What MEMS is
MEMS, the Modular Engine Management System, is the engine controller Rover Group used through the 1990s on its own cars and on Rover engines fitted by other makers. It could be built up for electronic carburetion, single-point injection or multi-point injection, with or without ignition control. SPi and MPi are those two injection forms. The ECU was a joint design: Rover wrote the software, and Motorola Automotive and Industrial Electronics Group manufactured the unit. Lucas had done this work for Rover before. MEMS is the in-house replacement, not a Lucas 14CUX or a Lucas 6CU.
Rover decided on the new system in 1985. It had to cover later engines, use less power, and take less space under the bonnet than the systems it replaced. The first production fitment in the published account is the Austin Montego 2.0 in 1989. Over the following years it was used on the Mk VI and Mk VII Mini and the MG F and MG TF, and with the Rover K-series in the Lotus Elise and in Caterham models.
Hardware
The processor is an Intel 8096 at 12 MHz, with 8 KB of on-chip ROM and 232 bytes of general-purpose RAM. The main connector is a 36-pin TE Connectivity 344108; the harness mate is a 344111. The modular hardware, as listed for 1990, could add a second injector, batch-fired injectors, automatic-transmission control, pulse air, EGR, a purge valve, knock sensing, air-conditioning control and oxygen sensors around a base of programmed ignition and one injector. Not every car has every module. The connector population is part of the identification.
Earlier ECUs carry the manifold-pressure sensor inside the case, so a vacuum hose has to reach the ECU. On MEMS 1.6 and 1.9 that sensor is the Motorola 5141550T02, and the hose passes through a vapour trap so liquid fuel does not reach the ECU. A split or a liquid-filled hose is a load signal fault, not an air leak at the manifold only.
What it calculates
The ECU reads engine speed, manifold absolute pressure, coolant temperature, intake-air temperature, throttle position and battery voltage. Fuel and ignition each start from a three-dimensional map. Wide-open throttle and a cold start enrich. Crankshaft speed and position come from a magnetic reluctance disc. Closed-loop control needs a lambda sensor; without one the system stays open-loop. There is a rev limit, an overrun fuel cut, cranking and after-start enrichment, and a battery-voltage compensation on the fuel. The idle-air valve position is learned, so two engines of the same calibration do not need the same step count.
A failed sensor is replaced in the calculation by a nominal value. Rover literature calls that the limited operating strategy. The engine can keep running with a poor signal. The substitute is not the measured temperature, pressure or throttle angle, and it is not a reason to ignore the fault.
Revisions
| Revision | What the published accounts record |
|---|---|
| 1.2 | First production version. Not for cars with catalytic converters. One 36-pin connector. |
| 1.3 | Can operate emissions equipment. A 36-pin connector and an 18-pin connector. |
| 1.6 | Finned aluminium case. Usually one 36-pin connector; some also have an 18-pin connector. On the K16, the 1.6 ECU is the 1990–94 unit. A specialist diagnostic guide describes the common SPI and early MPI cars of 1993–99 as having a round 3-pin diagnostic plug, lines in the lid and a vacuum hose at the ECU. |
| 1.9 | From mid-1994. Revised idle-air control and multi-point injection. On the K16 this is the ECU from 1995, still with a distributor on the inlet camshaft. The same diagnostic guide describes a 16-pin connector, a cross on the lid and a vacuum hose beside the main plug. |
| 2J | Sequential injection, falling back to batch firing if the camshaft-position sensor fails. Also the controller for Rover VVC variable valve timing, which is why the VVC engine could drop the distributor: that engine has cam belts at both ends. The diagnostic guide describes two 36-pin plugs, black and red, a large X on the lid, no vacuum hose, and a 16-pin diagnostic plug. Early Mini MPI cars are in this group. |
| 3 | Introduced with the Rover 25 and Rover 45 in 1999. Twin coils and sequential injection on the K16. Described as supporting European on-board diagnostics. The diagnostic guide describes a 16-pin plug and no vacuum hose. Specialist mapping of MEMS 3 has been published without the enclosure changes used to reprogram a 2J. That is a statement about access, not a calibration method. |
The K8 started with a carburettor and a breakerless distributor on the camshaft. MEMS single-point injection is what the K-series account says became standard with the Rover 100 in 1994. A K16 of the early 1990s can be single-point or multi-point on a 1.6 ECU. Match the revision to the engine and the year. Do not fit a 1.9 map unit in place of a 1.6, or a 2J in place of a MEMS 3, because the connector or the case looks similar.
Diagnostics
Early MEMS predates a standard on-board-diagnostics socket. The protocol is ROSCO, Rover Service Communications. Early cars use a round 3-pin TE Connectivity 172202 plug. Later cars use the 16-pin ISO J1962 connector. A stored fault is cleared by a command on that port. Disconnecting the battery is not the clear procedure given for this system. Dealer testing used COBEST, Microcheck and Microtune, and later Rover TestBook.
ERIC
A carburettor system developed at the same time was called ERIC, Electronically Regulated Ignition and Carburetion. The published comparison says ERIC did not use the MEMS Intel controller. It is a different ECU. A carburettor Montego or similar Rover is not diagnosed as MEMS because the era matches.
Sources
- Modular Engine Management System, Wikipedia, citing Haynes, the 1990 ISATA paper by Crabb, Duncan, Hiljemark and Kershaw, and Rover training material. Accessed 2026-11-13.
- Rover K-series engine, Wikipedia, engine-management section. Accessed 2026-11-13.
- James Pearce, Rover MEMS diagnostic and remapping, connector and revision identification used above for 1.6, 1.9, 2J and 3. Specialist site, not a factory manual. Accessed 2026-11-13.