Motronic 1.0, Pin 20 Fuel Pump Relay - SOLVED

[Canuck YYC]

TLDR: ECU pin 20 is not grounding the fuel pump relay when it should. Has anyone solved this previously? If you couldn't repair the ECU, what was your work-around?

Trying to track down likely spots for failed components or cracked solder joints that others may have definitively found. This car has been somewhat temperamental from the time it was acquired, frequently leaving the driver stranded in inopportune places like rush hour traffic and the drive-thru. To say it wasn't one thing would be an understatement but the car is slowly but surely having all of these little issues addressed. Fuel pump relay, fuel pump (it was tired), master relay, braided ground to the firewall, new ground on the motor mount, new electrical feed to pin 86 on the fuel pump relay. Yet the random stalls persist. Once it died in the middle of our street while moving it for street cleaning and no amount of persuasion would bring it back to life - eventually we towed it back to the driveway and a few days later it fired up just fine again. Recently, while moving it into it's winter storage and service bay, it died once more and has refused to come to life ever since.

Trouble-shooting it yet again, the very first thing I noticed was the lack of fuel pump noise despite being in a quiet garage. When the key was turned, there was no prime cycle before shutting off again. Once again, back to the relay, the fuse and so on. Power at terminal 86 (switch side), power at terminal 30 (switched side input), jump power to terminal 87 and the pumps fire right up. Try a new relay - no dice. Also, no click. Main relay clicks when it turns on, fuel pump relay - no clickity click clicking. But there's power at 86...why no click? Even if there were no VR sensors in the car giving a run or start signal, the ECU runs the pumps momentarily when you turn the key. Not happening. Feed ground to terminal 85, relay clicks, fuel pumps fire up.

Pop out the ECU and remove the connector to check for wire continuity between both ends - no issues at all. Which leaves us at Pin 20 on the ECU. This is not grounding despite the correct inputs. Fortunately I have a dedicated electronics shop in which to work on it. Unfortunately electronics are not my AOE. Any pointers beyond "find and repair cracked solder joints" available?

[Ju@n]

Hi!
Motronic 1.0 (and 1.1-3 as present in 528e) doesn't pre prime the fuel pump with the key in the ON position.
It starts the fuel pump when it detects rotation of the crank sensor.
Usual suspects for the fuel pump not running when the engine is cranking is either no power to the ECU or the VR sensors malfunctioning (if you already changed the fuel pump relay and fuse).
 

[Mike W.]

This is something I wrote recently for someone having fuel issues.

30 gets 12V straight from the battery, unswitched.
87 is the load, the fuel pumps, plural.
86 gets power from the main relay when the ignition is turned on.
85. This is where it gets a little trickier. 85 is grounded thru the Motronic ECU, and only when cranking or when it sees that the engine is turning, so unless it's running or cranking you will not see a ground on 85.

And there is no prime function. Some cars like 520's had it, but not the motronic cars.

The full thread is here, viewtopic.php?t=162739

[Canuck YYC]

Thanks both,

For reasons only google search knows, that thread did not come up in my digging, but 30 seconds after posting this, I came across it in the subforum contents. I must be getting old - I'd have sworn there was a prime cycle

The VR angle (ha) is my fear honestly. No power is an easy solution assuming the root cause is outside the ECU box. The VR thing is a bigger pain although I suppose a strong battery and an AC meter should reveal whether it's generating a signal. Or perhaps a scope to see what kind of waveform it's generating. Is there a spec for peak-to-peak voltage at a given speed?

[Ju@n]

Thanks both,

For reasons only google search knows, that thread did not come up in my digging, but 30 seconds after posting this, I came across it in the subforum contents. I must be getting old - I'd have sworn there was a prime cycle

The VR angle (ha) is my fear honestly. No power is an easy solution assuming the root cause is outside the ECU box. The VR thing is a bigger pain although I suppose a strong battery and an AC meter should reveal whether it's generating a signal. Or perhaps a scope to see what kind of waveform it's generating. Is there a spec for peak-to-peak voltage at a given speed?

I don't remember a vp-vp voltage, however I would remove them, see if they have a lot of crud on them, and then clean the connectors that are on the support of the intake.

[Canuck YYC]

Tested:

• continuity from reference sensor to ECU
  • 5, 25, 26
• continuity from engine speed sensor to ECU
  • 23, 8, 27
• continuity from fuel pump relay 85 to pin 20
• resistance check reference sensor (~950)
• resistance check speed sensor (~950)
• All ground connections from ECU connector
  • 5, 10, 16, 17, 19
• switched power to 35
• Persistent power at 18 

Again, no fuel pump. Jump terminal 85 on the fuel pump relay to ground and they fire up instantly so it’s nothing in the main or fuel pump relays (or pumps). Even with the jumped pumps, no dice. Lots of cranking but no firing. 

Remaining to do:

•  capture the waveforms from the speed sensor in particular and the reference sensor. 

I’m somewhat resigned to something having failed in the ECU itself. Also, stumbled across this little tidbit in the Fuel Injection portion of the Bentley Bible: The fuel pumps operate for a few seconds when the ignition is first turned on, and continue to operate only if the engine is running. Seems I haven’t completely lost my mind. 
 

[Ju@n]

Did you try starting the car with a bridge in the fuel pump relay? If it starts it would point to the ECU part that triggers the fuel pump as an issue (probably a cold solder joint or a blown transistor)

[Canuck YYC]

I have. It did not change the symptoms though the pumps do run that way. Cheating the pumps to get them working hasn’t been a problem. Came across yet another reference in the Bentley describing the pumps running momentarily when the key is switched on: …turn the ignition on and listen for the sound of the fuel pump. It should run for a few seconds , then stop. 

That doesn’t happen of course, but there’s also no spark when cranking. Me thinks the ECU is on permanent vacation. 

[Ju@n]

If you have an oscilloscope you can try reading the VR sensors, just in case.
Another small silly thing to check is that they are correctly plugged, since they'll fit on each others plugs (though now that I think about it, the car did start and work in the past).

[Canuck YYC]

Well...I don't know what to say. We pulled the speed sensor, cleaned it, wiped down it's seat on the bellhousing, plugged it back in and the damn car fired up immediately. I'm not convinced that was the problem but it's up and running. Trying to track down a spare ECU now so if it decides to die in the middle of the road again, at least that part can be ruled out. It is possible the VR was the cause. It's also possible that somewhere in the 40-year old wiring, there's just enough wire fatigue to be intermittent. I genuinely don't know, but it's running so we'll take it.

[Ju@n]

Awesome! Hopefully it was just the grime there (or in the connector) that caused it

[Canuck YYC]

This car…I swear it’s going to be the thing that puts me in the asylum. Fortunately it‘s @DorytheBlue’s all-the-time problem and only mine sometimes

He drove it home last night, first day on the road. Come out the next morning, no start. Crank til the cows come home but still no start. Lots of fuel pressure, but no spark now. Tach needle doesn’t flicker, absolutely nothing happening at the coil. Run through a million tests again - harness, bell housing sensors, coil, even swapped the ECU. Built a new ground strap for the valve cover to firewall (why is that even a thing a shouldn’t the big strap at the motor mount take care of that?), cleaned all the battery connections, double-checked power was getting where it’s supposed to. After 3 hours I had to step away. 30 minutes later he gives it a crank and it fires right up.

[Canuck YYC]

TLDR: Voltage from terminal 87 of main relay to pin 35 of the ECU failed as soon as the ECU load was connected to it but worked as long as the ECU connector was unplugged for testing. 

The only thing more daunting than an intermittent electrical problem, is an intermittent electrical problem layered into additional problems that is also load-dependent. 
While troubleshooting this no-start / sudden stall problem, I tested what I believed were all the things, numerous times. My son took it upon himself to replace both crank sensors (questionable) along with a new cap (manufacturing defect, didn't fit, thanks Rock Auto) and rotor, plugs and wires. We tested fuel pressure, current draw, literally every sensor on the ECU harness, tested the harness from every input and output, tested them together gathering readings from every single pin at the ECU connector. Basics - do we have constant power? Do we have switched power? Are we grounded? Are we getting the right readings from the sensors? Is there a short in the cluster? On and on and on. I tested every pin in that ECU connector not less than 6 times. Still wouldn't start, then it would, then it wouldn't. Ground the fuel pump relay, get lots of pressure but definitely has a check-valve failure, still no start. No attempt. 

Under the hood, the main and fuel pump relay sockets were no longer attached the fuse box but hanging loose. The car, in one if it's momentary good moods, fires up and runs like a top. We are both very pleased and very confused. @DorytheBlue grabs the fuel pump relay and puts in back into position on the side of the fuse box. Grabs the main relay to do the same and everything shuts off like you turned off the key. AHA! A CLUE! Cycle the key, relay clicks, battery voltage at the injectors now - I twist and turn and push and shove and punish that relay socket and absolutely nothing happens. The voltage doesn't flicker a mV, yet the car WILL NOT START again. Bash head against wall repeatedly. 

Staring at the schematic, the main relay portion in particular, and it occurs to me to check the other pin 87. Switch on the key, relay clicks, battery voltage on the "outside" pin 87, 3 volts on the inside terminal 87. What?? I've tested this circuit countless times - pin 35, are we getting battery voltage at pin 35 when we turn the key on? Check!  Clearly we are not getting that now. 

Disconnect the battery, pull the ECU connector, re-connect the battery, turn on the key, probe pin 35 - battery voltage. Cue incredulity. Check the relay - battery voltage. What is happening?? Plug the ECU connector back in, check voltage at relay, 3 volts. No replacement relays in the stash. Plug everything back in, jump battery voltage to the internal pin 87, hit the key and everything roars to life like nobody's business. Off, start, off, start, off, start - the damn thing works!  A quick trip to Lordco and $16 later and that problem is finally licked. DorytheBlue is up and running and cruising the streets without issue. 

Look, I like to think I'm pretty good at problem solving. I'm methodical, I don't throw parts at the problem (the VR sensors were not my call) and I at least thought I was thorough. Surely if I'm testing at the delivery end of the wire, and the power is there, I don't need to test the relay to make sure it's providing the power at the start of the run. Wrong! The one thing I couldn't do, which frustrated me to no end, was test everything in-situ as 35-pin Motronic connector sets aren't exactly growing on trees.

I assumed 
- If there's power at pin 35 in the connector, when it's unplugged, it will have power when it's plugged in. 
- If there's power at the injectors from from the main relay, there's also power 2nd output terminal. 
Those two assumptions kept me troubleshooting and pulling my hair out far longer than needed. 
 

[Mike W.]

Rereading it now I was just thinking, you know, that could be a main relay. I went thru madness with mine when I got it as it had a flaky main relay and a bad starter drive which would only engage sometimes. So if the starter engaged, it might or might not start. That about drove me nuts. But anyway on the main relay, there are at least 2 types, one of which actually has 2 different contacts inside it, so if there's 2 different contacts it easy for one to not be good while the other is. And of course when you test intermittent things, they always work.

In other words, I have felt your pain and glad to see you hopefully have it sorted out!

[Tiit]

I don't have answers, but I can offer some ideas. 
* From my experience, speed/reference sensor can't be conclusively tested with an ohm meter. I had a sensor with correct resistance values, but the waveform was weak tested with a scope. 
*I have a M1.0 535i with intermittent electrical issue. It shuts off around 10 minutes after cold start.  I haven't solved it yet, but I'm planning on doing the process of elimination. I'll wire up a volt metre to main relay and set it up on the dashboard. Then drive until the issue occurs. If the main relay is still live, I'll repeat the test witn the next suspect. To speed up the process I'll do the test with multi input scope, but volt meter works just as well. 

[Canuck YYC]

And of course when you test intermittent things, they always work.

 

Thanks Mike.

Always. Always.  It definitely kicked my ass along the way. Hopefully my troubleshooting game is a bit sharper now, and with any luck my tribulations will help someone else. 

My car (as opposed to the car in this topic) came with the factory ECU and the MS3X.  I'm going to take advantage of this "bonus" ECU and modify it with a bunch of flying leads out the side so one can drop it into a misbehaving car and have access to all of the signals and outputs in-situ. Not quite as ideal as an actual breakout that would keep the offending car's ECU installed for the testing, but at least it will allow for testing everything outside of it in an operational status. 

[Canuck YYC]

I don't have answers, but I can offer some ideas. 
* From my experience, speed/reference sensor can't be conclusively tested with an ohm meter. I had a sensor with correct resistance values, but the waveform was weak tested with a scope. 
*I have a M1.0 535i with intermittent electrical issue. It shuts off around 10 minutes after cold start.  I haven't solved it yet, but I'm planning on doing the process of elimination. I'll wire up a volt metre to main relay and set it up on the dashboard. Then drive until the issue occurs. If the main relay is still live, I'll repeat the test witn the next suspect. To speed up the process I'll do the test with multi input scope, but volt meter works just as well. 

 

I have a small scope I used to test both sensors and they both have a strong signal. I've claimed the old ones for my troubleshooting kit as payment

There's a handful of scenarios that can stop the engine in it's tracks (as opposed to causing erratic or partial operation). The following based on late E28 535i's:

• Loss of power:
  • switched and constant power to ECU terminal 18 (un-fused battery direct)* or 35 (main relay terminal 87)
  • switched power to main relay 86 (from ignition switch, through OBC relay box on OBC-equipped cars)
  • broken ground at main relay 85
  • battery direct to main relay terminal 30
  • switched power to fuel pump relay 86 (from main relay, same branch that supplies 12V to the injectors)
  • switched ground to fuel pump relay 85 (performed by ECU)
  • battery direction to fuel pump relay 30
  • failed fuse #1 to fuel pumps (not really applicable here)
  • switched power to ignition coil (ignition switch, un-fused)
• Loss of VR sensor(s) input (although I've never experimented to see what would happen if the ref sensor signal disappeared while running) 
• Loss of grounding at ECU
• Short circuit to ground (or power) between ECU pin 1 and ignition coil negative terminal. 
• Obviously - failure of any of the individual components themselves in this list

I believe any other failures external to the ECU would not cause the engine to shut down, short of a significant harness failure. There are some known issues in the ECU around solder joints at various transistors however I am not that guy who knows unfortunately. 

* I spent numerous hours chasing down an intermittent electrical problem on a customer's motorcycle some years back. The bike would start and run fine but suddenly shut down for no reason while you were out riding. Of course it would not come back to life necessitating a rescue pickup and a lot of cursing. If you left it to sit and idle, it would run without issue until it ran out of gas (if you let it). Then for no reason at all, it would run fine again and you'd be out in the middle of traffic somewhere and the whole process of shutdown/refuse to start/curse a lot/rescue pickup would start again. It wasn't until I started literally yanking on wires that a failed connection right at the battery revealed itself. A brass ring terminal had failed underneath the heatshrink between the crimp portion and the ring portion. Sometimes the heatshrink would pull the two parts together and everything would be fine. Later, likely due to vibration, the terminal would separate but remain wonderfully hidden inside the heatshrink and harness. What I was convinced was a heat-related problem had nothing to do with heat at all. 

[Ju@n]

And of course when you test intermittent things, they always work.


 

Thanks Mike.

Always. Always.  It definitely kicked my ass along the way. Hopefully my troubleshooting game is a bit sharper now, and with any luck my tribulations will help someone else. 

My car (as opposed to the car in this topic) came with the factory ECU and the MS3X.  I'm going to take advantage of this "bonus" ECU and modify it with a bunch of flying leads out the side so one can drop it into a misbehaving car and have access to all of the signals and outputs in-situ. Not quite as ideal as an actual breakout that would keep the offending car's ECU installed for the testing, but at least it will allow for testing everything outside of it in an operational status. 

I did try (and succeeded... but very shortly), to make my MS3X work with the stock speed/reference vr sensors, but without some board in the middle there's no way to make the 135 tooth wheel on the flywheel work with the MS3X. So unless you have Motronic 1.1/1.3 on whatever E28 you are trying to diagnose, you won't be able to use the MS3X to diagnose it.

[Canuck YYC]

No no , the subject of the thread is a stock ‘87 535i that my son built. The modified ECU I mentioned is for this (or others), not my car. 

I went down more than one rabbit hole in trying to get the 116-tooth flywheel to talk to the MS units a decade ago but the 116/360 configuration made more problems than just sourcing a B35 front pulley. 

My MS3X-backed B35 / E28 build is slowly coming together rather nicely. Who knew I’d have this little free time as an adult…