84 SC no start after rebuild
 

A buddy and I rebuilt an 84 3.2 this winter that he bought as a basket case with hacked electronics throughout the car. That said, car drove off the truck and into the garage.
He's successfully done 996 and 997 engines after taking Raby's course but says the 3.2 is more of a PITA.

Finally have engine back in and trying to fire. Car barely coughed. It coughs at first attempt the nothing. Let it sit, a small cough at first revolutions, then nothing. We have spark, we have fuel pressure at the rail, but injectors are not pulsing as evidenced by dry plugs and no gas smell. (I don't think my DVM is sensitive enough to measure pulses) I swapped in my 87 ECU and the car coughs more, it seems close, but #6 plug still dry.

Both CPS and CHT Ohm OK
Swapped in known good DME rely, no difference
ISV is humming
12V to injectors at connectors
Injectors Rebuilt and tested before reinstalled.
Cap, rotor, wires, plugs new.

I am not getting any voltage at ECU connector from Crank sensors. Bentley says I should see variable DC voltage with cranking but I see nothing.

Next steps/ideas welcome.

Set your DVM to AC, not DC. Then you should see a small voltage (around 2V AC) on the engine speed sensor, and a lower voltage (0.2-0.5V AC) on the ref sensor.
If not, then you found the culprit.
But since you said you have spark, these sensors should be OK. How do you know for sure you have spark?

Also check ignition dizzy and rotor are set correctly, as the coughing makes me think of inverted positions or wires.

There is zero transfer of knowledge from the 996/997 engine build to an old 3.2L air cool engine rebuild.

Do this.

1. Check compression
2. Check leakdown of each cylinder.
a. Leak down of Cyl 1 at the C1 TDC position should be less than 10%
b. While in this position, leak down at Cyl 4 should be 100%. Do you get that?

3. Ensure that cam timing is correctly done on one bank, then rotate engine 360 degrees to do cam timing on the other bank

4. Make sure you have the distributor rotor pointing to the right position for Cyl1 when Cyl1 is at TDC.

Obviously there is no fuel since plugs are dry after cranking. I would start with simple checks and leave compression and leakdown tests for later troubleshooting if needed.
Two ECUs have been tested in the car with same results, so that leaves it to something wrong in the car itself. Both flywheel sensors should be tested for AC voltage. Also make sure they are correctly set (0.8mm gap) and that there is no inversion between ref and speed.

I tend to agree here. @ ECUs and 2 DME relays. Confirmed today, Fuel pump running while cranking. Pressure is 33 PSI+-

I believe this means the crank sensor is working as safety feature = FP will not run with no signal? I tried to test the 2 CPS via LED test light but could not confirm also as my DVM has no AC setting. Still no fuel moisture on plug#4 or fuel smell. I'm guessing firing signal to the injectors is not getting through. As just a hired engine assembly hand I'd still have to guess that even way out of whack a cylinder would catch here or there but nothing. Engine is moving a lot of air as measured by exhaust pressures.

If any of the two sensors does not send the correct signal to the ECU, the ECU will not send commands to the injectors nor the ignition. Hence the engine does not start. When you release the key after cranking, power to the fuel pump is not maintained. This is a safety feature.
When cranking, power to the fuel pump is allowed momentarily by the ECU no matter if it sees sensor signals or not. If it sees correct signals it maintains power to the FP. Otherwise it quits and does not allow the FP to run as the engine is not running anyway.

So you say you have spark so the crank sensors are working.
You said you are not getting injector pulse but reading through your post it does not seem you have checked for a pulse signal at the injector plugs with something like a Noid light.
If you have spark, and you have fuel pressure and you have actual injector pulse at the injection harness to individual injectors then your injectors are stuck. See if the car will run on a squirt of ether and if it does you can tap on the injector bodies with a tiny ball peen hammer or something similar while cranking the engine to free up the injectors.

Richard

Yes, that's the heart of it. I have pressure at the rail and can easily apply bypass power to the FP. Even if car only got fuel during cranking, all else being fine it would catch sputter and die. Only reason that I can think of now for there being no evidence of fuel in the cylinders is injectors' not letting it in. Injectors were sent to Dr Injector for cleaning and test and installed upon return. Trying to find what cycles the injectors.

I have 12V at the injector connectors, how do I check for a pulse? I do have and LED test lead with a 630 Ohm resistor build in for testing other cars of the era systems. I'm picking up starter fluid later today and will report back

Bentley says one of the crank sensors is for factory tech diagnosis. The fact that I'm getting power to FP during cranking tells me the other sensor is OK.

Google Injector tester Noid light, your auto parts store might possibly have one.
There are many different styles of terminals so pay attention to that.

Richard

That’s not correct, the two crank sensors on the LS are speed and reference mark signals.
The sensor the Bentley manual is referring to is a TDC sensor used by the factory and the connection port is below the ICV in the center of the engine.

Richard

No, the TDC sensor referred to by the Bentley is on top of the engine case and is not one of the two flywheel sensors. It is indeed only used at factory and never used afterwards.

As I said before, the two flywheel sensors, those that are attached to the same bracket on the upper left of engine, are sending proper engine speed and ref position signals to the ECU. If one of those signals is failing to be received at startup, the ECU does not keep power to the FP. Again, during cranking, power is always sent to the FP no matter the sensors.
The sensors are in no way a condition for sending power to the FP during cranking. They become indeed a condition once you release the ignition key. The ECU senses this by seeing an engine speed of at least 200 rpm at the speed sensor. But for this, it must see this signal as well as the ref signal at startup (the ref signal becomes less indispensable once the engine is running).

If either the speed or reference mark sensor is not working you won’t have spark or injector pulse.

Richard

Thx Richard,

No confirmation of injector pulse from both LED and Noid lights on #4 injector harness. Recall injectors came directly from Dr Injectors service so presumed OK.

We installed New on hand Fuel Pump same pressure as old one ~34.* No fire.


Spray ether, immediately caught, ran for a second died.* Confirmed 3 times.* Using my known good 87 ECU and DME relay.* Bentley seems to show nothing between ECU and injectors. So need to check "input signals"? To those wires. What wiring/circuit do I need to check to see why injectors are not firing?
Thx

The injector wiring is pretty simple. Inside the ECU is an output stage that provides GND on pin 14 and 15 of the 35-pin ECU connector to fire the injectors. The injectors are two-wire devices with an internal coil. One side of the coil sees +12V once the ignition is ON and the other side of each bank is connected to either pin 14 or pin 15 of the ECU.

With the ignition ON and engine not running (duh) pull one injector connector and confirm +12V on each socket of the connector. Both sides with read +12V. If not you might have a blown fuse in the engine bay or some other wiring issue. See here for how things are wired. Pin 14 and pin 15 are internally connected inside the ECU. Pin 18 and pin 35 are powered by +12V as soon as the ignition switch is in ON or START position.

This configuration is called a low-side switch.

http://forums.pelicanparts.com/uploa...1756069566.jpg

Thx, I have that diagram. As mentioned in the first post, I have 12v at the injector harness. (Both connectors) i assume to get a pulse one side gets grounded to complete a circuit and fire the injector. If so, that signal us not getting to the injector.
This side of the diagram is fixed per se. I think I need to know the origin of the signal to ground out one side and pulse the injector as that is not happening.

During cranking the GND signal coming from pin 14 and 15 is only about 1 - 2 milliseconds in duration. A NOID light (an LED in a fancy package) will flash very dimly but it's hard to see unless the environment is really dark.

You may want to get yourself a simple oscilloscope (less than 30$ on evil bay) to make it visible. A multimeter will not help here. If it's not present check the wiring.

Disconnect all injectors and ohm out the GRY/WHT and GRY/BRN wires on the injector connectors back to the ECU connector pin 14 and pin 15 respectively. That's all there is to it.

Something is broken or disconnected, You will have to start backtracking with a continuity test of the wire harness. As a quick check look at the multi pin injection harness plug to the right of the left side shock tower and make sure you don’t have a broken wire or out of place pin.

Richard

Hopefully, this hasn't happened to your DME ECM:

http://forums.pelicanparts.com/uploa...1756101314.JPG

Systems Consulting indicates they received two of these in the last month for repair.

OK so now we know your ECU is OK (since you tried both ECUs and got identical results with the known good ECU from your own car).
We now know your flywheel sensors are OK too, since the engine caught when you tried to start it with ether. We also know or "suppose" your injectors are OK, since they have been professionally serviced recently.
The wiring between the ECU and the injectors is indeed a suspect and should be tested. Seeing 12V on both injector connector terminals is normal when ignition is ON. But that doesn't mean that the signal line (pulsed ground) between ECU and injector loom is good, because the 12V voltage can be seen "through" the remaining five injectors themselves, as they are pretty low impedance and show almost no DC resistance being connected in parallel.

As suggested above by Ingo I would disconnect all six injectors from the loom, disconnect the big connector from the ECU under driver seat and ohm between this connector pin 14 and each injector connector on cylinders 1, 2, and 3. Do the same between connector pin 15 and each injector connector on cylinders 4, 5, and 6. For one of the pins of each injector connector your ohm test should show good continuity. Less than one ohm, not taking into account the test leads that must be pretty long in this case since you must ohm between engine compartment and cabin.

Another easier test that doesn't require extended test leads would be to leave all six injectors connected in the engine bay and simply work at the ECU big connector under the seat, of course disconnected from the ECU. Don't switch the ignition on. Now from the ECU connector, harness side, you have access to both injector loops. That is, you can ohm test between pin 14 and pin 18 (pin 18 is battery power, but 12V is not present here as ignition is off, and this pin is permanently tied to the +12 leg of injector circuit). Do the same between pin 15 and pin 18.
In both cases you are measuring the total resistance for 3 injectors connected in parallel, plus the line resistance for these 3 injectors. That should roughly be equivalent to one ohm or less, excluding the ohmmeter test leads.

If these ohm tests all pass, then there is this rare but real possibility that something is wrong with the six injectors together, that makes the ECU injector driver stage "flatline" because of a too low resulting inductance. That has already been shown in past threads where the engine wouldn't start on six injectors but would run on 5, and reconnecting the sixth injector would make it die again.

Also, a shorted injector might prevent the other five to receive any signal, since the ECU will also flatline in this case. You may want to try disconnecting them one by one and see if the engine starts on the other five. Event disconnect two at a time and crank to see if the engine starts on the other four.

Lastly, I can't remember if you have tested your CHT, but that can also prevent the engine from cold starting if it is shorted.

Thanks for the step by step. Unfortunately probably the weekend before I can get back to it. CHT is new and I can try the other tests mentioned as well. Kostal connector for injectors is seated properly at firewall but as mentioned PO tore through lots of wiring when converting the car to track use so it is all suspect.
I'm very heartened by the engine catching with starting fluid. I feel we are really close. Someone mentioned that if one injector is shorted, none will fire. I disconnected 3-6 and tried to fire, nothing. I can try the same on the other side but likely unnecessary if I test each alone. It's a little tough to get test leads firmly on the mid and firewall cylinders.

Without having a scope to analyze the injector signal complicates the troubleshooting process.
A simple test while using a node light, is to disconnect the temp sensor. This will enhance the brightness of the injector signal significantly.
This is done first with the injectors disconnected. If there's no improvement in brightness, there maybe a short in the injector wiring.
If the brightness improves, you can connect & test each injector one by one, while using the node light.

Hopefully, you've tested the resistance (~2.5 ohms) of all the injectors before installation?

Just to make sure that you have spark and valve timing correct, spray some easy start into the intake while someone turns the key, if the engine starts and runs, then you know its definitely fuel/injector related.
Its a simple test, and will eliminate a multitude of checks.
Ant.

Ant, read post #14 on previous page. OP has already sprayed ether at air intake and successfully got an engine catching for a few turns. So we know at least he's getting spark, meaning sensors are OK. As for the valve timing, that can be checked later on when he gets fuel. Right now only fuel is missing. He's already tried two ECUs, among which his own known good ECU from his own car (he's helping on a pal's car if I remember). Injectors are newly cleaned and checked. So what's left, wiring between ECU and injectors, and possibly one or two injectors electrically out of specs that could make the ECU flatline. This has already been seen in previous threads. Can't find the old thread where I read this, but the car wouldn't start on six injectors and would start on five, then die again when reconnecting the sixth injector. I mentioned that on previous page.

Ahh, my bad, I did go through the previous page briefly, but obviously missed that, I'll search the old grey matter, and see if I can come up with something useful to add.
Ant.

Ok, how about getting a test light, remove all six injector plugs from the injectors so no chance of any injector issue messing with the test, plug test light into one of the injector plugs, then turning the engine over to see if there is any pulse light, if still no light then we can at least rule out the injectors causing the issue...
Ant.

Yes, I've heard one fried injector will cause all not to fire. that's why I discoed one rail and tried but no fire. I'll try the other. Working in shifts with the owner as he's retired and I'm still 9-5.

Tested;
CPS x 2 working, new CHT, (known working ECU and DME relay) new fuel filter, now new fuel pump, ~34PSi fuel pressure at rail, 12v to each injector lead 4-6, will test the rest. , spark at plugs, momentarily fires on ether.
As of yesterday did have noid light pulsing on a few injector connectors though still no evidence of fuel when pulling #4. maddening

Some suggested checking AFM output, will do as well.

Have got loaner oscilloscope to check other items. Will continuity test injector leads back to ECU. will also try to confirm valve and spark timing though when catching with ether sounds very normal.

I know we hate open loops on these types of threads but will update as squawk list items checked off.
thx

https://i.imgur.com/uHfJ1kE.jpg

I am wondering if the 12v supply to the injectors is dropping under load, there will be a big difference in the current needed to open the injectors in comparison to being able to light up a noid light.
Just a thought.
Ant.

Again, disconnect the temp sensor and you should get the MAXIMUM amount fuel - max injector pulse width results while cranking!
The test light should be very bright. The AFM functionality doesn't affect this test. Avoid wasting time with naive "guessing" tests.


Start with a one injector load (connected) at a time with the node light.

Just to clarify: disconnecting one bank of injectors at a time isn’t the same as disconnecting all injectors. All 6 injectors are driven off the single low-side driver. Pin 14 and pin 15 inside the DME are connected to each other.

If you have a flashing NOID light with one of the 6 injectors disconnected you can look at either side of the injector pin connector with an oscilloscope and verify constant +12V on one side and correct injector signal on the other while cranking.

:)

YES! That's why you start by ONLY connecting ONE injector to begin with using the noid light, and then add additional injectors.
You can also add a VOM to monitor the 12V injector supply voltage at another injector plug.

:)

Update. My buddy, the builder, said while checking timing yesterday, car fired up on 2 cylinders. He said 1 and 6. No idea how he could tell which ones.
I went at lunch time today. FP rate was 1.5 liters in 30 seconds. (Bentley spec is .875ltr)
I tested AFM resistance, linear response to moving flap. I could not measure any voltage change across the AFM so I thought we found the culprit.

When replacing AFM harness, the male end kept dislodging the keeper clip on the plug and I could not get it to properly seat. I flipped the keeper clip 180 and refitted the harness clip. To see if we could replicate 2 cyls running we cranked it over. After 2-3 seconds the damn thing fired up like it had never been apart! No smoke, no smells, no immediate leaks just pure flat 6 noises. Wish I had a concrete answer, but it seems great.
Thanks a ton for the ideas and check lists. A huge shout out to Sal Carceller knower of all things electrical 3.2 for lot of PM help.
Comp #'s below
https://i.imgur.com/ZJ8XMvqm.jpg

Good news! :)
Ant.

Damn. Dealing with the exact same thing on my 3.2/3.4
I hope it’s this same problem.

Aaarrrgghh, so close. Got a leaker!

https://i.imgur.com/2KGVosJm.jpg
https://i.imgur.com/9GFJ5cam.jpg

You win some, then always with these old cars, its something else! :D
Ant.