Making a Jaguar Trip Computer Work Accurately With a GM TPI
By Wes Channell and Dick Vandermeyden
Two years ago I purchased a "converted" 1983 Jaguar XJ6 powered by an IROC 350 Tuned Port Injection engine with a TH700R4 transmission and new leather seats. I knew it needed front shocks, paint, and a cruise control. The price was right and I had wanted one configured that way for years so I found myself the proud keeper of a new cat. In the intervening time it has been completely rebuilt from the radiator to exhaust tips. The only things not worked on, replaced, or renovated are the engine itself, mechanical fan, and the seats. The transmission is now a 4L60, all the mechanicals are new or rebuilt, and a bare metal paint job makes it look as pristine as it runs. At this stage I decided I had put enough in it that I wanted the last piece to work properly. The fuel consumption portion of the Trip Computer.
Equipped with Haynes manuals for both the Jag and the Camaro, the XJ6 ROM, and Kirby Palm's book, I started with the archives, FAQs (esp. Mark Chiampi's and Charlie Nowlin's), and questions to the lists. After about a month it was obvious this is one of the persistent unresolved issues in conversions. I began to suspect I was way out of my league in terms of the electronics. Fortunately, Dick Vandermeyden from the other Jag lists came to my rescue. Dick is trained in electronics, owns several Jags, and is the brains behind this project. Without him, I doubt it would have gotten further, let alone finished. What follows is what we did and what we've learned. The first step is understanding generically how the system works.
The basic wiring is:
Trip Computer (Two Connectors)
Red - Lights "ON" power supply
Brown - Full-time power supply (Fuse No.. 16 @ 2A)
Black - Ground
Green - Ignition "ON" power supply (Fuse No 4)
Orange - From Interface Unit
Yellow - From Speed transducer
Interface Unit (One Connector)
Orange - To Trip Computer
Yellow/Green - From ECU (EFI injector pulses)
Black - Ground
Green - Ignition "ON" power supply (Fuse No 4)
Both Jag (Bosch) and GM fuel injection systems function the same way. There is a "computer" (Jag=ECU, GM=ECM), which receives a variety of information from sensors about engine operating conditions. Based on this information and the circuitry inside, the "computer" sends a pulse signal to the injectors to open for some appropriate duration. This allows metered fuel to be available to the firing cylinder. Unnecessary fuel is returned to the tanks via fuel return lines. In both systems the injectors are all fired simultaneously once per crankshaft revolution with one-half of the necessary fuel. Thus, the appropriate amount of fuel is available when the intake valve opens every other revolution. The amount of fuel is a function of the injector flow, fuel pressure, and the pulse determined by the "computer". Both systems operate at approximately the same pressure with the Jag at 36.25 PSI and the GM at 34-39 PSI.
The Trip Computer in the dash is a nice clock without information from the speedometer, and the fuel injection system. The speedometer feed provides speed and distance information but the ECU doesn't connect to the Trip Computer. Rather, the ECU is connected to the Interface Unit which is located adjacent to the ECU in the trunk. The Interface Unit actually creates and sends the fuel flow information to the Trip Computer.
The Some what Technical
The three units work together in sequence. The Jag ECU generates an injector pulse, which is sent to the injectors and the Interface Unit simultaneously. The Interface Unit takes the voltage of the injector pulse and derives a consistent measured voltage, which is linearly related to the full duty cycle of the injectors. That "control" voltage in turn drives a variable frequency generator, which then sends square waves at variable frequencies to the Trip Computer. The Trip Computer decodes those frequencies into pulses, which increment a fuel consumption register. Given enough increments the Trip Computer reads out fuel consumption in 1/10ths. of a gallon increments. MPG is derived by dividing fuel consumption by the distance information contained in a separate register.
Testing the system in place was the next step. To test the function of the interface the ignition was turned "ON" but the vehicle not started. This caused the Trip Computer to increment at a rate of approximately 18.5 gph. This meant that, absent an ECU pulse, and operating on a nominal 12vdc line feed, the Interface Unit was sending a constant signal to the Trip Computer. Dick determined the signal was a fixed 250 HZ. Obviously this was the base frequency representing constant full duty cycle flow of six Jag injectors. Implicitly, the default is full duty cycle frequency generated by constant "ignition on" voltage. The interface sends a signal to the interface whether the engine is running or not. Subsequent vehicle road testing with the ECU/Interface Unit connection eliminated, gave an indicated flow rate of 20.65 gph. This difference between 18.5 and 20.65 gph was thought to be due to higher voltage (14.2) with the vehicle operating. Bench testing was the next step.
Dick was able to acquire an Interface Unit he could disassemble for research and testing purposes. After tracing and checking the circuitry in the unit, he was able to confirm it's operational parameters and design components. Once the circuits were understood he was able simulate higher flow rates by modifying the frequency generated by the interface. Logically, he would need to alter the base frequency to one which would increment the Trip Computer fuel register at a rate consistent with the flow of eight GM injectors.
After tracing down part numbers for both Jag and IROC injectors the big search began. The first thing learned in this step is that manufacturers don't part with injector flow rate data. Various listers from several lists helped and some Internet sites both helped and confounded. In summary, injector flow rates are stated in cc/min or lbs/hr not gallons per minute or hour. Furthermore, rates may be stated at different pressures and pressures may be stated in PSI or kPal. The following are useful conversions:
To convert cc/min to lbs/hour divide by 10.2
To convert lbs/hr to gal/hr divide by 6 for US gallons
To convert cc/min to gal/hr multiply by 0.015873 for US gallons
To determine rates at different pressures use (sqrt(new pressure/old pressure))Old Rate=New Rate
To convert kPal to PSI multply by 0.145
After a great deal of computation what finally became apparent was the 20.65 gph flow rate we found in operational testing all but matched the full duty cycle flow rate of six Jag injectors. This supported all the circuit testing and bench work done to understand the system component functioning.
The impact of substituting eight GM injector flow rates to establish a new base frequency should have been easy. Unfortunately, we had three different flow rate specifications for the GM injectors. (218 cc/min @ 300kPal; 245 cc/min @ 45 PSI; 22 lb/hr @ 43.5 PSI) Thus, we decided to plug the GM ECM injector pulse into the Interface Unit in place of the Jag ECU injector pulse and see what happened. Once completed, the Trip Computer began functioning for fuel related data. The Interface Unit was generating a signal to the Trip Computer based on injector pulses. However, each injector pulse meant a different quantity of fuel was going through the GM system than would have been going through the Jag system. The solution - test it on the road to find out how much more fuel was flowing.
The tests were run over a period of two weeks using the simple procedure of filling both tanks and resetting the Trip Computer to zero. Drive, refill both tanks (to eliminate cross flow errors) record Trip Computer readings and actual fuel at the gas pump, reset the Trip Computer and start over. The end result was a co-efficient by which to increase the base frequency of the Jag Interface Unit. This approach eliminated fuel pressure vagaries, injector inconsistencies, and anything not accounted for. After over 1200 miles of testing, a co-efficient of error was determined. The plan was to modify the base frequency of the Interface Unit by the amount of the co-efficient and re-install for further road tests.
Prior to modification, Dick bench tested the Interface Unit and several others to verify base frequency. The results revealed wide variance within and among all units at voltages below 15 volts. That is, the frequencies generated by the interfaces varied contingent on voltage until about 15 volts was reached. No two units generated the same frequencies at the same voltages and the stable frequency output at 15 volts also varied. In short, each interface generated a unique frequency output contingent on voltage. Thus, the frequency change for any interface would be unique.
To determine the actual base frequency, the source voltage in my Jag was measured (14.18 V with the engine running) then applied to the Interface Unit on Dick's bench test bed. At that voltage, the base frequency worked out to be 278 HZ rather than the 250 HZ resulting from the preliminary tests. Thus, 278 HZ became the reference frequency for my Interface Unit. This was the frequency to which we applied the co-efficient we derived from the test data. It resulted in a new full duty cycle frequency of 367 HZ and a flow rate of 27.7 GPH. These were the specifications to which Dick modified the Interface Unit.
The second test series procedurally replicated the first and resulted in a co-efficient of error indicating we had adjusted too much. While we had overshot the mark, we had it bracketed. The Interface Unit came out and went back to Dick for a second modification. This time the frequency was set to 331 HZ which produced a flow rate of 25.15 GPH.
In the Mod 2 test series the refill procedure was changed to include visual verification that the fuel level was actually at the top of the tank refill orifice. This took approximately 13.2 gallons for an empty tank. Testing took place over the course of nearly 2,300 miles and seven refills. The trip computer readings totaled 138.3 gallons used and 137.279 gallons were pumped. These figures yield an error rate of less than 1.0% over the entire trial. One refill was from an old fuel pump with mechanical metering in 1/10th increments. If that entire refill is deleted from the data, the cumulative error is less than 0.5% overall. At these levels the system is probably more accurate than the display. It remains subject to the accuracy of the pumps used at refill, injector inconsistencies, and the vagaries of temperature related fuel density. In short the frequency is "high" by one or two HZ and the GPH is several hundredths of a gallon higher than actual flow. In real world driving this means when the fuel display says 25.2 gallons have been consumed, one tank is empty and the other is nearly so.
We learned several things from the project. First, the Trip Computer can be made to accurately reflect fuel consumed and MPG, if it is provided the proper frequency square wave signal, which is consistent with actual fuel flow. Second, we made the project more complicated than it had to be. In our zeal for consistency and mathematical accuracy, we ended up with Dick doing more work than absolutely necessary with voltage stabilization. The short version of the project could have been:
1. Disconnect the Y/G wire from the ECU to the Interface Box.
2. Do the fuel flow test and record the results of indicated fuel flow in gallons per hour.
3. Run a jumper from the injector pulse line at the GM ECM to the Y/G wire going into the Interface Unit.
4. Measure the constant voltage at the Interface Box (green wire) with the engine running and record.
5. Do the base fuel flow trials to develop an average and cumulative co-efficient of error.
6. Modify the interface to generate a GPH rate at the voltage in step four which corresponds to the GPH rate after applying the average co-efficient to the GPH in step two.
7. Road test to determine remaining error rates and repeat step five.
In the final analysis, our efforts over a six month period ended with a frequency which produced exactly the same results.
Another thing we learned is that the procedure is not limited to fuel injected Lumps. From the other lists it is clear that Trip Computers may or may not reflect accurate fuel flow. This procedure will work to correct fuel flow rates in any functioning Jaguar Trip Computer/Interface Unit combination. I also suspect the same procedure might work with any fuel induction system controlled by electrical impulses. With a TPI injection system the procedure will work every time.
For those with a non-electrically controlled fuel induction system, several Lumpsters have experimented with a fuel flow sensor of some type. If such a unit produces a square wave, I would suggest looking for a "Black Box" to modify its frequency for calibration purposes. Similarly, TPI owners might be able to use such a box between the Interface Unit and the Trip Computer rather than modifying the interface unit. The Dakota Digital box designed for calibrating electronic speedometers might be a starting point for either approach.
Dick Vandermeyden adds:
Some Technical Details
Both the GM fuel injection system and the Jaguar (Bosch) system share many similarities in their operation, and from the aspect of electrical signals to the injectors are identical. The Jaguar Interface Unit originally monitors one of the injectors and uses that information to determine the total consumption of the six injectors for a 4.2 L engine.
The Jaguar Interface Unit (DAC-2686) consists of two distinct circuits. The first part determines the instantaneous duty cycle of the injectors, that is, the ratio of the time that the injector is open versus the time that it is closed per injection cycle, times 100%. Thus, when the injector remains entirely closed, the duty cycle is 0% and when open continuously, it is 100%. During engine operation, the duty cycle varies between about 3% at idle and 80% when at full throttle and full load. It is worth noting that the duty cycle is independent of the injection frequency, which varies in step with the engine RPM. Once the duty cycle has been established, a linear DC Control Voltage is generated. The second part of the Interface Unit consists of a square wave generator, the output frequency of which is directly proportional to the Control Voltage. When the duty cycle is 0%, the output frequency is 0 Hz. When the duty cycle is 100%, the output frequency is maximum and in the case of the original 4.2 L engine, about 250 Hz. The relationship is linear, thus when the duty cycle is 50%, the output frequency is 125 Hz and so on. Note that each output pulse represents a fixed (very small) amount of fuel consumed.
The Trip Computer receives the pulses in two registers, one (Fuel consumed) accumulates the pulses and displays Gallons until reset by the driver. The second one accumulates for a fixed period (about 3 second), then resets and restarts. This information together with the distance covered in the same period produces the average MPG reading.
When the Interface Unit is connected properly to a GM injector, the Trip Computer starts to operate directly, but the indication of fuel consumed is way off. That is not surprising, as the injector flow is different, plus there are 8 injectors at work instead of 6. The nice thing about the Interface Unit is that it can be scaled up to accurately reflect the new conditions. (This is why the same unit, suitably re-calibrated, resurfaces also in the XJS, in a 12 cylinder application) The Duty Cycle circuitry remains the same and 0% remains 0 GPH flow. But the 100% flow needs to produce a frequency that is commensurate with the 8 GM injectors at their nominal fuel pressure.
The Interface Unit suffers from some inherent design deficiencies, the most serious one is that the square wave frequency is battery dependent below 15.5 V. Since Lumps seem to float at 14.8 V, it is of some concern. Thus fuel consumption measured can vary, depending on battery voltage. This effect varies in severity from unit to unit, but can be ameliorated if so desired by a regulator modification in the unit before re-calibrating.
When embarking on this Modification, the first point of order is to find out what the original 100% duty cycle produces in Gallons Per Hour on the Trip Computer. This is best done by leaving the input lead to the Interface Unit disconnected. Since this is now 0 V permanently, the Interface Unit assumes that the Injector is open permanently (100%). The Trip Computer rapidly accumulates Fuel Consumed. Using a stop watch, determine precisely how much Gallons are added in one hour exactly, preferably under operating driving condition (that will take prevailing battery voltage into account). About 21 GPH is the norm for a 4.2 L Interface Unit.
Then fill up both fuel tanks and reset the Trip Computer to 0. Connect the Interface Unit to one of the GM injectors. Drive at your leisure until both tanks have been filled 10 times, keeping accurate record of how much each fill-up was. Let us say that it took 200 Gallons for the 10 fill-ups according to the gas pumps. Compare this with the Fuel consumed on the Trip Computer. That will be far less, say 140 Gallons. Dividing the two readings will tell us exactly how much the Interface Unit needs to be speeded up.
Take the Interface Unit and the Trip Computer from the car and connect both to a variable power supply and a suitable frequency counter. Adjust the variable power supply, so that the GPH on the Trip Computer is the same as measured while on the car (About 21 GPH). This method will assure that the prevailing battery voltage is used while re-calibrating. Now measure the output frequency of the Interface Unit (About 250 Hz). Multiply that with the correction factor, 1.43 in the above example, which gives a new frequency objective of 357 Hz.
Open up the Interface Unit by carefully drilling out the 4 small rivets at each corner with a small drill bit. If so desired, install the voltage regulation modification first. Then reduce the value of the square wave generator feedback capacitor until the new frequency of 357 Hz is produced under prevailing battery voltage and open input to the Interface Unit. This requires items such as a desoldering tool, soldering iron suitable for PC cards, small pliers and cutters. Carefully remove all solder flux with an appropriate solvent. Using 2-56 screws and nuts, close the unit back up again and affix a label on the outside to reflect the modification. Re-install together with the Trip Computer. Re-connect the input to the GM injector.
The Trip Computer should now accurately reflect the fuel consumed. Note that the error is solely determined by the accuracies of the original GPH at 100%, fill up record and the gas pumps used..... If it is determined later that the readings are still off by a small amount, appropriate fine tuning of the square wave generator will eventually get the error to below 1% as reported by Wes Channell.
Those folks that are accomplished electronically should be able to do so with the above information at hand. More details are not available, as I have no wish to provide unscrupulous people with another way to cash in on other people's efforts without recompense.
The following, is NOT a "Lumps" offer to sell information, parts, or kits.
It is offered by the author, only as a courtesy. Jag-Lovers.org , or their associates, should not be considered as party to any offer on these non-commercial "Lumps" pages. Offers, are made solely, and completly, by the person /persons named.
I, Dick Vandermeyden, can be available to individuals modifying their own interface units for off list procedural advice. For those who opt not to mess with delicate electronic circuits, I am willing to do this modification for a small consideration, provided an accurate original GPH and scaling factor are supplied by the sender together with his Interface Unit mailed prepaid. It will be returned prepaid in the Continental US via USPS. Other areas at cost.
Contact me, privately, at: firstname.lastname@example.org.
Improve your Jag-lovers experience with the Mozilla FireFox Browser!
View the latest posts from our Forums via an RSS Feed!
©Jag-loversTM Ltd / JagWEBTM 1993 - 2018
All rights reserved. Jag-lovers is supported by JagWEBTM
Use of the Jag-lovers logo or trademark name on sites other than Jag-lovers itself in a manner implying endorsement of commercial activities whatsoever is prohibited. Sections of this Web Site may publish members and visitors comments, opinion and photographs/images - Jag-lovers Ltd does not assume or have any responsibility or any liability for members comments or opinions, nor does it claim ownership or copyright of any material that belongs to the original poster including images. The word 'Jaguar' and the leaping cat device, whether used separately or in combination, are registered trademarks and are the property of Jaguar Cars, England. Some images may also be © Jaguar Cars. Mirroring or downloading of this site or the publication of material or any extracts therefrom in original or altered form from these pages onto other sites (including reproduction by any other Jaguar enthusiast sites) without express permission violates Jag-lovers Ltd copyright and is prohibited