EFI Build Diary
I sourced the following components for the conversion to fuel injection:-
Twin dual Jenvey 40mm IDF flange throttle bodies
Jenvey idle air manifold (4 to 1)
RAT custom air horns and 4 individual K&N air filters
RAT centre-pull linkage
Megasquirt (MS1extra /V3 board) ECU
Bosch idle air valve
Bosch Hi-Z 30lb/hr injectors
K&N air filter for idle air
EDIS ignition module
Late style Ford EDIS (zetec) coil pack
36-1 toothed wheel, machined to fit crank
Billet pulley, machined to accept toothed wheel
Mallory EFI Pump
Mallory EFI Filter
Mallory return style regulator
Air & coolant sensors
MAP restrictor valve
Fuel collector pot with integral filter
DTAfast S40 Pro ECU
Set of Custom Magnecor Spark Plug leads
Various AN-6 and AN-8 fittings and adaptors
Various barbed hose adaptors (for vacuum lines)
Various electrical connectors
Ford Zetec coil connector (3-pin)
I also bought the following tools:-
Small Propane Torch & HTS-2000 Aluminium brazing rods
Firstly an admision that i've decided not to use the Megasquirt ECU system but have switched to using a commercial system instead - The S40 Pro from DTAfast. At the end of the day it was the thought of wiring the car up to a 25-way D-type connector (normally used for computer RS232 serial comms!!) that made me think that Megasquirt was a bad idea from a reliability point of view and that my car deserved better!
I've also found a decent engine services company in Reading (CC Engine Services) who machined the billet pulley and toothed wheel for me, to my specifications, to fit the crank hub. The toothed wheel now bolts directly to the crank with the pulley in front (the thickness of the toothed wheel was machined away from the hub of the pulley to ensure belt alignment was not affected)
The toothed wheel has to go behind the pulley due to the close proximity of the exhaust system at the front (photo below is taken from the floor looking up).
I've now made bracket for the VR sensor to "read" the toothed wheel (photo below). I made this from 1.5mm thick aluminium angle iron with a 10mm thick tapped aluminum block to take the sensor. I brazed the two together at the correct angle using HTS-2000 brazing rod. I was impressed with the joint made - stronger than the original alumnium (hard to grind with the angle grinder) although this technique is not suitable for complex multiple material joins as you cannot work on just one area of the material - the whole work piece heats up and all the HTS-2000 melts.
My car currently uses a CB Performance hex-bar throttle linkage sat atop the engine. This mounts on brackets fitted to the air cleaner bases specific to my Weber IDF carbs. When moving to throttle bodies I won't be able to use the same linkage (well not without fabricating new mounts) so I am swapping to a superior centre-pull one (I'd had swapped to it for my carbs as well but it seemed too much work). Unfortunately this involves re-drilling the throttle conduit through my DTM shroud which also involves raising the oil cooler & housing 3/4 of an inch - not much but enough to make it interfere with the inner skin of my engine lid (again - see my original DTM install in my sites Outlawification build section). I fear to remove the oil cooler i'll need to drop the engine from the car due to the rear bodywork - Indeed Jake recommends engine removal for installation of this linkage. All in all, a bit of a pain!
First step is to drill and tap the alternator stand to take the bracket for the linkage. This was straight-forward enough except my linkage appears to have been manufactured 180 degrees out... I'll worry about this later. Quality of construction of the linkage is excellent.
You can also see in the picture above that i'll have to relocate the oil breather hose fitting on the oil filler so that the hose won't interfere with the linkage (1/4 NPT thread)
Well, as you can also see, I've managed to remove the oil cooler, housing, and shroud without having to remove the engine - by about 3mm of clearance and lots of wiggling! I've also stripped out most of the fuel lines and also removed the brake servo feed lines - I plan to relocate these outside of the engine bay to make them more accessible and to give more room in the engine bay itself.
I've also relocated the throttle conduit and having a little extra degree of freedom in the speedster body rather than a beetle I've managed to do this without having to raise the oil cooler housing at all which will save me work down the line....
You can see the conduit just squeezing under the main part of the housing prior to inersecting with it at the rear.
Because of my desire to have a "choke" for when the engine is cold I need to install an idle air valve. Jenvey sell a manifold and adaptor kit that allows you connect feeds from each cylinder, taken from the bypass valve, together. This is alot of plumbing but I can use it to my advantage to get a better MAP signal for the ECU. Here are the component parts beginning at the manifold (I'll show the bypass adaptors with the throttle bodies later).
After some thought, I decided to mount the ECU in the engine bay on the left bulkhead, with the fuse/relay box on the side of the engine bay next to it. I'll keep the existing 8-way plug into the engine bay (carrying tacho, oil warning, batt warning, earth, oil pressure, oil temperature, ignition switched 12v, and (obsoleted by LC-1 digital signal) analog lambda) and use the obsolete wire as the fuel pump power feed (the ECU controls a relay that enables the fuel pump and turns the pump off when the ignition is on but the engine not running - I used to control the fuel pump from the oil warning lamp with a push button overide for start-up, and the relay for this is at the front and will be removed later)
The DL-32 (Innovate Datalogger) is currently only logging AFR, RPM, and G-Force (internal sensor). I'd always planned to add a wheel speed sensor to log MPH and was going to do this on a rear wheel - However, if I do this on an undriven wheel I can then also feed this signal to the ECU which will enable Launch Control (When put into launch mode by pressing a button, you fully depress the throttle pedal and the ECU will automatically keep the car at the set "launch" RPM (say 2750rpm); As you let the clutch out, the ECU manages the power to ensure optimal wheel slip off the line without the driver needing to lift off the throttle. Fairly pointless for normal use, but a cool feature, and free as I have the wheel speed sensor already!). I'll also connect the TPS and MAP signals from the ECU sensors to the DL-32 (i'll connect MAP voltage rather than use the DL-32's internal MAP sensor to save running a hose to the DL-32 which is in the battery compartment).
The fact that i'm sensing MAP, of course, means I plan to use MAP as primary load rather than TPS. The DTA S40pro supports both modes but MAP is considered optimal. (If using TPS you'd use the MAP sensor in open air for barometric correction).
Using the launch control feature does mean i need to run wires for a switch (enable launch mode) and telltale light (gear change point and, when flashing, launch mode engaged) from the ECU to the dashboard as well as the already mentioned fuel pump 12v and wheel speed sensor signals.
Sensor wise I need to wire TPS, MAP, air temperature, water (oil) temperature, and VR (crank) sensor. Outputs are the four injectors (on one of the ECU outputs), idle valve (on other other output), and coil (two outputs for wasted spark).
The relevance of all of this is planning the engine wiring loom. Knowing the location of each item, the electrical connector for it and its pinout, I drafted out and then made the loom. This took much much much longer than I expected - a good 20 hours over 1 1/2 weekends.... but I am very pleased with the result. The £250+ charged by DTA for a custom loom actually looks pretty reasonable value after all!!
I sourced most parts for the loom from Vehicle Wiring Products with odditys like the Colvern TPS pot connector coming from Jenvey directly, the GM map sensor connector (& sensor) from CC Engine Services, and the Zetec coil connector from Webcon.
This photo (above) also shows the newly installed fuel lines within the engine compartment (with AN fittings), a pair of plastic idle air tubes running in p-clips from the left bank towards the idle air manifold, and the air filter for the idle valve tucked up above the CSP breather box. Just to the right of the inlet manifold at the bottom of the picture you can see one of the feed lines for the brake servo exiting the compartment through a grommet. (The large hole to the left of the ECU is for an AN-8 oil feed line from a yet-to-be-fitted ACCUSUMP)
On the other side of the compartment you can see the other fuel lines (AN fittings again), idle air tubes, and brake servo feed (the two feeds tee after a NRV for each bank). The idle manifold (connected to the large blue mocal hose) is clearly visable but the idle control valve itself is tucked up about the breather box. The LC-1 wideband oxygen sensor controller hasn't been moved but it now wires to the ECU via its analog outputs as well as to the DL-32 datalogger and XD-1 gauge via RS-232. Finally you can see in the top right the new return style fuel regulator (high pressure) with a gauge attached.
Having shown the new fuel lines and fuel regulator, I should explain the overall fuel system, which the following diagram should do:-
So up at the front I only need to swap out the existing pump and filter for EFI equivalents, put the filter after the pump (it was before), and add the accumulator. The accumulator accepts the return line and acts as a small buffer of fuel if the outlet of the tank is ever uncovered by hard cornering or braking. It is fed by gravity from the main tank. I also need to add a second fuel line from tank to engine bay alongside the existing one for the fuel return - I'll use 3/8 aluminium tubing for the main run, with flares to connect to AN-6 bulkhead fittings.
The completed under tank plumbing is shown below. The only line missing is that which will connect the accumulator to the fuel tank. I had to move the fuel filter to the right slightly to make room for the larger AN-8 feed hose for the fuel pump required by the 12mm outlet on the fuel accumulator.
Back in the engine compartment, I temporarily hooked up the air and oil temperature sensors, and the TPS, removed all the fuses from the ECU fusebox except for the ECU main fuse itself, and switched on the ignition. Hey, presto! My laptop could talk to the ECU which switched on and off the fuel pump relay as expected (with the engine not started). It responded to throttle position changes, MAP changes (vacuum generated with my mityvac hand pump) once I fully pushed in the 5V power in the MAP plug, and temperature changes (body heat). I calibrated the ECU for the sensors and set the throttle stops. With that confidence gained, I decided to reassemble the cooling system atop the engine.
The relocated servo lines and idle air hoses meant I needed to trim the sealing tin infront of the engine when refitting. I also decided to relocate the CHT sensor cables to under the engine - this wasn't possible without removing the shroud so I'd routed them over the top when first fitting them (This made removing the shroud a pain). I also needed to make a suitable aperture in the sealing tin for the relocated throttle conduit (for the new linkage). All this done, the shroud could go in (I degreased it earlier), together with the oil cooler housing, throttle bodies and alternator.
You can see i'm still trying to work out what has been assembled back-to-front on the linkage and have also discovered I need to clearance the manifolds for the nut that will be fitted on the end of the throttle arm from the linkage. Damn. You'll also see i've trial fitted the air cleaner bases on the foreground chokes of the throttle bodies.
I chewed up a whole day this weekend tweeking the throttle linkage and throttle body lever arms. This involved removing and refitting the manifolds so I could shape the outside of the them to ensure clearance for the lever arm, adding additional holes in the throttle body lever arms (twice) to ensure that fully accelerator travel resulted in full throttle at the throttle body (I ended up tapping the holes M5 x 0.8 so that a nut didn't need to be placed behind the arm after all), trying two different mounting positions for a helper return spring for the linkage (as the return springs on the throttle bodies were insufficient to return the pedal as well), slotting the mounting holes on the centre bracket as the left hand linkage was slightly shorter than ideal, and redrilling the linkage arm mounting holes on the pivot the correct way around.
After all of this I have a nicely synchronised linkage, which returns well, and on which the travel of the pedal matches the travel of the throttle bodies.
You can also see in the photos above that I've installed the toothed wheel and pulley. Firstly I fitted just the toothed wheel and ran the crank oscilloscope in the DTA software to ensure it could be read.
So I fitted pulley, fan belt, and tinware
In the middle of the car (passenger compartment) I've installed the fuel return line and replaced the supply line as it had a gouge in it from the seat runner (I'd trust it with 3psi but not 40psi). I'll make a protective cover for these front and back as before so the passenger can't accidently damage them.
You can also see the wiring for the fuel pump and front wheel speed sensor (4 wires in all)
With all the fuel plumbing in front to back, I pressure tested the fuel system with a car tyre footpump. I got it to 40psi, found (heard) a small leak in the engine compartment where I hadn't tightened an AN fitting, but no other problems. So I installed the tank, put in some fuel, drew through some fuel with my vacuum pump at the regulator to ensure the pump was primed, and with a couple of fire extinguishers on standby ran the fuel pump (switched on via my laptop using the DTA software). No leaks, so I adjusted the pressure to 40psi at the regulator.
Time ran out this weekend, but i'm now ready to try and start the car. I have put in initial fuel, ignition, and target AFR maps. I generated the initial fuel map using the megasquirt VE table generator and translated for DTA (ms rather than % of req fuel).
One of the final jobs was to fit the coil and HT leads. I had a custom set made up by Magnecor to suit the late Zetec coil and the DTM shroud. Because the Zetec firing order is 1 3 4 2 and the Type IV firing order is 1 4 3 2 the leads for cylinders 3 and 4 need to be swapped on the coil.
On trying to fire up the engine it started more or less straight away. Hurrah! I was about a factor of 4 out (high) on fuel in the map so it started eye wateringly rich. A bit of fiddling around revealed that the analog output of the LC-1 wideband sensor was broken and that the vacuum at idle may be insufficient to use MAP as the primary load - I may need to switch to TPS and indeed i've yet to discover a Type I or Type IV with ITB (Independant Throttle Bodies) running on MAP.
Replacing the LC-1 with a new unit (Thanks to Graham Goode Racing for quick delivery and a healthy discount on the replacement) solved the analog output and will enable closed-loop tuning with the DTA ECU. Now that the engine is proven to run i'm focussing on bolting the car completely back together before I start trying to drive and tune it. I have checked the timing and linkaged balance however. So, the breather hoses are now connected to the air filters, the air filters are in place and oiled, all the tinware is back in place, the engine lid cover is back on, the air temperature sensor is mounted on the left throttle body, I've removed the distributor (and IMPORTANTLY it's driveshaft and thrust washer) and i'm in the middle of making protective GRP covers for the fuel lines in the passenger compartment.
I'm still waiting on a mount for the water sensor which i'm mounting in the oil cooler lines, The unit I ordered 6 weeks ago with AN-8 fittings was shipped on Friday so should be with me soon. I'll change the rest of the oil lines to AN-8 and fit an oil thermostat at the same time as fitting this.
The sensor housing finally arrived allowing me to fit the oil temperature sensor. I also took the opportunity to fit a oil thermostat with it in the lines to the oil cooler as the engine could take a long time to get up to operating temperature. The housing and thermostat's bought had AN-8 fittings so I also replaced the existing hose barbs with 3/8NTP to AN8 adaptors to allow the use of AN8 hose throughout.
Along with the replacement LC-1 I bought a TC-4 to allow me to log all 4 EGT temperatures alongside TPS, MAP, RPM, and (in the future) MPH. Installation was straightforward (in the battery box above the DL-32) but in setting up Logworks I discovered that two of the DL-32 channels were measuring 5 volts as 3.6 volts and 4.1 volts respectively. Why Why WHY do Innovate Motorsports have such problems with electrical reliability??? Fortunately I can offset these errors within logworks by taking a ratio of 5/3.6 or 5/4.
I initially tried to setup the fueling using MAP as load, but quickly realised that the engine was going to generate insufficient vacuum (only down to 70kPA) especially when the idle valve was part open. Plus everyone I spoke to with a fuel injected beetle was using TPS and DTA themselves recommended TPS with MAP compensation. So, I switced to this approach.
I also discovered that the Bosch idle valve flowed a great deal of air, even when "closed" (at ~40% PWM). So I switched to an idle valve from a Ford Zetec which fully closes and appears more controllable. I'm not yet convinced that it flows quite enough air for a cold engine - but this may be my cold start mix being too rich and itself keeping the rpm down as experiments with leaning it out cause the RPM to pick up.
Note you can tap the holes on the Ford Idle valve with a 3/8 NPT tap WITHOUT needing to drill (The valve was designed to bolt down onto a flat casting). A 3/8" NPT to 1/2" hose barbs then screwed straight in.
After alot of time tinkering in the garage mapping around idle and blipping the throttle I finally ventured up and down the drive, then round the estate, and then convinced my better half to drive the car (for her first time!) whilst I sat with the laptop in the passenger seat tweaking.
The LC-1 has been an essential aid in getting the fueling right, and a bout 150 miles later, and after a trip to the Hampshire Pageant of Motoring, the car is already running better than on carbs. No more black exhaust tip, better cold starting, and leaner cruising. Result!
So, A final picture for this section of the completed conversion.
P.S. Don't forget the 5 part technical article also describing this in the technical article section