The problem is that the car has no choke, and there is no idle until the engine has warmed up for a few minutes.
It hesitates when cruising at 1500 - 1800 rpm and I give it some more gas, even when fully warmed up.
People behind me gets some serious cough attacks if I step on it and let it redline.
It seems to me that the carburettor is far from optimally tuned.
Now I haven't had a car with a carburettor in 15 years. Most of my cars the last 20 years have had fuel injection, just to avoid all the negative issues with carburettors.
Looking into the headers I see that some cylinders runs richer than others:
Cylinders 2, 3 and 7 seems to run richer than the rest. Checking the plugs also confirms that at least 3 and 7 runs rich.
This injection system looked beautiful on picture. It had a Kinsler intake and throttle bodies plus a Hastec control system.
I read the article with great interest as I was very tempted to get rid of the Holley instead of trying to tune it.
The article talked about a lot of different sensors and what they do, and then they said that this particular installation only used four inputs: Air Temp, Coolant Temp, Throttle angle and Crank RPM. What? Is that all?
I sent the engine builder a mail asking if this was correct. After several months (!) I got an answer that this really is the case. He said that Sweden really is so flat that this doesn't matter. Thanks, but no thank...
I sent a question on a mailing list asking how a system like that would
compensate for variations in air density, and someone pointed me to the
eec mailing list. I got in contact with Peter Bysterveld in Australia,
and he has retrofitted an EFI system from a Mustang to his Australian
Ford Falcon UTE, with a 351C.
Everyone agreed that the system most suitable is the EEC-IV system in 5.0 liter Mustangs, 89 - 93. These all have MAF (Mass Air Flow) sensing, which easily adjust to changes in cams and other things, which the non-MAF doesn't. It is also a SEFI, sequential firing EFI, which is uncommon in aftermarket EFIs.
More EEC info available here.
A good book I found in the Ford Motorsport catalog is the "Ford Fuel Injection & Electronic Engine Control : All Ford/Lincoln-Mercury Cars and Light Trucks 1988 to 1993" by Charles O. Probst. It contains a lot of good background info and also wiring diagrams for all the interesting donor cars. I have seen this book occasionally on Ebay.
According to the author, Ford used the EEC in the Formula 1 cars! If it's good enough for F1, it's good enough for me.
I was recommended to get the real wiring diagrams from Helm Publications. I ordered the diagrams for a 91 Mustang and it is money well spent.
But where to get a system?
Mustangs are not that common in Sweden, so another source had to be found.
I tried contacting a few wreckers I found on the web, but the best source I found was Ebay.
Searching for "EFI" and "fuel injection" for a few weeks, I finally found a system that was described as "complete", with the exception of a few sensors. The seller was also prepared to ship world wide.
I placed a bid, no one outbid me and after a few days I was the owner
of a early 90:s Mustang EFI system. The seller sent the package with FedEx
on a Friday, and I had it in my hands on Tuesday ;-)
When I tried to identify and mark all the wires and make sure that everything fits, I noticed that the wiring harness wasn't from a MAF car, although the EEC module was for a MAF system and a MAF sensor was included. I tried getting the MAP to MAF harness from Ford Motorsport, but they said it isn't sold separately. Luckily they are frequently auctioned on Ebay.
This is the conversion kit I bought on an auction.
Peter Bysterveld used a standard single plane carburettor intake for his conversion. He created a plenum that sits on the intake instead of the carburettor. The throttle body is then attached to this.
Peter recommended me to start with a good EFI intake for a 351 Windsor and modify it to match the Cleveland.
The deck height of the Windsor is 0.3", or 7.5mm, higher than that of a 351C. A 351W intake would have to be narrowed a few mm:s to match the height of the 351C. The Windsor also has water passages in the intake, which the Cleveland lack. The thermostat house in front of the 351W intake would have to be removed as this is where the thermostat sits in the 351C also. The rear horizontal part of the block is different and would have to be modified.
Peter recommended the Trick-Flow intake. I went with this, and here's a comparison of the lower intake with a 2V 351C turkey pan:
This picture shows that the rear (left on picture) of the intake will have to be modified to fit the 351C. The thermostat housing has to be cut off as it obstructs the Clevelands thermostat house.
The good thing is that the important stuff like ports and most of the bolt holes seems to be reasonably in the right place.
The not so good thing is that all the bolts on the Windsors are vertical, while on the 351C only the mid four bolts are vertical.
The TrickFlow 5.8L EFI intake has a 50 x 30 mm port size at the head, and this is reasonably close to the 2V port size. I measured the ports to 50 x 34 mm in the head and the turkey pan to 55 x 39 mm for the 2V and 66 x 47 mm for the 4V. The standard 302 port size is 42 x 22 mm.
This is how the lower intake fitted the engine before modifying it. The ports matches the heads quite well. The ports are bit rounder in the heads, but I don't think this will be a big issue.
I found a good shop in Uppsala, Ells AB, that welds light alloys and has done heavy intake modifications in the past.
This is how it looks when finally bolted on:
The bolts are of slightly different length and the only strong UNC bolts I could find were Allen screws.
The top of the lower intake has been milled down 6 mm to add some room between the top of the upper intake and the engine cover.
There's not much room for the wide distributor cap.
This picture shows the ACT (Air Charge Temp sensor) between #5 and #6 injectors. It senses the temp of the air in the #5 runner.
This is how it looks with all parts mounted and cables and hoses connected:
The engine cover actually fits, but it is really tight.
I had to buy a longer piece of silicone hose to reach the BBK pipe. Unfortunately the shop were out of blue 75mm silicone hose, which would have matched the blue shorter pieces included in the BBK kit.
View from the wheel well:
The BBK kit included a mounting bracket that fits nicely against the
The best option is to get a distributor from an EFI 460. This drops right in.
The second best option is to use the distributor from an EFI 351W and change the gear to a 351C gear.
I have now located a couple of sources for 460/7.5L distributors. I entered 88 - 95, "Ford pickup", and "Distributor". I got about ten pages of distributors and three or four sources that had 7.5L EFI dists.
I ended up buying the dist from Harvey Gunderson for $100 plus shipping.
When the dist arrived, I found that it wouldn't fit the rest of the EFI directly as It used a different connector and used an external TFI module.
The one on the left is the 460 distributor and the right one is the 5.0 Mustang dist, but with the shafts and gear swapped.
Before committing to this setup I tried both houses in the block and I dound that the 5.0 house was about 1.5mm to narrow, ick!
I realized I had to use the 460 house, time to compare:
The major difference is that the 5.0 house has a hole through which the TFI module connects to the interior of the dist.
This is what the 460 dist looked like after half an hour with the power drill and a file:
The TFI module needs silicone grease for proper thermo coupling. Be
liberal, you don't want the module to fry.
The finished distributor
I have used Polyamide fuel lines on the high pressure side, which I believe is the same as the factory uses. The fittings are easy to mount, drop a bit of lubricant (I use Sonax MoS2 spray) into the line, insert the fitting and give it a whack with a hammer. I have also used Polyamide lines for the vacuum lines and the line between the tank and the carbon canister. I like the "stock" appearance of this.
I manufactured a mounting bracket out of aluminum sheet metal and used four small rubber dampers to isolate the chassis from pump noise.
The pump is mounted in two standard can type ignition coil brackets and I drilled holes in the ears for the rubber dampers.
The pump is a Bosch 0 580 254 957 unit, rated at 128l per hour @ 5 bar, which theoretically is good for a 90% duty cycle with 30lb/h injectors. A better pump is the Bosch 0 580 254 044, which is speced at 172l/h @ 7 bar.
The bracket is mounted with four M5 screws to the cross member.
Apparently earlier cars used some M22x1.5 drain plug. YMMV.
Peter Bysterveld used a T-fitting on one of the heater outlets. This produced some comments on the EEC list about the sensor sitting in a pocket of non-circulating coolant, which would make it read the wrong temp.
The tube running from the thermostat to the pressure tank has a 3/8" NPT fitting just a bit from the thermostat. I have used that for the ECT as it will simplify the wiring and it will not sit in a pocket of cold water. The only problem I see with this is that the water circulates slowly before the thermostat opens. There is a little hole in the thermostat that allows some water to pass, this will hopefully be enough. The thermostat will open in a few minutes anyway. The ECT can be seen facing backwards in the picture above.
The picture above also shows the chromed fuel return line in front of the distributor. It reaches to within 10cm of the fuel supply pipe in the tank sender. I use the old supply pipe for the return.
The EEC actually fits between the trim and the firewall.
The rubber boot above the EEC is from a Euro Escort, genuine Ford ;-)
I fed the wires one by one through the hole after inserting a short piece of plastic tube, 30mm dia, into the boot. When all the wires were through I pulled the tube out and cut it to pieces to remove it.
I used an 83mm hole saw to cut through the firewall.
I taped the wires here and there to keep them together and dressed them up with new sleeves:
The end result:
I have also used the A/C-clutch control of the EEC. I disconnected the
wire to the A/C compressor clutch and connected it to the EEC and then
connected the corresponding output to the clutch. The EEC now compensates
for compressor drag by increasing the idle slightly. It will also disengage
the compressor briefly at WOT, when cranking and I believe it will also
disengage the compressor at high engine RPM:s to protect the compressor.
The standard ball on the pedal was too large for the Lokar ball end, I replaced it with the Lokar ball, which fit directly.
Did you know that the pedal has a needle bearing? Serious stuff!
I have been having problems where the engine frequently dies coasting to a stop at a red-light or just stopping hard.
This is apparently a well known problem and the fix is to install a Vehicle Speed Sensor (VSS), which makes the EEC increase the idle RPM:s a bit when the car is moving.
In a Mustang this signal is taken from a Speed Control Amplifier, which creates the differential 8000 pulses per mile signal the EEC wants.
I found a VDO Hall sensor, X39 397 106 191, that fit directly between the angle drive for the speedo cable and the cable:
Using three 10k resistors I interfaced this to the EEC. I used one resistor as a pull-up between VREF (+5V) and the VSS+ and sensor output. The other two resistors were connected in series between VREF and GND and VSS- to the EEC was connected between the resistors to provide 2.5V. This was apparently acceptable to the VSS as I could see that the idle RPM:s increased when I ran the sensor with a cordless drill.
Sensor power was taken from VPWR (+12V) and GND.
The VDO catalog lists the sensor as having 8-pulses per rev, but text on the plastic body seems to indicate that it is a 6-pulse per rev sensor. This turns out to be 7260 pulses per mile nominally, which is close enough for this application.
I went over some of the electrical stuff one last time before attaching the EEC, which probably was a good thing as I noticed that I had forgot to connect the HEGO ground wire.
Connected the battery, carefully, listening for suspect pops and crackles,
but nothing alarming happened. Verified that all supply voltages were in
some of the sensors ACT and ECT had reasonable resistance for the temp and that the relay I connected for the HEGO heaters worked.
I disconnected the battery and connected the EEC to the harness. Once again carefully connected the battery, still no smoke ;-)
I pulled the fuse for the fuel pump and switched the ignition on. Got
a yellow light (I'm using the "heater" lamp as the check engine lamp),
the fuel pump
relay klicked once, and then again about a second later. Looking good!
Read out the error codes from the EEC:
31 (EVP below minimum voltage) I have no EGR, ignore
67 (neutral switch open) Not connected, ignore
81 (secondary air injection diverter) not connected, ignore
82 (secondary air injection bypass) not connected, ignore
85 (canister purge) not connected, ignore. I will eventually connect this.
84 (EVR) Still no EGR, ignore
95 (fuel pump failure) likely due to the removed fuse.
Nothing alarming here.
I filled the cooling system, and noticed I had a leak around the thermostat
:-( Drain, remove the tube, clean, new gasket and some Permatex. Filled
system again, now tight ;-)
Connected the timing light, switched the ignition on, turned the key
to start, motor turns over, no start. Checks the timing light, doesn't
flash. Smell of
Verifies the signals at the TFI (distributor), all OK. Verifies the
signals at the coil, all OK. What? I bought a new coil yesterday! Pulled
out the old coil
and used some lab cables to connect it. Removed the fuel pump fuse to avoid flooding the engine. Turned to start, varoom, its alive! Dies when it runs out
I see now that one of the spade connectors in the new coil is bent and
doesn't make contact with the connector. Fixes this, inserts the fuel pump
starts it up. Starts immediately :-) and dies... New start, stretches for the throttle and give it some gas to keep it running.
Warms up enough to idle properly. Removes the SPOUT plug and adjusts base timing to 10 degrees. Tighten the dist and insert the plug again. I now see a small leak in the water neck where the pipe is welded to the neck :-(. I also see a small leak between the left head and the block at the rear, arghh. Laziness is punished immediately. I pulled the right head as I saw that there had been a leak but didn't pull the left head. Oh well, I guess that will be enough work to do while I wait for the throttle cable.
Three small mpegs that proves it is alive:
Peter Bysterveld told me to ground the NDS wire to get better cold idle, which I have done, but I didn't notice any improvement in idle. Given the problem below I feel that this problem might have had other causes.
I have now pulled the head and replaced the head gasket. I saw that cylinder #7 definitely had had a problem:
I couldn't see any problem with the head or the cylinder. I did notice that the head bolts were pretty loose and I hope that it was just a leaking gasket or a problem with the old Holley setup. I have been losing coolant previously and this might actually be the reason it had cold idle problems. When cold it probably didn't run on all cylinders and after a few minutes it started running on all eight and all the idle problems were gone. I will have to read the plugs now and then to verify that the problem is gone.
Mounting the engine cover:
There is no problem with the height of the cover. The problem is instead the large distributor cap. The front part of the cover goes between the distributor cap and the firewall and there wasn't much space there to begin with. I rotated the distributor one tooth as there was one cable post dead north and rotating the dist the 24 degrees one tooth makes helped quite a lot.
2001-04-02 I took it for the first test drive.
The trip was uneventful and everything worked according to plan ;-)
There was no sign of the bucking between 1500 - 2000 rpm it used to
have with the Holley. Just smooth acceleration.
I have been using the car during the summer, and while the EFI seems to work fine, I have been having some performance problems.
I have taken compression tests, without finding out why the #7 cylinder fouls the plug.
Finally I built a leakdown tester and did a leakdown test on the left bank, and I definitely need to rebuild the engine. The leakdown was between 14% and 32%. The odd thing is that #7 was the cylinder leaking 14%. My engine builder said this was probably due to excessive amounts of oil in the cylinder and that I should have run the test longer.
The performance has been noticeable less than last year, even though it runs smoother. I have attributed this to the narrow throttle body, MAF and small injectors.
While shopping for performance parts I felt that I really should check the performance using datalogging. The problem was that the A9P EEC-IV I used wasn't supported with datalogging, only the A9L family was. I therefore bought and X3Z EEC-IV, which was supported.
I had expected the MAF to be pegged at maximum and the injectors being wide open, but this was not so. The datalogging showed that the motor sucked at most 750kg/h of air, and the MAF maxes out at 835kg/h. The injectors were open around 20ms at 5000rpm, which really is wide open, but the oxygen sensor showed that it was running rich, when I had expected lean.
I bought a 75mm Pro-M plastic Bullet MAF calibrated for 30lb/h injectors and it handles up to 1600kg/h of air.
The noticeable difference with this was that the engine started and ran much better when cold, and that it felt like it was running lean during wide open throttle runs.
I bought a used set of 30lb/h injectors. Later I found out that they were actually 19lb/h injectors, so I couldn't use them.
I did a vacuum test, and it was very low and irregular and I thought that perhaps the lifters or the cam was bad. I replaced the lifters, which didn't make any difference.
At this time people had actually developed a datalogging payload that worked with the A9P EEC-IV, and as they needed testers, I switched out the X3Z and put the old one in again. The data logging worked fine, but it was still only one bank that was logged.
Later someone had developed a new logging payload that would log both banks and I switched to this, in a hope that it was a problem with the bank that hadn't been logged up until now. This did not show any problem,
I upgraded the throttle body to a 70mm MAC, and once again I can't say I noticed any major difference, other than good looks.
One night while hanging over the logs from a run, I just happened to notice that I had a maximum of 26 degrees of ignition advance during WOT acceleration, while I had around 38 - 40 during part throttle cruise.
I have heard several Pantera owner say that the Cleveland really likes timing, lots of it. I asked and most people agreed that I need 30+ degrees. I set it to 34 degrees from 2600rpm and went on a test drive. I immediately knew that this was the problem I had had. It felt at least as strong as it did last year.
I'm now running with 36 degrees max, and I have just started to tune
the MAF-transfer function of the Bullet. I have tilted the curve so that
it better matches the stock MAF in the upper regions, and it seems to
run really well. I will work a bit on the lower parts to optimize idle
and low speed operation.
After the winters rebuild I have been on a trip to France with the car.
I had a lot of problems with a bad connection between the EEC-tuner and the EEC, but opening up the box and wiggling the connector fixed that.
The higher compression 302C heads together with flat top pistons gives a compression ratio of around 10.6:1. This has proved to require retarding the ignition from last years 36 degrees at WOT. 24 - 26 degrees seem to work fine. I have also had to lower the part throttle timing from 40 degrees down to around 30 as I had audible high speed pinging.
The 75mm plastic Pro-M Bullet MAF seemed to have a weird transfer function. Logging a few WOT runs showed that I was running extremely lean, and I could feel that I was lacking power. I richened the curve up about 37% (!) and it was still running lean.
I replaced the Bullet with the stock Ford MAF and I immediately felt that it had more power. I logged 850kg/h worth of air, more than 100kg/h more than I ever saw with the Bullet.
I found a Ford paper describing all the emission stuff, and for the MAF it said that MAF-contamination shows as rich idle and lean WOT. I figured the sensor was destroyed and I was about to ship the MAF to Pro-M for recalibration when I found a document on the web describing how to check if the sensor is contaminated and how to clean it.
Seeing that the hot-wires were whitish on the intake side and dull grayish on the filter side, I removed the sensor from the MAF body and very gently cleaned the wires with some cotton swabs.
After reinstalling the Bullet and configuring it with a stock 30# transfer function it was running stronger than ever ;-)
Six years since I upgraded the web page... I did ceramic coat the headers in an attempt at reducing underhood temp. Not sure if it did much difference, but they are prettier now :-) In 2004, during our trip to Austria and the factory in Italy, the engine decided to eat a cam lobe on our way home. In 2005 I had a loaner 351C in the car, and in the 2005-2006 winter I built a new engine, this time with a hydraulic roller cam.
I have installed a wide-band lambda sensor, and interfaced it to the EFI, so I can compare the actual AFR now with what the EEC commands.
In the spring of 2008 I replaced the fuel pump with a high performance Bosch 044 pump (mentioned above). I also replaced the stock fuel rails with the larger rails from the later GT:s, and I also reamed out the fittings to get better flow to the drivers side rail.
I finally got rid of the K&N filter, which was constantly fouling the MAF, and replaced it with an airbox and MAF from a super charged Lightning truck, which uses a proper paper filter. I have mounted the airbox such that it draws air from below the car, and the air temp should be lower than it was before.
I have now upgraded to a TwEECer RT. The logging features are much better in the TwEECer, and being able to switch between four different tunes, besides the stock, is a real winner.
I did an experiment with E85 this summer, and I had four different tunes, one for plain gas, one that adds 10% fuel, one that adds 20% and one that adds 30% fuel. E85 has a lower energy content and the engine needs more fuel for the same amount of air.
During closed loop cruise it seemed I could chose any of the settings, the EFI would quickly adjust. I checked the lambda reading on the display I have on the dash at WOT, which runs in open loop, and I selected a switch setting that would give the best AFR for the given gas/E85 mix.
Unfortunately I didn't get very far on 100% E85 before it blew a head gasket. I doubt that the fuel was the real problem though. The gasket was probably going soon anyway.
It also has improved the mileage. On shorter drives the mileage is around
20% better than before and on a longer run to meeting, the improvement was
around 35% or so.
I have now eliminated the distributor and use Ford EDIS ignition. The EEC
has code for EDIS and it only required changing three variables in the tune.
HP_CID=1 (CID sensor present)
HP_CIDSEL=1 (CID sensor type, 0=hall, 1=VR)
HP_HIDRES=1 (High Data Rate Electronic Spark Present)
The 36-1 timing wheel and sensor:
I have adapted a cam sensor from a 3.8L V6, which have the same diameter oilpump shaft. The block hole is slightly narrower, and the sensor sits slightly higher, so it was easy to make an adapter:
The EDIS-8 module and coil packs:
The plug wires are a Taylor 72623 wire set for a 1998-2001 5.0 Ford Explorer. These have normal plug boots and EDIS coil boots.
I've used schematics for a 4.6L 1992-1993 Crown Victora from the Probst book to connect
the CAMP and EDIS module to the EEC-IV. I've also used the
MegaSquirt documentation, which describes the system nicely. Unfortunately the coil wires are swapped
on the EDIS-8 wiring diagrams so that the cylinder id numbers on the coils will be incorrect.
I swapped wires 8/9 and 11/12 in the EDIS connector to correct this.
If you have comments or suggestions, email me at thomas@Hax.SE
Last update: 2009-05-26