Where did Elon Musk go wrong?

I am periferally involved. The comments about “automation failing” and “cars being built by hand” are extreme exaggerations. The main issue they are having, in my opinion, is that fixture vendors from all over the world are being tasked with rapid delevopmemt of equipment that is only vaguely reminiscent of the products they have been building for other car mfrs, so everything is custom and all these disparate vendors insist on using their own control platforms in order to do it fast. So when it arrives, there is a huge mish-mash of platforms; some Siemens if the machines came from Europe, Omrom and Mitzi if it comes from Asia and A-B if it comes from the US or Canada. Their MES system is trying to coordinate all of these systems and THAT’S the nightmare. On top of all that, last month they discovered that a disgruntled Engineer working on the MES system purposely sabotaged it. So the “hand” aspect is that they had to stop production and restart without the MES system, meaning that Engineers were “manually” overseeing the total production line while they dealt with the mess this guy left behind.

The press, unable to comprehend the complexities, boils it down into little nibbles of stuff that John and Jane Q Public can understand.
 
Last edited:
In my many years of experience in the automation field, I have seen a reoccurring cause of failed automation. A large majority of these failures result from a company expecting a new machine/process to function perfectly the first time. Thus ignoring the need to re-engineer things that did not work correctly because of the unforeseen.

I'm sure many of you have seen the situation in which a company/management will order 10 of a new machine that has never been built before. The result is 10 failed or poorly functioning machines.

The example I like to refer to is the modern automobile. These are well refined machines that have been re-engineering and revised for over 100 years. You can be sure the first automobile would have been lucky to make it the first mile without some kind of failure. It is now expected to be able to insert the key and drive for thousands of miles without any issues. Imagine this translated to an automobile assembly line that was refined for 100 years. I would be confident to say you could press the start button and have new cars roll of end of the line for years without touching a single thing.

There is a reason most OEMs have well functioning machine. It is because they have refined their equipment over many revisions.

I do not know the case of Tesla, but I would dare guess that if they are truly having automation problems it is because they are on revision 1 of most equipment and it is now difficult to make major changes now that they are in full production.
 
I've never heard of Tesla till sometime early last year. Never looked at anything they make until this post.


I'm too old (and broke) to buy a new gizmo, but could make an argument that electric motors could whoop some tush at the drag strip, if they had enough juice to do it.


Most cars today, you can't tell what they are, and this car is no different.


I like my Challenger SRT 392, and being totally stock, kicks some butt at the local drag strip. Hellcats have to play catch up, except for one, 650HP Camaros give up if they don't cut a good light. Mustangs have a Voodoo image to poke at during their rituals. One gal in a Pontiac, strictly race car... Both had 0.074 lights, so didn't tree her, but every time I saw her after that, her eyes were throwing pitch forks at me.


Yep, I like to have fun and practically deaf, so if that engine don't cackle...
 
Last edited:
In my many years of experience in the automation field, I have seen a reoccurring cause of failed automation. A large majority of these failures result from a company expecting a new machine/process to function perfectly the first time. Thus ignoring the need to re-engineer things that did not work correctly because of the unforeseen.

One of our Engineers has a favorite saying when things don't go perfectly while bringing a new system online.

"That's why they call it 'Start-up' and not 'Run'"

If more customers understood that and allowed just a little more time in the early phases to work out issues, 'Run' would probably come sooner.
 
That fits. One integrator I am working with now has never heard of Studio 5000, and they are a thousand miles away. Another one in Colorado I had to explain this to, still doesn't get it. I really don't know what to do. I'm tired of sending emails. Leave me alone!!!
 
This sounds so much like companies that run solely off contracts. “Here is what I want, I want it now, here’s the check”. It always looks good at first, and may checks all the boxes, but rarely ever lasts especially when everything has to come together. But I know nothing about Tesla other than the picture printed by mainstream media.
 
Every electric vehicle maker (outside China) faces a critical point the day that they deliver their 200,000th unit in the US, and trigger the phase-out of the Federal EV tax credit.

In every market where a major credit has expired, the sales rate for EVs crashed. Georgia and Denmark are great examples. The sales went to virtually zero, then stabilized at about 25% of their former level. It's survivable if the company sells combustion-engine vehicles or if the market is small.

But Tesla sells only EVs. And the market is the whole United States. So playing it safe is not feasible.

So they're being audacious, largely in the person of their CEO. They've created vehicles that are more than just cars; they are objects of desire and technological marvels. That's how they got over 400,000 deposits for Model 3 vehicles. I put down my money before anyone had ever seen the vehicle, on March 31, 2016.

When you go over a precipice, you plummet. Unless you're going fast enough to spread your wings and fly.
 
but I can tell you that from what I've been seeing lately, the pendulum has started swinging away from "new modern efficient" methods of programming – back toward "plain old-fashioned" Ladder Logic ... note that this is being driven by the CUSTOMERS' demands (read: "money") and we all know that "money talks" ...

I think this is somewhat of a trade off between paying more for someone to implement ladder logic that would be done in half time with another language versus providing training and or having a well defined standard for the logic.

Also, as much as training is good for employees, one cannot reasonably expect that a week or three day course is enough to catch the complexities of certain languages like SCL or IL.

As much as I dislike programming in Ladder, it is the best visualisation language for Boolean logic. You simply cannot beat it's simplicity for electrically minded people as you can see the path or why the path to a certain coil isn't lit.

However, there is a balance that can be had by specifying functionality that sits inside its own block and does just that... and inside that block, the logic is whatever suits it best.

What I would say is that in certain places, the PLC is always to blame when in fact the technician didn't fancy a trip down to whichever sensor may be having an issue... and in others, the technician needs to know the PLC program so he can bypass some safety because we don't have "stock" or to "tune" the PID because no one cleaned the system and it is now not working as it should.

Additionally, one can argue that PLC programs in generally are poorly described for the operator. I lost count of how many machines I've seen without a description of conditions or interlocks for functions and sequences. If the operator/maintenance has one of those, does he really need to go into the PLC logic?
And the same with alarms... if an alarm is properly descriptive, the tech can take action without looking into the program. However, time and time again, alarms are an OR of 3 or more conditions or have ridiculous descriptions.
 
I am in the automobile business, my company makes machines for the subsuppliers to the actual automobile manufacturers.

The foodchain that goes from parts to subassemblies to finished cars is MASSIVE.
It is amazing that cars can be manufactured at the prices that they do, and that the endproduct is so reliable as it is. To get to this point, both subsuppliers and manufacturers have had decades and decades to optimize their production.

The big automobile brands have a big advantage that makes it hard for a newcomer to get in. Every now and then the big brands will close down an entire line and start a new line, taking advantage of all experiences that they have gathered until that point.
Tesla has only 1 line that they are trying to optimize on the run. They cannot shutdown and start all over. Tesla has an advantage in that they are not burdened by old habits, and they can possibly implement procedures that are totally different to what everybody else does. But to get to the point where it will run smoothly and the new different ways of doing it will pay of, maybe that will take longer than what Elon Musk has anticipated.
 
Last edited:
I'm too old (and broke) to buy a new gizmo, but could make an argument that electric motors could whoop some tush at the drag strip, if they had enough juice to do it.

Interestingly, from what I've read, the limiting factor on the Tesla isn't the motors, it's how much juice the batteries can push out without overheating.

I was reading an article about a new record set last week at the Nurburgring in Germany (famously difficult, multi-mile track). They mentioned that the super speed version of the Tesla had amazing 0-60 times, but couldn't make a full circuit of the track at top speed without overheating.
 
Tesla has only 1 line that they are trying to optimize on the run. They cannot shutdown and start all over. Tesla has an advantage in that they are not burdened by old habits, and they can possibly implement procedures that are totally different to what everybody else does. But to get to the point where it will run smoothly and the new different ways of doing it will pay of, maybe that will take longer than what Elon Musk has anticipated.

I think this is a big part of it. My understanding is that Tesla technically has two lines. They bought and repurposed an old GM facility, and the existing line was repurposed for the Model S & X (same platform). The Model 3 was getting a brand new line in the same facility.

From what I heard, they were starting from scratch and re-imaging both what the car and what the process could be. The plus side of that is that you aren't saddles with years of baggage and compromise; the downside of that is that THERE WAS OFTEN A GOOD REASON FOR THE COMPROMISE.

They have the opportunity to find new solutions as they approach the problems everyone else has faced, but they still have to pass those hurdles one way or the other.

It sure doesn't help when your HR strategy seems to be the common Silicon Valley idea of "constantly find new people, burn them out, repeat". It works fine for Mcdonalds, but that's not how I'd treat people building my critical infrastructure. I think that whole area is crazy.
 
I just finished reading this book (yeah, real paperback): Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future by Ashlee Vance.
It is a biography of Elon Musk. Interesting read about why he does things his way.
He does have a background in programming and I kind of think he would shy away from ladder logic. But he is a bit removed from "the line" and is very demanding on results.
The MES issue may have more to do with the problems and current bad press reports.
 
Maybe it hasn't met the expectation yet, they are producing a lot of cars and they are improving. I wouldn't classify it as a failure. They are learning and improving.

I don't see it as a failure either. I guess it depends on what articles you read. I assume everyone knows by now that everything you see on the internet isn't necessarily true.
 

Similar Topics

I'd bet more than pocket change that once-upon-a-time I saw an official spreadsheet type chart that listed all of the features for the various...
Replies
4
Views
2,195
I'm having a tough time trying to get used to old habbits in S7 and them not being the same in TIA portal. First impressions are its cr*p what...
Replies
7
Views
8,320
I have a lot of different equipment that I inherited from my father's company, he allowed me to sell it, how can I do this? Siemens Simatic S7...
Replies
4
Views
258
Can anyone help me find where is recipe plus or recipe manager in View desginer(studio 5000) software. i want to create an recipe but i am unable...
Replies
2
Views
143
Hi everyone: I went to AB.com, found only firmware, no EDs file showing up. i took a pix of device properties of RSLinx. Thanks
Replies
1
Views
100
Back
Top Bottom