What if I told you I found a vampire draining our battery during sleep??

The F-SAM has a setting used to keep poling the windshield rain sensor in case it needs to safely close the sunroof on rainy days.

I think this is outstanding for a convertible to cover up in case of rain - but on my sedan I do mind my big battery doing unnecessary overtime during sleep.

That one sensor thing when bundled with other favorites give me over 3 weeks of normal parking without CTek on. Before it would be slow poke start after 4 to 5 days depending if one module git jammed ON.

Longer parking (6 weeks) may be possible but cruises are only so long.

I am not withholding any unpublished information... LOL - To this day I still indulge in Reboot to please the F-SAM-Gateway at least.

 

Sweet info. How often does it wake up per hour and for how long ?
Have you measured its extra drain in milliamps ?

ADD:
Your post got me thingking about my friend's W204 which his rain sensor is declared defective by Xentry.
Was his higher amperage spikes during sleep....in this post #4 https://mbworld.org/forums/e-class-w...ml#post8836714
due to othermodules trying to request rain condition from a dead rain sensor ?


DAY 1 test - W204 with dead rain sensor


The lower "zoomed" file are the consumtion spikes which I separate from the normal 4-5 milliamps and merge them each with 1 second extra interval as zero milliamps separator,
to see how long in seconds these up to 75 milliamps spike last.

.
Day 2 test - W204 with dead rain sensor



.
My W212.065, with all healthy modules. Yep, not the same ECM and level of options compared to W204, but for comparison sake :
Here is a simple log I did. Below is 15 minutes worth from total of like 25 ish minutes log I did. On average the spike is at 10 milliamps only.




My W204 friend does not want to fix his rain sensor, so I guess I can never have another battery drain session to compare if having a proper working rain sensor ...will 75 milliamps still be the peak ?

 

Wait... my W212 (MY 15) will close my sunroof if it starts to rain? I had no idea...

 

Yes and that's active with no opt out menu option. You need to use a OBD LAUNCH Elite or MB
Personally I say "no thank you!".

You will find that under F-SAM > settings options. I was in a rush with other things beside this nugget.

I will be back on there looking for other troublesome features.
When CAN-B is loaded with bugged up door modules, this rain-protection help dump your battery with extended power cycles!

The easiest shortcut is to use a CTEK float charger. The battery is a popular target.



​​​​

 

Why disable it? I imagine it has resistance detection so it won't chop off a finger? I'd like to test it, maybe keep it open and pour water on the rain sensor?

 

Yep, pano roof and windows on mine has the anti chop-off-our-finger feature. Under 10kg of load will activate them, that is my estimate when I use my hands as brake

 

yes, normally door and roof controllers have auto-stop load detection if it was trained well.
It watches for a high motor current value to know when it has stopped.

​​​​​​
Disabling the "rain sensor polling" is an effort to spare AGM battery charge. It does not interfere with basic "chop-chop" detection... same module different function.


I mentioned the 4x DCU or other CAN-B controllers being soft-crashed(*) that slow down the polling process to become more heavily taxing for each cycle.

The battery trade-off for rain detection while sleeping makes it a no brainer. I'd much rather have a great battery.

***: this is what's materialized by crazy Keyless delayed door locks lock/unlock.

 

Is there a cutout when all windows and sunroof are shut? Seems odd to ping something when everything is shut and no danger abounds.

 

The F-SAM Power management has plenty of circuits, timeouts and sleep levels. Perhaps this is only active 6Hr and expires during extended inactivity.
yes and even while driving with rain when sunroof is opened nothing happens, wipers active or not - Roof stayed open


I have to retrace my original steps. Yesterday I found other pieces related to rain self-closing...

 

6 hours is the max limit a healthy no issue W212 like mine.... will wait before going to proper DEEP SLEEP for the modules and K2 relay disengaged..... this is when and if
the has been run form some distance.

K2 relay disengaged is a simple confirmation the W212 is in DEEP SLEEP.

If the car been sitting for exceeding a day in the garage,any door opening will wake up modules and activate K2 relay, but car will fall to DEEP SLEEP very fast, under 30 minutes.

Hood open is OK, no need to trick its locks (2), MB computers understood that as Workshop Mode and will ignore the hood being open.
My hood is always open when car at home.

Bear in mind my model is a facelift and there could be a difference to older non facelift deep sleep strategy for open hood concern.

 

CaliBenzDriver, I have been reading your posts in this thread and others. Thank you so much for sharing your knowledge and helping us! I have three Mercedes and do all my own work. The most recent addition to the family is a 2016 w212 sedan and it appears to have this battery draining issue. I have replaced the key fob battery, have an auxiliary battery on order and will likely replace the main battery. I will check the primary battery voltage at rest (engine off) and I’ll check the voltage at higher rpm to ensure the voltage regulator is happy. And, I have a CTek charger on order. I’ve become a CTek fan and have one on my other Mercedes which uses a Li-Ion battery.

If you come up with other ideas on how to address this battery vampire, I will read them and adopt suggestions.

Thank you again for taking the time to help us!

 

Dave, thanks for your interest in normalizing bugs out of your Mercedes.

It's fun to fix them better-than-new once we understand where the opportunities are hosted.

Researching topics in-depth with MS is really fun too.


> Basics turned Advanced...:
-- Let me summarize what I have tested to help your chassis quit draining MAIN then AUX when voltage deeps below 12.5V threshold.

-- Here we are dealing with the combination of SOLDERLESS + PAINTED GND applied to numerous CAN-B modules. They do a very amazing job as well on high speed CAN-C... not for here.

-- It is basic because poor GND built by MB are amazing.

-- It's advanced because when you network data in partially crashed modules, it's tough to understand what's happening on a large scale.

-- I don't believe we can visualize the error rate per module. So the challenge Bosch gives us is to prioritize repairs with missing information.

-- We are going to deal top-down: from VIP module to dumbo. The count of soft-crashed dumbo is whats overwhelming the Gateways yet we need them in top shape to handle normal flow volume.


> 2 birds with 1 stone!
-- Here we are going to incrementally improve both the network performance and battery life.

-- There is no good reason Mercedes must quickly drain their battery more than other cars do. It's a nasty defect we are going to retire.

-- Dealing with CAN-B we have the most populated network with all secondary misc. modules that are not engine related. There's like 20 units that must work side by side with organized talk/listen. Unfortunately that part does not work well as implemented by the best German experts. The timed out modules endup flooding the gateway stacks with requests. The Mgt matrix to gauge the network error rate network is not used to trigger a fault.... the best or nothing. We have a network without anyway to administer it.

The F-SAM power manager keeps polling regularly this whole string of modules then back to low power.
The issue is it doesn't happen fast enough and too many live modules are awaiting to go back to sleep.

All right then, let's make that work normally.


> PRACTICALLY...:
-- We all love dismantling our car into pieces, right guys... not unless we have to!
Let's keep the overhead time low: I always recommend to disconnect all chassis when pulling the modules in questions. It's not plug and play at all.

-- reWork method:
We as owners do not have the same objectives than mechanic shops. We can fix multiple things together at the same time across different topics.

When you visit something you don't want to have to come back. Fix A to Z else you may end up doing busy work for nothing. Inside the doors grease up the window crank and solder the seat memory buttons as well as windows buttons.... at once!
When pull the headCOMAND Unit to access the EZS Key switch you then clean comand fan filter.


> Suggested Work Order:
1 - solder EZS Ignition key hole
2 - solder KeylessG module
3 - solder the 4xDCU door control modules
? - (check what is SCM connecting to)
4 - Solder Rear-SAM as a target partner!
5 - Trunk lock module
6 - Solder the AAC HVAC controller (manages humidity)
- (Bus-bar distributors have no built-in problem)
(OCP module is built with ok soldering)
- Fix painted GND (Footwell, trunk, engine bay)
- deal with main chassis GND strap (clean + 2nd)


> EASY TESTING:
Once you deal with the top 3 items above, chassis should begin reacting much faster and without random stumbling delays to the lock/unlock command including the short horn actuation.

These nice doors modules are architected to generate delays (*). Rear trunk lock module depend on your chassis.
*: DCU check status one slow assynchronous step at the time instead of having a status with a wake-up trigger. Making 4x DCU CAN-B faster is a quick win for "team batteries".


> THE TARGET:
your goal is to help the whole bus of devices transact in a snap then re-enter low power in Not minutes but 1/2 second.

Fixing solderless modules leads to radical network speed performance. Conversely degrading performance is easily achieved by network chaos caused by poor GND, solderless modules, software settings.
You can't buy your way out of this... unless clever on-line services or sponsors offer basic refurb & soldering $100 or $200 per box according to work extent (Big SAM vs. tinny KeylessG).

> HORIZON:
Once you have the basic CAN-B mastered you can get real tackling high-speed CAN-C.
It is some of the same classics in a different cocktail.

For know pay attention to how your tranny is working with your engine in relation to accelerator. ECU+TCU timely work squarely rest on effective transfers. It's totally obvious once you know it!
Cancelling CAN-C latencies is a truly amazing prize.
​​​​​​

 

Dave,

You can start to verify during car deep sleep and before deep sleep , what is the actual current drain of the car.
From there you can expand to see the direction of the next investigation, when and if your car ( all 3rd party accessories logged first and then removed ) own modules are
consuming more power than it should during K2 active/ON and after deep sleep K2 OFF.

 

You know troubleshooting is a little bit like fishing. You never know exactly what size you're gonna catch but you have to start with a setup in relation to what you're out for.

So here we are going to split big drain vs. small drain. There are different ways to catch them


> WHAT KIND'O DRAIN:
Everything is about testing to help figure what you're dealing with most likely... it's never-ever easy to deal with chaos in organized system.

-1- If the drain is accute, test for fast type.

-2- If the drain makes your new battery crank slowly and tired all the time, it's the slow type.

The best thing you can do is to define what the car is doing wrong. Slow crank... how soon? Dashboard Voltage display read high Amps. - See story at the bottom -

> FAST DRAINS in Amps:
Use Surya's method above to help find high drain from failed swampy modules.
This is the kind of faults that drain Batt. very shortly in 1 or 2 days, not 2 weeks.

When any of the unprotected circuit boards see water, copper traces develop green crusties and the module functions go nuts. That's one way of finding bad : F/R-SAM, OCP, ....

DO NOT DISCONNECT Modules from powered car and expect everything to act normally. Modules in low power condition are still powered and active to answer regular polling (Headlights, ECU, SAM, DCU, KGO,...)
Use DC-Amp clamps and DVM to measure drop voltage in fuses ie. current ie. drain activity.


> SLOW DRAINS in mA:
The accelerated drain from extended polling delays makes any new battery act old quickly.
This type I have described in my previous post#12.

GUESS WHAT... this drain by design comes from factory solderless so count how many ppl have it: eveybody minus one!

Currently I sport my ten year old OEM Battery ....
It's good, no reason to swap.
It gets drained zero like a best car.
It gets full after 10mn driving
It never see significant Neg--Amp while driving!!!
(It does have reduced capacity: good for testing!)

> Factory Built-in Features:
Fix in that order:
Single GND Strap + SolderlessBs + Yo-Yo:GND (+ Pump solenoid).

When I was canceling the crazy Keyless solderless game, I was also after fixing crazy Yo-Yo... Instead I got reduced latencies everywhere applied.
Out of all of that drain game receded so I kept experimenting.

Peeling off layers of bugs with like minded friends is fun

++++ CHASSIS MILK +++
I was going to do an quick add on... a thead, yes!