A different kind of startus interruptus

[pswiatki]

Got hit by the original, well known issue last year and replaced the wire feeding power to the ignition relay (relay no. 3, pin 3 on this schematic)

by going straight from the battery terminal with a short piece of AWG12 (with inline fuse). That cured the problem for a while. It re-appeared later and I had the starter refurbished (the solenoid was replaced). Everything was perfect until just now. But this time the starter problem is somehow different. The original was that the starter would not even attempt to engage (solenoid would not be activated), but this time the moment I put the switch into "crank" position it tries hard to activate the starter (a firm knock of the solenoid) and immediately does something like a total RESET of the system. That is: no crank at all, I guess the solenoid gets released, instrument cluster (i.e. dashboard) re-initializes - sometimes in the manner it does when the key is turned to "ignition" position, sometimes backlit stays on and no gauge check is performed. Fuel pump is heard again and after some 3 seconds I can hear (and feel) auxiliary injection relay (no. 30) clicking:

By the way: the main injection relay is that bigger, square one, right?
Now, with the tank taken off (and fuel pump not present, obviously) I can hear another quiet sound,
something like ticking. Can that be the stepper motor being re-adjusted (say, put to its initial position)?
I tried the following to find out if there is a problem with the battery itself, the ignition/starter circuitry and the starter:

• used a piece of wire to connect BATT+ directly with the solenoid - a firm crank was observed and the engine started
• removed ignition relay (no. 3) from its connector and measured voltage drop on its contacts on my bench at work; a current source was used and contact resistance could easily be derived - it was 6 to 8 mΩ; I was told the result was GOOD
• tried to start the engine by shorting the contacts in the ignition connector - nice crank and engined started; the dashboard was showing the error and the red triangle was lit - apparently the absence of the relay caused it
• reinstalled the ignition relay and tried to start the engine by pulling down (to GND) pin 2 of this relay (so, activating its coil manually) - also a good crank and engine started; the dashboard briefly showed an error and lit the red triangle but it disappeared the moment I removed the pull0-down connection from the relay's coil; my guess is ECU's internal diagnostics may have detected some unexpected state of pin 1 of the grey connector

So, I don't understand what is going on. Can the problem be a faulty ECU, or it is something less serious? I could try to check the connection between pin 1 on the ECU (crank pulse) and pin 2 of the ignition relay, but I doubt I will find any issues with that.

By the way, this is my ECU:

and the label on it:

Side question: is the ECU isolated from the GND at its mounting bolts and the only power ground (as opposed to signal/sensor GND) it is connected to is fed to it through the harness straight from the grounding post where the high current cable from the battery is connected (next to the starter)?
When I worked on my problem I forgot to attach the ground wires to the engine block with that post next to the starter and turned the key to ignition position. Got some very weird rattling sound - possibly all the relays were constantly clicking. Dashboard did not initialize and remained dark.
Immediately removed the key and realized my mistake - attached ground wires back to their designated grounding point. Later on, I didn't notice any abnormalities apart from the one described above (with the crank / started activation). But that one was present before this GND mishap.

My bike is: BREVA 1200, MY 2008.

 

[Brian UK]

Dud Battery. Typical symptom.
But check the cleanliness of the battery connections too to make sure that isn't making things worse.

 

[pswiatki]

Hmmm...... The battery is OK, though. I had a bad one before and the behaviour was different. Typically, it would not allow me to retry the engine start sequence, or would quickly get noticeably weak. This one is less than a year old (YUASA) and it starts up the engine without any effort if (and I quote myself) I try to start the engine by pulling down (to GND) pin 2 of the ignition relay (so, activating its coil manually). The solenoid works, starter works, the engine gets proper crank and all is good in this case. So, I would say: this shows the battery is capable and its terminals are clean.

 

[Brett]

Use a multimeter to measure the voltage across the battery terminals. Measure the same voltage when engaging the starter. What are those 2 readings?

 

[GuzziMoto]

What you are describing sounds like the starter tries to engage, but the current draw causes a lack of voltage to where the system resets. That is either a short in the starter or a dude battery (as mentioned) where the battery may show enough voltage but it lacks the required amperage.

 

[pswiatki]

Dud battery... OK. I will check that. This should be covered by the warranty, I hope. After all, the battery is less than one year old, so it should be good. However, what part of the total current draw is contributed to the starter load as opposed to the load from the rest of the electrical system? You see - if I initiate the sequence somewhat manually, there is no issue whatsoever and the battery is more than capable of starting the engine without much effort. Also, any short in the starter would probably drain the battery quickly, but after several "manual" cranks the battery is still OK and capable of moving the crankshaft and starting the engine again and again. This is like a normal startup in the normal condition of all parts and systems in the vehicle... We will see, I will post again when I have news.

 

[Trout]

Also check all grounds especially the one from the battery to frame.

 

[pswiatki]

Aha! This is something I've been wondering about. As I said above: I accidentaly put the key into ignition and turned it without ground cables connected to the engine block. One of them was the high current cable for the starter and the other was actually a pair of medium size wires which - I guess - go up the frame to the main wire harness. The result was the rattling sound (relays clicking fast) and dashboard not initializing at all. So, somehow the ground connection had to be completed someplace. Question: is there another lone wire directly from the negative battery terminal going to the frame? I am not 100% sure, but I don't remember seeing such connection.

 

[Kevin.NZ]

A large earth wire will go from the battery -ve to either the engine block, the frame or both with no sign of corrosion.

 

[pswiatki]

Exactly - and I have a single thick wire (earth wire) going to the engine block next to the starter. Clean all contacts and reassembled. Checked the operation - still no good. However, I was able to hook up a really good scope to my bike and captured the voltage at the battery when cranking. Here are the results:

• pressed the start button and got a short shot of the solenoid:
  
  
  
  
  
  The voltage drop is very short - here the zoom in:
  
  
  
  
  
  Voltage actually starts dropping at marker a and comes back to the original level at marker b.
  Here is yet another look and the actual drop:
  
  
  
  
  
  So it is less than 14 ms long and already causes an abort of the startup sequence. Strange.
  OK, I understand the minimum voltage looks low, but why so much drop and no attempt to move the starter.
  Perhaps it is that ECU abort mention in this thread
  https://www.guzzitech.com/forums/threads/2006-breva-1100-schematic.17864/page-2#post-138969
  
  
• I started the bike using a piece of wire (helped ECU a bit and activated the coil in the startup relay). Here is what I got:
  
  
  
  
  
  This is the entire sequence until the voltage stabilizes at 14.2V and - if I am not mistaken - when the increased revs (~2000 RPM) go down to proper idle revs (1100?).
  Here is a closeup:
  
  
  
  
  
  So, the time it takes to start up the engine is less than 1.6 seconds.
  I am not sure why there is that instability of the charging voltage for a good 9 seconds:
  
  
  
  
  
  Is it the voltage regulator that does this (quite fast, as you can see above)?
  
  
• Found something interesting in the next attempt to start up my BREVA.
  I am not sure if it is just the oscilloscope, or there really are very short spikes like these
  (brightness adjusted in post-processing):
  
  
  
  
  Here is the single spike:
  
  
  
  
  
  It is extremely short, so maybe it is just that the scope is overly sensitive

So, I got all the scope captures, turned everything off and just wanted to go back to the office to return the scope. Before I did, I gave it the final try.... and..... well, it started up without any issues.
Yes, from the handlebar switch. I don't get it. Let it run for a while to charge the battery a bit after all the testing. Turned off and retried. Worked again. One more try - again it worked nicely.
This is really strange. I did not do anything prior to this final test that could help it work. Well.... I just made sure the screws on the battery terminals were properly holding the terminals. They did. Neither one moved, so I don't think anything changed that could affect this final result.
Needless to say: I don't trust the bike now. Not being able to clearly identify the malfunction makes me worry that the next time it happens will be somewhere on the road. Heh - looks like a piece of wire is a critical accessory every Guzzista must keep at all times.

 

[Trout]

Not strange at all.
It is a Guzzi.

Just Remember: Guzzi=making mechanics out of riders since 1921.

LOL

 

[pswiatki]

I can see your point, Trout. Well... quite disappointing I have to play with my bike like this. I'd much rather ride it - last days of really good, sunny weather here in POLSKA.

 

[Earl Bowman]

I'm having a similar issue with my 2007 Breva V1100. It will start when it feels like it, maybe once every 30 attempts or so. Once it does start it runs normally and when shut off, it will always restart, I've tried it 10 or 12 times in a row and it starts every time. Once I walk away and come back the next day "Click" solenoid makes a healthy snap but the starter doesn't engage. I preemptively installed MPH Cycles relay fix harness a year ago, added a very healthy Antigravity Li-ion battery and rerouted my ground cables to the larger bellhousing engine bolts. Since the issue developed I have played musical chairs with the starter and other relays no change. I have disassembled and cleaned the contacts in the start button and checked the kickstand switch and tip-over kill switch. It's the occasional starting without changing a thing that I don't understand. I don't know what to do next! Anybody have an idea?

 

[V700Steve]

I would run a hotwire to solenoid, strait from 12V. Put a fuse in it and never worry again. But, thats me, modify it once.
Put a small button somewhere just for that occasion.

 

[pswiatki]

Sorry, kind of forgot to come back with results.
So, it seems the ECU kills the crank cycle as soon as it detects the voltage drop which it doesn't like (when the battery voltage falls below some threshold, I presume). I charged the battery, put it back in my Breva and everything started working as expected again.
I guess the next step would be to check the sleep current. Something (immobilizer perhaps?) is drawing enough current to cause problems with the start-up cycle when the bike is not used for a week or more. When it is ridden more often, the issue doesn't pop up at all. The battery is relatively new (2018, AGM from YUASA), so I don't think it is already "spent". Hope this helps a bit.

 

[pswiatki]

I know the ECU has to monitor its own power supply in order to make sure it can operate properly. When the voltage is out of the correct range some action is taken (depending on the state the ECU is in). My buddies designed ECU's that way (I work for in the automotive industry, but mainly deal with electronics for cars, trucks, trailer trucks and such - not bikes, unfortunately). Everything has to meet customer's requirements which are quite strict in this area. There are also internal requirements, so I am pretty sure some reqs came from MG and others are internal of Magneti-Marelli. I would imagine the ECU just interrupts the crank cycle when it determines the power supply is out of range. My buddies say the voltage drop during crank should not go below 6 volts or so, but the exact figure depends on the manufacturer and their requirements. Some of my measurements (above) show the voltage went well below 5 volts and - apparently - the ECU in my Breva did not tolerate such deep drops.

This makes me wonder - if you run a hotwire and force the crank cycle bypassing the ECU - can anything bad happen? I mean: yeah, the bike runs normally when the engine starts, but I occasionally saw the DTCs (error codes) getting registered when the hotwire was used. Who knows what else might be happening in the software when the ECU is subjected to a severe power drop like that.

 

[john zibell]

Just because the battery is relatively new, doesn't mean it is good. If the battery isn't prepped correctly, it won't last long. Guzzi even put out a service bulletin on the issue Once acid is installed into a new battery, it must sit at least 3 hours (I like about 12 hours to be sure) for the acid to fully permeate the glass mat before charging. Also the battery should be fully charged before going into service to have the expected service life. So, take the battery out and have it load tested. If good, check for a parasitic drain.

 

[Earl Bowman]

On my '07 I've load tested my battery and my load tester says it's good to go. Measuring battery voltage at the battery I get 13.2 after pulling the charger off. I know you can lose up to 1 volt between the battery and the starter just to resistance in the wiring harness. I can jump start it by shorting the solenoid wire to the hot terminal and she cranks over, so I'm stumped. I've been told that MG programed the ECU to interrupt starting at 11.8 v or less. True?

 

[Mackers]

Earl, I had similar problems with my "06 Breva some years back. This is what I've learned;
1) Check the tab on the starter solenoid. Mine was a poor fit and loose. Simply moving the bike on and off the side-stand was enough to cause starting issues.
2) She needs a YTX20CH-BS battery, not a YTX16CH-BS. The latter works for a while but starts playing up. Keep it charged.
3) The 30amp fuses can cause strange issues. Clean the fuse holders and give her new fuses.

 

[Earl Bowman]

Earl, I had similar problems with my "06 Breva some years back. This is what I've learned;
1) Check the tab on the starter solenoid. Mine was a poor fit and loose. Simply moving the bike on and off the side-stand was enough to cause starting issues.
2) She needs a YTX20CH-BS battery, not a YTX16CH-BS. The latter works for a while but starts playing up. Keep it charged.
3) The 30amp fuses can cause strange issues. Clean the fuse holders and give her new fuses.

Thanks, I've checked the solenoid tab and it seems snug. I switched out the 30 amp fuses for the spares, cleaned the holder and no change. My Antigravity Li-ion battery is rated at 480 CCA which, I think, should be more than enough to do the job and it is fully charged. I know it spun the motor much better than the stock battery it replaced and still does when the starter decides to engage. I've also had the starter out and smeared some light oil on the plunger and bore as well as the drive gear shaft. I'm wondering if the starter itself could have some intermittent internal issue.