MS in 27 LITRE RR / ROVER METEOR

[attleej]

MS1 ON A METEOR IN A CONQUEROR ARV
Dear All,

BACKGROUND. I thought the Forum would be interested in reading about MS1 fitted to a V12, 27 litre Meteor petrol engine. The REME Museum’s Conqueror Armoured Recovery Vehicle (ARV) is fitted with a mechanically fuel injected Rolls Royce (Rover) Meteor M120 petrol engine developing around 800 BHP. The Conqueror is a heavy ARV equipped with a winch that can pull 45 tons on a single line pull. She weighs just under 60 tons and is 13 feet wide.
Unfortunately, petrol was leaking into the engine oil sump via the FIP. In addition, it was hard to start due to lack of cranking enrichment or ASE. Ironically, it would have been easier to have fixed the FIP despite lack of literature or test equipment!
Out of interest, watching a U Tube video of a carburetted Meteor Mk IVB running with exhaust stubs only, it was clear that some cylinders were emitting hotter exhausts than others. This could be due to slightly unequal breathing between cylinders and is a problem less likely, or easier to solve, with EFI.
It was decided to fit MS1 based EFI using self-assembled V3 ECU kit from Eddie Walsh in Essex, England and I will describe below the current iteration of the system:
RPM INPUT. In place of the petrol injection FIP an RPM transmitter was fitted. It consists of a steel chopper wheel and a Hall Effect switch. Interestingly the FIP was driven at crankshaft speed so only six segments were needed for the chopper disc.
IGNITION. One of the two magnetos was converted by fitting an array of six photo transistors and a rotor with two segments slightly non-symmetrical in order to give 12 equally spaced impulses to a separate module with 6 IRGS14C40L Ignition IGBTs. The angle of the segment is designed to provide the correct dwell at 2500 RPM. The photo-transistors are powered from one of the two 12 volt fuel pump circuits. This stops any current flowing through the ignition coil if the engine is not turning and avoids burning out the coils. The power for the coil packs comes from the tank’s 24 volt ignition circuit via a power resistor to keep current correct. Three dual wasted spark coil packs from a Renault Clio provide the spark for the 12 ‘inlet’ sparking plugs. The ‘exhaust’ magneto was left unchanged apart from fitting commercial spark plugs in place of the military / aircraft type plugs which are very time consuming to remove. The high voltages go nowhere near the ‘RPM’ sensor circuits. I call the photo-transisors and rotor in the magneto housing the “Discombobulator”.
It is anticipated that the MS EFI system will give significant advantages without interfering with the ignition advance characteristics. Without access to an engine dynamometer, it is not likely that I will be able to improve on the current ignition advance, as designed. The automatic advance units are incorporated in the drive units of the magnetos so the ignition system will still have automatic advance for both inlet and exhaust plugs. The auto advance units looked to be in good condition and it never occurred to me that there could be a problem with them. Eventually I found out that they were both absolutely seized solid with rust. ‘Special Measures’ were needed to disassemble them.
The ‘new’ ignition system will not limit engine RPM but the MS ECU does so by cutting the fuel off at 2500 RPM.
INJECTION. It was decided to use the ERC 3620 injectors for the best of reasons (they were what I had!). 24 of them would deliver 667 BHP within their rating. Whilst not quite enough it would probably not be noticeable, especially with the relatively low RPM. Opening time at idle is about 4 or 5 ms which is about right. Unfortunately the upper limit off about 25 ms on REQ FUEL was a nuisance. I think that I have 50 ms per cycle at 2500 RPM. It would be good to be able change that limit. I had not read the MS small print that indicates a severe limit on the number of low impedance injectors that can be controlled with the PWM of the MS. Fortunately, we were able to get the tank back indoors before the second MOSFET blew up in protest!
To safely control the injectors, a ‘peak and hold’ (P&H) system is now used. The two signals from the MS signal the pulse length to the 24 off LM1949 P&H chips. Each injector current passes through a low resistance, when the current has risen sufficiently, the resulting “current sensing voltage” tells the P&H chip to reduce the current from a ‘peak’ to a ‘hold’ value. After about 3 ms the P&H chip will, in any case, automatically reduce the injector current to a ‘hold’ value. The current is controlled by a 2N6044 or a TIPxxx Darlington pair transistor. The data sheet for the LM1949 gives full details.
FUEL PUMPS, FILTERS & PRESSURE REGULATOR. A pack comprising two electric fuel pumps, two big fuel filters, the pressure regulator and a gauge is mounted near the engine. I am reluctant to fit a pressure smoothing tank as it would take a while to build up and lose its pressure. Unfortunately the pressure does fluctuate. I discuss how this might be cured later under “improvements”.
MAP SENSOR. I found the MAP sensor and vacuum connection to MS to be vulnerable when working on the MS PCB. I mounted the MAP sensor in a separate small enclosure. This is connected to the main ECU enclosure by a three way military pattern connector. In addition, there is a changeover switch and potentiometer of which more under “Test & Simulation”.
LAMBDA SENSOR. As it is an MS1 system it is only fitted with a NB Lambda sensor. The system does correct the mixture once closed loop is enabled. I can fit a WB sensor in the other exhaust ‘up’ pipe if I want to.
POWER & ENCLOSURES. The MS and P&H circuitry are mounted in a relative large military pattern enclosure and connections are made with military pattern plugs and sockets. A ‘high power’ box is mounted on the outside which contains circuit breakers protecting the MS itself, each of the two pumps and each of the two injector circuits. The various relays required are mounted inside this box. The complete assembly, which is located in the winch compartment, next to the engine can be removed in about three minutes.
The tank’s 24 volt ignition circuit energises a relay in the ‘high power’ box which turns “on” the MS system. There is a facility to switch off the ignition and to operate the starter on the outside of the MS enclosure.
The MS system is powered by a 12 volt car battery which also operates the starter on the ARV’s winch engine. The 12 volt battery is kept charged up by a car alternator driven of the main engine prop shaft driving the main gear box. Thus it is hoped that the supply to the MS is relatively clean.
TEST & SIMULATION FACILITIES. With so many circuits with the potential to go wrong, built-in test facilities have proved invaluable. BNC sockets are provided so that an oscilloscope can be connected. A thirty-way rotary switch allows the scope to be connected to each of the 24 “current sensing voltages” arising from the injector currents flowing to earth in the main enclosure and the 5 volt circuit. If a particular P&H circuit is not working correctly the waveform will be different and if there is a break or high resistance in the circuit this will be obvious. All the injector circuits can be checked in about two minutes with the engine running or under simulation.
I have a test lead with the appropriate multiway military connector, one injector, a thirty position rotary switch and small battery. This can be used to bench test each of the P&H circuits in ECU enclosure.
The main enclosure has a 12 way military pattern connector for all the sensor connections. I can connect a ‘simulator’ that has the appropriate resistances for the TPS, IAT and MAT. The RPM signal comes from the signal generator on the scope. Since the MAP sensor has a simulation function I can convince the ECU that it is controlling a 27 litre engine powering a 60 ton tank when in reality it is sitting on my bench at home.
ARMOURED FIGHTING VEHICLES (AFVs). I think that it might be worth pointing out that AFVs are extremely difficult work on. For instance, very often it is extremely difficult to access components or to view them directly. It is very difficult to cut-out one cylinder to observe if it is firing or not. They are also expensive to run with a fuel consumption of 3 gallons per mile.
OPERATING EXPERIENCE. The first secured advantage of the MS EFI is that the engine starts after the necessary few revolutions for the pumps to pressurise the injector fuel rail. If it does not start after those few revs, it is not going to start at all and something is wrong.
There is one oddity with the pump circuit. When the ignition is switched “on” for the first time, the fuel pumps do not start up to prime the fuel system as they should. Nor will they start if the engine is rotated. If the ignition is left “on” for a few minutes, suddenly the injectors will go “Clack!” and the pumps will run and then the engine can be started. Subsequent starts are not a problem. After placing test LEDs in suitable places in the system I am sure that the fault lies in the ECU or firmware.
The engine has not yet reached full potential with the MS EFI system. The compressions are satisfactory and even but it is easy to foul the plugs on this type of engine. The idle is pretty good with plenty of low speed power. The mixture automatically adjusts with the MS stepping the fuel up or down until it is correct and you can feel this happening. When driving the tank you need to know to try to hold the revs and MAP reasonably constant so that the MS can adjust fuelling. Readers should not underestimate how difficult it is to keep a constant load on a tank engine when rolling along on the flat around a camp (the camp authorities are not too keen on bare steel tracks either!). With a car you can just pull up a hill or press the foot brake a bit. It is easy to deduce that the tank’s VE table is not correct at all. It is much harder to develop it! I suspect that a WB O2 gauge running on the other exhaust bank would be a great help but the inability to develop a constant load is a problem. Has anyone got a VE table for a Meteor that would be a good start point?
FUTURE IMPROVEMENTS. The tank (ARV) is now going well enough so that it could unload from a transporter, drive a mile off-road and do a winch recovery. However, it is nowhere near full potential.
Longer term, I am seriously considering building an engine dynamometer rig specifically for Meteors both carburetted Mk IVB and fuel injected M120. It is a lot of work to fit a Meteor to a Centurion only to find that it is not running properly. Unfortunately, a Meteor cannot be run without water in it, even for a few seconds.
If I could run the Meteor in the Conqueror at a certain MAP for ten seconds I could develop a great VE table in an hour!
In the short term, I am developing a discombobulator with 12 photo-transistor switches, driving 12 IGBTs and with a single sector rotor. This would mean that I could have a wasted spark coil for each of the 12 cylinders or a coil-on-plug for either just the inlet plugs or both the inlet and exhaust plugs with another 12 IGBTs. A significant advantage of this arrangement would the ability to easily momentarily ‘switch off’ any of the 12 cylinders or even any one of the 24 spark plugs.
ACKNOWLEDGEMENTS. I would like to thank Nige Barker at “Megasquirt V8” for helping me with various software problems. Andy Trimmer has tirelessly helped me with the heavy lifting and by demanding perfection in my engineering.

[DaveEFI]

Crikey.

[jsmcortina]

Sounds like an interesting project. If you want per-cylinder ignition control and per-cylinder fuel adjustments, do consider MS3. With some DIY upgrades it can run V12 full sequential.

James