News:

Due to a technical issue, some recently uploaded pictures have been lost. We are investigating why this happened but the issue has been resolved so that future uploads should be safe.  You can also Modify your post (MORE...) and re-upload the pictures in your post.

Main Menu

70’s ECU Wiring: Bruce Roe kit install on a 79 Eldorado

Started by 79 Eldorado, December 02, 2018, 09:06:38 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

79 Eldorado

It was warm today so that made it a good day to install the relay and wiring kit I bought from Bruce. I wanted to make some observations while the install was fresh in my mind. It went well other than having to run back into the house for a couple of items I didn’t anticipate needing… like clothes pins! There were a couple of things I wish I had known ahead of time so I will try to point them out for others.

For anyone who is not familiar with the kit it is a preventative measure which protects an ECU from future failure. On Cadillacs from around 1970 through 1979 a terminal at the ECU could easily burn out with the factory solution. Bruce’s kit adds an external relay for the fuel pumps and shares the load for the fuel pumps between two of the ECU terminals as opposed to a single terminal like the original configuration. This is a great kit and great effort Bruce went through to understand the issue.

Going into this I wanted to complete the job still having a factory look to the wiring with minimum splices and wire color changes. That being the case I made most of the connections in the multi-cavity terminal connector itself by unpinning the terminals from the plug (I hope Bruce will approve). This was the first time I tried to unpin this particular type of terminal/connector and of course GM has a new solution for every connector I think I’ve ever seen. For these you can slip a tiny flat head screw driver down a channel made for the locking tab from the FRONT of the connector (Front being the side which interfaces directly with the ECU contacts). Once a terminal was removed there is room to solder on one additional wire but there isn’t a lot of room so if you go that route try to keep the second wire tight to the first wire because there is limited space in each cavity. I tinned the terminal on top of the existing clinched wire and then I used a wooden clothes pin to hold the second wire to the first (I also tinned the second wire before attempting to join them). Be careful to not extend your new wire too far as it may interfere with the ECU terminal in that position.

Tips:
-Remove the screws holding the ECU in place to gain better access to the plugs.
-Terminals can be unpinned with a tiny screw driver from the FRONT of the connector once the connector is unplugged from the ECU.
-For the terminal to fit back in the added wire MUST be added to the side of the terminal where you can see the clinch tabs holding the existing wire.
-When ready to install the terminals back into the plug start them from the back but then PULL them through the FRONT. You can hook the terminal and pull it into position. You will feel it click into place. Keep an eye on the back if you added a wire because if it is very difficult to pull through the second wire is likely catching on the housing. The factory did place two wires in some locations but it’s tight after soldering a second wire.

Suitable ground question:
Finally I completed the job with the exception of the relay ground because I wanted to ask one more question regarding a suitable ground for the new relay. Having been a long term 79 E-Body owner I know that the dash assembly comes out completely and that it is a poor area for grounding. Close to the ECU are two large bolts which secure the right side dash mounting bracket to the frame. I was wondering if that could be a suitable ground. There are at least two “official” ground points but as I know they are on the left side and buried under the dash (wish I knew I was doing this job when I was changing the heater core).

I hope this helps others. I was not aware there was a potential issue but I feel safer with it installed.

Thanks Bruce!

Scott

bcroe

Great!  Does it work?  The 79 Eldo is somewhat unique with a
factory purple wire directly connecting terminals F and H
together, and to a fuse block fuse instead of an in line fuse.  The
California (350 engine) 1980 is probably the same. 

Bruce Roe

79 Eldorado

Once I attach the ground it should work.. In Bruce I trust :) That purple wire does look strange though. Before touching anything the purple goes from "F" to "H" and there is as well a brown wire going to both "A" and "F". It gives the appearance that A, F and H are all in communication. I didn't go back to the Electrical manual to see where the brown wires are going to but being on the same plug and the same color normally means they are related to the same function.

I see that ground is only for the relay but I wanted to be certain there are no special requirements like the ones I see sometimes for fuel pumps. I'd like to find a nice factory looking location for it even if there is no special requirement other than "a good ground". I'll likely do it on Friday.

BTW: I think I've seen more of the 7mm screws with a single stripped thread than any other single car at one time. Although if you are removing the entire dash, as I did with the heater core, that's a good time to encounter them. I found another one when I detached the ECU from the mounting points. I have a box of replacements I bought from clipsandfasteners.com though. They're a great resource for a lot of things like that... They didn't have those CTS bullet connectors as far a s I could see though.

Scott

79 Eldorado

I installed the ground today. I attached to a screw which I placed between the 2 large bolts which attach the right side main dash bracket to the side of the car. I soldered on a ring terminal and used one of the twisted tooth anti-back-out washers under the head. I don't know what they're called but I've seen them several times for ground screws.

The car started with no issues.

Thanks Bruce,
Scott

78Elegante

Just trying to understand what is going here.

The purpose of this mod is to release the ECU of the full current that the fuel pumps draw? Is that correct.

I cannot understand what is the purpose of the second part in step 1 (connecting the 3 inch portion from 'C' to 'J').
'J' is going to the Fast Idle Valve (I assume controlled by the ECU). By connecting 'C' to 'J', the Fast Idle Valve always has +12V from the moment the ignition is switched on (not controlled by the ECU anymore).
Why the "3 inch portion from 'C'" isn't just used to activate the relay? (and directly connected to the coil of the relay?)

Then a small remark of caution.
I think it would be safer to switch terminals '85' and '86' of the relay.
In this application, where the relays supplied do not have a fly-back diode, there is no polarity, but when this relay breaks and is accidentally replaced with a relay with a fly-back diode inside, in the best case only that fly-back diode will burn out.
I think it is better to always use terminal '85' on the 'ground side' of the coil and '86' on the driving side of the coil.
Also see https://en.wikipedia.org/wiki/DIN_72552

bcroe

Quote from: 78Elegante. Just trying to understand what is going here.

The purpose of this mod is to release the ECU of the full current that the
fuel pumps draw? Is that correct.

I cannot understand what is the purpose of the second part in step 1
(connecting the 3 inch portion from 'C' to 'J').  'J' is going to the Fast
Idle Valve (I assume controlled by the ECU). By connecting 'C' to 'J',
the Fast Idle Valve always has +12V from the moment the ignition is
switched on (not controlled by the ECU anymore).

Why the "3 inch portion from 'C'" isn't just used to activate the relay?
(and directly connected to the coil of the relay?)

Then a small remark of caution.
I think it would be safer to switch terminals '85' and '86' of the relay.
In this application, where the relays supplied do not have a fly-back
diode, there is no polarity, but when this relay breaks and is
accidentally replaced with a relay with a fly-back diode inside, in the
best case only that fly-back diode will burn out.

I think it is better to always use terminal '85' on the 'ground side' of
the coil and '86' on the driving side of the coil.
Also see https://en.wikipedia.org/wiki/DIN_72552.

There are a couple objectives, the external relay circuit attempts to
address them both, while still using any stock original ECU.  The first
is to move pump power to the external relay (same as all later design
EFIs do) to avoid the so common burnout of the ECU ctr connector pin
C, or the internal relay contact or its connecting traces.  The second is
to allow an ECU with a burned out out pin C to continue to serve, a way
to sidestep the lack of replacement connectors and the considerable
labor of replacing it. 

The Fast Idle valve IS NOT switched on (continuously) the moment the
ignition is turned on.  My drawing shows that inside the ECU terminals
C and J are already connected together, BOTH the Fast Idle Valve and
the fuel pumps are under control of the ECU.  By connecting them
together externally an ECU with a damaged pin C may now control the
pumps again.  Part of the idea here was that one circuit design fixes
all, no deciding which is needed for this ECU or another later. 

The original ECU relay sustained the kickback from disconnecting a
couple of fuel pump motors.  With the external relay that kickback is
reduced to a tiny amount from the relay coil, and that is mostly
surpressed by the couple dozen ohms of the Fast Idle Valve heater. 
I deemed that acceptable, while avoiding the risky problem of the
owner installing the relay coil (and diode) backwards.  It is another
reason to have the Fast Idle Valve connected to pin C, it was originally
anyway. 

I show and supply a relay with no coil kickback diode.  Having sent a
lot of drawings with PIN numbers AND relay WIRE COLORs out, I
wonder if it would be advantageous or confusing now to reverse my
drawing coil connections?  Bruce Roe

79 Eldorado

Bruce,
As long as there is an explanation with the instructions and a "rev" level with a date on the instructions I don't think it would be confusing regarding the instruction.

The only thing which might seem confusing is I guess that means the white wire (because it is pre-installed in the relay?) becomes the ground and normal with DC circuits is a black wire as you had it. I didn't look very closely at the relay itself before installing, and too long ago to recall, but can the white and black wires be unpinned at the relay and swapped relatively easily? That would keep everything the same color, soldered/paired to their current locations, and only require swapping at the external relay.

Scott

78Elegante

Bruce,

First I want to apologize when I offended you because that is the last thing that I want and for sure that was not the intention of my question and remark.
I just have a question about something that I don't understand (yet) and I see a point for improvement.

Regarding my question:
I should investigate the circuitry of the ECU myself first, because the chance is big that I will figure out what you explained.
At this moment I don't get the point that, internally or externally, terminals 'C' and 'J' are electrically connected.
As I understood, the fuel pumps start running when you switch the ignition. In my world that means that terminal 'C' has to become +12V (to supply the pumps with power and the pumps themselves connected to ground).
So, assuming terminal 'C' becomes +12V with ignition on, this means that terminal 'J' also becomes +12V.
Given the fact that the Fast Idle Valve only has a connection to terminal 'J' and ground, what you say is that the FIV always (with an original ECU and with your mod) gets power when the pumps get power.
That means that my assumption that the pumps always receive power when the ignition is on cannot be correct. But what is the signal that the ECU uses to decide to switch off the pumps? Or to activate or deactivate the FIV?
I don't understand the relation between the fuel pumps and the FIV. And why the FIV should be activated when the pumps are running or why the pumps have to switch off when the FIV is deactivated.
I should measure this first.

I understand your explanation about both preventing a problem to occur and solving a problem after it occurred (given the fact that 'C' and 'J' are internally connected).

Then the terminals 85 and 86. I understand you provide relays without kickback/flyback diode. And if, in the unlikely case that the relay fails, the relay is replaced with again a relay without diode, nothing will happen. But you cannot be sure about this.
That is why I mentioned the problem.

I think the solution from Scott is one. Another could be adding a warning that the relay cannot be replaced with one with a diode is another and changing the design is again another.

I will also install the relay.
Given the fact that my terminal 'C' is not burned out, I can use the simple way of just cutting the wire to the pumps (in a Seville that seems to be DK GRN) and add the relay (and leave terminal 'J' untouched).

Antoon

bcroe

Quote from: 78EleganteBruce,

First I want to apologize when I offended you because that is the last thing that I want and for sure that was not the intention of my question and remark.
I just have a question about something that I don't understand (yet) and I see a point for improvement.

Regarding my question:
I should investigate the circuitry of the ECU myself first, because the chance is big that I will figure out what you explained.
At this moment I don't get the point that, internally or externally, terminals 'C' and 'J' are electrically connected.
As I understood, the fuel pumps start running when you switch the ignition. In my world that means that terminal 'C' has to become +12V (to supply the pumps with power and the pumps themselves connected to ground).
So, assuming terminal 'C' becomes +12V with ignition on, this means that terminal 'J' also becomes +12V.
Given the fact that the Fast Idle Valve only has a connection to terminal 'J' and ground, what you say is that the FIV always (with an original ECU and with your mod) gets power when the pumps get power.
That means that my assumption that the pumps always receive power when the ignition is on cannot be correct. But what is the signal that the ECU uses to decide to switch off the pumps? Or to activate or deactivate the FIV?
I don't understand the relation between the fuel pumps and the FIV. And why the FIV should be activated when the pumps are running or why the pumps have to switch off when the FIV is deactivated.  I should measure this first.

I understand your explanation about both preventing a problem to occur and solving a problem after it occurred (given the fact that 'C' and 'J' are internally connected).

Then the terminals 85 and 86. I understand you provide relays without kickback/flyback diode. And if, in the unlikely case that the relay fails, the relay is replaced with again a relay without diode, nothing will happen. But you cannot be sure about this.
That is why I mentioned the problem.

I think the solution from Scott is one. Another could be adding a warning that the relay cannot be replaced with one with a diode is another and changing the design is again another.

I will also install the relay.
Given the fact that my terminal 'C' is not burned out, I can use the simple way of just cutting the wire to the pumps (in a Seville that seems to be DK GRN) and add the relay (and leave terminal 'J' untouched).    Antoon 

Antoon,

Not to worry, I have had a career of countless design reviews where
peers try to pick a design apart and I try to justify it.  Just the actual
facts, and do not assume. 

Here is the relay fix drawing, and the first page of a 79 ECU schematic
I created.  The entire set for this and all 5 versions 76-80 are available,
rather poor resolution versions  are already on my PHOTOBUCKET. 

Here you may see, IGN power closes relay RLIG which powers up the
ECU direct from the battery.  There is a second fuel pump relay RLFL,
which is operated by a control circuit on the page.  This circuit will only
operate for a second, OR until a starter crank signal comes in from input
START, OR from speed sensor indication of continuous engine rotation
at FUELOFF from sheet 7. 

The reverse engineering thing is never 100% finished, trying to read
the original designers mind.  After investigating my own 79 EFI in the
1990s I started to see other owners ECUs with the burn out problem. 
Eventually I noticed that pin C was always the connector failure, never
pin H.  Perhaps that is because the PC board acted as a heat sink,
and the connector lead (to that heat sink) was much longer for pin C
than pin H. 

The next observation was that the pin C signal was also on pin J, which
never burned out.  That inspired the 2009 pump relay fix article.  The
wire from pin C had to be cut, or the relay would never release.  My
thought was it should not be left flopping around and causing questions,
tying it to J solved that and added a bit of reliability. 

The 70s Cad EFI had a terrible reputation for being unreliable, but perhaps
the the specific issues could be targeted and solved one by one.  This
fix might avoid 30% of all the failures I had knowledge of, if owners could
be convinced to install it.  Scott has already added another solution with
an improved temp sensor. 

I did try to find a standard relay harness for consistent colors, and
put in pin numbers just in case.  My source had a fifth (red) wire which
I remove before sending out, to avoid confusion.  BUT I was unaware
of the DIN 72552 standard, so the coil wiring was not chosen with
respect to a relay with an internal kickback diode.  The problem with
releasing a design, is you cannot take back any errors.  In industry
there is process for revisions, but not to the public.  My thinking here,
is to advise relays with diodes are not to be used, otherwise reverse
the coil connections.  Changing the drawing is easy, but will it result
in confusion in the field?  Bruce Roe

78Elegante

Bruce,

Thank you so much for explaining this in so much detail! For me, as a relatively new fellow HAM, you are the kind of people that I look up to.
It took me some time to read your text, understand what you say, compare what I understand from your text to the schematic and finally totally understand your words.

Basically there are two situations where people will install your relay conversion:
1. When no problem occurred yet and when everything is still working as designed. You do this to 'save' terminal C from future burn-out by moving the full current draw of the pumps to an external relay.
2. When terminal C is burned (and the pumps are no longer powered), to get the car running again without having to buy a new ECU.

Now, the last thing that puzzles me is this Fast Idle Valve and the fact that this thing gets power at the same moments as the fuel pumps.
What is this FIV doing and why can it be switched on or off at the same moments the pumps are switched on or off? Or, the other way around, why the pumps can be switched on or of when the ECU decides to activate or deactivate the FIV?
Why there is a relation between these two? Do you know this? Can you explain this?



bcroe

Yes, the external relay also protects the internal relay and copper
traces from burning out.  But at least those faults are a lot easier
to fix than a burned out connector pin. 

The FIV has a time constant of something like 5 minutes, so any
short on off cycles starting the engine have little effect. 

Bruce Roe

bcroe

Based on discussions and experience, I have recently revised
my Fuel Pump Relay Kit design from the 2010 article.  I have added
a second schematic wiring drawing that more accurately reflects the
different fusing arrangement of the 1979-80 front wheel drive cars,
with the 5.7L gas engine.  In both drawings I have also reversed the
relay coil wires, to be consistent with wiring standard DIN 72552. 

One relay coil wire is grounded, but I wanted to avoid cutting another
wire by connecting to the ECU ground wire, which is actually on the
drawings.  So just a ground lug is shown.  In fact the ECU case might
be used for ground.  Beware that some brackets are insulated and
may NOT be a ground. 

79 Eldorado

Bruce,
I was just checking the diagram again. I was recently working on another project where I was using a diode protected relay and I now catch what Anton wrote. Before looking at your revised diagram with color changes, to me, the simple solution would be to unpin the white and black wires and simply change their positions. The white wire would go to terminal 86 and the black to terminal 85.

There was a good youtube video showing how, in absence of a diode, can result in really high voltage spikes.
"Back EMF High Voltage Spikes on the Oscilloscope – Part 1" by electronicsnmore
https://www.youtube.com/watch?v=43HFng0CVKg

The person who made the video explained that when the field generated by the relay coil collapses a substantial, and potentially damaging, voltage spike is produced without some sort of protection. He then goes through an actual example where he measures the spikes on an oscilloscope. A diode containing relay routes the power from the spike in the opposite direction and effectively cancels the voltage spike. Really pretty clever. On the other project I was referring to I switched to diode protected relays and so I was thinking I might do the same for the Cadillac.

Scott

bcroe

#13
Its good to understand real world problems like
spikes.  In theory an inductive V spike could go
to infinity, but none of the components used in
the real world could do that.  Note in the ECU
internal drawing above, the coil of the inside
fuel pump relay RLFL does use a kickback diode
CR7 to protect the transistor driver Q3.  But
the ignition turn on relay RLIG does not, it is
driven by the extremly robust ignition switch. 

In the case of the RLFL contact controlling the
fuel pumps, a 30A contact was used to control
the pumps, most DC motors have a lot of inductive
energy (more than a relay coil).  That was
operated with no protection diode, causing some
wear from the contact arcing when opening.   

Once the external pump relay is added, the V
kickback to the ECU is reduced from a pair of
motors, to just a relay coil.  So the energy
involved is reduced an order of magnitude.  So
without a diode, it is far better than the
original design.  I deemed that "good enough"
in the interest of keeping things simple, and
avoiding field problems with reversed diodes. 

The pump motors are still kicking back, if a
diode was added, it should go directly on the
high pressure pump terminals.  I did not want to
go there, instead be satisfied with saving the
ECU and possibly needing to change out the cheap
external relay in rare cases.  Keep in mind, the
kickback current is not greater, and the duration is
extremely limited.  So it is no wiring overload,
unlike the earlier problem. 

Another problem with the design, sometimes the
fuel pump relay would vibrate briefly when it
was opening, this does dump a lot more energy
into the motor controlling contact.   And
several of the relays have been observed to be
open circuit or stuck closed.  The interaction
cause was rather subtle, I modify the internal
control of any ECU I see to give single positve
operation.  Bruce Roe