Experiment: Force exerted by Model S door handles on closure

[cantdecide]

I think there is one significant factor we are missing in this thread... Whether the force is a shearing force or not.
Take a random 4lb object, bottle of apple juice...
Place your hand on the table and the bottle gently on top. It doesn't hurt.
Now place your finger to extend over the edge of the sharp edged table and place the bottle on your finger precisely on the overhang. I expect a significantly different experience due to the shearing force.

What does that mean for a model S? If the handle is closing and like me you react by pulling then you are fine, it is like pulling a microwave door. If you just let the door close on you then you will get a shearing force and you will feel it more.
How much? That will depend on factors like the width of your finger, size of bone, etc...

In our family my wife screams when she gets caught, but I naturally pull back so never have an issue, even on the same handle. Even if I just let is close on my I might be fine due to different finger size.
These aren't the only factors, but I expect they are significant.

 

[jaguar36]

The industries I'm intimately familiar with (aerospace, medical device, chip manufacture, pharma, biologics) require a First Article sample for the first part received from a vendor, with a sample size of 1, to ensure that the vendor is capable of meeting print. But that's only First Article. With every single shipment, sampling is done based on size of shipment, past vendor history, criticality of component on device performance, and other factors dependent upon industry and usage model. I'm most familiar with C=0 acceptance testing, but MIL-SPEC-105, for instance, will set a minimal acceptable quality level (AQL) for a sample size.

In the aerospace industry we call it First Article Inspection, and its just that an inspection. As you said parts are usually rigioriosuly checked against the drawing, and may be checked with other Non-Destructive Inspection methods as well. Some parts may also be proof tested, but that is typically only parts that will see constant pressure over their life time, such as hydraulic or fuel lines.

I was referring to qualification testing which is usually done very early on in production, and is generally only repeated if the part fails or if the part is significantly redesigned. When the part/assembly being tested is a high dollar item, or the test itself is expensive or time consuming it is common to only test one of them. Obviously testing multiple ones is better, and allows you to use a less conservative load.

 

[essaunders]

maybe we need to measure the force of someone yanking their hand out of a spring-constrained gap? I'd postulate (as I think someone else already did) that the potential for 'injury' is more due to hand movement than handle. Would it be prudent for Tesla to consider the edges and 'entrapment' issue in handle design?

 

[bonnie]

In the aerospace industry we call it First Article Inspection, and its just that an inspection. As you said parts are usually rigioriosuly checked against the drawing, and may be checked with other Non-Destructive Inspection methods as well. Some parts may also be proof tested, but that is typically only parts that will see constant pressure over their life time, such as hydraulic or fuel lines.

I was referring to qualification testing which is usually done very early on in production, and is generally only repeated if the part fails or if the part is significantly redesigned. When the part/assembly being tested is a high dollar item, or the test itself is expensive or time consuming it is common to only test one of them. Obviously testing multiple ones is better, and allows you to use a less conservative load.

Aerospace still does Acceptance Testing in addition to First Article, though (unless they are taking zero gov't dollars - then they can do what they want - but whoever is buying their product would be foolish to do so). What you're talking about with qual testing (and I agree with) is First Article, supplemented by performance testing. That only is part of the verification cycle: Is the vendor capable of making the right part? Is this component designed correctly against the initial requirements? And if it fails during product validation testing, there are then no concerns that a particular component may be off.

In no way, however, is that used to determine that every other part in the lot has no variances or that any future part will be acceptable. (I'm assuming your company has to meet Dept of Defense requirements). That's the role of acceptance testing - which is a sample only, based on a number of factors (listed above) to determine the probability of the entire lot falling within tolerances. And because that's sampling-only, there is still a requirement for in-process testing and final acceptance testing.

To make it more clear, if you were a manufacturer, would you spend all that money on assembling a product and shipping to your customers without some minimal assurance that the parts you're using were good? If you've ever inspected parts, you know there are frequently outliers.

And for the sake of this discussion, let's pretend that all parts were fine, completely within norm (usually are). What about a new person on the line, who adjusted a couple cars incorrectly until realizing he was doing it wrong? Those things happen a lot. Hopefully process or final testing would catch something like that, but you get test-weariness (meaning if someone has run a test 1000 times and never had a failure, odds that they will fail to catch a problem go up significantly).

There are so many things that could contribute to one car being different that it dismays me to see conclusions drawn here. Maybe you're all right. But maybe the OP has handles that withdraw a little too forcefully. My only point (which I've posted before) is with all this data, we only know how the tested cars did. We cannot rule out an issue with his car.

(And your comment that a test might not be repeated if it's a high dollar item? Required testing should not be determined by cost, but rather by need to do so. If the test is necessary to demonstrate safety and/or performance, then it's non-negotiable. And if not, don't waste money ever. Are you sure that cost is a factor?)

 

[wk057]

And for the sake of this discussion, let's pretend that all parts were fine, completely within norm (usually are). What about a new person on the line, who adjusted a couple cars incorrectly until realizing he was doing it wrong? Those things happen a lot. Hopefully process or final testing would catch something like that, but you get test-weariness (meaning if someone has run a test 1000 times and never had a failure, odds that they will fail to catch a problem go up significantly).

There are so many things that could contribute to one car being different that it dismays me to see conclusions drawn here. Maybe you're all right. But maybe the OP has handles that withdraw a little too forcefully. My only point (which I've posted before) is with all this data, we only know how the tested cars did. We cannot rule out an issue with his car.

When you're referring to OP here I assume you're referring to the other thread...

In any case, I invite you (and anyone else) to dismantle one of the handle mechanisms some time. I had one here from a salvage for a bit that I was able to tinker with a little (friend got one on ebay and wanted to be able to actuate it with an arduino for some project that I still don't know what it's about exactly...regrettably I did not document this quick project/favor with photos...). If I still had it I'd invite you to check it out. If you saw how this mechanism actually worked you would see that there really is no way for there to be a major discrepancy in the force applied on retraction. It would require a different design entirely to be able to apply more force on retraction. If anything I'd think that the force would be biased towards less than the force I measured on my P85D since the spring will wear over time and use.

There is just no way this device could apply substantially more force than I have measured. Having seen the mechanism, and after having done testing, this is one point I will not concede without evidence to the contrary. I'm certain beyond any reasonable doubt that such evidence does not exist, but I welcome it if it does and will be the first to hang my head in shame while apologizing for prior comments if it were ever presented.

 

[bonnie]

When you're referring to OP here I assume you're referring to the other thread...

In any case, I invite you (and anyone else) to dismantle one of the handle mechanisms some time. I had one here from a salvage for a bit that I was able to tinker with a little (friend got one on ebay and wanted to be able to actuate it with an arduino for some project that I still don't know what it's about exactly...regrettably I did not document this quick project/favor with photos...). If I still had it I'd invite you to check it out. If you saw how this mechanism actually worked you would see that there really is no way for there to be a major discrepancy in the force applied on retraction. It would require a different design entirely to be able to apply more force on retraction. If anything I'd think that the force would be biased towards less than the force I measured on my P85D since the spring will wear over time and use.

There is just no way this device could apply substantially more force than I have measured. Having seen the mechanism, and after having done testing, this is one point I will not concede without evidence to the contrary. I'm certain beyond any reasonable doubt that such evidence does not exist, but I welcome it if it does and will be the first to hang my head in shame while apologizing for prior comments if it were ever presented.

Consider there has been no need to call others into question. Nor does someone have to prove you wrong or right. It's not a contest here. Someone said they had a painful experience. Who are we to say they didn't?

I still remember losing skin off a hand when peeling off a wetsuit in extremely cold conditions. I was shocked - I'd never had skin peel off my hand happen before. Ever. But it did. Skin behaves differently in diff conditions. Health factors diagnosed or not, such as diabetes, can contribute to either real pain or perception of pain issues.

Not a contest to be won.

 

[wk057]

Consider there has been no need to call others into question. Nor does someone have to prove you wrong or right. It's not a contest here. Someone said they had a painful experience. Who are we to say they didn't?

I still remember losing skin off a hand when peeling off a wetsuit in extremely cold conditions. I was shocked - I'd never had skin peel off my hand happen before. Ever. But it did. Skin behaves differently in diff conditions. Health factors diagnosed or not, such as diabetes, can contribute to either real pain or perception of pain issues.

Not a contest to be won.

Your wet suit situation sounds pretty normal. I had a similar experience a while back while clamming in Absecon Bay in NJ with my uncle in a suit that was a hair too thin for the weather at the time.

In any case, I was referring to the handle force and handle force engineering tolerance point in my last post and said nothing of pain.

 

[jaguar36]

(And your comment that a test might not be repeated if it's a high dollar item? Required testing should not be determined by cost, but rather by need to do so. If the test is necessary to demonstrate safety and/or performance, then it's non-negotiable. And if not, don't waste money ever. Are you sure that cost is a factor?)

While in an ideal world that may be true, cost and schedule are always a factor. The most extreme example are the static ultimate tests done on a n aircraft wings (Like this one: http://www.wired.com/2010/03/boeing-787-passes-incredible-wing-flex-test/ ) That test will likely never be repeated, regardless of any manufacturing changes, or even design changes. If a major change is required, there is a very small possibility that it would be re-done, but its highly unlikely due to the extreme cost of the test. That is why the test is run to 150% of limit load, to account for any statistical variation in manufacturing.

 

[bonnie]

While in an ideal world that may be true, cost and schedule are always a factor. The most extreme example are the static ultimate tests done on a n aircraft wings (Like this one: http://www.wired.com/2010/03/boeing-787-passes-incredible-wing-flex-test/ ) That test will likely never be repeated, regardless of any manufacturing changes, or even design changes. If a major change is required, there is a very small possibility that it would be re-done, but its highly unlikely due to the extreme cost of the test. That is why the test is run to 150% of limit load, to account for any statistical variation in manufacturing.

Since I wasn't part of that decision-making process, I can't speak to that particular example. I can, however, speak to countless others.

There are times that a test is only needed one time, even if other attributes change. For instance, if I have a concern about material strength and engineering extrapolations, I may test to the point of destruction to ensure that all calculations were valid associated with the material strength. I may put it through an artificial aging process over the course of several months to collect data on stress factors, etc. And then, for any future changes, I would determine if those changes could possible impact the original data in testing a particular requirement (which is what your example is about).

Federal regulators & inspectors look very poorly on testing decisions made solely upon cost, and for good reason. I've first-hand experience over the last 20+ years in this particular area. And if you argue that a test 'cost too much' to perform, you will lose. If, however, you can show why it was not necessary ... then you have a good discussion.

 

[thecloud]

An attempted delivery was made this morning but I missed it. Should hopefully have it tomorrow.

For deonb and others who were interested, the gauge finally squeezed its way out of DHL's delivery system, and I did some measurements by pulling just enough to move the handle out slightly from the retracted position. Summary: my handles seem to require somewhere between 4.2 and 4.5 lbs of force to counteract the spring trying to hold them closed. This is slightly more than wk057's measurement of 3.87 lbs, but obviously still in the same ballpark.

Since the numbers are close, this suggests to me that there isn't any problem with the handles being out of spec, and some people (myself included) may just be more sensitive to that amount of pressure than others.

Here's a lame attempt at making a video with my phone. Apologies for the jerky quality; there's probably some frame rate issue I need to figure out, but I didn't feel like doing it all over again once I'd finished.



<br>