My 998 project - MS3 cam sensor built
Turbo Mini Forums

Thought I would add something on my project, not sure it's quite in the same league as the others on here tho!

I bought this Mayfair as a non-runner in 2007 and got it on the road in 2012, after much work and moving house a couple of times. It needed a new front end, sills and the other usual rusty areas replacing, engine rebuild, rewire and so on.

If the pictures work this is a pic of it just after passing its first MOT.



Anyway, toward the end of last year I laid it up for winter and have been (slowly) making some improvements. The original plan was to fit an EDIS based Megajolt setup, things have snowballed a little from there!

Got a trigger wheel kit fitted for the jolt, while the radiator was off for that I made up some brackets to attach it to the inner wing.



Theres an electric fan on the inner wing from a Corsa, seems about the right size for the rad. The rad fits but removing the cap is a bit fiddly.



Hopefully moving the radiator there will force a bit more air though it on its way out of the engine bay. I started reading up on electric water pumps as well and decided to have a play with the pump in the pic. I wasn't convinced it would flow enough to replace the mechanical pump, instead it's been plumbed in as a booster pump that can be turned on when the engine is shutdown. Got some more pics of the setup later.

Found a Daihatsu in a scrapyard a while back and took the Nippon alternator off it. Decided to mount it quite high up, similar to where they are on an MPi, in the hope that an oil cooler would fit underneath. So, had some bits laser cut and welded this together.




It's bolted to the bracket for the water pump mounted engine steady. It seems solid enough despite it's appearance, will see how it holds up! This bad photo shows the turnbuckle used on the lower mounting.



Didn't like the bulkhead bracket supplied for the water pump engine steady so made this up, using a repair panel for the original engine steady to form the lower half.



Had this bracket laser cut to fit the starter solenoid to the old alternator bracket mount.



I had the exhuast off to reseal a leaky connection and to replace the rubbers, thought I'd plan ahead and welded in bosses for two lambda sensors and three EGT sensors.



I have a spare narrowband that I'm planning to use to help set the carb up and some K type thermocouples, mainly to plug the holes up for now as I need to sort out the electrical side of things.

Needed to cut off the breather canister to fit the extra steady so decided to upgrade the breathing and fit a catch can. Bought a cheapy one off ebay and made this bracket for it.





Catch cans with internal baffling were a bit pricey so I improvised and modified the ebay can with some stainless mesh (scourer pad!). Will see how well it works, from looking around this mod has been done quite a bit before.

This is were I got to earlier this week with everything back together.



The electric pump plumbing took me a while to decide on. I did play around with putting it in the lower hose but wasn't happy with how that looked. The pump has 19mm connections and I was concerned it would restrict flow. The main reason for fitting the electric pump was to keep water circulating when the engine is turned off, in conjunction with the electric fan.

So, iIn the end what I've done is put a T-piece in the bottom hose which connects to the electric pump inlet. the outlet connects to the 22mm copper pipe in the pic which is attached to the rocker cover. That the tees off at the end to the heater valve outlet on the head and to a modified Ford Ka heater valve. The Ka valve connects to the heater matrix and the return runs along the back of the rocker cover, under the big copper pipe and into an MPi thermostat sandwich plate.

The thinking is that when the engine is running the electric pump is off. Water can flow out of the heater takeoff on the head to the heater valve and back to the sandwich plate. The big copper pipe has a one way valve in it (just in front of the carb) which stops any flow going down that line and bypassing the radiator. When the engine is off the elctric pump turns on and pumps water out of the bottom hose and into the 22mm copper pipework. Some will go into the heater, some into the head
and hopefully promote some flow through the engine. I think this will be more useful if/when I go turbo as a water cooled unit could be plumbed in, probably in the return from the heater, which should guarantee it pumped flow when the engine is off.

I want to instrument things up a bit more to prove to myself that it actually does what I think it is!

Anyway, the coolant was filled up yesterday and the engine, surprisingly, started first time. Just waiting on some hose to finish connecting up the catch can and then I can think about booking an MOT.

Edited by beejay on 22nd Aug, 2015.

Some interesting ideas goin on here, i like ur idea of movin the rad over... let us know how this goes

Cheers, will see how things work out on the road!
Had a mixed couple of evenings in the garage. Got the narrowband rigged up to have a play, heater seemed to work and was using a digital multimeter to measure the output. It was giving a 6V output when the heater was turned on, which then dropped down over a few minutes to about half a volt. Output didn't seem to change much as I was adjusting the mixture, although I'm not sure how fast the sensor can react? Don't know if this suggests a broken sensor or if I'm doing something wrong? The sensor is an secondhand one thats been on the shelf for a few years so don't know for sure if it works.
Then to round off the evening it pissed coolant all over the floor due to a couple of loose hose clamps :/
On the plus side I have managed to wire up the rad fan and give the car a bit of a wash ready for MOT tomorrow!

Is it me or does the altantor look like its being pulled over a bit?

Good luck on mot

I am impressed by your plumbing skills! I hope you didn't use self cleaning flux though. Good luck when you get it MOTed :)

They don't die, they just get faster!

your doing ok

Cheers guys, MOT passed this morning :) Was great to drive it again, been too long! Temperatures were ok too so looks like the new rad position is working. Started creeping up when it was sat in the MOT til I turned on the fan, then you could see the temp gauge falling. Which is nice. Bit noisy mind.

Chalkie, you may be onto something, I hope its just the angle of the pic tho! Will have a gander tomorrow

Shauna, thanks, got a lot of practice in replacing the central heating! Whats the issue with self cleaning flux?

So, the original plan was to fit a megajolt as the igniton timing on the dizzy was a bit all over the place, Got a trigger wheel kit fitted on a 1275 damper pulley when the radiator was out. Then a slight change of plan and I got a megasquirt 3 built up for the engine instead. Very interested in trying injection next and in the meantime it lets me do some data logging :)
Got the MS reading the trigger wheel ok (after wiring it up the right way!), an MPi temperature sensor in the sandwich plate and a wideband lambda. Thi has been quite useful getting the carb set up a little better so far.

Sorted out a bracket for a Ford coil pack.



Made up a bit of a dashboard to provide somewhere to mount the MS, relays, fuses etc and got the rev counter running off the MS tach output



Then bit the bullet and switched over from dizzy to coil pack. Left the dizzy in for now since I've got nothing to plug the hole!



Ran it on fixed timing to check it was correct on the timing light then tried the MS base map as a starting point. Seems to run fine but is probably not optimal, will have to experiment maybe. Certainly seems smoother at idle now. There is an occasional sync loss which I'mhoping a new crank sensor will fix.
Next thing is to get the MS controlling the fan and logging data on the SD card, then to think about sorting out some fuel system upgrades to work with injection.

Cool beans! Liking this one.. i reckon will evolve into awesome machine haha

Bit of an update on this! Been driving it as much as possible until the starter motor packed up toward the end of last year, which wasn't much of an issue as I don't use it much over the winter months. Changed it over in the new year and been out and about as much as possible. That was until a few weeks ago when the engine just lost all power and just about got me back home. Seems that the carb had sprung a leak, which seems like a good excuse to take it off and implement the injection plan.

Have a swirl pot, pump, filter and return line ready to fit this weekend. Also have an SPI and MPI manifold on the shelf and now trying to decide which one to fit.
The original plan was to put the SPI on at the moment to get the car mobile again as that only needs an air filter and IAT sensor to finish it off, so it should be the cheapest way forward. Issues are that its a big old lump and from an initial trial fit there isn't much clearance to the bulkhead. (I thought it was the MPi cars that had the engines moved forward a smidge?). Also wouldn't be able to drive the idle stepper as my MS3 isn't wired up for it. Not sure if not plumbing up the water and/or electric manifold heating will affect driveability either?

The MPI one doesn't have the injectors or fuel rail, which all will take time to work around given my usual glacial progress in the garage. However it would seem to give me a more compact manifold, and I've already started adapting a 2 wire idle control valve to fit. My MS3 doesn't have the siamese injection options in tunerstudio, although they are mentioned in the MS3 manuals, I'm guessing a firmware update may be required (I haven't been able to find at what firmware version they were included as yet)

Need to see what fits best at the weekend and make sure I'm clear on the hardware and software mods required of each option, some more reading on here required I think :)

A number of us on here use the MS3/3X without the siamese options. It is not as flexible in terms of injection timing as the MS2 but gets the job done.

Easiest solution for the IACV is to use a 4 wire one of a Peugeot/Citroen. Direct bolt in replacement for the Rover 5 wire item.

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."

Since the last post I decided on going for the MPI approach. Got manifold fitted and plumbed up, 2-wire idle valve adapted, modded the MS3 for second lambda and cam input and spent last week sorting out the software setup when I've had any spare time.

At the moment I have it set up in semi sequential mode with both injectors connected to injector channel A. This would fire both injectors every 360degrees and, with the right timing, would fire either onto an open inlet valve on an outer cylinder, or onto closed valves ready for the inner cylinder. Not ideal I know, but was interested in seeing if this would be good enough to get the car driveable whilst I sort out a cam sensor.

The MPI manifold I had didn't come with injectors so some Bosch ones from my old 318 were fitted. It was only once I'd read up on the intricacies of Siamese injection that I realised these are a bit small at 154cc/min. Got some bigger ones on the way though. In the meantime I pressed on with my comically undersized injectors to see if I could at least get it to idle and prove everything works. Got it running earlier, briefly mind since it was quite late. Hopefully have some time tomorrow to play a bit more.

Had a bit time to play today and got some pics.

modded the MS3 a while back with this breakout card. Takes the second lambda input, fan control output, CAN lines and the MS3 cam input to a DSUB on the back. Also got a Bluetooth dongle thingy for shadowdash




re-plumbed up the fuel tank with return line and all the rest of it



It had the alloy tank feeding the carb with a low pressure pump originally. Added the other tank for extra capacity and cool points. the low pressure pump now feeds that and theres a high pressure pump and filter drawing from the new tank mounted to the side of the battery box underneath.

This is the manifold all connected up.



Idle valve blanked off for now while I get things sorted.

Had a bit more of a play with settings and got it warmed up. seemed to be getting somewhere with getting AFRs where I wanted them, until there was a weird smell which turned out to be the damn MAP sensor line melting on the exhaust manifold Called it a day there and left it looking a bit pissed off with me.



and finally a file photo of a time back when I could actually drive it



Also, does anyone know for sure if the MPI air temperature sensor has the same characteristics as the MPI coolant sensor? I found measured resistance data for the coolant sensor in a thread on here, using the same calibration for the air gives me sensible temperature figures (ambient when cold, mid 50's from memory when running). Just wondered if anyone knew for sure before I get too far.

been making up a cam sensor so I can move to full sequential injection recently. Just tested it and it seems to be working, in that I can see the cam tooth trace in the tunerstudio composite logger. Is the timing of the cam signal critical relative to the missing tooth on the crank? It was timed up according to the MS3 manual and the cam signal is a few teeth before the missing tooth on #1 TDC. Will have some more time to play later and post some pics if anyone wants to see :)

If you have timed the cam signal as per the MS3 manual, then you will not have a problem. That's how I do it.

I saw your post on the MS forum about the siamese options on MS3. I just use the standard sequential setup with a single long pulse to feed the inner and outer cylinders. You can also use a pulse per cylinder, wiring two injector outputs to each injector.

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."

PhilR: will post pics later then, I will warn you that my approach isn't quite as slick as the others I've seen on here!

Paul S: Thanks for confirming. I'm following your approach of using outputs A & C from the MS3 for the single long pulse approach. Do you have any pointers on where to start with injection timing, is it in your build thread?

Looks good! Doesn't need to look slick, it won't make it any more efficient! Personally I kinda like things a little rough around the edges..

Looks good! Doesn't need to look slick, it won't make it any more efficient! Personally I kinda like things a little rough around the edges..

Having siamese options on the MS3 would make it more flexible and add a few things that are not possible with the current standard setup.

As is possible on the MS2, you could have a single pulse or 2 pulse mode with a single output per injector and so without wiring changes. Also, you could have a hybrid mode where you can change from a 2 pulse mode to a single pulse mode at a certain RPM. And you could also have split injection timing in the 2 pulse mode where timing is adjusted differently for inner and outer cylinders; that can be useful at higher duty cycle to avoid pulse collisions and leave the injector enough time to do a complete on/off cycle to keep injection more linear.

Having said that, I'm not sure there is that much interest from the MS3 developers in implementing this. It seems that the MS3 development is going more towards satisfying the needs of DIYAutotune customers and leaving the rest as secondary at best. That doesn't make the MS3 a bad choice, far from it, but rather you have to be satisfied with what you have now or what will be developed for those users.

Jean

http://www.jbperf.com/

so, for the cam sensor I used an opto sensor with the rotor arm of the dizzy as the target. Removed the guts of the dizzy and welded the centrifugal advance up, then chopped off one of the HT posts and attached the opto sensor using a short piece of aluminium.





wrapped up with loom tape for now to seal it off. Like I said, a bit agricultural but its taken an afternoon of chopping and drilling, and a few quid for the sensor. Will probably work up something cleaner as I'm not convinced on the longevity of the sensor I've used in an automotive environment.

Also made up new injector drivers to allow me to use the MS3 injector outputs for full sequential control



these replace the onboard injector so I haven't had to do anything with the loom. Got the MS3 back in the car today and tried to get it going, however after a few cranks it became apparent it didn't want to fire. After a bit of head scratching I thought I should check the injectors, given all the messing around I've just done with that side of things. Embarrassingly, it turned out I'd forgotten to break one of the tracks on my prototype board, so one driver was always on. I've managed to fix that tonight and check the injectors work properly using the tunerstudio diagnostics (which I should have done in the first place, lesson learnt!). Not sure how much fuel made its way into the engine while all this was going on but I might change the oil tomorrow just as a precaution. Fingers crossed things go better tomorrow!

I don't know why you did not just fit the MS3X expansion. That would have given you all the extra I/O that you need.

Anyway, injection timing table for a 998 running an MPI manifold and injectors:


We have not ran it at boost yet so the top 4 lines are unproven but based on my earlier boosted 998.

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."

Thanks for the pics That's one approach that I'm considering, but as you've noted, durability is questionable. I may try something very similar but with an Audi hall based cam sensor in place of the opto.

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