Hello,
I have a problem with my freelander 2 09. The car takes a strange start. The first firing is always with a slight white smoke and wavering engine speed at idle. This only happens the first time the engine is started in the day. It can choke sometimes. I noticed that this could be related to the humidity in the air. EGR showed errors but was replaced. The symptom is still there, but it feels smaller.

Anyone had something similar?

cheers
T. https://www.instagram.com/halotuswiat/

If:
1. The problem was because of EGR,
2. You recently changed the EGR (1-2 weeks),
3. The garage didn't performed the "Change EGR" procedure on the engine ECU,
then wait a while.
It can take as long as one month (!) for the engine ECU to adapt to the new EGR and it's particularities.
Shorter if you take long journeys, like 100+ Kms (1+ hours) with constant cruising speeds.
The engine ECU had slowly adapted to the old, defective, EGR and now it's still running on it's particularities/adaption.
It needs several conditions to be met (engine load, temperature, requests, etc) in order to start the procedure to adapt to the new (for engine ECU) EGR parameters.

ok, that sounds logical. I will go and observe his behavior. Thanks for the advice. https://www.instagram.com/halotuswiat/

alex_pescaru wrote:

3. The garage didn't performed the "Change EGR" procedure on the engine ECU,
then wait a while.

Yes, if you have SDD, then there is a replace EGR procedure which will default the EGR learnt values to 0.
From there the ECU will learn the characteristics of the new EGR, which are more close to the default (zero) values than the values of an old, defective, EGR.
So the time for the system to stabilize is much shorter.

Thanks Alex.
So will the ECU not learn a new EGR anyway without resetting the value to 0 (albeit slowly ). ?
Or is the learning progression one way only?

What happens if you simply clean and free up a sticking EGR valve, should you recalibrate it then also? Jules

The ECU is continuously adapting to the new conditions and new characteristics of all components.
So, if you have a slowly carbon fouling EGR, it will continuously adapt to its changing characteristics in order to still give you an as functional as possible engine.
For example, it will increase the EGR opening values in order to compensate with the decreased flow through the fouled EGR.
And these new, increased values, are stored into ECU's NVRAM.

If you are replacing the EGR and you don't perform the "replace EGR procedure", it will adapt to the new EGR slowly.
Meaning, it will see an increased EGR flow, because based on the old values the EGR opening is greater - remember - to compensate with an fouled EGR, and it will slowly reduce the old NVRAM stored values, in order to reduce the EGR flow.
But this process of learning new parameters is a slow one, because in order to test/verify/adapt the ECU needs a certain functional parameters to be met, parameters that are met only through driving cycles and on certain moments of the driving cycles.
If those conditions are not met, the learning is postponed until the following driving cycle.
And so on and so forth...

For having an as correct as possible running engine, the "replace EGR" procedure will speed up the learning process, because the default EGR values are close to the values of a new or clean EGR, and therefore the learning process is a quick one and the slight differences, up until the learning is complete, are almost imperceptible in the engine performance.

In the OP case, those "open the EGR more to compensate with the fouled EGR" old values are hindering the performance of the engine, introducing too much burnt gases on the intake and generating those RPM variation.

BUT....

The OP should also take into account that, sometimes, those RPM variations are also a sign of a partial or complete failure of the oil temperature sensor...
But because he said that the EGR was showing errors, I've supposed that the above explanation is a more close to reality than the oil sensor possibility.
In the end, it is pretty difficult to diagnose a problem from behind a keyboard and screen, based on the little info you have and the ability of the user to explain the problem...