Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Model X Crash on US-101 (Mountain View, CA)

This site may earn commission on affiliate links.
That interchange doesn't look too confusing to me, but it does look ripe for folks trying to merge at the last minute on the 101 (maybe trying to cut ahead or just really, really not paying attention).

Here is the signage at the start of the intersection:
View attachment 288725
Screen Shot 2018-03-25 at 2.57.27 PM.png


I count thirteen (13) lines of text (and symbols) in that first picture you posted. Those 13 lines are composed of 24 words and five (5) symbols. Those 24 words + 5 symbols are spread across four (4) high hierarchy sections, with 2 subsections of which each has subsections of its own, three for the left pane and two for the right pane. In addition, the left subection has a yellow subsubsection in it. The top of the two subsections has its own inverse subsection, with small text. There is no way a driver can read all of that and get any sense of what is going on, and realize that the lane drives them straight into the sword tip of a wall, simultaneously. It is quite near the worst type of signage possible.

The signs are too big, with too much space for too much crap on them. They need simple small signs, like this:
Screen Shot 2018-03-25 at 2.47.38 PM.png

In this example, there are two panes, with four symbols: a leftist diagonal arrow, a rightish diagonal arrow, and the two freeway symbols with numbers in them. Simple.

---

Another type of sign that I find very simple and straightforward is the list of next three on the left of the freeway so you know you better get right for your exit; it has a simple pattern, so if you know you don't want to know that, you can skip it. Here is an example:
Screen Shot 2018-03-25 at 2.47.42 PM.png

This tells you two things. First, it gives you an exact ordering of the next exact three exits. Second, if you happen to catch it, it tells you the distance to each.

Then, before you get to Evelyn Ave or Moffett Blvd, a sign to the right of the side of the road tells you you are about to exit on the correct exit, for instance, Evelyn Ave instead of 237 or Moffett, once you are hugging the right side.

In this example, there is just three lines (not 13) of text in a table format of only two simple columns: exit and distance. Very simple. No extra text or crap. Just information. Total of 11 words, but because it's lined up in a standard format that is simple and concise, all you have to do is scan left column for your destination and then read the right side for distance. Even if you don't have time to read the right side, the standard format tells you that, for instance, Evelyn Ave is second exit from now, so you can count exits, getting over right and slowing down, without knowing that it's exactly 1/2 mile. Or if all you could muster is reading the first column of the first line, then you know, for instance, that your exit, Evelyn Ave, is not the next exit, so you have at least one more exit to figure it out. (In that case, if you don't see the Evelyn exit tucked behind the from-237 merge onramp, you might find yourself at NASA wondering how you got there since you had to exit late, but at least you'd have not ended up on 237 stuck in Traffic from Hell.) I actually don't like my example as much as usual examples that don't have some confusing "237 to 101" wording, but I happened to pick one I use every day, and that was it, so whatever.



But in my crop of your screenshot, you can see an additional 6 signs: 1 blue, 1 white, and 4 green. What the hell?! That's just insanely intensely wrong. Way too many signs, and way too large crappy confusing signs at that.
 
Last edited:
TEG did you notice the barrier was reset in this day. I edited my post to include this fact.
Also did you hear like a thump? Wonder what that was from. Maybe the white car clipped or got clipped by the gray SUV that was in the 85 HOV? Don't see the white car after that point while the gray SUV continues up the 85 ramp.

Yes, saw that the "smart cushion" was retracted/reset in that one.
My take on that "thump" was some controls in the Model S, like maybe disengaging auto pilot or something like that.
It seems like maybe the white car got back into a proper lane after having slowed down and threaded the needle between those other cars.
 
Earlier I asked if anyone knew if there was a traffic cam for that area. If so and maybe it gets recorded, it could shed some light on what actions the Tesla made. So here's the list of Caltrans live traffic cams. I opened the San Francisco-Oakland Bay Area listings and see SB101 cameras with Junctions listed (1, 12, and 116) but can't check out since I don't use the Adobe Flash Player plug-in. Can someone else check it out?

Live Traffic Cameras
 
Just watched it 8 times. White car hit the brakes before starting the lane change, then saw the car with dash cam hit the brakes and still moved to the right. Based on brake lights behind it, it was braking before that. Also no turn signals. White car was driving poorly (may have been new car test drive). Dash cam car would have had to slam on brakes to let them in. White would have been fine if they used the accelerator instead of brake (Or no brake).

Exactly, the dashcam driver (a Tesla, I can make out the mesh material used on the dashboard) was accelerating and even speeding. The carpool lane slowed down for a reason, and this driver sensed it (that's why it changed lane) but chose to be god or Darwin. Sad.

Great power comes with great responsibility.
 
  • Like
Reactions: Ulmo
Earlier I asked if anyone knew if there was a traffic cam for that area. If so and maybe it gets recorded, it could shed some light on what actions the Tesla made. So here's the list of Caltrans live traffic cams. I opened the San Francisco-Oakland Bay Area listings and see SB101 cameras with Junctions listed (1, 12, and 116) but can't check out since I don't use the Adobe Flash Player plug-in. Can someone else check it out?

Live Traffic Cameras

Unfortunately no camera exists from Foster City to San Jose. Below is the screenshot from QuickMap app. It surprised me they put a lot of cameras between Santa Cruz and Watsonville.

upload_2018-3-25_15-0-59.png
 
  • Informative
Reactions: SMAlset and Ulmo
Exactly, the dashcam driver (a Tesla, I can make out the mesh material used on the dashboard) was accelerating and even speeding. The carpool lane slowed down for a reason, and this driver sensed it (that's why it changed lane) but chose to be god or Darwin. Sad.

Great power comes with great responsibility.
What??
Dash cam car was staying on 101, was going slower than the car in front of it, and was passed by a pick up at the beginning.
Left lane slowed due to white car much later after dash cam car lane change.
 
Then this as you get closer:
View attachment 288727
Your second image that I quoted above shows that the exit lane the dead Model X driver was possibly in leads directly into the wall tip that killed him. You can see the silver car driving directly straight toward that wall.

Then, there are nineteen (19!!!) lines of text in signs you might think you have to read in no less than 10 hierarchical elements, while unbeknownst to you at the last moment the lane you're in has pavement that stays straight but some wild paint marks that jag viscously left. At that moment, you crash into the death-causing tip sword of a wall that is the near same drab gray color of literally everything else around it with some white sprinkled in, unless you were super wise and smart and happened to know not to be reading any of that damned signage. That's if you're lucky and didn't have to deal with other drivers on the road, too, which is almost never the case.

And, that's if you're also lucky enough not to have any malicious drivers around, which thank god, we only run into around twice a week around here.

That's why @TEG has trouble finding images of that accordian slowdown system (or its predecessor) in the uncrashed state.
 
Last edited:
You know @privater, looking at this frame from your 2-week old footage it looks to me like the barrier was compressed back then, maybe never repaired before this accident. Sort of looks like you can see the metal track exposed that it is suppose to slide on on impact. What do you think?

View attachment 288756

Compare to TEG's posts about the barrier:
Model X Crash on US-101 (Mountain View, CA)
Model X Crash on US-101 (Mountain View, CA)


From the KTVU Fox livestream broadcast today:
View attachment 288758
You're right: in the last picture you included, there is 0 impact debris in the track of the compressable slowdown, but lots of debris starting at the compressed state position.
 
Caltrans deffinately has responsibility, but they were not driving the car, so 100% seems a bit high. The accident resulted from the driver changing lanes too late (per witness report). Caltrans signage is not good, but still driver's decision. Being a fatality is high linked to the barrier not being functional.
It just takes 1%fault for caltrans to pay 100%.
 
I was prepared to see an amazingly effective stopping system. The first video showed a few crashes that the accordion system saved the driver's life, but also some that spun out the vehicle twisting the people's necks and putting the vehicles into traffic to get hit again on the side which would probably brain damage or kill the drivers. I was dismayed by the second video that showed a disposable tin can tiny car take huge impact and get spun into high speed rear ending traffic which would re-impact the car.

The best solution to this at the 85 South HOV onramp is move the lane back to the old lane markings that had the left onramp to 85 be far seperate from 101 for much longer, have cross hatches between the solid white lines in the middle, and do away with the amazingly large and excessive signage (replacing it eventually with petite and meaningful simple small signs small in number, properly placed in reading and comprehension distance before decision points). As I've said, I hate the new markings, and it almost killed me a few times.
 
Last edited:
Just got back from the hardware store. We have a brand new $$$ interchange at the I-15 and 91. It has toll roads, HOV lanes, transitional area local exits, multilevel (4?), flow metering, etc. 4 out of 5 gore points I saw had no idiot cushions AFAIK.

Prior to the rework (everything pretty much was redone for the 2nd or 3rd time) there were barrel/water/sand idiot cushions which are no longer there. There is a single telescoping one like you see in the incident at a HOV transitional. They are all too new to be destroyed yet. Some gore points are galvanized guard rail then concrete. Nothing really to collapse and speeds are 65 mph.

Are the idiot cushions becoming obsolete by road design? Or are we hoping Darwin makes up for our lack of Driver's Education?

And of course they are wildly confusing. You've seen NOTHING until you've seen what they did to the Toll Road access points. Demolition Derby was the idea it appears. And the transitionals between the two freeways is like a house of mirrors at the carnival. Tons of signs but they don't always mean what you'd assume.

Is it hard or unsafe to drive in this kludge?
The kamikaze lane Fastrak entrance/exit... yeah, too high of speed deltas are possibly required to switch to or from Fastrak.

But all the ramps and transitions are safe to drive. You just might not end up going in the correct direction. That is irritating, but not unsafe.
 
  • Informative
Reactions: FlatSix911
Man, that thing looks like it gets hit a LOT. At what point do you redesign the interchange? And, yes Caltrans shouldn’t be leaving it in the collapsed position for any length of time. If it takes a while to reset, they should be supplementing it with the barrels. Like the same day an accident takes it out, just bring the barrels by on a truck after the accident is cleared.
Agreed. Or even try a hybrid approach.

Or, just fix the intersection by moving the paint lines back to where they were, with the offramp being away from 101, not leading straight into that concrete wall.
 
Let me be pedantic. Here's how it is now versus how it was and should be, once again showing that older is often (not always, but often) better:

Bad, current setup, with notes of where the prior good setup used to be:
Screen Shot 2018-03-25 at 4.33.23 PM.png

In the following picture is the good, prior setup, marked by yellow and white line I drew in GIMP over the picture of the current setup in about the places the yellow and white lines used to be and where the pavement cracks still are (in other words, all Caltrans has to do is go back to putting the lines where the road cracks are rather than driving people straight into the wall tip, and they'd get rid of lane divers at the same time):

Note that I left the current bad yellow line so you can see how far awful and murderous the new line is compared to my drawn in line the way it used to be and should be:
Screen Shot 2018-03-25 at 4.33.23 PM copy.png
 
Last edited:
View attachment 289169

I count thirteen (13) lines of text (and symbols) in that first picture you posted. Those 13 lines are composed of 24 words and five (5) symbols. Those 24 words + 5 symbols are spread across four (4) high hierarchy sections, with 2 subsections of which each has subsections of its own, three for the left pane and two for the right pane. In addition, the left subection has a yellow subsubsection in it. The top of the two subsections has its own inverse subsection, with small text. There is no way a driver can read all of that and get any sense of what is going on, and realize that the lane drives them straight into the sword tip of a wall, simultaneously. It is quite near the worst type of signage possible.

But it doesn't. Both of the lane markers are over actual, functioning lanes. The death lane doesn't appear until a couple of hundred feet after the bridge that you see beyond the sign.

The problem here has nothing to do with the signage other than the fact that the sign should happen much sooner. That's a fairly typical traffic sign for California, and folks are used to reading them.

The problem is that the lane split is too wide and doesn't happen soon enough. There should really be about a one-foot-wide double set of lines with stripes between them starting approximately where that sign is, forcing people to make a decision by then. And the striping should continue beyond the point where the death lane appears (where the stripes get far enough apart to look like a lane).
 
There will likely be lots of "blame" to go around for this accident. I will address the ones that bother me from what I've read so far.
  • This entire intersection seems like a great example of how NOT to stripe, mix surfaces, erect signage and execute an intersection!
  • The "bumper" appears to not have been reset. If this is the case, the accident likely went from bad to worse and could have been the biggest contributor to this accident being fatal.
  • The "bumper" is evidence that Caltrans recognizes that this exit has a dangerous gore point. With that evidence comes responsibility to do something about it. Not resetting the bumper certainly seems "at fault" to me.
  • Unknown if driver was speeding or using EAP, but those are certainly possible contributing factors in the accident
I'm really disturbed if this accident was fatal because someone in our government didn't do their job (did CHP notify Cal Trans? How often do these things get checked? etc). I continue to be upset that our government seems to make mistakes that result in loss of life and no one seems to be responsible.
 
  • Like
Reactions: bhzmark and mongo