Jim Klein recently posted a junkyard post on a 1976 Cadillac Seville. Not only was this an awesome junkyard find, but it stirred up lots of conversation in the comments section. Of particular interest was the fuel injection system. When the Cadillac Seville was introduced in late in the 1975 model year as a 1976 model, it included an electronic fuel injection (EFI) system on its Oldsmobile 350 engine. This made it the first American car to come equipped with electronic fuel injection as standard equipment, a fact commonly forgotten.
1976 Cadillac Eldorado with an EFI equipped 500 V8
Cadillac did not limit its electronic fuel injection to the Cadillac Seville with its Oldsmobile 350 engine; it was also available as optional equipment in other Cadillacs equipped with the 500 cubic inch Cadillac V8 engine beginning in the 1975 model year and later the 425 V8 engines through the 1979 model year (there was some availability in 1980 on 350 Oldsmobile powered Cadillacs). This fuel injection system delivered a much more precise fuel and air mixture, allowing the engines to burn cleaner, use less fuel and have considerably improved drivability. A positive side effect of the cleaner and more efficient fuel metering was that the Oldsmobile and Cadillac V8 engines were able to be tuned to produce more power while meeting emission standards than their dirtier carbureted versions.
Note: The 350 engines equipped with carburetors were not installed in Cadillacs and are used for comparison purposes only.
1979 Cadillac Seville with an EFI equipped 350 Oldsmobile V8
Despite this system being the first mass produced EFI system offered by a domestic manufacturer, there is little historical information about this fuel injection available today. Likely part of the reason is that it was a dead end technology, being an open loop analog fuel injection system. Cadillac completely abandoned this system in 1980 with the introduction of its digital electronic throttle body fuel injection. Nothing was carried over from the analog EFI.
The analog EFI used by Cadillac was a joint effort between Bendix and General Motors. Bendix’s history with electronic fuel injection dated well before 1975. It introduced an EFI system in 1957, called the Bendix Electrojector system. In 1957, this system was optional on the Rambler Rebel, with estimates of 0 to 6 cars equipped with the option. In 1958 it was offered on some Chrysler products, with 35 Chrysler, DeSoto, Dodge and Plymouths so equipped.
This original EFI was very complex and the electronics of the time, in particular the semiconductors, were not advanced enough to create a durable electronic control unit. Although I couldn’t find any primary sourced material, other sources claim that Bendix sold the intellectual property for its electronic fuel injection to Bosch. With this, Bosch developed the much more successful Bosch D-Jetronic, introduced in 1967.
The fuel injection system that Bendix developed with Cadillac in the 1970s shared many similarities with the Bosch system. I couldn’t find any definitive sources that explained the finer details of Bendix reusing the same technology that it had sold to Bosch. That said, while the Bendix Cadillac EFI had many similarities to the Bosch EFI, the Bendix-Cadillac EFI was was an independent design that used almost all Bendix or General Motors parts.
The Bendix Cadillac EFI is a multi-port electronic fuel injection system. Unlike later fuel injection systems, it was an open loop system. This meant there was no oxygen sensor which could read the exhaust and monitor the fuel air mixture to adjust accordingly. This open loop system simply had to rely on preprogramed fuel mixture charts which were used based on the inputs from the engine sensors. The Bendix Cadillac EFI consisted of three major systems; the electronics system, the fuel delivery system and the air induction system.
The electronic system was controlled by a computer called the Electronic Control Unit (ECU). There were a series of inputs fed into ECU. These inputs were interpreted by the ECU and were used to calculate the appropriate amount of fuel to be metered into the engine. As seen in the diagram above the input data came from a number of electronic sensors. These included a coolant temperature sensor, a manifold absolute pressure sensor, an intake manifold air temperature sensor, the engine RPM and the throttle position sensor.
The ECU could interpret the data (voltage) from each of these sensors to understand the engine’s operating conditions and the driver’s desired response. For example, if on a hot day the driver was going up a steep hill at high speed and had downshifted to second gear, the ECU might see a high coolant and air intake temperature, low manifold pressure, high engine RPM, and the throttle will be opened significantly. From this data, the ECU would calculate that the engine would require a large quantity of fuel to be delivered.
This EFI system was from a 1975 Cadillac 500 V8 engine.
The ECU controlled a series of output devices to properly deliver the correct fuel mixture. It controlled the eight fuel injectors to spray a predetermined amount of fuel through a precisely time burst. The quantity of fuel was controlled by the length of the burst. The fuel injectors were batch fired, meaning that numerous injectors fire simultaneously, unlike the later sequential systems which fired each injector in sequence (similar to spark plugs). For the Bendix Cadillac EFI, the system was divided into two injector groups: Group 1 was cylinders 1,2,7,8, and Group 2 was cylinders 3,4,5,6.
This ECU is for a 1975-76 Cadillac 500 with CA emissions. It was listed on eBay for $1000.
The ECU controlled other components in the fuel injection system besides the injectors. This included the fast idle valve, which was used when the engine was cold. It also controlled the EGR solenoid, limiting the conditions that the EGR valve was opened. Other cars of this era typically used thermal-vacuum switches to control when the EGR valve operated. And finally, the ECU controlled the dual electric fuel pumps. The ECU would activate the fuel pumps when the ignition was turned on and the engine was cranking or running. It deactivated the pumps when the engine stopped running, or if the driver turned the ignition on but did not start it within a short period of time.
The fuel delivery system was significantly more complex than the typical mechanical fuel pump with a single feed fuel line of the era. The Bendix Cadillac EFI had two separate electric fuel pumps. One was a pump mounted in the fuel tank. This was a small boost pump that simply feed a larger externally mounted pump. The small pump helped prevent vapor lock by keeping the large pump from having to suck the fuel from the tank. The large pump was a constant displacement, roller-vane pump mounted to the chassis. This pump feed the engine and was responsible for producing the high pressure required for the EFI. It also incorporated a check valve to maintain fuel pressure when shut off. From there the fuel flowed through a fuel filter, and then to a fuel pressure regulator that maintained the 39 PSI required. The fuel pressure regulator feed the fuel rails and injectors, while excess fuel was bled off and returned to the tank via the return line.
The air induction system was similar to that used by the carburetor equipped variants, but there were a number of significant differences. Air was fed into the engine with a dual bore throttle body. Like a carburetor, the throttle body controlled the air into the engine via butterfly valves actuated by the accelerator pedal. The throttle body also had a fast idle air valve. This valve would open when the engine was cold to allow more air into the engine and increase the engine speed. The valve had a thermal element within that warmed up and slowly closed the valve as it did so. When the thermal element reached approximately 140 degrees, it was fully closed and ineffective. At this point the engine would be warm enough to not require a fast idle.
This intake manifold is from a Oldsmobile V8 engine used in a Cadillac Seville.
The throttle body is mounted to an intake manifold and it is similar to the intake manifolds used by the carburetor equipped engines. The biggest and most obvious difference was the addition of fuel rails with injectors mounted near at the end of each port at the cylinder head. The EFI intake manifold did not use an exhaust crossover. This is a passageway under the carburetor where warm exhaust would heat the incoming fuel air mixture from the carburetor to help improve atomization in colder conditions. The EFI manifold was dry, meaning only air passed through, making the exhaust crossover not required.
Many of the sensors and electronic components used on this EFI system became common place in the 1980s, but during this time in the 1970s they were uncharted territory for the American manufacturers. Like other technologies in their infancy, this made for some growing pains. Cadillac had to manufacturer many variations of the ECU, each calibrated for different models and/or drivetrain configurations. This resulted in a much more costly and complex ECU part supply. It wasn’t until the 1980s that GM began to use ECUs with removable PROM chips, which would allow for one basic ECM to have many different variations by using different PROM chips.
GM and Bendix supplied most of the EFI components other than the throttle position sensor which was a Bosch unit, meaning most of the components were new and unproven. Unusually, the MAP sensor was mounted in the ECU itself, which was mounted in the dash of the car. The MAP and the other sensors and components used in this EFI system were known to fail. There were many reports of EFI problems such as no-starts, hard-starts, high fuel consumption and other driveability issues. This combined with many technicians being completely unfamiliar with EFI meant not only did the systems fail, but more often than not they were not able to be repaired correctly. And to top it off, parts for these systems were extremely expensive and often difficult to obtain. Surely this must have soured more than a few Cadillac customers.
This 1990 Cadillac Brougham was powered by a 307 Oldsmobile V8 and was one of the last US market cars to be equipped with a carburetor.
Cadillacs of the 1970s are heavily criticized for being substandard cars compared to past models, and much of it is justified. Despite all of the short comings of the era, Cadillac’s EFI was somewhat of a valiant effort to introduce the latest technology. The Cadillacs equipped with EFI were much better running cars and had improved power, fuel economy, throttle response and emissions. Unfortunately, like much of what GM and other domestic manufacturers did at the time, the effort was good up to a point.
The EFI was failure prone while being costly and difficult to repair. The technology limitations were partially to blame, but GM certainly didn’t put enough development effort into this technology before it was released to the public. Ironically, after numerous other technology failures such as the variable displacement engine, HT4100 engine, Cadillac actually reverted back to a carburetor in 1986 when most other American cars had switched over to EFI. The big Fleetwood Brougham was powered by the 307 4-bbl Oldsmobile V8 engine until 1990. In 1975, it seemed like a good idea to install EFI into Cadillac’s flagship cars, but its poor execution was one of the first of many hits to its reputation. Perhaps that might even qualify it to be a GM deadly sin, but I will leave that up to you to decide.
Related reading:
1957 Chevrolet Fuel Injected 283 V8: Ahead Of Its Time And The Competition PN
This is fascinating. I had no idea Cadillac had EFI on anything other than a Seville until the 80s.
Vince, I don’t know you or your background? I’m dismayed though at the huge generalisations you make. Before going on, consider the Rolls-Royce Silver Shadow from 1965, that their V8 had a block made by Cadillac, that they had Citroen suspension? These stories have been around since the sixties, both completely untrue. Perhaps fake-news around earlier than we thought?
I’ve been involved with Cadillac over 45 years now and been lucky in that I had many friends at (old) Cadillac on Clark Street, including Bob Templin, Cadillac’s Chief Engineer and Father of the ’76/’79 Seville.
Firstly, the Cadillac/Bendix EFI was introduced late 1974 as an option on 1975 Cadillacs. It was not born through the ’76/’79 Seville. The Seville did not appear in late 1975 either, it was introduced May 1975.
You glibly use the term ‘reliability issues’? Where have you got this from? I have all the paperwork Cadillac produced, workshop manuals, Serviceman bulletins and the other ‘internal’ dealer bulletins. I care for a number of Sevilles around the world, I know many of them around the world, even in Iran, most are still running perfectly well on their original EFI system?
It is true early Sevilles had an ECU fault, Bendix jumped on this, it’s believed production was slowed if not halted to get the corrected ECUs fitted. Dealers were instructed to replace them as a precautionary measure.
You say there is not much information on the system? What about Cadillac manuals, including the two versions of the special diagnostics books, the first up to ’77, the second ’78/’79? These are designed to take you through the system in order to correctly and efficiently diagnose problems? Regarding parts, not really any more of a problem today than on any other 40 year plus old car? There’s a superb electronics expert dealing with ECUs and several guys making new parts, one the all-important Fast Idle Valve unit.
Sometimes I discover people trying to solve problems with this EFI with little experience and very often, no Cadillac books? No wonder there are problems. I don’t know where your ‘many reports’ statement regarding failing MAP sensors comes from, if you know do us? If I’m wrong, I’m wrong?
Why use the word ‘abandoned’ when saying Cadillac moved to something new in 1980? Cadillac quickly understood the power of EFI, not just for performance but to deal with the endless emission and gas-mileage regulations your country had. The problem was the ’75 onwards system was what was described as zero-intelligence and that was its real problem. GM spent a fortune developing the digital electronics. Not entirely new though, digital computing goes back a long way, much further than some might imagine.
Again you slate technicians? Cadillac service department provided training in all of this and accompanied by superb technical publications. Do you really believe Cadillac would introduce their EFI system, then on their shinning new car the Seville and not have trained their dealer technicians on how to care for it? I’m sure here and there some people may not have been as good as others, but…?
You close describing the Cadillac/Bendix EFI as ‘failure prone’ but in fact it was technically very advanced, not perfect, but it led the way to what we have today.
Paul, I am aware that the EFI was offered on 1975 Cadillacs and that the ’76 Seville was introduced late in the 1975 model year as a 1976 model. I had typed that the Seville was introduced in late 1975 (I meant to write late in the 1975 model year). This was a typographical error. I didn’t overtly state that the EFI was introduced in the 1975 model year on other Cadillacs. However, the title of the article shows it was introduced in 1975 and my chart that I created showed the 1975 Cadillac 500 had EFI available. Nowhere did I state that EFI was introduced with the 1976 Seville. I said that the Seville was the first American car to have EFI as standard equipment, which is true. The text has been slightly revised to clarify these points.
Further my comment about lack of information today, was a reference to a lack of historical information. This is why few car enthusiasts remember that EFI was offered as an option on Cadillacs (other than the Seville). I am aware that there is still all of the original service documents available, many of which I used to write this article.
While you may dispute my claim that the Cadillac EFI was unreliable, you have not presented any evidence to the contrary. Do a Google search, and watch how many stories of problems related to Cadillac EFI pop up. The post below by 79 Eldorado, who appears to be a member of the Cadillac LaSalle Club, specifically addresses several reliability issues the systems had. He mentions how “a retired electrical engineer who has identified and found fixes for most of the ECU related issues.” This clearly demonstrates that the ECU’s had issues that GM did not ever resolve when these systems were new. It took an enthusiast from many years later to come up with a solution.
He also goes on to discuss sensor problems when he states “I have recreated/ re-engineered the air/coolant sensors to eliminate the problems associated with the originals. Prior to my effort there was a sensor company owner who also owned one of these EFI Cadillacs, frustrated specifically with the poor sensor reliability, his company manufactured replacements for a few years…The original sensors had a reputation for going bad; sometimes in less than a year.” Sensors failing in a year and requiring re-engineering to eliminate problems isn’t a sign of reliability.
Finally, I never once berated the EFI system as being poorly designed, rather, I stated that the system was a good idea but poorly executed. Cadillac was clearly trying to follow the new leader in the luxury world, Mercedes-Benz, by offering EFI. However, its effort wasn’t very well executed.
Vince,
Just a few comments from a retired product validation engineer that also dabbles with these systems. You mention poor execution on the part of Cadillac, but I don’t necessarily agree. The system was robust enough that it was continually produced with only minor changes for the entire run of the first gen Seville. No major design changes, no carb retrofits. It’s dead simple in operation and troubleshooting. It offers all of the advantages of D-Jetronic with fewer moving parts and increased performance.
In that vein, I am puzzled about the Mercedes comment, as the Cad EFI was much more reliable than their K-Jetronic system, which incidentally got worse fuel economy with a smaller engine. Something dealers liked to point out to prospects.
The engineer you mention above happens to be a friend of mine. Just to clear a few things up, the ECU reliability issues weren’t a problem back in the day, in fact the design was quite robust being Bendix had decades of off-highway/OEM experience. The repair work my friend performs now is due to the electronics being 40+ years old. The electrolytic caps have dried out, connector fingers have oxidized, and other age-related issues have set in. Some of his re-engineering is also due to component obsolescence. Through careful design he can salvage an ECU with a failed integrated MAP sensor using a cheap, readily available modern part. AFAIK, “drift” has not been a problem. The laser-trimmed resistors have been remarkably stable and he has a significant sample size to draw from.
The sensor issue is a bit conflated. I’m not aware of these being any more unreliable than the competition, who also sourced theirs from the same supplier(s). The kerfuffle is that a decade after production, you could no longer drop into the parts store and pick one up. These use a PTC design unlike modern NTC sensors, so there was no simple retrofit. The reference below is to the grassroots effort that’s reintroduced the design. So when Seville owners complain about sensor failures, up until recently, it got a lot more attention.
Looking at your comments, I know you want to hate on this system. Were there 1st year pains? Absolutely. But calling this design anything other than a success would be disingenuous.
BlueBiarritz, that is fine you don’t agree with me that the system was poorly executed. This world would be boring if we all had the same opinions. Your argument that the system was left in service for numerous years without major change doesn’t hold water with me. GM left many badly executed or designed things in service for years without major change during this era and into modern times. I appreciate you offering further insight on the ECU rebuilds, however, I used reputable sources that show ECUs were problematic when these cars were new on the earlier cars. One is a Popular Mechanics Owner’s report, which BTW, shows the EFI was by far the number one trouble spot with the ’76 Seville and it cites some owners reported ECU and sensor problems.
I didn’t say that MAP sensors were expensive, or hard to come by, but the fact is, the owner of a Deville with a Q-jet didn’t have to replace a MAP sensor or any fuel system sensor ever. Q-jets were cheap to repair, and rarely required adjustment or service. Most Seville owners were former Cadillac and GM owners used to reliable Rochester Q-jets, so these sensor issues were new to them. The Mercedes comment was not about comparing the EFI systems. I simply said Cadillac added EFI to the Seville to chase after Mercedes, who was clearly the leader at this time. The Seville itself was aimed directly at Mercedes.
My intention was never to cause so much friction with Cadillac owners. The bulk of this article is explaining how the system worked. I used multiple sources for this article and not one claimed the Cadillac EFI was reliable; all said it was troublesome. I know Cadillac enthusiasts such as yourself are passionate about your cars and that’s great. However, that doesn’t mean that they were perfect. Looking at them objectively for their faults isn’t a bad thing. Contrary to your belief, I don’t dislike the EFI system or late 70s Cadillacs. In fact I actually like the late 70s Sevilles and Devilles. In this article I never once state that it is poorly engineered or a terrible system, but I do believe it had reliability issues and that it didn’t help boost Cadillac’s reputation. I also think that your claim that it was a success is a stretch, but hey that’s just my opinion and I am okay if you don’t agree.
So how come a lot of Seville systems were replaced by carbs in the day.
Private repair shops and shade tree mechanics unfamiliar with the system or and/or the cost of components to fix a drivability problem pinpointed to it. With these Cadillacs carburetors and intake manifolds for the same engine existed, were widely available and a known entity to a repair business, so they could convince an owner to convert as an investment for future repairs(and guarantee return business for tune-ups carburators need).
Same reason many 60s era performance cars that came with milticarb setups, be it dual 4bbl or triple 2bbl, were often replaced by a regular single 4bbl back then too. It takes extra time, steps and components to get them running right. it’s not that they’re impossible to get running right but whether it’s worth it over simply going with the familiar tried and true.
wonder how much better was the fuel economy on 500 fuel injection. the carb versions got about 9 mpg
Hi Mr Bedford,
My name is Marc J.R Forestier, I own a 1979 Cadillac Seville Elegante. I would appreciate to be in touch with you because you look to be an expert of this car and of the Cadillac brand. Here is my e-mail address, marcforest65@gmail.com and my phone is +34640974911. (I am on Whatsapp and Telegram). I live in Mallorca, Spain where my car is actually in process of restoration. I hope we will be able to communicate because I will really to have your opinion on several matters.
Best Regards
Marc J.R Forestier
Not quite Deadly Sin territory – mainly because, until reading this article, I never realized that 70’s Cadillac’s offered fuel injection. Somewhere along the line, I always assumed carburetors.
In the early Nineties a guy I knew bought a ’76 Seville with the intent to restore it into a weekend car. He wanted to upgrade the stereo, so he removed and threw away the head unit and what he thought was an amplifier. Yup! He threw away the ECU.
Thank you Vince for this great writeup on a mostly forgotten piece of GM history – I had never heard any grumbling about the Cadillacs with this era of EFI from my Pap or Dad, but I’m not sure how many Sevilles were in Mifflin County PA at the time. Pap for a short time did own a 79 Sedan deVille with the EFI 425 that he said ran like hell, but not long enough for any repair nightmares.
I don’t think this qualifies for “DS” since they didn’t fail out of the gate like the Chrysler Imperial 318 EFI or the 1st Gen Olds Diesels – but it does show why they were resistant to put EFI in the D Body until 1990…and even then, you had to get the trailer towing package or coachbuilder package to get the 5.7 TBI.
I sure as hell wish they put TBI on the 307 Olds – I love my 88 Brougham, but not when pulling a hill or trying to pass on anything but flat road…
EFI was in the C/D-body Caddies before 1990. The HT4100 powered Caddies all used DFI, which was last used in 1985. For 1986 the 307 was the sole engine option but in 1990 the 350 Chevrolet TBI engine became available as you stated. In 1991 the L03 305 replaced the 307 and it was a big improvement in performance.
Ford and Buick had sequential port fuel injection on most 5.0L V8 and 3.8L V6 by 1986. Both systems were very reliable and effective. Chevrolet had port fuel injection on high output versions of the 5.0 and 5.7 by that time. I don’t know why GM didn’t have port fuel injection on the regular output 305, 307, and 350’s by 1986.
GM was slow to get multiport EFI on its V8s, other than the High-performance and Cadillac engines. The GM TBI system was hardly advanced or cutting edge but it was simple and reliable. It was a big improvement in drivability and reduced maintenance over the E4ME electronic Q-jet. GM was very lackluster to update any of the RWD platforms during the mid and late 1980s. Ford went to CFI in 1983 on their Panthers, GM waited until 1989 to introduce TBI on the B-body. At least the pickups got TBI in 1987, but they didn’t get mutlipoint injection until 1996.
Ford first put their vomitous* CFI system on the 1980 Lincolns with the 302 engine.
Totally agree with you that GM’s TBI was simple, reliable, and a whole lot preferable to a feedback (or any other) carburetor.
*That’s my opinion and I’m stickin’ to it
Right, I forgot about the Lincoln’s getting CFI first.
Chrysler spent big bucks on recalling every Imperial to convert from digital EFI back to carbs.That was in 1981!. P
I should have clarified and said “trouble free EFI” since everyone hated the HT4100 and the V8-6-4 (unless you cut the trans wire to make it run on 8 all the time).
My first car was an 1984 C/D Body RWD Sedan DeVille, and the EFI was the only nice feature of the HT4100. Had better throttle response than the ElectronicQJet 307 but not as much power.
A co worker moved up from a 1978 Coupe Deville to a 1981 Coupe Deville w/V8-6-4 motor, both bought new.
He never complained about the cylinder switching but admitted it ran noticeably better with a ‘wire disconnected’ that allowed the motor to run on all 8 cylinders all of the time.
He left our employer some years later still driving the same car.
My current 2020 Chrysler 300s 5.7L is the first car I’ve ever owned with MDS (Multi Displacement System).
Seamless switching between 8 to 4 cylinders, with excellent highway fuel mileage. I’m being told the Chrysler 6.4L with MDS is even better – hoping to drive one at some point.
Thanks for researching this Vince.
While I was aware of this system’s existence on the 1976 Seville, I didn’t know much else about it, or that it was offered on the Cadillac 500/425 engine.
Why didn’t Cadillac use the fuel injection system from the Corvette? Tried and proven technology that worked and a real world performance boost.
The Rochester Fuel Injection used by Chevrolet was a mechanical fuel injection, which even by 1975 was old tech. With emissions regulations and fuel economy becoming more important, EFI’s more precise fuel metering was the way forward.
This page from Bosch states that Bendix and Bosch shared patents; not sure if there was licensing involved: https://www.bosch.com/stories/50-years-of-bosch-gasoline-injection-jetronic/ Note that some of the text in that page is unrelated to the topic; fortunately in my experience Bosch’s product QA is at least slightly better than their website proofreading. By the way I know that Bosch electronic fuel injection remained analog in production for quite some time, even after digital variants were available. I assume the Cadillac ECU’s were also analog.
Fascinating! Thanks,Vince and for the shoutout too – one note, the junker was actually a 1976, first year model.
Thanks for this. I’ve never read anything in detail on this until now, so you’ve filled a hole in my head.
Given that car radios and tape players were all analog, and GM’s Delco units had a solid rep, I’m a bit surprised they weren’t able to make a more robust system. Bosch seems to have been able to do so.
These cars were before my time when I worked at a Cadillac-Olds-Chev dealership, so I don’t have any first hand experience (the oldest stuff I remembering in our shop being HT4100 Cadillacs). So I only have anecdotal stories about the issues. Many of the EFI electronic components were not overly robust, including the ECU. It seems that the Bendix and GM components were just poorly made and prone to failure. Heck, as I mention below, GM couldn’t even get the GM o-rings for the injectors leaked. There were problems with the fuel lines leaking too, due to the higher fuel pressure. Then there was also the analog drift issues that Daniel mentions below. Combine that with the techs who probably couldn’t (or didn’t want to even try) to properly diagnose and fix the cars, lead to a unreliable system.
Interesting. I was thoroughly familiar with the Bosch system in squarebacks, so your detailed comparison is meaningful. The Cadillac ‘lookup table’ approach is surprising in the ’70s. I guess it made sense as a way to calibrate changing EPA standards easily. Change a few diodes instead of recalculating all the sensors.
The look up table is what was changed to calibrate for the specific application. That is why they switched its location to a PROM with the advent of their TBI system. That way one ECU could cover applications from the 4.3 V6 to the 454. Note those were minor changes in the look-up table the majority of the tuning to a specific displacement was in the change of injector sizes. That is why the 4.3 used the same injectors as the 5.7. Both had the same size cylinder and hence the same req fuel for the next intake event.
Fuel Injection was pretty ubiquitous and reliable in Volkswagens in the late 70’s. By 1978 most VW’s were equipped with Bosch K-Jetronic. I have it on my 1981 Scirocco and its been reliable with good drivability. Instead of teaming up with Bendix, they should have teamed with Bosch
At this time, GM was plagued by “Not Invented Here (NIH) syndrome. For the GM of the time, buying any technology from outside the company was heresy.
From what I have heard in the car business (and it could very easily just be rumour) is that the Cadillac EFI experiment made GM wary of EFI and therefore it didn’t develop a real FI system until it absolutely had to.
I love the K-Jetronic system. One you take the time to learn how it works, service is a snap. Just be careful to replace the fuel filter on time.
Stealing it, on the other hand, was just fine; John DeLorean says in his book GM went in desperation to Holley, who bent over backwards and put all hands on deck to devise a carburetor that would solve severe problems holding up the Vega, whereupon GM went »yoink« and gave Holley’s work to Rochester. I don’t know if this is actually what happened, but it seems believable to me.
…and to use the correct material when replacing underhood fuel lines. About 25 years ago a friend lost his ’75 Volvo 240 when some dillweed replaced the lines from the fuel distributor to the injectors with SAE J30R7 fuel hose—that’s the kind suitable for low-pressure carbureted systems, and nowhere near adequate for fuel injection pressures. The new lines held for a few days or a couple weeks, then failed with the engine hot…car-B-que.
SAE J30R9 fuel injection hose would’ve been well more than adequate, but even now it’s relatively hard to get; back then it was much harder.
“But that trick never works!”
‘This time for SURE!’
Based on Michael Porter’s account, one of the contributing factors to stuff like that was that the corporation could overrule the divisions on any outside supplier contract, so a supplier might negotiate a price in good faith and then have senior management demand further cuts or concessions. According to Porter, Bendix president Michael Blumenthal negotiated the Cadillac deal directly with Robert Lund, Cadillac general manager, and then after Lund had agreed all points of the deal (which was quite generous for Bendix), insisted that it be approved directly by Ed Cole, then GM president, concerned that GM senior management would pull a Lucy-with-the-football move on the deal.
Hi,
I always thought that had GM developed the fuel injection system introduced in the 57 Chevy and Corvettes, they would have owned the market in the 60’s through the 80’s.
The tuned port fuel injection system in the 96 Roadmaster is excellent and very reliable.
Gary
That ’57 system was a rip-off of an aircraft design that is still in use today. It used metered continuous flow, not pulsed injection. Excellent for maximum power output, not so good for cold starts, idling or hot starts. Aircraft engines usually operate at over 50% power most of the time. Road cars, except in racing, do not.
Another reason why more than a few American luxury car buyers switched their allegiance from Cadillac to Lincoln…..and never returned to the GM family.
Ironically, Bendix nearly made a deal with Ford in 1971 to supply this system for Lincoln-Mercury, but according to Jack Campbell, head of the Bendix carburetor division, Bendix Automotive Group senior management managed to talk their way out of the deal by asking for too much upfront.
If Wikipedia can be believed, D-Jetronic was the Bendix Electrojector as debugged by Bosch after Bendix got their hands burnt and sold the IP to Bosch. So that would imply the Cadillac system was somewhere between Bendix buying back from Bosch a licence and Bendix having another go at things to make what might reasonably be called “D-Jetronic II” for the Cadillacs. Round ‘n’ round ‘n’ round we go, wheeee!
Except, as you mention, the throttle position sensor—which is the D-Jetronic item with, IIRC, 16 positions plus a throttle-closed position, rather than the later stepless continuously-variable-resistance type.
The magnet/reed switch engine speed sensor in the Cadillac system was closely similar in concept (monitor engine speed by looking at the distributor) with an updated implementation versus D-Jet’s trigger points. And as you describe, the manifold pressure sensor was moved and redesigned.
Beyond that, TTBOMK there’s not much to distinguish the Cadillac system from D-Jetronic. Same fuel pump, very similar injectors differing only in flowrate and minor physical configuration details, etc. Same analog ECU architecture with components subject to drift (with resultant running problems), same dirty gasoline constantly clogging injectors, same severe knowledge deficit amongst mechanics, same disappointingly poor fuel economy.
The D-Jetronic used in Type 3s was drop dead reliable. The only real problem was leaky injector hoses which were braided rubber lines. Easy to fix. Just cut off the factory crimps and put new hose on with worm clamps. I think it’s pretty amazing that such a reliable FI could use analog processing. Again Wikipedia:
As in the [Bendix] Electrojector system, D-Jetronic used analogue circuitry, with no microprocessor nor digital logic, the ECU used about 25 transistors to perform all of the processing. Two important factors that led to the ultimate failure of the Electrojector system: the use of paper-wrapped capacitors unsuited to heat-cycling and amplitude modulation (AM radio) signals to control the injectors were superseded. The still present lack of processing power and the unavailability of solid-state sensors meant that the vacuum sensor was a rather expensive precision instrument, rather like a barometer, with brass bellows inside to measure the manifold pressure.
25 transistors: wow! As I recall, the Type 3 ECUs used tantalum caps.
Consider me skeptical.
I can only go from my experience, Daniel. Never had to replace an ECU nor any other D-Jetronic component when I worked in a VW repair shop and several VW dealers from the mid-1970s through the early-1980s. At that point those FI units were ten to 15 years old. But I did fix lots of leaky fuel hoses on them.
They were great for skiers who changed 5000′ altitude twice/day. A definite improvement over carburetion.
You have to consider the state of technology in West Germany at the time. They had the talent and the DM and the worldwide market to make very reliable vehicles. And they did.
I don’t doubt your experience, I just question its universality.
No debate there, on either point. Certainly if we compare the two with both systems in perfect order and adjustment, the D-Jetronic wins. But if we factor in roadside repairability and the (un)likelihood of competent diagnosis and repair at any given service station or garage, I think things get murkier.
I don’t think German vehicles of the late ’60s to early ’70s—the D-Jetronic era—were categorically more reliable than (say) American cars of that time. If the topic were whether German cars of that time were more carefully built with better workmanship than American cars, I’d have a much easier time getting onside, but that’s not the same question.
Paper-wrapped capacitors were a death knell to car ECUs just as they are to 15-year-old desktop computers today. Going to synthetic capacitors increased ECU reliability by an order of magnitude.
In the mid-60s US domestic car manufacturers didn’t much consider reliability. VW had an edge there and they exploited it, in the same way Toyota did in the 1970s.
Late-60s/early-70s Type 3 Volkswagens were drop dead reliable. They didn’t break. They wore out, but they didn’t break.
I don’t agree. What makes you say so?
I don’t think German vehicles of the late ’60s to early ’70s—the D-Jetronic era—were categorically more reliable than (say) American cars of that time.
US automobile market share:
https://en.wikipedia.org/wiki/Automotive_industry#/media/File:Motor_Vehicle_Prod_share_RITA_T1-23.svg
Were VW’s cheaper? Not really. They were more reliable on a day-to-day basis.
That’s a lovely graph of market share. It addresses nothing about reliability, though. You keep on saying VWs were more reliable, and I understand you believe that, but repeating a belief doesn’t really make it into a fact, eh! I’m not trying to pick a fight or start a quarrel here; you’re entitled to believe whatever you want. If there’s dependable data showing German cars of (say) ’66 to ’73 were particularly reliable, though, and more so than cars of other nationalities, I would be interested to see it.
(Cheaper…? When and where did that enter the discussion?)
“Cheaper” on a supply/demand curve would mean a bigger market share. But VWs weren’t cheaper. If anything, they were more expensive for their respective markets.
Daniel, I can’t give any data whatsoever that mid-60s/early-70s VW’s were more reliable than the Big 3 cars. Even anecdotally, I was just a kid back then and had no car experience. I just know that VW’s market share took off and held steady back then which meant that the demand for their products was high. They only actual experience I have is with D-Jetronic fuel injection used on Type 3s and from my experience it was as close to perfectly reliable as FI could be. In fact it was every bit as reliable as today’s FI.
+++++++ Real POS.
Unless remark = Reliably Drop Dead
Nope. My memory failed me. It looks like Type 3 ECU’s used polyester caps. TheSamba(dot)com:
Beyond that, TTBOMK there’s not much to distinguish the Cadillac system from D-Jetronic. Same fuel pump, very similar injectors differing only in flowrate and minor physical configuration details, etc. Same analog ECU architecture with components subject to drift (with resultant running problems), same dirty gasoline constantly clogging injectors, same severe knowledge deficit amongst mechanics, same disappointingly poor fuel economy.
That pretty much sums up my thoughts too, but you are a lot more knowledgeable about the Bosch systems than I am.
Sort of, but the Bendix IP was licensed to Bosch, not sold, and the license involved a technology sharing agreement that gave Bendix and Bosch access to each other’s EFI technology and manufacturing developments. Bendix didn’t have to buy it back because they never sold it, although until the Cadillac deal, they had been more or less sitting on it because Bendix management didn’t believe there was enough of a market to be worth any substantial investment.
GM’s digital throttle body fuel injection was a step up from this but still was problematic in its day. It was frustrating chasing check engine lights, counting blinking lights to try and diagnosis problems only to find out the problem was programing in the system and you couldn’t fix it. whentThe ability to reflash or download updates arrived that was a real improvement.
Another odd one was the Ford Throttle body system on my ’84 302 T-Bird. All you had heard about starting a car with fuel injection was “do not touch the throttle when starting”. Well my ‘Bird didn’t start very good when cold. Turns out the ‘Bird needs to have the throttle floored before cranking to set the fast idle. Other than that odd thing it was a great running car, a bit Dr Jekyll / Mr Hyde on fuel economy 15-16 around town and 25-26 on the highway.
Didn’t have a lot of experience with the Chrysler products. The one odd problem we did see was the throttle body system on the V8’s. The actual problem was the intake manifold design. The intake manifold had a large plate on the bottom side of the manifold and as the engine aged the gasket sealing the plate on the bottom of the intake would fail and get sucked into the intake. Massive oil consumption, vacuum leak, fouled plugs and if they ran it long enough fouled catalytic converters. Oops
All in all tho they did get it figured. Fabulous mileage, smooth running, minimal maintenance, HORSEPOWER and quite possibly the cure for making turbo’s work.
Having owned or driven more than a few of the early 1980’s fuel injected Mopars, General Motors and FoMoCo cars; I found the FoMoCo TBI system the much more trouble free and easier to live with than the competition. The GM units had off and on, questionable reliability and dayum expensive for parts and labor to “fix” them. Mopar’s V8 Imperial system seemed to be immediately DOA when leaving Highland Park.
A Ford dealer mechanic alerted me early on to pumping the gas pedal once to “set the choke” on their TBI system. No drive-ability issues when I remembered to do that.
Yup, Ford used the standard electric choke coil to engage a fast idle step for warm up on the CFI system. Not a bad idea since it allowed them to use existing carb parts and tooling and allowed for a fast idle that slowly came down vs the early Fidle valves used on Bosch systems that gave only one high fast idle speed which led to far higher idle speeds than needed in moderate temps to be sufficient at the coldest expected temps. See Audi unintended acceleration.
There was no counting blinking lights in the DFI system which is the genesis of today’s modern OBD system. It gave us access to live data, sure you could only look at a single parameter at a time but none the less it made diagnosis much much easier.
Sorry, what kinds of vehicles did you see this on?
You’re talking about fuel injection in general here, not TBI, though, right?
That was the 318 and 360 Magnum engines that had that Plenum Plate on the bottom of the intake manifold. That plate covers the majority of the bottom of the intake and is made of roughly the same steel gauge as a valve cover and is secured by 15 small bolts and had a thin paper gasket from the factory. Needless to say that gasket would degrade over the years and create a vacuum leak into the lifter valley. The first one I ran across in the mid 90s was when I pulled the oil filler cap off while the engine was running for some reason lost to time and the engine idled up considerably!
There are much thicker aluminum plates made now to replace the thin steel factory item that are supposed to give much better sealing.
Okeh, thanks, that makes more sense. XR7 looked like he was saying the Mopar TBI V8s (which would’ve been 318s and 360s plus the 239 V6 in ’88-’91 Dodge trucks and vans) had this manifold bottom-plate problem. Their intake manifold was as poorly engineered and cheaply made as the rest of that miserable system—crude, halfassed Holley hardware sloppily driven by crude, halfassed Chrysler firmware—but there wasn’t a bottom-plate situation with it.
GM’s digital throttle body fuel injection was a step up from this but still was problematic in its day.
Sure there were some teething problems, but the GM TBI became pretty bulletproof once it became more ubiquitous. When I work at the GM dealership, GM trucks with TBI engines were our bread and butter. They were pretty rock solid, and many made very high mileages with little issues (at least related to the EFI). That said, power, performance and fuel efficiency weren’t its strong suit.
VinceC, your comment here reminds me of your comment here.
I forgot about that comment. Actually it reminds me when I read your Crapiece COAL, as soon as you mentioned lack of power on its 350 TBI, my first thought was timing. That was a really common fix on those engines when a customer complained of poor performance.
Any guess how/why timing got retarded? Sounds like common issue, what made issue common?
Regarding the DS status of this system. Many years ago I had a discussion with the then head of the Montessori school my kids attended. Not sure exactly how he got there from where he started but he did share that his first job out of College as an engineer for the Rochester Products division of GM. He noted that he “knew” it was the beginning of the end when GM resorted to building these injectors under licensee from Bosch. They are a direct interchange which is nice for those people who want to do an aftermarket port EFI for their Olds or Cadillac engine.
It was open loop because the O2 sensor did not exist when they started designing the system. Yes it did come to market during the time the system was in use, but too late to make it worth while to do a complete redesign, considering it wasn’t necessary to meet the emissions standards of the time. That of course changed as standards got more and more restrictive and soon the O2 sensor was everywhere.
Setting aside the merits of this argument (automakers buy a lot of components from out of house!), this isn’t what happened. Bendix went into the Cadillac deal not having any injector manufacturing capacity at all. They expected to buy injectors from Bosch, but Bosch was too heavily committed to its own customers, so the injectors, at least for the first couple of years, were supplied by Nippon Denso and Diesel Kiki, the Japanese companies to which Bosch had sublicensed the EFI technology.
Not true — Bendix was well aware of developments in closed-loop feedback control (they had access to all the work Bosch was doing in this area, and Bendix engineer J.G. Rivard, who was involved in developing the Cadillac system, published a paper on closed-loop control in 1973), although it was very new technology that hadn’t yet been used in production.
Not exactly. For the early systems, Bendix was scrambling to produce them on time at all, let alone making any major changes to them, but Bendix engineers gave a presentation at the SAE conference in February 1975 proposing adding closed-loop control, pushing its benefits in idle control and fuel economy as well as emissions. My guess is that Cadillac didn’t bite — the EFI system was already costing them a LOT of money — and there were probably supply concerns as well as questions about sensor durability.
Every time I see an article on the domestic car companies activities in the seventies and
(most of) the eighties, I have to imagine the Three Stooges in charge. Excellent article on
yet another blind alley cruised by GM.
Excellent article. I remember it being a big deal when Cadillac began offering fuel injection in the mid-1970s. One question – I recall some early units causing under hood engine fires on Cadillacs. Or am I not remembering correctly?
Thank you kindly. You remembered correctly, there were fires associated. The o-rings on the injectors would fail causing a fuel leak leading to a fire.
O Rings are such a challenger.
I wonder if many were converted to carbs when GM support gave out. Maybe the Eldo convertibles would be worth the trouble.
I see what you did there!
#DeftlyDone
That’s really interesting. Seems like one of the easier things to get right, doesn’t it? I don’t know of any such problems with D-Jetronic…
I agree. If GM couldn’t even get the O-rings right, what else did they skimp on.
And Ralph, the Challenger reference is awesome!
D-Jet – Temptation, when presented with fuel leak, to leave the vehicle isolated & running. Let it brew up & burn down.
Excellent.
Thank you.
I had a couple of f/i early Datsun Zs. They were very reliable and fuel pumps and injector hoses were easy to replace if they started leaking. I had these when they were almost twenty years old and they weren’t a problem. My ’94 Cadillac Northstar used an electric throttle motor for idle speed, which did give some trouble, That car was only three years old when I got it. My ’96 4.6 Mustang, and ’97 5.0 Explorer, were already modern systems that were pretty much trouble free. I messed around with my FIL’s 450 SEL and I was amazed at the analog system. The throttle body sat in the middle of the manifold like a carb and a fuel distributor looked like an ignition distributor to direct the fuel to the injectors. There were several vacuum modulators to fine tune fuel delivery. It looked pretty complicated to me. Now every motor is fuel injected and pretty trouble free.
Cosworth Vega was FI also. Same system?
Yes, the Cosworth Vega used a Bendix EFI system as well, and it was very similar to this Cadillac EFI system.
The FI may or may not have been a deadly sin, but the expression “the death of a thousand cuts” comes to mind.
My late father’s last Caddy was a late 1975 Coupe de Ville with fuel injection. Had to replace the fuel line on the frame rail once & several of the sensors. The car made it to over 140,000 miles & ran when we sold it in 1993. One of the cars I learned to drive on.
Wonderfull writings as I deal with 2 of these early boxcar Sevilles was able to deal with most ! but at time wen I found a guy that rebuilt computers he did All for a $100 ! But this was the only one he couldn’t ! and wanted $400 And this was the only one he could not guaranty ! so I quit ! sold them ! but in previous writings no one said what a Great handling car this was !! I had more fun embarrassing the wealthy ones in snooty expensive foreign cars who could not drive them the way they were meant to be driven !! No one mention its a 4 dr Camaro / Firebird underneath ! so I tighten up things urethane & etc … ran 38 psi. got rid of the wire wheels & white walls ! put 59 chevy dog caps on ! lol if wen I drove slow from the beach I had people come up on me & back off thinking it was an unmarked as in NY at the time trooper cars were brown & it looked similar to the fury / diplomats to the non car people ! lol this only came alive around 45-50 Aah the memories !! Now have a 24valve 2000 Taurus wagon another great sleeper & great handling car !! but another nitemare to work on !! lol if only Tesla made a wagon version as it said to be faster & better handling than a corvette !! Thanks for all info.
Very good explanation of how it worked , BUT, if it went wrong how many guys knew how to diagnose the fault and fix it, thats what it was really up against at the time it was new technology and did all the dealerships have a guy who was up to speed with it,
Ive seen that problem recently with a truck repair and dealership I was getting something fixed and just casually asked about the new Hino hybrid trucks and how good they are, very good I was told except only one tech knows how they work and can diagnose them if theres a fault and if your new hybrid truck is off the road revenue will suffer apparently they are far more complicated than a Toyota Prius from the same company.
It must be pretty tough to keep one of these Cadillacs going these days on its original injection system; I imagine that more than a few have been converted to carburetors. What will happen to a 2021 model year car when its electronics become obsolete? Saving the cars of one’s youth will likely be an even bigger challenge, but people can be creative when they want to be.
Nicely done, Vince.
They bothered to swap out the diesels on a fair number, though I never understood why anyone bought luxury cars with them to save a few dollars on fuel. At least FI gave a drivability & power benefit instead of big disadvantage.
Someone in Britain is marketing conversions to EV for ’60’s Jags/Rolls/etc.
There are new electronic fuel injection systems available for older engines with carburetors or aging fuel injection. I don’t see a problem here.
Vince,
It was a nice surprise to see a 2021 article on these systems. There are still a lot of these systems in service today. Some key parts were getting difficult to find/service but there are a few CLC Forum members supporting the systems. One individual is a retired electrical engineer who has identified and found fixes for most of the ECU related issues. I have recreated/ re-engineered the air/coolant sensors to eliminate the problems associated with the originals. Prior to my effort there was a sensor company owner who also owned one of these EFI Cadillacs, frustrated specifically with the poor sensor reliability, his company manufactured replacements for a few years. To clarify for others the air and coolant temp sensors were the same part number. The original sensors had a reputation for going bad; sometimes in less than a year. Especially the coolant temp sensor was vital to cold start. The sensors would often break internally (corrosion and lack of thermal freedom). If they failed open-circuit, infinite resistance, the ECU thought the coolant was extremely hot. When that occurred the car wouldn’t get the enrichment required to start.
All of the original diagnostic trees are published on a CadillacSeville org website. There’s also an EFI manual which was published around 1976. I don’t know if I can share images but the cover is white and blue. Some refer to it as the “blue book”.
The o-ring material was mentioned. As well if someone services the in-tank pump the short length of flexible hose going from the pump to the sending unit hardline needs to be submersible fuel hose. The original line eventually degrades/cracks/leaks and not everyone is aware that the standard for submersible fuel hose is not the same as standard EFI fuel hose.
Someone commented about the Datsun Z car system. Kent-Moore sold a diagnostic tool, J25400, which worked on both the Datsun and Cadillac systems. There was a specific overlay and connector change for each manufacturer (Cadillac/ Datsun). The aforementioned CLC member above commented on the forum that the two systems didn’t share a lot despite the common Kent-Moore tool however but I thought it was worth mentioning. I’ve never obtained a Vega temperature sensor but I understand they are different. I read through some topics on a Vega forum and it sounded like the sensor interaction part of the system was misunderstood but I’ve not seen information comparable to the body of information published on the Cadillac system.
Thanks for your comment and I appreciate some insight into how CLC members have addressed the EFI problems. I was aware of the material service documents on the website and reviewed some of them prior to writing this article. My comment on lack of information was in reference to any historical information about the EFI today. It’s great to hear that there are enthusiasts who are keeping these old EFI Cadillacs running.
Oho! I missed this bit last July. You’ve jogged my memory and almost certainly explained a thing.
In the summer of 1995 I was an intern on the investigative news team (the “I-Team”) of Denver’s № 1 television newsroom; that was KUSA 9News. One of the stories centred round a woman who was being harassed by the state over her 1980 Cadillac, which she had owned since new: It always passed its emission tests, but then the testing was made more stringent: not only did the exhaust have to put up acceptable numbers on the sniffer, but there was also a more detailed equipment check.
Her car flunked the enhanced inspection; the state refused her a sticker because no Check Engine light lit up on the dashboard when turning the ignition on. Told her she’d have to have the light fixed before she could get a sticker. Problem was, the car never had a Check Engine light. She had a Cadillac dealer attest in writing that the car was not built with a Check Engine light. In response, Colorado said she’d have to fix the light before she could get a sticker and renew her plates. She got a letter from GM stating no Check Engine light. In response, Colorado said she’d have to fix the light before she could get a sticker and renew her plates. She had a state-certified emissions repair shop take apart the dashboard and attest that there was nothing to repair and no evidence the car had ever had a Check Engine light.
I don’t remember how many rounds this fight went or how it wound up working out. I don’t know which 1980 gasoline-powered Cadillacs did and didn’t have Check Engine lights, either. And I don’t know for sure that the analogue EFI system described in this post was never commercialised with feedback mixture control, but I think it was not. And if I’m right about that, then it seems likely her Cadillac (Seville, maybe?) might’ve been one of those “some availability in 1980” cars.
(Me, I might have quietly rigged up a Check Engine light connected to a binary oil pressure sender so the light would come on with ignition-on, then switch off promptly on engine startup)
Hi Daniel,
The women you are referring to may have a California emissions 1980 Cadillac. For CA emissions cars ONLY the 1979 version, of at least the Eldorado, was carried over to 1980. The CA emissions cars were the only cars where the 5.7L was carried over. That would make it different from almost any other 1980. I just checked my 1979 FSM and I can confirm that none of the Cadillac dashes have a “Check Engine light”. All of them have a “Stop Engine” though. It seems like any reasonable authority could say that counts without admitting they are incorrect. If you have a FSM for 1979 the info can be found in section 8B on pages 16 through 21.
It sounds like nobody within the state was actually reading the information; simply replying with the standard response. You are correct these “70’s EFI” systems had no feedback, there was no O2 sensor, but as I know they were quite clean if running properly.
One other interesting thing is the E-Body Eldorado new smaller generation started in 1979. Both the 1979-85 and prior generation were FWD. The “Bustle back” Seville, which launched in 1980, however was the first FWD Seville. It is a close cousin of the Eldorado but wasn’t launched until the 1980 model year (one year after the Eldorado). I don’t know which generation of Seville was used in California for 1980. The Seville would have theoretically had the same issue as the Eldorado but as I know the FWD Seville never used the 5.7L so Cadillac would have needed a different solution.
What an interesting story Daniel. I suspect that that woman may have had a 1980 Cadillac with the 350 EFI Olds engine. As 79 Eldorado posted above, the 1980 Cadillac Eldorado with CA emissions carried over the Olds 350 EFi engine. The 49 state models used the 368 with Digital EFI. The Seville came standard with the 350 Olds, but could be upgraded to a gasoline engine for no cost. In 49 state configuration this meant the 368 with Digital EFI, but like the Eldorado for California it was equipped with the 350 Olds with analog EFI. The 1980 Cadillac brochure essentially states that the analog EFI doesn’t have the check engine light when it states the following:
“A computerized system that helps eliminate guesswork from servicing. (This advanced engine is standard on Eldorado, except California…available at no extra charge on Seville).”
Here is the 1980 Cadillac engine availability:
The 350 EFI Olds engine which GM claimed was “produced by GM- Cadillac division” in this chart you show. Same ol’ GM, fulla laffs!
The word “produced” is doing a lot of work there, but it does raise a more philosophical question: There were a number of instances of one division supplying a version of its engines for exclusive use and final assembly by another division. How much divisional content or work does an engine have to have before it’s considered that division’s engine?
For instance, the 1964–65 Pontiac Tempest came with a 215 cu. in. six designed and manufactured came from Chevrolet — it was a 194/230 derivative with a unique 3.75-inch stroke — and then sent to Pontiac for final assembly. (This wasn’t yet the OHC six, which had a completely different Pontiac-designed head.) Was the 215 “produced” by Chevrolet or Pontiac? I’d probably say “Chevrolet,” but the opposite conclusion isn’t TOO absurd, and while it was obviously a Chevrolet design, the fact that you couldn’t get a 215 cu. in. six in a Chevy II or Malibu did provide some cover for the idea that it was a Pontiac engine.
When ever GM introduces new technology, there should be a warning label stating: “Let the Buyer Beware”.
What an interesting thread of comments on an interesting article! Thanks to all involved.
I was surprised by the 35 hp drop from the Olds 350 to the Cadillac 368. Anyone know why?
The arrogance of GM, Cadillac in this case, never ceases to amaze me. The “not invented here” syndrome. But I also don’t understand why all the references to D jet. Yes, it was similar, but it was 60’s tech, by the mid 70’s it was Bosch L jetronic which soldiered on for another 20 years or so. But my god, yeah, it was a basic pushrod engine, but even with FI they only got 180 rated HP out of 350 inches? I know smog was difficult, which is surely why they played, and I mean played, not used, EFI, but my god, that was down in flathead territory in terms of output per cubic inch. Cadillac was a premium brand, maybe they couldn’t spend money on tech on a Chevy, but for their flagship? Yeah, I know, they were a grandfathers car, boy were they, but even old guys like a little punch under the hood. Just think if they’d actually done a little work, used Bosch L jet and gotten 220HP out of their 350. That would have been hot stuff back then. Not to mention a modern system instead of one that was already being abandoned by other manufacturers.
Because the Cadillac system is very similar to D-Jet, as described in the post and comments—much closer than to any other system.
Bosch licensed Bendix EFI technology to produce D-Jetronic, and the license agreement gave Bendix (which hadn’t manufactured any EFI systems since the old Electrojector) access to Bosch technology. There was a clear relationship between the systems, which, given the contractual arrangements between the companies, was neither coincidence nor plagiarism.
In the seventies, Bosch was building three distinct fuel injection systems. The L-Jetronic system ended up being the most successful, but they were also doing a lot of business with K-Jetronic, and some of their customers and licensees remained interested in D-Jetronic into the eighties. For instance, Toyota used both D-Jetronic and L-Jetronic, built under license by Nippon Denso as EFI-D and EFI-L; they insisted MAP metering was preferable for some applications, and adopted EFI-D for some engines after they’d been using EFI-L for years. There was not yet any clear consensus. (The earliest production cars with closed-loop injection used K-Jetronic!)
Apologies if this is a stupid question –
With all the apparent demand, has anyone ever produced repro versions of the Rochester 1957-1965 fuel injection for small-block chevy engines; and if so, why not? Seems there would be a ton of demand to make a performance engine with period-correct technology, without spending a ton for an original unit, assuming one could be found. Counterfeit concerns would be dealt with easily by adding identifying marks.
thanks in advance –
It’s not a stupid question, but the answer is likely to deflate your balloon. The second problem with this idea is that particular period-correct technology was severely and constantly problematic. This could be worked around, at least to some degree, by making a setup that looks like the ’57-’65 hardware while actually consisting of modern technology.
The first, bigger problem is that to design, engineer, debug, produce, and support this what you describe would cost an enormous amount of money. The price of such a system would be extremely high, and there would be few takers, which would drive down production volume, which would further increase the price. This is the death spiral that kills most ideas like this one. It’s what led me to pen this haiku on the Slant-6 forum in one of the innumerable discussions about a new-design cylinder head:
If only we could
Make aluminum heads by
Talking about them
Thanks for the response – makes sense.
Gentlemen,
Your posts are one and all valid. The incidents of technological failings typically decreases with time, with the exception of planned obsolescence. You and I will age and have experience physical failure. The same holds true for any material object regardless of age. Granted, newer technologies are more inclined to fail at a higher pace than those with established non-planned histories of obsolescence. I submit that given the rigors of automoblian service, innovations like electronic fuel injection will have initial as well as historical deficits.
VinceC and Paul: I found recently that there’s some useful clarification of the relationship between Bendix and Bosch electronic fuel injection in Case in Competitive Strategy, a 1983 business text by Michael E. Porter (starting around p. 300 of the 1983 Free Press edition).
As is sometimes the case with business-specific texts, I think some of Porter’s comments on the technical particulars (which are not the main focus of that chapter) should be regarded cautiously — the SAE papers are more useful in that regard — but that chapter discusses the business relationships in some detail.
That chapter also discusses at length how the Cadillac deal came about and why so many things went wrong.
I looked over Porter’s Bendix chapter in more detail, and it’s very clear Porter is not very well-versed in automotive engineering or the longer history of fuel injection — one only reading that chapter could come to some very erroneous conclusions, and it appears Porter may have read a couple of SAE Technical Papers from Bendix and misunderstood some of their points. However, Porter is a highly respected Harvard professor who had direct access to Bendix personnel and records in writing that chapter, so the business stuff is probably reasonably correct.
From that, it’s possible to address some of the questions and confusion raised in the article and comments:
– Bendix licensed, but did not sell, its EFI technology to Bosch in 1967.
– Bosch D-Jetronic was indeed essentially a modernized and refined version of the original Bendix Electrojector system.
– The deal with Bendix included a technology-sharing agreement, so any resemblance between the seventies Bosch system and D-Jetronic is not coincidental.
– Bendix had managed to get themselves into a weird corner where they had wide-ranging exclusive rights to a technology they weren’t using, had almost no production facilities to manufacture, and had routinely gotten cold feet about making any significant investment in.
– Bendix went into the Cadillac deal (which only happened because Bob Lund and Ed Cole agreed to some rather extravagant demands) with a guaranteed production commitment, an 18-month lead time, and basically no production facilities.
– Bendix had to do a mad scramble to find the various components: They apparently expected that they would be able to buy a lot of components, including injectors, from Bosch, and then discovered that Bosch couldn’t spare the capacity, while also not having allowed enough lead time for most other suppliers.
– They decided to make their own ECUs in a plant in Virginia whose management and staff had any particular experience with complex electronics and which wasn’t set up for the kind of manufacturing environment needed for complex, relatively delicate electronics.
– Bob Lund had wanted the EFI system to be standard on the Seville, and Bendix had talked Cadillac out of offering a carbureted version of the Olds 350 for the Seville, apparently concerned that Cadillac would get cold feet on the EFI system and leave Bendix holding the bag. So, Bendix was under a lot of pressure to meet their initial production commitments, resulting in a mad scramble.
– The initial cost of the system to Cadillac was very high: The 1976 projection was $327.85 per system.
– The design and manufacture of the system appears to have been almost entirely Bendix, although obviously Cadillac would have been heavily involved in developing the air-fuel maps.
– The phrase “learning curve” appears frequently in Porter’s account, which seems like a good summation of the situation and a substantial contributor to the system’s checkered reputation.
A revealing quote from Automotive Group president Douglas Crane:
Bendix engineer Jerome Rivard also noted in a 1974 technical paper on the system:
Also, Porter’s account suggests that Cadillac’s initial interest was indeed focused mostly on the Seville. The likely reason the Bendix EFI system was also offered on other Cadillac models was that Bendix insisted on a guaranteed production commitment of 60,000 units per year for five years. First-generation Seville production never hit 60,000 units for any model year; the closest it came was in 1978, with 56,985 units, and a fair number of those had the misbegotten Olds diesel.
I own a 1978 Seville 5.7L gas Bendix EFI model. It has ran flawlessly over the years of it’s original purchase. 1 fuel pump replaced because the original owner ran it out of fuel. I have heard the false “same as a Nova” claims/the faulty non repairable EFI claims. Unless it is a faulty connection, wiring frayed,shorted parts, etc. These are pretty much sensor/actuator plug and play repairs for this system, in comparisons to Ford Carb EEC 1-3 and Chrysler Lean Burn carb attempts. The actual EFI was miles ahead of that game. These cars originally ruffled a lot of barge owners feathers because they were the smaller version for the top price of any Caddy and any other US built car of the day. Those days are long gone. Gm was also dealing with the 5.7L D blocks (Diesel Fuel) That was a huge disaster. So the gas EFI was a much better option. The repair flowcharts were very primitive compared to CanBus technology of today. But a smart tech of that decade could get one on the road in short order. Most of the car’s EFI issues were mechanic opposition to change, and learning new technology. And pre-training to release was not even close to ample. IE: the same people that complained about an O-ring leaking fuel on an injector had no troubles watching a QuadraJet 4bbl flood over and cause a fire. biggest issue was EGR vaucum leaks undetected by technician error.
Vince nice article…. You got underserved heat from the old schoolers. You can always tell an old GM guy, you just can’t tell them a damn thing. That said, the early Cadillac EFI with the 350 olds was an unmade bed. In the simplest terms, it lacked feedback and your generous observation of drift was too kind.
Eight low-impedance injectors firing into a manifold with no runners. This was switched to the TBI injection for this very reason. Nobody could control those eight injectors. But what owners got was black smoke out the tailpipes with a rich mixture that had no feedback forming a vicious cycle, fowling plugs and throwing gasoline into the oil. This is hardly the exception. All these early EFIs smell like gasoline. No way to control any of it.
The “ECU” was nothing but op-amps. They didn’t even have any memory that could attach an RPM to MAP sensor table. No feedback, no control over the mixture. The coolant temp sensor was a positive RTD. Modern days are negative thermistors with a curve. The RTDs are linear.
The training was laughable. The EFI was connected to a Kent Moore J-25400 EFI tester
It looked like the dashboard from LOST IN SPACE… DANGER WILL ROBINSON… and it was useless. Did it actually check the function of the injectors? NO. Did it test the fuel oxygen mixture? NO. It did nothing.
The Service manual gives broad rules of thumb such as testing with a multimeter. The onliners are always complaining about the coolant sensors. I never saw a coolant sensor failure on one of these cars. NOT ONE.
Think about this… Ethanol gas causes the viscosity of gas to increase. That would require more rail pressure to atomize the fuel or it would squirt like a squirt gun. Since the engine is a runaway Rich Burning smoker, there is no way over time to isolate the many problems that arise. But one thing remains certain… there is NO feedback in this system. There is no Oxygen sensor, no MAP RPM tables to adjust it manually.
And let’s talk about borrowed technology. This was the FIRST TIME Cadillac borrowed an engine from Oldsmobile.
So what are the options? If you are low tech, change to a carburettor and call it a day. If you are high tech then build your own ECU or add a Holly Sniper which is basically a TBI like the ensuing Cadillac systems.
But even if you build your own ECU, which is very doable, you still have to cope with low-impedance injectors in a modern world of high-impedance injectors. These old system were batch fire injectors… Hows that work? there are two banks of injectors group 1 and 2. They fire at 90 degrees, two at a time. These are the companion injectors and cylinders. So they fire in compression and exhaust. This is called wasted spark because only the compression side ignites. There is nothing wrong with this, it works. But it still has no feedback. So the ECU builder must have RPM MAP tables to alters the pulse width modulation of the injectors.
I recommend a broadband oxygen sensor that can feedback directly to the MAP and RPM throttle sensor and reduce the gas flow to oxygen ratio.
I have the luxury of speaking as a man from the future looking back at the arcane Caddy EFI system. And to all the GM EXPERTS… what is the waveform and voltage of the timing magnets on the Cadillac Distributor? None of these experts know. Imagine that the very device that controls the fuel injector groups and the spark are in the distributor and none of these old-schoolers know exactly how it works. Modern fuel injection uses a cam sensor and distributor-free coil system like Ford’s fabulous EDIS. But at the heart of crank and cam timing on these old Old engines is a plastic timing chain. And if that is quality engineering for an EFI system, I will eat the seville op amp ECU in my cereal.
So as I said. If you are low-tech, change the manifold, bag the EFI and put on a four-barrel. The guys that run on a stream of BS can talk about the Holly Sniper but all it is is a TBI so you need a new manifold too. And its not a cheap conversion. There is no reason it won’t work but if you are low tech, don’t BS the crowd. I know you are frustrated but unless you understand EFI then you are behind the curve.
Other high-tech options are Minisquirt 3 and Speeduino. But even with those you still have low-impedance injectors that will need increased resistance.
Another option is to drop an LS engine in the mess and it should bolt to the Turbo 400. Now you have some eye candy and EFI.
Everything in life depends on how you adapt to things that don’t exactly work out as you expected as time passes. You can cry about the good old days but baby they are gone gone gone…
And here is a quote of my own… “Most old dogs learn new Tricks… Most humans don’t!” And remember you heard it first from Hot Rod Lincoln… that’s me.