35D Active DEF tank forensics

[ZetaTre]

Originally Posted by ard

Which (I think) is why a 'bad urea' warning can be due to a clogged DEF nozzle: the duty cycle goes to its maximum limit; the DDE does not see the NOX values respond as they should; it then declares 'must not be def in the tank'

You get the idea, but there's an additional complication that doesn't allow a diagnosis by "maximum duty cycle"

The NOx sensor is not capable of distinguishing NO and NO2 (the NOx) from NH3 (ammonia, which is what the urea converts into once it hits the hot exhaust gas).

There is a phenomenon called "ammonia slip" that can occur when too much urea is injected. If there isn't enough NOx in the gas stream than some of the ammonia has nothing to react with and is emitted through the tailpipe which is something you don't want.

So the DDE changes the duty cycle to minimize the reading it gets from the NOx. A "high" reading could mean both too little urea (some NOx passes the SCR) as well as too much urea (some NH3 has slipped) and iteratively tries to target the low point. This is registered in an adaptation value to compensate for various tolerances in the system. The quality of the fluid is diagnosed by observing how marginal changes in the duty cycle affect the marginal changes in the NOx sensor reading, but also as sudden (meaning above a certain threshold) changes between the closed loop readings (the action-reaction of duty cycle to NOx sensor readings) and the adaptation tables.

[bawareca]

Originally Posted by ZetaTre

....

Thanks for clarifying these tricky points. I just didnt have time to go thru the technology explanations.
There is also a control unit attached to the active tank, do you the purpose?

[ZetaTre]

Originally Posted by bawareca

Thanks for clarifying these tricky points. I just didnt have time to go thru the technology explanations.
There is also a control unit attached to the active tank, do you the purpose?

That is basically the pump that pressurizes (and evacuate) the line going to the injector in the exhaust.
It has another heating element in it (there's one inside the tank and one in the delivery unit).

It also evacuate the line before shutdown: this avoids DEF to christalize or freeze in the line going to the injector.

[Txchoicemotors]

I am willing to be a test guinea pig on this if anybody wants to point me in the right direction. I can get any type of resistor if i know exactly what to buy, and if anybody can give me a some what of step by step direction, I have full access to the tank with the bad sensor and front bumper off the car for the next 2 weeks, i'm more than happy to experiment

[Txchoicemotors]

One question, does the pump on top of the tank cycle 100% of the time or only at certain intervals? My pump does not show any codes whatsoever but also doesn't appear to be moving any fluid at idle ( I cannot test on the highway as the car is currently disassembled and this part is cracked

[bawareca]

The pump only transfers DEF from the passive tank to the active tank when the level goes below certain treshold. My pump is bad and i have code for it, but it doesnt trigger CEL.

[Txchoicemotors]

Oh, well this one is cracked from the accident I came across one earlier on eBay for $170 so I grabbed it vs $400 from BMW, I just want to avoid future problems

[Txchoicemotors]

I have a 2 wire ntc thermistor in hand 50k ohm at room temp and goes down with heat applied, as well as a simple 2w 10k resistor

I just need to know what wires to splice now

[bawareca]

^^^White and black wires on the big connector(smaller size). Thicker white and red are for the heater.
If it doesnt go much below freezing in your neck of the woods i would go with the resistor.

[Txchoicemotors]

So just cut a splice both white and black wires and install the resistor between both wires vs in line with just one wire?

it hardly freezes here very often but if you have an ohm meter to read the resistance of the sensor in your damaged tank maybe using different temp ice water we could try to find a thermistor that would work properly in the system, if not no big deal.

I will try the 10k ohm resistor first and report back.


One other thing, this code is new. I am not sure what this code is. Driver detection?