When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
I am keen to further understand how amazing chaos can snowball issues across CAN networks: B, C, E1/2.
Such as the solderless MFK Gateway arguing with Instrument Cluster Gateway to display road signs and speedlimit info... makes two busy gateways tied-up unable to transact timely data for others.
The dedicated Bosch Central Gateway hosted in the F-SAM module may very well be the chief troublemaker - Now we have a chance to find out by only resetting it alone!
Today I was helping un-confuse people about battery issues (link) when I got a quick shortcut idea....
What if we just pull the power away from the power controller... reset the crazy N93 gateway ?
We need to know what Fuses # power the CGW Gateway and the F-SAM... I guess N93 + N10/1, right?
CAN modules are allegedly able to be distressed under the radar without any published fault code until is wonky disrupted.
> How is that...?
I am talking about improving the latency performance.
CAN resiliency is built upon retransmits.
When a node can't get no satisfaction, it keeps retrying.
The gateway literally keeps track of each individual outstanding message connections... that queue depth has limits. Limited in size and limited in slices of CPU time.
The time necessary to go through the gateway grows and keep getting longer to a stumble.
When everything is fresh, CAN messaging flies. When connection skeletons piles up, then gateway management crawls.
Not something Bosch architects ignore!
Most modules are instantly reset by cycling the ignition power.
However a small portion of the car is powered 24/7 non-stop to manage the wake-up signaling. I know F-SAM is one and I am going to guess its roommate CGW is two.
So let's streamline the reboot process by only targeting these two VIP modules fuses. What do you think?
More power to ya, no soldering this time.
Last edited by CaliBenzDriver; Jul 21, 2023 at 09:32 PM.
Reason: ⚡
Make good sense Cali, thank you.
The CAN is half duplex I read, so imagine so many modules taking turn to broadcast message and then N93 gateway got choked like airport during Christmas
Let me take a look deeper into the N10/1 Front SAM and figure out, where does N93 take power from within inside N10/1 ?
I basically have traced all fuses on N10/1 and none claimed to be for N93, well I never knew back then when doing that fuse check that N93 is inside N10/1 hahahaah.
N93 existence is only visible on a scanner or Xentry by it being visible as N93 in a full scan and the topo by Xentry.
Otherwise in all wiring diagram, we will never see N93, we will see only N10/1 front SAM.
Ok, let me re-visit N10/1 ........................
N93 close partners
I guess if you follow the dataline from the DLC connector it will lead you to what the diagram references as CGW N93... F-SAM N10/1?
Same thing with the A2 gateway, the diagram is going to reference what they are calling N93... N10/1?
It's time I drill my fuses to put them on a leash exactly like you did. Hopefully tomorrow so I can drive/report this Sunday...
N10/1 fuse would make an excellent "kill switch" for security and a "deep sleep" on-demand to discourage battery vampires wake-ups
Oke, I think I know how the N93 gateway and the N10/1 CPU gets the power.
First we re-read the description of Front SAM to get a better idea on how these two processors are different in duty and function.
We will call one as CPU and the other as N93.
Let's imagine the electronic section of Front SAM is a single PCB mother-board, not 2 separate PCB boards where N93 lives say on west side and the other one where the CPU lives being east side.
I been seeing this TRANSISTOR label thingy 2 years ago but never understood what it meant, except it is "electronic" thingy.
Now that you raised this N93 question, it all make sense
The good and bad news. Good is that now we learn, the bad news is that there is no way to reset N93 on its own, without removing battery power as in DISCONNECT battery
No reset button like our wifi router at home.
The CPU and the N93 gets double power, from 30g and 30. This set up is similar on my yachts for throttle and transmission controller which is managed by 1 module.
The power sources are 2, one from Left engine battery and Right engine battery, hence the diodes are there to prevent paralleling between 2 battery banks, while the device still get dual power source.
My "imaginary" duty separation between CPU and N93 . N93 wont be allowed to handle power, only data. CPU does the power switching as described by MB.
Here is a re-drawn CPU & N93 for its CAN BUS "imaginary" wiring, if I translate it correctly from the MB WIS.
What I wonder is, older model W212 with small 1.2ah battery inside the dashboard powering the N73 Ignition Switch as back up power, that one technically never sleep ever, even with main battery disconnect.
Will this N73 be the reporter when power is lost to N93 & CPU of N10/1 Front SAM and created stored codes which may confuse people ?
Here is my naked F-SAM getting a protective coating
facelifted SAM + CGW
I recall it's not a friendly module to repair.There are dozens of heavy solders that prevent access to the other component side of PCB.
Above picture shows two identical microcontrollers side by side.
Ouch, if only both side can get conformal coating .
When I can remove my Front SAM, I will conformal coat it too
I can get a good brand locally , spray type which I wanted to spray my UPS/Inverter for Bali home. I lost 2 UPS in 4 years.
1st UPS was insect sucked in and exploded, also corossion too.
2nd UPS, I filtered all air intake...... the coastal air killed the logic board hahahah. 18 months only.
I total loss total 4 TVs already in Bali, all from death by minor corossion. and none exceeded 4 years old. repaired 3 units too.
In memoriam of a great commercial sea diver: Fred Hurt who was diving regularly all the way to 79 years old.
He was the leader featured in White Water Gold-Rush. Thank you Sir 🙏
His relentless quest despite handicap was really inspiring.
I used to be licenced drysuit, search and rescue open water diver. Today's deep digging dedicated to his hard earned legacy.
So here is what I got for you... a big CAN MAP
Excuse the screenshots, the VM is hardly configured,
the Bookmarks don't work, got take pictures instead!
So I found the N93/7 they call it "chassis gateway" (bingo!) as well Flexray controller.
Its fuse power comes from N10/2... (why take the short route N10/1 ) .... of course this could be an entirely different device than N93 CGW.
FYI: I have not managed to get results from search by short names ("N93") - The tricky search tool is beyond my understanding to use at all.
To get anywhere I have to surf from diagram label to diagram label (the one with a pointed hand!) - It's so uneasy it feels results are random. I got to see lots of interesting stuff like ...
The CAN-E2:
Steering rack!
TPMS
RFK
Static LED L+R HL
N93!
Diagrams do show CAN-E2 going into N10/1, meaning N93 its being confused as F-SAM directly powered by 2x 150A pre-fuse links.
So, yeah my head is spinning now from all the deep WIS diving. R-SAM related assignment
I need to do more digging around N93/7 before I start pulling fuses, my reboot goal here.
++++ Questions to self regarding N93/7...
What is "DistroPlus" module ID ??
Search through CAN-E1 to find CGW??
Digest whole CAN gateways vs. solderless chaos.
Solderless MFK-Gateway crimpings are not DIY friendly for reassembly. Built-in high load heater supply helps its loose pins oxidize, gateway chaos distinction. Even single front camera has the same heater ordeal.
++++ Troubleshooting... Lottery!
There are so-so many different build options according to countless factors and preferences that each car is virtually a unique combination.
That makes it harder for anyone to diagnose issues without knowing what's involved from car to car.
One thing that always help is knowing what to expected. For that reason, diving around WIS interface is tough but excellent learning opportunity.
Last edited by CaliBenzDriver; Jul 22, 2023 at 09:21 PM.
I think I got an extra reason to celebrate 4Th o'July this year !!
I found the elusive "5mn Reset".
I was searching for a way to power-cycle CGW by pulling its fuse...
I could not so I had to reboot whole chassis.
I found out that VARIANT coding does reboots modules quickly like I wanted.
exiting Coding reboots CGW
Why does CGW gets unstable?
The whole chassis wants to talk to CAN-C VIP modules.
CGW manages every transaction across the gateway.
scan report of faults
This shows how busy CGW gets to work network translations.
All these modules keep shooting transactions at ESP + ECU + TCU...
The greater volume of transactions the more left-over junk memory leak result. CGW runs out of available processing space then starts dropping low priority packets.
This starts a repeat-traffic storm that guarantees a traffic jams like "voltage yo-yo".
> OVERALL UNDERSTANDING...
MB suffer from multiple factors that in the end detune timings of both engine + tranny.
Use as many technics available as possible to reduce known chaos so it does not compound into bigger headaches (misfires).
Last edited by CaliBenzDriver; Jul 5, 2025 at 08:52 PM.
- Coming out of VARIANT Config reboots module to load new settings (even if no settings were changed!)
- We only want the reboot part to clear unstable CGW.
> SOFT vs. HARD:
-- This method is a "soft-reboot" ie. power stays ON.
-- A hard reboot is whole chassis power-cycle
-- WHY use soft/hard :
soft reboot is faster in the field but targeted to one specific module.
Hard reboot is 100% all modules but takes longer to switch batteries on > off > on!
By targeting only CGW, we're able to conclude it seems like the #1 chaos troublemaker + recipient.
If you've come to appreciate the difference in stability REBOOT brings.... try the soft version to see if results are similar.
The prefered way is still Reboot + Floatover CGW only.
> ALL IN ALL...
What does reboot + float + CGW have in common... is chassis instabilities.
MB uses multiple factors to create live chaos traffic storms. Our best interest on the other end is to preserve basic performance by limiting chaos.
rogue firmware versions
mismatched variant configs
soaked modules/ bus-bars
voltage stock swings
dismal voltage yo-yo
salted chassis strap
painted GND nuts
Solderless random glitches
Undersized capacitors
nodes on different thresholds < nxt
... misc more unknowns.
All carefully managed for luxury ownership.
> ECU LOCK QUESTION...
I am unable to "Variant" code my ECU !?!
My ECU config has a legacy "power steering pump" option I really would like to disable for my electric rack.
This is the type of mismatch setting that creates overhead for real time processes.
Bosch MED17.7 module seems "locked" or disabled.
How can I go pass that to further test???
Last edited by CaliBenzDriver; Jul 6, 2025 at 04:55 PM.
So in Vediamo 4.2.2 I open the cgw231.cbf file and just perform a SOFT Reboot and then observe if there are any differences in the cars behavior?
What shall I specifically be on the lookout for, remember that I am on a OM651 Diesel.
Yes, this sounds like it.
I don't know if your OM651 is impacted by Network chaos... Diesel has no multi-shot spark ignition. Lets find out quickly if direct injection alone register a difference?
Try with the CGW soft-reboot to begin with else whole chassis power-cycle as a base step.
Go on a 5Miles city drive to appreciate driveability from idle to 3000Rpm. It's all about response, not Max Dyno power.
See if you get ...:
stronger accelerations response
more torque at 1700.Rpm
how gears shift based on throttle
Repeat with "whole chassis reboot" given there are no active engine nor networking issues.
You have a winner if it makes your "chassis lighter" and responsive. If it stays heavy, nothing gained nothing lost.
Perhaps only tranny might lighten up. I see this beeing effective for a couple days up to a week. Then bingo another round.
This CGW Reboot is NOT to be confused with any ECU / TCU fault clearing that does trigger a mixture relearn.
This is purelly CAN NETWORKING!
Last edited by CaliBenzDriver; Jul 10, 2025 at 07:25 PM.
stronger accelerations response
more torque at 1700.Rpm
how gears shift based on throttle
Perhaps only tranny might lighten up. I see this beeing effective for a couple days up to a week. Then bingo another round.
I did it, but there was only a HARDRESET available so i did that one.
Dont know if its the placeboeffect but it seems like a feel stronger pull from start and the shifts are more correct.
I did it, but there was only a HARDRESET available so i did that one.
Dont know if its the placebo effect but... it seems like a feel stronger pull from start and the shifts are more correct.
Right, exactly!
At first this is a surprising discovery but then you begin to wonder how to ENABLE this FULL TIME!
I'm going to coin this result condition: "BEAST-MODe"
This is still early in the research & development.
Access is provided by a combination of module refresh:
Basic battery disconnect: REBOOT!
Simple LAUNCH Scanner: restart x,y,z
Zero computer necessary!
My current target is testing how to enable this full-time by disabling network stress sources.
One the important aspect here is to witness how the ECU/TCU excellent performance is disabled by network chaos.
These VIP should not be externally impacted do much, yet they are!
@FLASHBACK
With your computer tools can you see what on the network gets improved by "BEST-MODe" ?
So far I see targeted modules reboot is effective fir research. However complete chassis REBOOT step is more effective way.
Last edited by CaliBenzDriver; Jul 14, 2025 at 04:28 PM.
Calibenz, would you ever take your dry suit diving qualifications, and experience, and go to a remote part of Alaska, get into glacial melt water, barely above freezing, dive into a rapid current, that literally is creating whitewater, have car sized rocks all around you while you are in the water, ready to topple over, pin you down, and doom you to a pre dug ice water grave?
Not to mention almost vertical canyon walls with more loose rocks sliding down the slope, into the water, to hit you on top of your head?
I love that series. I wish those guys would strike it rich. But all they seem to be able to do, like so many other gold prospectors is barely find enough gold to scrape by to the next season.
Fred was a legend in the game, and his son is a chip off the old block.
Last edited by MB2timer; Jul 16, 2025 at 02:23 PM.
Reason: -t+y
Calibenz, would you ever take your dry suit diving qualifications, and experience, and go to a remote part of Alaska, get into glacial melt water, barely above freezing, dive into a rapid current, that literally is creating whitewater, have car sized rocks all around you while you are in the water, ready to topple over, pin you down, and doom you to a pre dug ice water grave?
Not to mention almost vertical canyon walls with more loose rocks sliding down the slope, into the water, to hit you on top of your head?
I love that series. I wish those guts would strike it rich. But all they seem to be able to do, like so many other gold prospectors is barely find enough gold to scrape by to the next season.
Fred was a legend in the game, and his son is a chip off the old block.
unstoppable by nature
Fred Hurt had a golden mind and always the best attitude to combat challenging obstacles. A true legendary inspiration like Grand'pa John Schnabel in the Gold-Rush show.
He stayed focused on overcoming all hurdles to achieve his goal of cleaning a plunge pool.
What a great guy on a mission! Always first in and last out... Never a problem too big to solve, only solutions.
Absolutely no limitation could keep Fred from working the hardest.
This TV-show is a great tribute to retired US service men!
Last edited by CaliBenzDriver; Jul 16, 2025 at 05:50 PM.
Mercedes SLR McLaren 722 S Is Extremely Rare Example Modified by McLaren
Slideshow: A one-of-one U.S.-spec Mercedes-Benz SLR McLaren Roadster became even rarer after a factory-backed transformation at McLaren's headquarters.
and none exceeded 4 years old. repaired 3 units too.
