Physical measurements of Variable Suspension show that you’re NOT getting the ride you call for.

[SteveSchweer]

It turns out that “Sport“ is the only Variable Suspension mode that actually delivers what you call for. All other ride modes are virtually the same, forcing the shocks to their stiffest possible setting. Watch this video to see physical measurements of the actual shock settings that result when various ride modes are selected. This is a software issue, and all cars with Variable Ride Control use the same software, so ALL vehicles have this problem.
I am aware that many of you believe that there is an obvious difference between ride modes, but measurement of the actual shock settings that result as different ride modes are selected doesn’t lie. I’m sorry to say that the difference you feel is an illusion. I call your attention to sub-thread 5 of the thread “Adaptive Suspension Damping Real or Ruse” where members Pvdoug and Maaz, one of whom saw an obvious difference and the other saw none, got together and compared their cars. After testing and riding in each other’s cars, they concluded that both cars were identical. Perception is a squirrelly thing. Physical measurement is the ONLY reliable indicator of how the system is functioning. Please also look at the thread “Plaid Suspension Accelerometer Test Readings in Comfort and Sport Modes”. The graphical results of this test show that Comfort Mode does indeed show HIGHER G-force readings than Sport…. Ie. Comfort is harsher than Sport, concurring with my readings.
Anyway, give the video a look… it’s only 9 minutes.

 

[scot]

Could you also get a logging scope (or log the suspension values/duty cycle as reported) and post up comparisons of the vehicle in use? It is unclear if you have done static only or dynamic testing from the video. The first pushback from Tesla is going to be "we show the same for all cars sitting still, but it is active so it requires use to start to show the difference".

Basically they will say the duty cycle adjusts as per immediate road conditions and will alway show the same at rest. So unless you are comparing the same road at the same speed with bumps that cause suspension movement and reaction, the data will always look the same. Of course if you see that the settings produce the same reaction while moving, that is a way different story. Adding this information should make it much, much harder to refute.

BTW while the suspension values might be easier to capture, I would still suggest capturing the pwm signal directly, just in case there is some smoothing or value constraints going on which make different behaviors look the same on the suspension screen.

 

[glide]

Can’t wait for the fanbois to invade this thread and say how they 1000% can feel the difference.

I’ve owned my MS LR for over a year and the adaptive suspension does nothing (except when in sport mode). Comfort is HARSH.

 

[SteveSchweer]

Could you also get a logging scope (or log the suspension values/duty cycle as reported) and post up comparisons of the vehicle in use? It is unclear if you have done static only or dynamic testing from the video. The first pushback from Tesla is going to be "we show the same for all cars sitting still, but it is active so it requires use to start to show the difference".

Basically they will say the duty cycle adjusts as per immediate road conditions and will alway show the same at rest. So unless you are comparing the same road at the same speed with bumps that cause suspension movement and reaction, the data will always look the same. Of course if you see that the settings produce the same reaction while moving, that is a way different story. Adding this information should make it much, much harder to refute.

BTW while the suspension values might be easier to capture, I would still suggest capturing the pwm signal directly, just in case there is some smoothing or value constraints going on which make different behaviors look the same on the suspension screen.

Thanks for the thoughtful comments! Your concerns are valid. I think I did address most of them in the video though. The problem is that I was trying to minimize the length of the video so a lot of information comes at you pretty fast. (You almost have to watch it twice to catch everything!)
The scope PWM readings were measured directly by tapping into the wires between the ECU and the Front Left Rebound solenoid. I powered the scope with a battery inverter and DID monitor it both at rest and while driving. (My video production facilities are primitive however so I was unable to balance the camera and the scope steadily enough to make a video while in motion.) The suspension system becomes “active“ the second you touch the brake… notice how the suspension data appears on the screen when you do so. While driving I noticed the exact correlation between the PWM duty cycle and the numbers labeled “Compression/Rebound Damping“ Those numbers are a digital representation of the analog PWM duty cycle. I say this because you can see them “twitch” between values… ie. they jump back and forth between 0 and 4. This is characteristic of digitizing an analog value.
In any case, you can watch the system in action. It starts out with a “base value“ pulse width (duty cycle) at rest which increases as rough pavement is encountered. This increase is RELATIVE to the base value… it is the same in all modes. In it’s current setup, since the base values are the same, the increases go up to the same values producing identical rides in all modes except Sport. The increased pulse width only lasts for a second or so however, so if the bumps/uneven pavement patches are spaced further apart (the usual case) then the ride quality becomes whatever that base value is. For this system to make any sense, it is ESSENTIAL that it be prepared to react to the first bump encountered, right from a dead stop. The bug here is that a proper base value has only been assigned to the Sport mode… (in the 20s). The rest are 0 or 4 or 8… effectively 0 and the same for each mode. As increasing pulse width increases power to the solenoids, there will be a threshold of power necessary to budge the valves… I believe that this is around 20. That’s why Sport, in the 20s, feels so similar to all the other settings which are way less than 20… none of them open the valves at all. A properly calibrated system will probably show “at rest” duty cycles in the range of 20, 40, 60, 80, 100 as you step through the ”Firm to Soft” slider in “Advanced Settings“. (Of course, these values probably will be slightly different when the system’s electrical and hydraulic linearity are taken into account... something only the Bilstein engineers will know.)

 

[Hayseed_MS]

@SteveSchweer What company do you work with in Michigan?

 

[SteveSchweer]

@SteveSchweer What company do you work with in Michigan?

I’m retired (that’s how I can waste all this time on this issue!). I did coding/systems analysis… mostly for the manufacturing industry. I am an electronic hobbyist and have always worked on/restored cars non-professionally. (When I got my first car I immediately disassembled it).

 

[scot]

Steve: I said what I said because it is not clear you have evidence from that video of the system that these conditions exist while driving, and why I suggest using an oscilloscope that has a logging feature. This is hard evidence, but right now it is just statements about what you saw. (Not fair and hope you don't feel it discounts your efforts, just work with enough engineers to know it takes a lot to get them to pay attention)

One thing I *highly* recommend, please reach back to your Bilstein connection and try to review the details on the PWM control. As others have pointed out there may be more communication than just the pwm signal and also would be best for them to confirm the duty cycle as compared to expected behavior (more or less damping). While it seems unlikely, it is a possibility as there seems to be some potential for multiple controls to be multiplexed together (compression and rebound) with multiple systems and Tesla only reports the pwm duty cycle because its easy. I honestly dont know, though looking at an actual shock out of the car and feeding it signals via a PWM generator would probably be best to confirm your thoughts, combined with feedback from an engineer about the design and function. That should leave little room for discussion.

 

[TLLMRRJ]

The real solution is to ride in a stock 2020 Model 3 LR for a couple of days. After that, the Model S feels like butter!

I had to change out the shocks in that car for MPP Comfort Coilovers, and it's still pretty harsh because they didn't change the spring rates from stock in their kit.

 

[Tdriver]

I have an 8/22 build refresh and I have no illusions about the ride quality. Comfort is very supple and Sport is not. I would not own the car if that were not the case.

 

[SteveSchweer]

Steve: I said what I said because it is not clear you have evidence from that video of the system that these conditions exist while driving, and why I suggest using an oscilloscope that has a logging feature. This is hard evidence, but right now it is just statements about what you saw. (Not fair and hope you don't feel it discounts your efforts, just work with enough engineers to know it takes a lot to get them to pay attention)

One thing I *highly* recommend, please reach back to your Bilstein connection and try to review the details on the PWM control. As others have pointed out there may be more communication than just the pwm signal and also would be best for them to confirm the duty cycle as compared to expected behavior (more or less damping). While it seems unlikely, it is a possibility as there seems to be some potential for multiple controls to be multiplexed together (compression and rebound) with multiple systems and Tesla only reports the pwm duty cycle because its easy. I honestly dont know, though looking at an actual shock out of the car and feeding it signals via a PWM generator would probably be best to confirm your thoughts, combined with feedback from an engineer about the design and function. That should leave little room for discussion.

The recording oscilloscope is a good idea! Do you have such an instrument? If so, it would be great if you could take the readings. I suppose one could also just video the scope while driving the same route in each ride mode. I would sure welcome the input. I won’t be able to do much more on this as I’m in Florida till April while my car is in Michigan.
As to my “Bilstein connection” … LOL! I don’t have ANY connections! NO ONE will talk! … certainly not Tesla, and Bilstein said that they build the struts to Tesla’s specifications and they are proprietary. I would LOVE to have even a 5 minute conversation with an engineer about this, but no luck!
Anyway, as to the communication between the ECU and the shocks, I believe that we’re WAY overthinking this. All they’re doing is opening and closing a valve by varying the voltage to a solenoid. There are separate solenoids for compression and rebound. What’s to multiplex? Also, Tesla’s system is not unique… this type of shock control is quite common in the industry and that’s how they all work. Occam’s Razor applies here…. The simplest solution is correct.
Sounds like you’re involved in engineering Scot.… Are you a professional?

 

[aerodyne]

I have an 8/22 build refresh and I have no illusions about the ride quality. Comfort is very supple and Sport is not. I would not own the car if that were not the case.

Wild guess. Are your cubby lights backwards?

If so, I think I know what the problem is...

 

[scot]

The recording oscilloscope is a good idea! Do you have such an instrument?

Unfortunately not. There are a bunch of very expensive options, but there are also some cheap ones on aliexpress like this :

at 80 bucks. Not great, but probably enough to do a PWM recording over at least a short amount of time to a laptop.

Or even better, an Arduino or any kind of signal capture system that could both be programmed to capture the signal as well as produce a signal. An Arduino person could probably whip something up pretty quickly. PWM reader based on Input Capture (Arduino Nano/UNO) you could check on the Arduino forum to see if anyone has done something like this.

I am *not* an electrical engineer though, i'm in a completely different trade though we do call ourselves engineers. I just know very well how hard it can be to convince someone in a technical field of a truth that they don't believe in. Cover the bases and make it 100% clear.

BTW I've looked at a bunch of bilstein devices and some use reverse duty cycles to change compression and rebound. (0 percent is firm or 0 percent is soft). With both a minimum setting or maximum setting for different drive modes and changing latency for updating the rates while driving. I dont know how this affects the Tesla solution, but it is there. Need to find a shock on the bench and feed it different settings and see how it reacts.

 

[MFrunkerOG]

I have been following this for a while now but never got the chance to test it until today.

I’ve gotta say that my comfort mode feels comfortable/ squishy. my sport mode feels much firmer. Also the advanced mode all the way to firm is firmer than sport mode.

Yea the damping % numbers are pretty similar to what @SteveSchweer has.

Here’s where I tested it - a 100 yards of non tarmac’d side road - lots of bumps and dips - test conducted at 30mph for each instance

 

[Tdriver]

Wild guess. Are your cubby lights backwards?

If so, I think I know what the problem is...

No, lights work normally.

 

[MFrunkerOG]

Wild guess. Are your cubby lights backwards?

If so, I think I know what the problem is...

Hmm. I’m curious now. My cubby lights were backwards… then they needed to replace the entire cubby to fix a simple light switch. Got that done many months ago

 

[aerodyne]

Hmm. I’m curious now. My cubby lights were backwards… then they needed to replace the entire cubby to fix a simple light switch. Got that done many months ago

I was making an attempt at humor. If the cubby lights are backwards, maybe the suspension settings are backwards too...

 

[SteveSchweer]

Check out this petition.

Can you spare a minute to help this campaign?

Make Tesla fix Variable Suspension bug. It should perform as advertised.

chng.it

 

[aerodyne]

A petition! Great idea for getting the rear seat gaming.

Sorry, I don't think I have the suspension problem.

 

[mcirish]

I have 2022 Model S LR, 19" wheels, delivered December 2021, cubby lights wired backwards, same poor suspension that's been reported.
Mostly too much travel and bounciness in comfort mode, Some stiffness improvement in Sport mode (less bouncy).

My suspension settings map can be viewed here as a Google Sheets Document:
SuspensionHeights

 

[pvdoug]

I have an almost identical car with the same date of manufacture and the same issue. There has been a lot of conversation about this problem, and it appears to affect only a small subset of owners. If Tesla ever comes up with a fix, I would be both shocked and overjoyed.