Anti-Skid Temporarily Off - $0 SAS DIY Technical Repair

[drlava]

I found a way to repair the common steering angle sensor or clockspring error which arises as a dash warning "Anti-skid temporarily off" or "Anti-skid temporarily disabled" and want to share it to hopefully help other avoid the $900 dealer charge. 

Tools needed: VIDA/DICE, standard socket and screwdriver kit, compressed air, whiteout/silver pen, patience and decent hands

Note: actions in this repair if done incorrectly may damage your SAS!  Follow at your own risk. Pictures are showing up in reverse order to my attachments, so references may be reversed in the text below.

Error codes fixed:

BCM-U042864 Invalid Data Received From Steering Angle Sensor Module - Algorithm Based Failures - Signal plausibility failure
SAS-C009404 Steering Angle Sensor Analog / Digital Sensor - System Internal Failures

Apparently this error is very common, and often requires a trip to the dealer who will replace your clockspring assembly which includes the airbag wire coil which looks like a clock spring (commutators not reliable enough) and the steering angle sensor assembly, which attaches to the clockspring with 3 screws.  They are sold as a kit together because they must be rotationally aligned and limited by the number of turns they can handle without damage.  The clocksping has a handy indicator built into it where you can see a yellow tab through a window when it's centered (and at/near its extremes).

The dash error of "Anti-skid temporarily off" would arise intermittently at first for my '08 XC90, and eventually stayed on.  The anti-skid BCM computer utilizes the steering wheel position to assess expected travel between wheels and calculate unwanted slippage under acceleration and deceleration conditions, modulating the brakes as needed. If it doesn't receive a reliable steering angle sensor signal, it will disable itself for the drive.

In the garage, VIDA showed normal data for the SAS angle, direction, velocity, and BCM steering angle and yaw and lateral accelerometer signals at rest (first attached image). However, during a drive, VIDA showed that the steering angle signal from the SAS and BCM locked up, froze at large left turn steering positions (over 300 degrees, attached image 2). This signaled that the SAS controller was receiving uninterpretable signals from the SAS at large left wheel positions, and it was giving up, throwing the error.

I had read that the SAS units utilize optical encoders, and suspected that when they 'go bad' it's really just some dust in the optical path, but couldn't yet see how that would occur only with large steering wheel rotations, so commence disassembly.

1) There are videos on youtube which describe in detail how to take out your clocksping, so I'll skip that here. It's not a hard job. Make sure you mark your steering wheel position relative to the splined post after removing the 18mm bolt and before removing the wheel.

2) Ensure that the clockspring is centered, and unscrew the 3 screws holding it on the steering column.  From now on your job is to mark and remember the orientation of how these pieces go together.  Flip it over and mark with whiteout the rotational position relative to the housing (3rd image, 2 screws already removed from center).
Vacuum the dust and dirt from all around it.

3) Unscrew the 3 screws holding the SAS to the clockspring on the back, and after removing the SAS, immediately mark on the back of the clockspring its rotational position (4th attached image). Set the clockspring assembly aside.

4) Set the marked SAS on a table and remove the 3 inner screws which hold the front and back coupling rings in place. MAINTAIN the position of the back coupling ring!  Carefully lift the front coupling ring off and mark with whiteout the relative position of the back coupling ring to the inner encoder wheel (attached image 5). The back coupling ring has 3 posts which protrude through alignment holes in the encoder wheel.

!! If you allow the back coupling ring to fall off, the encoder wheel will become loose in the enclosure.  In this state, if it is lifted and rotated it can skip teeth and loose its position relative to a linear encoder which is coupled with a gear. This can be bad news and prolong your repair as you try to find the correct position again.

5) With the front coupling ring off, you can now carefully pop up the clips holding the cover on, exposing the encoder electronics underneath. Be clean! this is static and dust sensitive instrumentation! ( attached image 6)

The design of the encoder is now clear.  There is a rotary optical encoder portion with diodes and sensors attached to the PCB (white silkscreened solder joints), and a linear 'raster' optical encoder portion which is the part that protrudes.  The rotary encoder can probably encode 360 degrees of travel with a 1.5 degree resolution, and the coupled linear encoder then encodes which rotation position the assembly is in with a wider rotational range but lower resolution.  It should be approximately in the center of travel.  The rotary and linear encoder results need to be roughly in agreement, or the plausibility error will be thrown.

6) You can now get your compressed air can and blow out the area of the optical rotary encoder and down the slit of the linear encoder. I suspect that my error was due to dust blocking a beam in the linear encoder area, resulting in incorrect data readout.  

7) You can now re-assemble the entire assembly, clear codes, and test.

 

Note: in my case, I purposefully rotated the rotary encoder wheel about a half turn while keeping it disengaged from the linear encoder drive gear.  When I went to test in the car, immediately invalid data error was thrown and the VIDA SAS angle readout was locked at 0.  This was corrected via a few trial and error gear relative adjustment attempts, keeping the wheel and airbag off to speed the process which of course threw more codes to be cleared on final assembly. It's possible this relative adjustment process also helped eliminate the error code, if for some reason the linear encoder had become slightly misaligned through use. It turns out there is a specific tool being sold which help savvy techs ensure the SAS is properly aligned!:

https://www.ebay.com/itm/153949744645?ul_noapp=true

This would no doubt save a significant number of unnecessary clockspring assembly replacements. I am not associated with the seller.

 

After this cleaning repair, codes no longer arise through turning the steering wheel lock to lock (attached photo 7).

I hope this helps some, and at least highlights ways that VIDA can help to identify which component is at fault via live-drive data plotting of SAS angle, yaw and acceleration sensor read-outs.

Calibration of the BCM with a new clockspring or SAS or yaw sensor can be performed with VIDA after this work by opening BCM telemetry, click the advanced tab, open BCM calibration, and then click the image of the controller box per instructions.

 

 

 

 

 

 

 

 

 

[gmsgltr]

This is awesome - thank you for sharing. I have the anti skid message on our 14 xc90. It comes and goes... I think a full lock left turn is where it trips on ours.

[Ihatespeedbumps]

Very cool write up, thanks for sharing!

[flyfishing3]

very nice.

 

thanks for sharing.

[Whyohwhy]

• My 2005 V70 often does this when parked with the wheels on full lock. Then goes into "Anti skid temporally off". Then it resets after 20 yards, and all is fine. I just live with it, like all the other problems it has.