1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

macaddict
MegaSquirt Newbie
Posts: 32
Joined: Tue Feb 17, 2009 3:42 pm
Location: Burlingame, CA

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by macaddict »

Ah, interesting! I've thought about building a VR sensor mount to go in place of the mechanical fan mount (similar situation with unused bolt holes) but that's quite a ways away from the crank. The factory mount would be easy to use if I installed the newer front cover, but it requires you to have the same odd design of trigger wheel where the fingers are bent over the damper assembly due to the placement of the VR sensor (forward about 1 cm or so from the body of the trigger wheel itself). I could design a pattern for whatever crank signal MS3 wants to see and get it cut, and then use the spare Marelli damper to hammer the fingers over. Or do what Philip suggested and use a Ford 3.0L EDIS trigger wheel and relocate the sensor. I can probably put the front cover back on the V12 currently on the stand to get some ideas, I haven't really spent a lot of time looking at it with all the other projects, so something might jump out at me.

I'm sorry to say I haven't watched any of the Megasquirt meet presentations, but I'll dig around for them online and check them out. I'd be up for starting a V12 development thread and signing up to test sequential injection on a MS3, do you mean on the MS3 development forum on msextra.com? One possible issue is that there aren't enough outputs on the MS3X card to fire more than 8 injectors and coils, but that could be overcome by using CAN to talk to a remote high current driver board. Or just using spare CPU outputs and designing a custom board which would probably be more compatible with the existing MS3 software. One of my shopmates is an engineer with NASA and so we have most of the board layout and fabrication equipment at the shop.

I'll upload the latest fuel-only calibration I have when I get into the office tomorrow, I've made some tweaks recently and have it running more or less as good as it's going to get without spark. Performance is good even when cold, although I'm on the lean side for warmup and accel enrichment since my smog pump isn't working. Let me know if you're feeling less lazy than I've been and start to play with spark control and I'll get off my rear and start working on that again. I don't drive the car much during the winter so it won't hurt to have it apart for a while.

The bungs on the fuel rail I have (presumably the same late XJ12 rail you have) are similar to those found on a GM TPI fuel rail, so something like this should work: http://www.summitracing.com/parts/SUM-220009B/. I wasn't able to get the return side to seal with an o-ring, so I ended up using a Parker hydraulic fitting with a square 14 x 1.5 mm fitting on one side and a male AN fitting on the other. I polished the face of the return bung on the rail and used a high-pressure metal/rubber sealing washer, and made my own lines using 5/16" and 1/4" thick-wall stainless tubing. The supply line makes a 180 degree turn and heads in front of the fuel rail over the A/C compressor (not very much room due to hood clearance) and the return line goes over the top of the fuel rail and heads straight for the fuel cooler. Here's an older picture: http://crosstalk.atomz.com/~ross/jagstu ... 0-1956.jpg with the original routing of the feed line, it's sense been re-done with a slightly lower and more form-fitting routing due to clearance issues but the path is more or less the same.

The injectors are 17 lb/hr at 3 bar, slightly less than the stock 5.3L HE injectors (17.2 lb/hr). They fit in the mounts about as well as the original 17 lb/hr or 18/lb hr Bosch injectors did except for the two front injectors which wouldn't seal using the original square o-rings. I used round o-rings for those injectors which were a bit of a pain to seat, but it does seal.
1984 Jaguar XJ-S HE (Rosie III) - Megasquirt-II v3.57 fuel+spark
http://crosstalk.splunk.com/~rdykes/jagstuff
StevenD57
MegaSquirt Newbie
Posts: 30
Joined: Fri Oct 14, 2005 12:14 pm
Location: Phoenix, AZ
Contact:

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by StevenD57 »

There already exists a I/O expander board. See the following link:

http://jbperf.com/io_extender/index.html

There is also a 4 channel ignition/injection driver board:

http://jbperf.com/quad_ign_inj/index.html

There are 5 - 6 different custom MS add-on boards there at that site. Don't spend time re-inventing the wheel. Re-use existing pieces.

There is a Jag V12 forum discussion over in the New Users forum at msextra.com

http://www.msextra.com/forums/viewtopic ... 94&t=37614

There was some discussion of using the CAN bus but I was told it was not real time enough to do coil drivers. That was where the suggestion of using the I/O expander board came from.
Jaguar E-Types, modified Merkur XR4Ti, Jensen-Healey
Linux = no virii or spyware, EVER
macaddict
MegaSquirt Newbie
Posts: 32
Joined: Tue Feb 17, 2009 3:42 pm
Location: Burlingame, CA

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by macaddict »

Ah cool, I'll take a look, i haven't looked at those specific boards before. The latest calibration has been uploaded to crosstalk if you want to get started with that.

- R
1984 Jaguar XJ-S HE (Rosie III) - Megasquirt-II v3.57 fuel+spark
http://crosstalk.splunk.com/~rdykes/jagstuff
macaddict
MegaSquirt Newbie
Posts: 32
Joined: Tue Feb 17, 2009 3:42 pm
Location: Burlingame, CA

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by macaddict »

The Jag started and ran for the first time with ms2-controlled ignition this evening, right now it's a bit of a proof-of-concept hack with the distributor mechanical advance still in place but at least it's progress! I used the MS-II I was previously using for fuel-only with the inverted VR-sensor input enabled and the coil driver transistor removed, and added the ignition coil driver module from a 1994 Chevrolet Corvette. This appears to be the same driver module found on all SCPI-equipped GM trucks as well, so something from a 4.3L S-10 or 4.3L/5.0L/5.7L Silverado should work too.

I started by cutting back and replacing the wiring for both the full load signal (originally connected to a mircoswitch on the throttle) and the vacuum advance cut-out (bypass the vacuum regulator in between manifold vacuum and the vacuum advance can for 45 seconds on start-up) as it had become brittle with age and I hadn't taken care of it when I did the rest of the harness. This allowed me to run the following wires over to the coil and ignition amplifier:

- +12v from fuel pump relay
- Full load signal wire
- Signal ground

I disconnected the distributor VR sensor from the original ignition amplifier and connected the positive lead (the one that's above ground when the trailing edge of the trigger wheel tooth passes the sensor) to ground and the negative lead to the ignition trigger signal wire, this will trigger the ECU on the negative crossing to help reduce the effects of noise (the original amplifier triggers on the positive crossing, hence the reverse polarity.)

The full load signal wire is connected to the ignition output line on the MS-II, I didn't use the high-current driver on the board (CPU I/O pin connected straight to the output pin on the DC37 connector) as the LT1 ignition module is driven with a TTL signal like the 7/8-pin HEI module.

The LT1 ignition module is wired in the following manner:

- Pin A: +12v switched (connected to +12v from fuel pump relay)
- Pin B: EST signal (full load signal wire, pull high to charge and low to fire)
- Pin C: Ground (connected to the heatsink internally)
- Pin D: Coil output (often labeled "TACH" on GM wiring diagrams)

The connector for this module is the same as the one found on most 1991+ GM IAC motors and the 8-pin HEI modules. The one I used was made by Pico and cost $6.

Since I used the inverted VR sensor circuit output and no MS-II coil driver, the basic ignition settings are:

- Ignition capture on falling edge (the VR sensor input inverts the signal, but the inverted output changes it back to match the polarity of the input signal)
- Spark on rising edge (inverted output as the code expects there to be a driver on the board, this causes the output to go low)
- 2.5ms dwell
- 1.5ms spark duration

I have not adjusted the dwell yet as I haven't properly wired and mounted everything, but I plan on using a scope to set that when I clean up the wiring and mount the LT1 ignition module this weekend. The ignition map only includes a minimal amount of advance based on manifold pressure as I haven't locked out the mechanical advance yet, I'll swap out the distributor for the locked-out version after things appear to be working properly. I may also use the stock amplifier to condition the VR sensor output and switch back to the opto-isolator input on the MS-II as the VR input is very sensitive to noise (I have not adjusted the threshhold or hysteresis yet) and it would keep the wiring nearly stock looking in case the emissions inspector decides to do more than glance nervously under the hood and run away screaming.

Right now it's still possible to switch back to the stock Lucas 6CU ECU by swapping the IAT sensor reference back to +5v (I have a jumper harness near the firewall for that), and plugging the trigger input, VR sensor and coil back into the stock ignition amplifier. After I lock the mechanical advance out the original ECU will only be suitable for limp-home mode, so I'll order a spare MS-II and keep that in the trunk.

Here's a link to the msq file:

http://crosstalk.atomz.com/~ross/jagstu ... 171503.msq

And the documentation of the install (currently only a few poor quality photos, will post more soon):

http://crosstalk.atomz.com/~ross/jagstu ... n_install/


Cheers!

- Ross
1984 Jaguar XJ-S HE (Rosie III) - Megasquirt-II v3.57 fuel+spark
http://crosstalk.splunk.com/~rdykes/jagstuff
StevenD57
MegaSquirt Newbie
Posts: 30
Joined: Fri Oct 14, 2005 12:14 pm
Location: Phoenix, AZ
Contact:

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by StevenD57 »

macaddict wrote:Ok, so the formatting didn't quite survive the posting process, here it is again:

Function | Megasquirt-II | Lucas 6CU/16CU
Ground | 7,8,9,10,11,19 | 1,16,17,19,23,34,35
+12V | 28 | 22
Tach | 24 | 18
Spark Out | 36 | 3 (Full load switch)
CTS | 21 | 5
ATS | 20 | 21 (make sure to change the reference for the Jag's ATS to ground, by default it's TPS Vref)
TPS | 22 | 7
TPS Vref | 26 | 20
Inj A | 32,33 | 13,14,31,32
Inj B | 34,35 | 8,9,27,28
Fuel Pump | 37 | 15
O2 | 23 | 6
OK, I have de-soldered the connector off of an old 6CU ECU board and I am about to start making up that little harness inside the ECU case to go from the back of the ECU connector to the terminal block inside the case.

Did much of the wiring inside the Lucas ECU change when you went to controlling the ignition with the Megasquirt?

Also did you notice any improvement or noticeable difference in the way the car ran after you made your mods for the ignition?
Jaguar E-Types, modified Merkur XR4Ti, Jensen-Healey
Linux = no virii or spyware, EVER
macaddict
MegaSquirt Newbie
Posts: 32
Joined: Tue Feb 17, 2009 3:42 pm
Location: Burlingame, CA

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by macaddict »

The wiring didn't change much, the only notable changes from stock on mine are the following:

- All grounds within the Lucas ECU shell are tied together to a stud mounted to the case itself.

- Only the injector "ON" lines are used for injector control (13, 14, 31, 32 for A-bank and 8, 9, 27, 28 for B-bank).

- The injector power resistors have been removed and the "ON" and "HOLD" injector lines for each bank have been tied together in the engine compartment (I switched to saturated coil injectors).

- The reference for the air temperature sensor has been changed to ground in the engine compartment (I added a jumper lead to the main harness so I can switch between the two reference sources).

-The full load signal wire (Lucas ECU pin 3) is connected to the ignition output on the MS-II.

- The engine speed signal line (Lucas ECU pin 18) is connected to the VR input on the MS-II (the negative side of the VR sensor is connected to this wire under the hood) if using the ECU to control ignition, and to the opto-isolator input if running fuel-only. The VR bias and hysteresis potentiometers are each turned 1.5 turns clockwise to help with noise rejection.

- Pins 2, 4, 11, 12, 24, 25, 26, 29, 30 and 33 on the Lucas ECU connector are unused (pin 24 is the B-bank 02 which does go to pad JS5 on the MS-II, but I'm not currently running dual VE tables so it's ignored).

A picture of the modified connector can be found here: http://crosstalk.atomz.com/~ross/jagstu ... 066320.jpg and here: http://crosstalk.atomz.com/~ross/jagstu ... 999857.jpg

There has been a noticeable improvement since switching to partial ECU control of the ignition. I'm still running the mechanical advance, but the ECU handles temperature and load advance. Here's are the base ignition settings (from http://crosstalk.atomz.com/~ross/jagstu ... 170936.msq):

- Trigger offset: 0
- Skip pulses: 18
- Predictor: alpha-beta-gamma 90, 80, 10
- Time mask: 0.6ms
- Percentage mask: 40%
- Next-pulse tolerance cranking, after-start, normal: 35, 35, 30
- Check tach sync: always
- Ignition input capture: rising
- Cranking trigger: return
- Coil charging: standard
- Spark output: high
- Base dwell: 1.9ms
- Max spark duration: 0.2ms
- Accel dwell compensation: 0.4ms

The distributor base timing was set to match the underhood label and the timing map was determined experimentally by listening for pinging during test runs. Interestingly enough the mechanical advance at max RPM is just about dead on, even a 1 degree addition to the timing map for 90 kPa and above results in some pinging when running 87 unleaded. The average fuel economy with ECU ignition control over the past 300 miles has been 16.5 MPG according to the trip computer (which reads approximately 10% low), whereas it's never broken 15 MPG on that drive cycle before (round-trip between Burlingame and San Francisco, CA)

I'm going to stick with this configuration until the weather improves so I don't change too many things at once, there are plenty of other repairs that need to be made that will keep me busy in the meantime.
1984 Jaguar XJ-S HE (Rosie III) - Megasquirt-II v3.57 fuel+spark
http://crosstalk.splunk.com/~rdykes/jagstuff
StevenD57
MegaSquirt Newbie
Posts: 30
Joined: Fri Oct 14, 2005 12:14 pm
Location: Phoenix, AZ
Contact:

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by StevenD57 »

Well, I have the grounds and 12V power wire soldered to the back of the connector that I was able to gently remove from the old 6CU control unit I was able to source from a local j-yard so I am on my way.

Thanks for the information. Are you eventually going to lock out the mechanical advance as well?
Jaguar E-Types, modified Merkur XR4Ti, Jensen-Healey
Linux = no virii or spyware, EVER
StevenD57
MegaSquirt Newbie
Posts: 30
Joined: Fri Oct 14, 2005 12:14 pm
Location: Phoenix, AZ
Contact:

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by StevenD57 »

macaddict wrote:- Only the injector "ON" lines are used for injector control (13, 14, 31, 32 for A-bank and 8, 9, 27, 28 for B-bank).

- The injector power resistors have been removed and the "ON" and "HOLD" injector lines for each bank have been tied together in the engine compartment (I switched to saturated coil injectors).
What is the purpose of why the ON and HOLD lines are tied together?
-The full load signal wire (Lucas ECU pin 3) is connected to the ignition output on the MS-II.
Why is the full load signal wire connected to the ignition output on the MS ecu?
- The engine speed signal line (Lucas ECU pin 18) is connected to the VR input on the MS-II (the negative side of the VR sensor is connected to this wire under the hood) if using the ECU to control ignition, and to the opto-isolator input if running fuel-only. The VR bias and hysteresis potentiometers are each turned 1.5 turns clockwise to help with noise rejection.

- Pins 2, 4, 11, 12, 24, 25, 26, 29, 30 and 33 on the Lucas ECU connector are unused (pin 24 is the B-bank 02 which does go to pad JS5 on the MS-II, but I'm not currently running dual VE tables so it's ignored).
Yes, I think pin 10 is also not used if I am understanding your directions correctly. What do these "O2 Sensor Integrate" connections on pins 4 and 10 do anyway?
Yes, I have3 seen those pictures. Thanks for sharing.
There has been a noticeable improvement since switching to partial ECU control of the ignition. I'm still running the mechanical advance, but the ECU handles temperature and load advance. Here's are the base ignition settings (from http://crosstalk.atomz.com/~ross/jagstu ... 170936.msq):

- Trigger offset: 0
- Skip pulses: 18
- Predictor: alpha-beta-gamma 90, 80, 10
- Time mask: 0.6ms
- Percentage mask: 40%
- Next-pulse tolerance cranking, after-start, normal: 35, 35, 30
- Check tach sync: always
- Ignition input capture: rising
- Cranking trigger: return
- Coil charging: standard
- Spark output: high
- Base dwell: 1.9ms
- Max spark duration: 0.2ms
- Accel dwell compensation: 0.4ms

The distributor base timing was set to match the underhood label and the timing map was determined experimentally by listening for pinging during test runs. Interestingly enough the mechanical advance at max RPM is just about dead on, even a 1 degree addition to the timing map for 90 kPa and above results in some pinging when running 87 unleaded. The average fuel economy with ECU ignition control over the past 300 miles has been 16.5 MPG according to the trip computer (which reads approximately 10% low), whereas it's never broken 15 MPG on that drive cycle before (round-trip between Burlingame and San Francisco, CA)

I'm going to stick with this configuration until the weather improves so I don't change too many things at once, there are plenty of other repairs that need to be made that will keep me busy in the meantime.
If I am not doing ignition control right now all of the ignition settings should not make a difference, right?

Also I plan on using high impedance multi-orifice injectors like I used on the Jensen-Healey since that turned out so well.
Jaguar E-Types, modified Merkur XR4Ti, Jensen-Healey
Linux = no virii or spyware, EVER
macaddict
MegaSquirt Newbie
Posts: 32
Joined: Tue Feb 17, 2009 3:42 pm
Location: Burlingame, CA

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by macaddict »

Steven,

I am going to install a locked-out distributor, probably this summer while I have the car apart for other maintenance items. The mechanical advance introduces enough jitter that the MS-II gets confused when cranking, I've managed to tune the next pulse tolerance settings to solve most of the issues but it still occasionally requires two attempts to start when cold.

The injector ON and HOLD lines are tied together because I removed the power resistors, it doesn't actually matter in terms of engine operation as I have the MS-II connected to the OPEN lines only, but the trip computer connects to the HOLD line on one of the banks and wouldn't measure fuel delivery if there wasn't a tie-in somewhere. Note that the trip computer will not work properly if you're using peak-and-hold to drive the factory injectors, although I believe you mentioned you plan on using saturated coil injectors so it shouldn't be a problem.

The full load signal wire is being used to control the coil driver module, it was used simply because it wasn't needed for fuel control with the MS-II setup. If I have to switch back to the Lucas ECU I can disconnect the wire from the ignition driver and connect it to the full load switch on the throttle pedestal.

The O2 sensor integrate pins on the ECU connector are routed to the 4-pin diagnostic connector on the right side of the trunk. They provide information on how much the Lucas ECU is deviating from the base fuel map for each bank by varying the voltage across each pin to ground. 2.5V indicates that the system is in open loop, less than that means it's reducing fuel delivery and greater than that means it's increasing fuel delivery. They are not needed when using the MS-II although you could simply use them for something else. The other two pins on the 4-way connector are +12 switched (from the ECU relay) and ground.

If you're not doing ignition control then the only settings that should matter are the trigger settings. If you're using the opto-isolator input (recommended if you're using the original Lucas CEI module with the resistor in-line with the tach signal line) then set the trigger input to "going high (inverted)" and set skip pulses to 1 (the lowest setting I believe.)


- R
1984 Jaguar XJ-S HE (Rosie III) - Megasquirt-II v3.57 fuel+spark
http://crosstalk.splunk.com/~rdykes/jagstuff
sdmorrison
MegaSquirt Newbie
Posts: 1
Joined: Sun Sep 05, 2010 3:57 pm

Re: 1984 Jaguar XJS V12 5.3L HE (11.5:1 CR)

Post by sdmorrison »

Hey macaddict,

Any developments to report with your ignition or tune? I really appreciate that you have made your files available to those of us planning to convert to Megasquirt.

Thanks,

Stephen
1986 Jaguar XJ12
Post Reply