70 EFI Cad ECU Repairs

[bcroe]

Here is a landmark in my 70s Cad EFI adventures.  Since I got the 79 Eldo
2 decades ago, I have reverse engineered the Engine Control Units that
control the fuel injection.  More than a hundred 70s Cads are on the road
today with an ECU I repaired.  Most of the analog computer components
are pretty generic and still available. 

The biggest problem has been with a few custom design parts that of course
are no longer available at any price.  There have been a lot of burned out
connectors, but this problem will go away once I get every owner to install an
external fuel pump relay in his car (parts: $10).  The "how to" article was in
The Self-Starter half a decade ago.  The second big problem was constant
failures of the MAP sensor mounted in the ECU.  Though functionally like
modern MAPs, it is completely incompatible electrically and mechanically. 
However I was able to put together a modern design that drops right into
the same spot. 

That leaves one final biggie.  In every 76 EFI Cad, and most 77s, is a big
ceramic circuit board.  This 2" X 3" part has 25 pins, 4 custom and
unavailable integrated circuits, and dozens of other parts.  Usually when
a ceramic fails, it can't be repaired.  So the only answer it to take one out of
another ECU.  Like a heart transplant, another must die so one can live. 
A big problem. 

I have been studying the ceramic for a long time.  TODAY (drum roll please)
I finished designing and ordered a circuit board which I hope, will directly
replace the ceramic.  It won't be ready for prime time for a while, but this
feels like rolling down the other side of the hill.  Bruce Roe

[76eldo]

Bruce,

Thanks for spending your time on coming up with new innovations for these cars.

I have an 80 Eldo that starts and runs just fine but who know when the day will come that the ECU decides to become a problem.

Please keep us informed.

Brian

[bcroe]

      Bruce,
I have an 80 Eldo that starts and runs just fine but who know when the day will
come that the ECU decides to become a problem.   Please keep us informed.
Brian   

Brian, The analog ECU was used on all Cads with the Olds 350 engine. 
Most 80s used a 368 with the next digital generation ECU, but a few
350s were sent to the left coast because the old system HAD BETTER
EMISSIONS.  Bruce

[Scot Minesinger]

Bruce,

Thanks for doing this, understand it is a quite a sacrifice.  It is really cool when an older product functions as new.  Please make all the records available after you get it figured out.

If I was not still in a very demanding career, might be willing to assist in documenting it.  Sounds a lot like a one man project anyway.

Again, thanks for your dedication to the Fuel injected 1970's Cadillacs!

[TMoore - NTCLC]

Bruce -

This is INCREDIBLE news.  With 5 of these cars in the garage, and the limited supply of spares that you have vetted for me, I have been sweating the day when these cars cannot operate as originally designed.

Thank you for all of your efforts.  When you are ready for beta testing, sign me up and I will happily (and with confidence) install one of the conversions for testing.

Tod

[TJ Hopland]

Cool news.   I'm sure it was quite a project. 

I'm guessing when the 80's and newer ones start failing the 'fix' will be a current ecu with the special programming and plugs to fit the existing system.   I would imagine the 80's will have a few parts that could be repaired but most components would be like the 'ceramic' that Bruce had to reverse engineer in this case, just there will be dozens of them instead of a couple.

[bcroe]

Cool news.   I'm sure it was quite a project. 

There is an issue looking for a mechanical type.  I have figured out how to
convert 76-77 Seville ECUs to substitue for the extremely in short supply
76 500 ECU.  Problem is the mounting brackets are different, so someone
needs to volunteer to make some 500 brackets that can be attached. 

I'm guessing when the 80's and newer ones start failing the 'fix' will be a current ecu with the special programming and plugs to fit the existing system.   I would imagine the 80's will have a few parts that could be repaired but most components would be like the 'ceramic' that Bruce had to reverse engineer in this case, just there will be dozens of them instead of a couple. 

There are some companies out there dealing with the digital ECUs for now; I
don't know of any that can fix the earlier analog.  The newer the car, the more
complex the electronics with proprietary software & customized chips.  I predict
that there will come a time when our current later models are wiped off the map
by failing electronics, but old time cars of today will still be on the road.  My
bumper sticker...

Scot, there is more time after retirement, but not as much as you think.  I am
making arrangements with Lars Kneller to get everything archived, but it may
take a bit longer while things get settled.  At work I sometimes got a reverse
engineering project, because my past includes analog vacuum tubes and relays. 

Tod Moore, if you have a 500 car or 76-77 Seville that gets driven, I would
be glad to set up some road time on all developments.  Do I have your email? 

All the parts might get here this week, and then testing.  Bruce Roe #14630

[TMoore - NTCLC]

Mine are all 1979 Sevilles, and one 1980 Eldorado with the 350 Olds  Analog Fuel Injection - you and I have traded numerous e-mails over the years, and I have sent a few ECU cores to you to cannibalize as needed.

Keep us all posted on progress -

Tod

[Jeff Rose CLC #28373]

Bruce,
Would you please, in a paragraph or 2, explain how the fuel injection worked in the early cars. I am just curious because I admit I don't know...... Throttle body, multi-port? What paramaters were measured and whar was adjusted by the "computer?"
Jeff

[bcroe]

I don't think much of anything preceding the Cad system.  On alternating crankshaft
revolutions, it fires 4 of the 8 injectors, onto the back of the most recently closed intake
valves.  The pulse width is adjusted to match fuel to the volume of air going in per
Manifold Absolute Pressure.  Since density of air varies, there is temp compensation. 
The mixture gets richer approaching WOT. 

The Cad ECU doesn't control ign timing; later systems do.  Only the last year had
combustion feedback using an OXygen sensor to fine tune the mixture.  The 70s Cad EFI
controls the fuel pump and extra air for cold high idle.  The idle speed control is open
loop; later EFIs actively adjust it.  Opening the throttle generates extra injection pulses
like a carb acceleration pump.  Starting sends in a lot more gas; flooring it cuts off fuel.  There is also an EGR valve control output.   More on the system is on my PHOTOBUCKET. 

     http://s93.photobucket.com/albums/L71/bcroe/

  click on an Album
  click on a picture to enlarge + description    Bruce

[Jeff Rose CLC #28373]

Jeez Bruce,
What an amazing setup you have. 
Thanks for sharing it.
2 questions:  Do the injectors fire into the intake manifold or the head, and when it fires 1 bank and then the other, does it fire all 4 at the same time?  Not the best timing wise, but seems like it would be easiest given the technology of the time.  Is there a crank position sensor or does it just pull its info from somewhere else..... OK, that was 3 questions.
Thanks,
Jeff

[bcroe]

Jeez Bruce,   What an amazing setup you have.  Thanks for sharing it.
2 questions:  Do the injectors fire into the intake manifold or the head, and when it fires 1 bank and then the other, does it fire all 4 at the same time?  Not the best timing wise, but seems like it would be easiest given the technology of the time.  Is there a crank position sensor or does it just pull its info from somewhere else..... OK, that was 3 questions.  Thanks,   Jeff   

Jeff,
The injectors are mounted at the edge of the intake manifold, but I believe
the orientation and pattern are chosen to pretty much hit the back of the
intake valve.  When guys just drill a carb intake for injectors, that engineering
isn't done, so I don't trust it, or even just changing injectors which may have
a different pattern. 

When the fuel is injected against a hot valve that has just closed, there is some
time for it to vaporize there.  At high rpm there isn't much time, but it matters
less.  The best systems use sequential injection to fire each injector at the
optimum time, but Cad found firing 4 at a time was close enough back then. 
Its done using a pair of "cam sensors" which are the (misnamed) "speed
sensors" on the HEI shaft.  I bought a rebuilt 70s HEI that had the magnets
reversed (turned 180 degrees on the shaft), so the engine probably would
not idle as smoothly.  You can check this using one of my diagrams, on
the bench. 

You can't do this with a CRANK sensor, because it doesn't know which 2 of the
4 strokes a cylinder is on.  Crank triggered ign works by firing 2 plugs at the
same time: one will always be ready for combustion; the other is finishing
exhaust and doesn't care.  Some EFIs do that: fire every revolution because
they don't know which is which.  Most people who add EFI have no idea
what mode it injects.  I made a special harness for my 79 Eldo, and added a
second set of sensors, so many other ECUs can be plugged in.  I did run it
in sequential mode, but there were other issues I didn't fix at that time. 
Bruce

[Jeff Rose CLC #28373]

Thanks Bruce.
Jeff

[bcroe]

  In every 76 EFI Cad, and most 77s, is a big
ceramic circuit board.  This 2" X 3" part has 25 pins, 4 custom and
unavailable integrated circuits, and dozens of other parts.  Usually when
a ceramic fails, it can't be repaired.  A big problem. 

I have been studying the ceramic for a long time.  I finished designing
and ordered a circuit board which I hope, will directly replace the
ceramic. It won't be ready for prime time for a while. Bruce Roe

UPDATE.  The PCB boards showed up and I built up a couple for test.  What
they showed, is that my efforts here are on track.  Everything fit and
functioned, mechanically and electrically.  This prototype isn't the final
version, but its getting closer.  Remaining is to design a high side dual peak
and hold injector driver, which would work not only to replace one on the 76
ECU ceramic, but also for the 4 later years without the ceramic.  Lets see
how far this gets before summer projects get underway. 

As a side benefit, the better understanding of these functions has brought a
whole extra page of explanation to my documentation & drawings, of each
of the 5 years of the 70s ECU designs.  Of course, that slows down the info
archive project.  None of this has been updated onto my PHOTOBUCKET yet. 
Bruce Roe