General Category > Engine related issues

Lambda/O2 sensor/probe stuck at 870mv

<< < (2/2)

javadude:
Here's an update following another test run.

In the first test, I used the fast idle to get the engine up to temperature quickly but the graph showed the ‘Idle switch’ turning on and not turning off again.

I ran this test specifically to test whether the ‘Idle switch’ was working correctly and got a completely different graph! https://drive.google.com/file/d/187Un1_JSCnUjVkPXQtLOAh4izD9A_d8d/view

From the very start, the GS-911 recorded the fast idle being turned on and off. I’d noticed this at the very start of my first test but when the engine was up to temperature in that test the ‘Idle switch’ remained on.

In the second test the ‘Lambda closed loop control’ remained off until the engine temperature reached about 58 C, when it flickered briefly. When the engine temperature reached about 64 C it remained mostly on, which means, if I’ve understood things correctly, that the ECU is using the Lambda/O2 sensor in its calculations.

At about the same point, the ‘Lambda sensor voltage’ is seen to drop and then follow its expected pattern of jumping between high and low voltages. Whilst this is much, much closer to the expected pattern, I’m uncertain whether it’s actually responding quickly or slowly when compared to other published graphs out there.

Interestingly, as the engine temperature continues to rise, the ECU switches back to open-loop and the ‘Lambda sensor voltage’ goes back up to the 870mv range seen on the first run. I suspect that the ECU is running a rich mixture in an attempt to control the engine temperature.

It seems that my first test was invalidated by keeping the fast idle switched on for so long during warm-up, which may have caused the ECU to remain in open loop, especially as the engine temperature continued to rise. It’s all down to the programming, and since we don’t have access to the spec or the code, all we can do is monitor the behaviour and comment, possibly speculatively.

Given that I’ve been able to observe a more normal ‘Lambda sensor voltage’ graph I’m going to leave it at that for now.

For the benefit of others who may come across this post:
My bike is a European spec R1100S manufactured in December 1998. It does not have a Cat Code Plug (CCP) fitted. I checked this out with an independent BMW business here in the UK, motorworks.co.uk, who confirmed that EC models were not fitted with a CCP. The GS-911 ECU readout shows the ‘Motorcycle Variant’ as ‘R1100 S ECE-Kat’. That reads like ‘EC model with catalytic converter’ to me, and I believe that this also refers to the set of maps that the ECU is using (please refer to http://forums.bmwmoa.org/showthread.php?58495-Coding-Plugs-R1100-R1150-Decoded&p=994823&viewfull=1#post994823 for R1100 & R1150 CCP variations).

Jughead:
Remember that your lambda sensors can only measure what is fed into the cylinders.  If one cylinder is running richer that the other one, how can one lambda sensor compensate for that?  It cannot, so it takes an average.

I will refer back to my earlier post and ask:

Have the valve clearances been set so that both cylinders match?
Have the throttle bodies been synced so that they match, not only at idle but also higher up the rev range?
Have the butterflies been synced?
Are the throttle cables (the short ones from the junction box) in good condition, moving freely and not frayed?
Have the idle bypass screws been adjusted to match?
When last were the bypass screws cleaned?  (As a matter of interest, how many turns out is the LHS screw in relation to the RHS?)

Before I do any exhaustive fault finding, I first check and double check all of the above.  When I am happy that everything on the 2 cylinders match 100% and I still have an issue, then only do I start fiddling.

99.9%of the time the issue is gone.

Navigation

[0] Message Index

[*] Previous page