I've had a 203D code (from a generic scanner) before and I just filled the AdBlue tank cleared it and it went away. But the tank wasn't really very low. I recently got a better scanner and it has the 203D17 Code Fill level sensor 'Adblue' has a short circuit to positive.

I looked at the live data and it's showing:

Absolute tank value of 25L
Relative level of 80%.
A103/1b2 (Fill level sensor FULL) - NOT WET
A103/1b3 (Fill level sensor RESERVE) - WET
A103/1b4 (Fill level sensor EMPTY) - WET

I haven't added any DEF but if I rock the car the FULL will change from NOT WET to WET and back.

So it looks like the sensors are working. I've seen on the forums that there are resistance values to check for the sensors but nobody has stated what they are. I pulled the cover off of the tank and got to the connectors and tried testing them and got weird readings.

If I'm reading the wiring diagram correctly the level sensors are read from pins 9,10,8 of the 10 pin harness connector and the supply to them is from the 8 pin connector pin 4. I was expecting to connect the meter to pin 4 of the 8 pin and then check resistance to the other three pins. But I was getting voltage! Not much like a half volt but with those 2 connectors unplugged there should be no electrical connection to the tank. So is this a galvanic response? Does the UREA act like a battery acid? If so is that what it's actually measuring not resistance. I get voltage between the 3 level sensor lines also.

Could this be from bad fluid being in the tank? I bought the truck about a year ago and the first time the light came on I drained (evacuated) and refilled it.

Also I scanned when I had the Level/Temp connector unplugged (the 4 pin at the tank end of the harness) and the code changed to "shorted to ground" all three were NOT WET and the temp read -9C. So leaving it unhooked wasn't an option.

Anything else to look at? I know I can replace the heater/sensor module for $$$. Are the sensor probes exposed and could be cleaned?

 

I heard the stories that DEF fluid can age, but technology is too new to get solid confirmation.
MB made the tanks big enough to add fluid only at regural services > 10k miles, so I guess aging was not concern.
Coming to your problem - short circuit should be self-explanatory, but how much we can trust the scanners? Per my experience, they are not precise at all and programmers who make the scanning programs often don't know what car has under the hood.
So my basic is always ... basic. Start with careful inspecting the wiring for damages.
Than take a can of MAF or other cleaner, pull the plugs and give them good bath - allowing to dry after.
Than add the darn fluid just to eliminate possibility of it being too low.
Bare in mind that at some point the system can put you into limited start-up number and then you will have to run to a dealer. .

 

I cleaned the three connectors I could get to without dropping the tank and inspected what I could see. My plan was to next top it off and see if that "fixes" the light again. Noting how much fluid it takes and checking the fill state with the scanner as I add it (a quart at a time).

I just filled it in Jan. - under a 1500 miles ago. I was hoping to not have to constantly keep it above the "FULL" sensor. I'll end up adding DEF every other fill up.

What gets me is getting a voltage reading across the sensors - just half a volt but there shouldn't be anything if it's measuring resistance. Can anyone confirm if it is just a resistance thing and if it's binary (that is each sensor is either closed (wet) or open (not wet/dry)) or if they somehow measure more levels (like 0-1k ohm to estimate level). With three sensors I would assume they are just simple wet/dry and they only have 4 levels that can be read. Above full sensor, between full and low, between low and empty and below empty.

DEF us corrosive to copper so my fear is that one of the sensors (or the lead wire to one or even the temp sensor or heating element) has exposed copper and if corroding and creating the voltage. I haven't tried dipping copper wire in DEF and seeing if it does produce a galvanic response yet.

 

It was only down 3qts. And after filling it says all three sensors are wet, but still only 80% full. The tank holds around 40 quarts I think so it was close to 93% full.

 

Well I read on a Jeep forum to drain/fill with distilled water/drain again/refill with new DEF to get rid of any crystals that might have formed on the sensors. I tried that and it didn't change anything.

I tested the wiring again and found nothing wrong with the connectors. But I did get voltage readings still from the sensors themselves (when unplugged) when testing to ground. I did this with the big connector removed from the AdBlue Control module so if the wiring diagram can be trusted there were no electrical connections to the tank assembly.

So I decided to go another route and "bypass" the level sensors and just make sure I don't run out.

I pulled the connector that has the 3 sensor lines and the signal wire (4 pin connector inside the fill area on top of the tank) and created jumpers for the 4 lines so I could check things out. Pin 4 (signal) also supplies voltage for the temperature sensor (that is on a different connector) so I left that connected all of the time (if unhooked it shows -30c as the temp).

I discovered the following:

Unplug them and you get a Code thrown for "Fill level sensor 'Adblue' has a short circuit to negative."
If you short them together and you get "Fill level sensor 'Adblue' has a short circuit to positive."

These codes would throw as soon as I displayed the values.

So it needs some resistance, so I wanted to check the resistance directly on the sensors, but the resistance would never settle on a number, it was constantly changing on my digital meter. I'm assuming that's because there is voltage (between 100-300mv) showing on the three lines. So I connected a 10Kohm variable resistor to each sensor line in turn and checked to see what value they changed from "Dry" to "Wet". I had to take all 3 out of the circuit because if I left any connected I got the shorted to positive. Since they all use the same supply line I had to do one then then put in a temp resistor then do the next. Then after getting all of the temp resistors in I went through them again and got different values that triggered the Wet/Dry change so I kept updating the resistances and looped through several times until I got the values I list below.

For the "Empty" and "Reserve" levels I used a resistor that signals wet. For the Full one I chose a resistance that shows dry. I haven't found the exact levels that each one is supposed to trigger at but the "reserve" is listed as 1 gallon on the e-class. That FAQ indicated that "Empty" = No start and "Reserve" = Low warning light. So I figured as long as the computer knows it's above that it will be good. But the computer might think that if it sits at full all of the time it might indicate that it's not being used and throw some code. Who knows it might throw a code if it doesn't see the values change change - time will tell.

I ended up with
Empty = 1kohm reads wet
Reserve = 1kohm reads wet
Full = 5kohm reads dry (2.2kohm reads as wet, the cutoff was just over 3kohm on the variable resistor)

I cycled displaying the values multiple times and done several engine starts and no code so far. I'll update in a few thousand miles if it's still working.

I put in 7.5 gallons of fluid after the drain/flush so it's almost full (listed capacity is 8.3 gallons) and I can see that the level is close to the top. I'll just check it every fill up for a few thousand miles until I get a feel for how fast it uses it. I'm adding it, DEF, to the oil/filter change schedule. It's supposed to last that long anyway according to the manual and be topped off or drained/filled at oil changes anyway.

 

I have the same issue on a 2011 R350 which is the same wiring as your GL. Just curious if your current setup has resolved the error codes?
I intermittently get NOx codes (P2201) which I hope are related to the adblue tank issue. At one point I had the adblue pump not pressurizing the tank. These were all due to bad connections with the harness for the tank.
I think the error codes generated are because of bad connections. I repaired the adblue tank heater and did not connect the adlue feed line properly which caused adlue to spray on top of the tank. Once adblue coats connectors and dries
it will short out the connections. I'm tempted to remove the tank and rinse/dry the wiring harness to see if this resolves this issue.

 

The code I was getting was just for the AdBlue level sensors not the NOx. Those sensors are in the exhaust pipe (I know because I had to replace one last year). So my "Fix" probably won't apply to that code.

 

Interesting thread, rosenhauer. Thanks for posting.

 

Actually there is a shelf life for AdBlue. Its about 18Months if stored properly. In bad conditions like high heat it goes down to 6 months.

Im not sure what exactly happening to it but it is what it is.

 

yeah, I googled it and indeed, 18 months under ideal conditions. I'll be checking the manufacturing date on the 2.5 gallon jug next time I buy one.

Curious though, would expired DEF throw a code or simply not diminish the Nitrous oxide emissions?

 

Peter, does it really matter what will it do to it?
I just realize, I usually top-up AdBlue each time when I do oil change, so it newer goes out completely so I have old & new mix.

Maybe I should let it run all out. ?)

 

Yes, Dan da Man! It does matter. If old DEF simply isn't as efficient with no other ill-effects, then that's one less thing for you and I to stress about *wink*.
On the other hand, if the now-antiquated crap causes the ECM to throw a trouble code, then you and I will bask in glorious stress on an even more regular basis.0

If that's even possible LOL

 

If you let it all run out your car wont start! If the level drops below the "Empty" level you will get a countdown for remaining starts.

Mercedes does state that every other oil change (B service I believe) that you are supposed to drain and refill instead of just refilling. I assume that it so get rid of the "old" fluid. But I've read on other threads that a lot of dealers just do refill at each service.

 

Not true. The car will warn you a total of (if I recall correctly) either 9 or 19 times, one each for every on-cycle of the ignition. Even just 9 is more than enough heads-up.

I was going to top up soon, having poured almost the whole 2.5 gallon jug in last fall. Instead, I'll wait 'till I get the warning.

 

Is there a drain plug for AdBlue tank? Wondering if I can do flush by myself.

 

No drain. I used a cheap hand pump that I will only use for DEF to avoid contaminating the tank.

I set the suspension to the raised level and once I got the flow going it siphoned out at a good pace to a jug in the floor. I had several of the 2.5 gallon jugs from prior DEF fills.

 

So, is there any evidence that DEF which is past the expiration date is bad for the engine and/or will throw a code?

 

Whati meant was that you can't run it out. You will get the countdown before it's completely empty. So you would need to drain it still if you want a complete removal of the old DEF.

But waiting until you get the countdown would get it as close to empty as you can.

 

Rosenhauer,
Did your bypass mentioned 0n 4/25/2018 work long term?
"Empty = 1kohm reads wet
Reserve = 1kohm reads wet
Full = 5kohm reads dry (2.2kohm reads as wet, the cutoff was just over 3kohm on the variable resistor)"

I have a 2015 VW Golf TDI with bad tank sensor (Issue with Level sensor G697) and I'm proposing a similar fix but all mechanic's want to do is sell the DEF Delete kit >$2000K! Cars not worth that just needs to last 12 more months, to get my son through college. Can anyone help?

 

I'm having a similar issue...

Code 203D17 read via Xentry for "The AdBlue fill level sensor has a short circuit to positive".

Sensor EMPTY (A103/1b4): 0.0V
Sensor RESERVE (A103/1b3): 0.0V
Sensor FULL (A103/1b2): 3196.00mv

That's the data I got from the stored fault (MIL came on today), but when I go to read the data live from Xentry it shows RESERVE & EMPTY as "IMMERSED" and then FULL as "NOT IMMERSED". The tank is not full, I haven't filled it up in quite a while. Anything to worry about? Any way to fix or replace the level sensor? Is it a separate part from the tank itself? I couldn't find anything on EPC that looked like a level sensor, but gawked at the price of the tank itself when throwing it into Google. Cleaning I suppose would potentially be an option if I knew what to look for and how to do it.

 

The level sensor on w213 is integrated into the adblue tank.

 

Do you know if it’s integrated for the W212 E250? I couldn’t find a part labeled level sensor online, however there is a large unit this goes into the top of the tank called the heater. Is it possible that is both a heater and a level sensor?

 

On w212 there is "repair kit" that can be boughted separately.

 

DEF level gauges on MB are ultrasonic. I believe they have to be mounted at the bottom, but as with new technology, ultrasonic gauges come with very limited knowledge on the forum.
Even less with repair shops.