAC set up correctly, so during the winter, it’s been starting fine, but ticking over at around 700Rpm for the first 2 minutes or so. Something I need to sort out.


Personally I find it much better as a daily drive in comparison to when it was carb'd: no messing about with choke and I have an electric fan fitted, which means I don't have the constant fan noise.
It just "feels" nicer to drive, especially around town.
My opinion is that it's a worthwhile conversion, so I'll now be getting a second MSII for the turbo build, rather than putting the 998 back to carb, as I had originally planned once the Turbo Cabriolet is ready.
One thing that does concern me for the Turbo engine is the AFR variance when warming up. Although I'm not too worried about it for the 998 NA, there’s a spread of around 2.5 AFR low down the rev range to about 1AFR higher up the rev’s until I've got the engine temperature up to around 71 -72 DegC...

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675

Graham,

Re. your last point, I see exactly the same (although my warmup is a fair bit longer as it still hasn't left the garage)

See about 1/3 of the way down here

http://www.turbominis.co.uk/forums/index.php?p=vt&tid=386739

Interestingly mine also stabilises at about 75 degrees.

I'm sure it's wall wetting as Paul says, and I would hope the warmup period of a turbo engine would be much quicker once on the road.

I can't think of anything obvious in the code to deal with it as I assume the warmup curve, although adjustable, affects both VE tables equally.

Schrödinger's cat - so which one am I ???

Graham, it's nice to see that it makes for a good daily driver.

Graham, Rod, I could look into adding a small curve that would be used to vary the amount of enrichment on the inner (or outer) cylinders by a certain percentage. I was thinking that a temperature dependent value could be used where you would put say 70 at 10 degrees and it would use 70% of the enrichment at that temperature.

So the current enrichment would be used for one pair and the modified (increased or decreased) enrichment would be applied to the other pair. Do you think this would mostly solve the issue?

Jean

http://www.jbperf.com/

Jean, I think that would help.
The inners are always richer than the outers during warm up, so maybe applying that curve to the inners would be better?
This would allow a reduction in the warm up enrichment applied to the inner cylinder, rather than adding to the enrichment applied to the outers.

When I originally started getting AFR's matching last spring, I was using a single VE table, and 2 injection timing tables with good results.
I have been for the last 3(ish) months been using 2 fixed injection timings and 2 VE fuel tables to match the AFR's.
I've not really looked at whether the single VE fuel table and 2 injection timing tables give the same sort of AFR spread during warm up.

I think, if the configuration was set using 1 VE table and 2 timing tables, the curve you are suggesting would not work?? as the curve would need to alter one of the injection timing tables, not one of the VE Fuel tables.
I'm just thinking that this could restrict the options available for tuning? Or maybe I'm over complicating things.
Either way, as I’m using 2 VE tables at the moment, I think it would be nice to see if this helps.

’77 Clubman build thread
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=618189

Siamese 5 port EFI testing
http://www.turbominis.co.uk/forums/index.php?p=vt&tid=611675