ECM Issues - No Start

[Macroblock]

I thought I should share some ongoing observations related to the No-Start condition of my 1990 ZR1. It's pointing to an ECM issue.

No negative-GND signal is present on the injector outputs from the ECM. Injectors therefore do not activate.

Vehicle turns over normally.

Fuel pressure is good. Both pumps are working.

Spark is good to each cylinder.

No diagnosis codes.

VATS works correctly.

Note: Changed the Crankshaft Sensor (right-hand underside next to dip stick entry point). No change in symptoms. Old sensor was tested and has an impedance of approximately 1 K-ohms. I suspect it was working correctly.

By removing the yellow harness from the ECM, I was able to ground each injector with a jumper wire, using the service manual as a pin guide. Doing this put a small amount of fuel in each cylinder and verified injector operation. Upon returning the yellow harness, I turned the engine over. It would ignite and/or run briefly.

Next - I will now open the ECM and look for signs of burning or corrosion on the circuit boards. After which, I will verify that each of the 8 Power MOSFETs (-) are operational.

[Macroblock]

I removed the ECM and opened it up and soldered a short wire from an internal 5 volt+ source.

While still apart, I returned the ECM to my car. Hooked up the 4 connectors and connected the battery. I turned the key to ON.

I carefully (and I mean VERY carefully) applied the 5 volt+ to the 8 GATES (marked G on the circuit board) for each of the 8 MOSFETs that correspond to the Fuel Injectors.

I heard each Fuel Injector click, one at a time, with the application of 5 volts+ to the GATE of its matching MOSFET.

This suggests the power transistors inside the ECM are in working order.

[Macroblock]

I turned my attention to the Ignition Diagnostic Connector. I identified "Terminal A", which is Circuit 430/Crank Reference High.

Using a signal generator, I inserted a spoof signal at Terminal A. This spoof was a square wave with a 0-5 VDC range and a 50-50 Duty Cycle.

The objective of the spoof is to simulate a refined Crank Reference signal that would normally come from the Ignition Module after odd-pulse-synchronization of the raw signal provided by the Crank Shaft Sensor.

Starting at .4 Hz, I observed the fuel pumps started to operate. For simplicity, I set the output to 1 Hz. I heard a click, resembling an injector becoming activated. At 10 Hz, the same injector was activating with a consistent on-off pattern.

After a number of subsequent tests, I discovered that only Injector 6 was operating. This remained true when I conducted the same tests with or without the DIS Ignition Module installed. None of the other injectors were activated by the ECM during testing.

Page 6E3-C2-2 of the Factory Service Manual indicates that all 8 primary injectors should be pulsed during start up (not just number 6).

It could be that this is a tell-tale sign that something is 'incorrect' inside the ECM and this is possibly the reason the car will not start.

More testing to follow...

[S.hafsmo]

:happy1:

[Ccmano]

:happy1: Wow, Your way ahead of me on electronics. Keep us posted. Haibeck provides an ECM testing and rebuild service should you need it. ECMs are bit scarce at the moment. But if you ask someone may make one available from their stash.
H
:cheers:

[Macroblock]

LOL - I built an on-bench emulator, so I can test further in the comfort of my LAB Chair!!!

It turns out that the injector for cylinder #6 correlates to Power MOSFET "Q1", the GATE of which is driven by U15 (16079435). This section is in working order.

Injectors for cylinders, 1,2,3,4,5,7 and 8 are correlated to Power MOSFETS "Q2 thru Q8", the GATEs of which are driven by U13 (23232). This section is NOT in working order.

I suspect, I could readily drive all the transistors with the 5v+ feed provided to Q1 from U15 but this would be a crude workaround because all injectors would fire concurrently (non-sequential-injection).

My preferred approach is to install a general purpose micro-controller inside the ECM and customize firmware to overcome the on-board deficiency and to maintain Sequential Injection capability.

Sidebar: If anyone has a defective, defunct or otherwise useless LT5-ECM, please drop me an email. I suspect I will need some test mules to further this project.

Found a photo here:

https://www.facebook.com/photo.php?fbid=10222963772627877&set=p.10222963772627877&type=3&theater

[XfireZ51]

I just recently spoke w Marc regarding injector operation and he is putting together an article outlining a test procedure for injectors while on the car. It does not involve modding the ECM.

[tf95ZR1]

From ZR1 Specialist website (Marc H)

http://www.zr1specialist.com/HAT%20Web/articles/Measuring%20Fuel%20Injector%20Resistance%20with%20 the%20Plenum%20in%20Place.pdf

Mb, I would talk to Jerry Downey of Jerry's Gaskets.
https://www.jerrysgaskets.com/
He is working on replacement ECM's (and he's
a good guy)

[Macroblock]

I just recently spoke w Marc regarding injector operation and he is putting together an article outlining a test procedure for injectors while on the car. It does not involve modding the ECM.

Awesome!

Verifying that the injectors are working is actually straightforward. Remove the Yellow or Gray 22-way Connector to the ECM. Turn the key to ON. One at a time, jump a wire from the (-) Battery Post to positions 1, 2, 3, 7, 8 12, 13 and 18 on the Connector. An injector that makes an audible click, is likely OK.

The injectors used in the early years of the ZR1 are susceptible to damage from ethanol in the fuel (not normally found in premium blends). The injectors develop shorts within their respective solenoid windings. Bad injectors will not produce a healthy click (a totally failed injector will not click). A current measurement can be used to verify the health status of an injector but the simple 'click' test provides discernible results.

Verifying the ECM is sending a (-) signal to each injector, without pulling the plenum, involves opening the ECM and ---CAREFULLY---checking each DRAIN PIN on the 8 Power MOSFETS. Use a 12 volt test light connected to (+) post of the battery. Have a friend turn the engine over (not literally). The test light should blink for each transistor. The ECM PC Board is marked "D" to guide you.

It is also possible to skin some of the jackets on the wires that feed the injectors and utilize a 12v test light, but by doing this you damage the wiring. To repair, coat the bare copper with heated solder and while still hot apply vinyl electrical tape.

In the case at hand, the ECM is unfortunately at fault. It looks like I.C. Component number U13 has failed (Sequential Injection Controller). As a result, only Injector 6 gets a (-) signal from the ECM.

[A26B]

...........
Mb, I would talk to Jerry Downey of Jerry's Gaskets.
https://www.jerrysgaskets.com/
He is working on replacement ECM's (and he's
a good guy)

Thanks for the referral Ted, but the ECM is not a project we are working on.

[Macroblock]

Problem solved:

ECM component failure U13 has been resolved by installing a general-purpose MCU, namely "Arduino" (a.k.a. an Atmega 328).

The hardware provides 7 outputs to the individual GATEs of transistors Q2 thru Q8. The DIS Crank Reference HI and the GATE from Q1 form the inputs.

The firmware monitors the RPM from Crank Ref. Hi and the injection pulse width from Q1. It then sends the same pulse to Q2-Q8 below 500 RPM. Above 500 RPM the Q2-and Q8 pulses remain the same but are slightly delayed (using some funky math) to permit optimum sequential injection.

https://www.youtube.com/watch?v=w9EcESdYibE

peter@zr1ecm.com (thought this would make a cool email address)

[A26B]

:notworthy

[Ccmano]

You know... someone could make a small business out of repairing these. Just sayin...
H
:cheers:

[Rodder]

Great job.

Anything can be built or fixed, the impossible just takes alittle bit longer.

[jss06c6]

Impressive diagnostic logic and repair!

Sent from my SM-G950U using ZR-1 Net Registry mobile app (http://r.tapatalk.com/byo?rid=90383)

[Ccmano]

So I suppose the next question becomes, is this what goes out on most of them or are there a multitude of electronics within the ECU that potentially fail?
H
:cheers:

[Marc Haibeck]

Hi Hans,

My ECM service has processed 35 ECM's so far. About 50% were repaired by fixing broken solder connections. About 30% were repaired with a fresh microcode download. 20% could not be repaired. Generally if there is a damaged component the unit can't be repaired.

[XfireZ51]

Macro,

Of course the secondaries are driven off the primary injector and powered by the secondary relays. How did u check to see if the secondary injectors were pulsing?
Or did ur procedure pulse the injector pairs?

[tf95ZR1]

I also say
:cheers:
and
:notworthy:

to Macroblock!

I'm impressed!
Do you have any photos of the surgery
you did inside the ECM?

[-=Jeff=-]

More questions.. will the Arduino hold to the heat? I assume you wrote the code for the arduino?

[XfireZ51]

Awesome!

Verifying that the injectors are working is actually straightforward. Remove the Yellow or Gray 22-way Connector to the ECM. Turn the key to ON. One at a time, jump a wire from the (-) Battery Post to positions 1, 2, 3, 7, 8 12, 13 and 18 on the Connector. An injector that makes an audible click, is likely OK.

The injectors used in the early years of the ZR1 are susceptible to damage from ethanol in the fuel (not normally found in premium blends). The injectors develop shorts within their respective solenoid windings. Bad injectors will not produce a healthy click (a totally failed injector will not click). A current measurement can be used to verify the health status of an injector but the simple 'click' test provides discernible results.

Verifying the ECM is sending a (-) signal to each injector, without pulling the plenum, involves opening the ECM and ---CAREFULLY---checking each DRAIN PIN on the 8 Power MOSFETS. Use a 12 volt test light connected to (+) post of the battery. Have a friend turn the engine over (not literally). The test light should blink for each transistor. The ECM PC Board is marked "D" to guide you.

It is also possible to skin some of the jackets on the wires that feed the injectors and utilize a 12v test light, but by doing this you damage the wiring. To repair, coat the bare copper with heated solder and while still hot apply vinyl electrical tape.

In the case at hand, the ECM is unfortunately at fault. It looks like I.C. Component number U13 has failed (Sequential Injection Controller). As a result, only Injector 6 gets a (-) signal from the ECM.

Used this procedure today to check injector pulsing. Worked well. It may have helped w a knock issue I have been pursuing for some time, but not sure if this was what solved it for me or not.

[Macroblock]

Verifying that the primary injectors are working is actually straightforward:

Remove the Yellow or Gray 22-way Connector to the ECM. Turn the key to ON. One at a time, jump a wire from the (-) Battery Post to positions 1, 2, 3, 7, 8, 12, 13 and 18 on the Connector. An injector that makes an audible click, is likely OK.

To check the primary and secondary injectors together, remove the Brown 22-way Connector to the ECM. Jump positions 16, 17, 21 and 22 from the connector to GND.

Then remove the Yellow or Gray 22-way Connector to the ECM. Turn the key to ON. One at a time, jump a wire from the (-) Battery Post to positions 1, 2, 3, 7, 8, 12, 13 and 18 on the Connector. Both the primary and secondary injectors that make an audible click together, are likely OK.

To check only the secondaries remove Injector 1 Fuse while doing the above tests.

See page 6E3-A-5 of Factory Service Manual.

[Macroblock]

The ECM is out of the way of most of the engine heat so, the Arduino should be ok, sandwiched between the ECM's 2 PC boards.

I've used Arduinos on hundreds of devices, many in harsh environments, so I am confident it will be fine.

As I test a few hundred iterations of the code I wrote, I will monitor the Arduino for signs of heat distress.

[Macroblock]

I do have great photos, but I wasn't able to upload them to this thread unfortunately. I suspect, I'll have to post them elsewhere and link them.

http://djrock.easycgi.com/Park/images/1.jpg

http://djrock.easycgi.com/Park/images/2.jpg

http://djrock.easycgi.com/Park/images/3.jpg

http://djrock.easycgi.com/Park/images/4.jpg

[-=Jeff=-]

The ECM is out of the way of most of the engine heat so, the Arduino should be ok, sandwiched between the ECM's 2 PC boards.

I've used Arduinos on hundreds of devices, many in harsh environments, so I am confident it will be fine.

As I test a few hundred iterations of the code I wrote, I will monitor the Arduino for signs of heat distress.

Great to hear. I have an Arduino Due in the dash with a 5? LCD to replace the DIC lights

[Macroblock]

Used this procedure today to check injector pulsing. Worked well. It may have helped w a knock issue I have been pursuing for some time, but not sure if this was what solved it for me or not.

I have found that if the LT5 has a dirty, clogged or malfunctioning injector, this manual 'force' activation technique can sometimes help.

[Macroblock]

Just discovered a serious error in the 1990 "RED" Edition of the Factory Service Manual that relates to the Camshaft Sensor.

Code 31 on page 6E3-A-50 describes the sensor as having a 0-5 volt output.
This is incorrect. It is Ignition Voltage which is usually 11-14 volts.

The correct voltage "(about 12 volts)" is referred to on page 6E3-C1-3.

This matters if you're reworking the ECM like I am at the moment! =D>

[dredgeguy]

Macroblock, have been watching this thread with fascination, no clue about the technical side of this but impressed that you certainly understand and have fixed the one component that is rare, expensive and could shut us down if it goes south. Well done and thank you for posting!! I think this would make for a great article with photos for the Heart of the Beast and as Ccmano (Hans) suggested, a great side business if you have the time and enjoy what you do.

[XfireZ51]

Verifying that the primary injectors are working is actually straightforward:

Remove the Yellow or Gray 22-way Connector to the ECM. Turn the key to ON. One at a time, jump a wire from the (-) Battery Post to positions 1, 2, 3, 7, 8, 12, 13 and 18 on the Connector. An injector that makes an audible click, is likely OK.

To check the primary and secondary injectors together, remove the Brown 22-way Connector to the ECM. Jump positions 16, 17, 21 and 22 from the connector to GND.

Then remove the Yellow or Gray 22-way Connector to the ECM. Turn the key to ON. One at a time, jump a wire from the (-) Battery Post to positions 1, 2, 3, 7, 8, 12, 13 and 18 on the Connector. Both the primary and secondary injectors that make an audible click together, are likely OK.

To check only the secondaries remove Injector 1 Fuse while doing the above tests.

See page 6E3-A-5 of Factory Service Manual.

What year FSM are u using? My 91 FSM doesn't have proc at that page.
So to clarify, are Brown pins 16,17,21,22 all grounded simultaneously before moving onto Yellow/Gray connector? Or are they done individually w corresponding pins on Yellow/Gray?

[-=Jeff=-]

I think he is using a 1990 FSM

[Macroblock]

My 1990 RED-FSM has the complete injector wiring laid out on page 6E3-A-5. My 1992 WHITE-FSM has it at 6E3-A-6.

The illustration should come up under "Driveability and Emissions" for the VIN J.

[Macroblock]

UPDATE:

I didn't like the original factory start-up procedure because it took too many cranks before it started.

The firmware now monitors the coolant temperature sensor and engine RPM and when appropriate, hi-jacks the start-up process.

For the first time ever, my ZR1 now starts with the first spark in cold, warm or hot conditions.

Also - got a website up to link photos to ongoing developments. Came with a free email: peter@zr1ecm.com

[-=Jeff=-]

Very cool.. using same Arduino? Or another added to the ECM?

[Macroblock]

At this point the single Arduino I installed directly to the Gates of the 8 Power Transistors can perform most of the ECM operations related to fuel management and ignition timing.

Ironically, the largest drain on computational resources is the Sequential Injection Function, which provides very little advantage (if any) to fuel economy, emissions and overall drive-ability.

My sense is it would be possible to take an L98, LT1 or an LT5 engine-control-module and make it work with 1, 2, 3 or 4 Arduinos. In addition, it would be possible to gut an L98 or LT1 computer and re-engineer it for LT5 use.

[-=Jeff=-]

At this point the single Arduino I installed directly to the Gates of the 8 Power Transistors can perform most of the ECM operations related to fuel management and ignition timing.

Ironically, the largest drain on computational resources is the Sequential Injection Function, which provides very little advantage (if any) to fuel economy, emissions and overall drive-ability.

My sense is it would be possible to take an L98, LT1 or an LT5 engine-control-module and make it work with 1, 2, 3 or 4 Arduinos. In addition, it would be possible to gut an L98 or LT1 computer and re-engineer it for LT5 use.

If you are interested in playing with that. I have a couple under hood L98 ECMs

[Macroblock]

OK - Great. Likely over the winter.

[ram_g]

Reading this thread with absolute fascination. Congratulations Macro and keep it coming!!

The only question I have is with regards to the following (reducing # of cranks for startup). I was under the impression that this was a deliberate move so that the engine could build up sufficient oil pressure before starting? I know that my 2016 C7 does the exact same thing as my 1991 ZR-1 and it has been explained as such.

UPDATE:

I didn't like the original factory start-up procedure because it took too many cranks before it started.

The firmware now monitors the coolant temperature sensor and engine RPM and when appropriate, hi-jacks the start-up process.

For the first time ever, my ZR1 now starts with the first spark in cold, warm or hot conditions.

Also - got a website up to link photos to ongoing developments. Came with a free email: peter@zr1ecm.com

[Ccmano]

Reading this thread with absolute fascination. Congratulations Macro and keep it coming!!

The only question I have is with regards to the following (reducing # of cranks for startup). I was under the impression that this was a deliberate move so that the engine could build up sufficient oil pressure before starting? I know that my 2016 C7 does the exact same thing as my 1991 ZR-1 and it has been explained as such.

That has always been my understanding. As a matter of fact its one of the factors that caused engine failures early on when plant personnel revved the engines to high rpm in cold weather conditions during the Transporter loading process. As a result the time to fire the engine was lengthened and the revs were limited in cold conditions. Its written up in the book Heart of the Beast. It caused quite a stir and weeks of shipment delays. Personally Im fine with the longer startup.
H
:cheers:

[Macroblock]

The stock ECM provides an injection delay of 1 or 2 seconds, which is not nearly enough time to prime oil into the LT5 engine. My inclination is that it would take closer to 30 seconds to prime and pressurize all the oil passages with a typical start-up crank of less than 400 RPM.

As I reverse thru the functions of the Engine Control Module (ECM) and the Direct Ignition Module (DIS), here are my observations.

The original starting process involves a fuel injection delay merely to allow the Ignition Module to synchronize with the crankshaft and to allow fuel pressure to build up.

When the engine is turned over, and the DIS is satisfied with the signal it's getting from the crankshaft sensor, it generates successive signals to the ignition coils and provides a stable 0/5 VDC square wave to the ECM on the Crank Reference line (D18). This square wave signal will change states 8 times per crankshaft revolution - this equates to 4 complete pulses per revolution.

Once the ECM sees an appropriate number of Crank-Ref pulses per second, it energizes the fuel injectors and in due course the engine starts.

This means the engine normally doesn't receive fuel for the first second or two after cranking begins.

To improve starting, my firmware injects fuel, when appropriate, before the DIS energizes the coils, which means the engine will come to life sooner.

[Macroblock]

That has always been my understanding. As a matter of fact its one of the factors that caused engine failures early on when plant personnel revved the engines to high rpm in cold weather conditions during the Transporter loading process. As a result the time to fire the engine was lengthened and the revs were limited in cold conditions. Its written up in the book Heart of the Beast. Cause quite a stir and weeks of shipment delays. Personally I?m fine with the longer startup.
H
:cheers:

In actual fact, the ECM does not have an input for oil pressure. I guess it wasn't a consideration for software control back then.

I find the the LT5 pressurizes its oil very quickly when cold because the oil cooler feed line is disengaged by a thermostatically controlled valve.

A potential problem occurs with oil pressure when the engine is in a warm state and the oil cooler is engaged. On a warm start-up, oil flow tends to fill the cooler network before traveling to the to top of the engine, including the timing chain area.

The workaround for the warm start scenario relies on the elevated placement of the oil filter. The LT5 oil filter has a check valve in it which stores oil and releases it after the engine starts. This provides oil to the top of the engine when the thermostatically controlled oil cooler valve is diverting oil away from the engine.

It is a very good idea to restrict engine RPM during warm up. The stock ECM for the LT5 does do this by refusing Full Power Mode (even when selected) until the engine is warm.

Notwithstanding, the original ECM firmware should have had "idiot-proof-code" in it to prevent abusive behavior like that of the car-jockeys. It is my understanding the newer Corvettes do have this feature.

[Macroblock]

Just an update to provide this link on troubleshooting the C4-ZR1-ECM:

http://zr1ecm.com/youfix.html

[-=Jeff=-]

Nice. Will you be going into more detail on the internals of the ECM? I have one I need to fix, I should probably pull it apart and look at the solder joints first.

Lastly these are conformal coated aren't they, what is the best way to remove the coating to get to the pins?

[Macroblock]

I normally use Acetone on a Q-Tip to remove the Conformal from the pins.