Fisker Fires

[markwj]

The Model S's drivetrain is liquid cooled, so we're talking pumps and heat exchangers. Might be a fan on the heat exchangers to cool the drivetrain when the car is not in motion...but I can't remember. The only air-based fans I know of would be the circulating fans for the climate control...but someone else can probably provide more detail (or correct me).

Of course the issue here is not that cooling fans were used, but that either the cooling fan design or its implementation in the system appears to be flawed.

Even liquid cooled, the heat exchanger part will need fans to force the air through. The roadster has two enormous ones in the front.

 

[strider]

Yea, every car has fans. It's not that fans that are inherently a problem.

I agree. I think Fisker's problem is they tried to put 10lbs of $hit in a 5lb bag.

 

[Larry Chanin]

Sorry to pull a TEGed here, but...

Larry

 

[smorgasbord]

Not sure how Fisker is going to get away with blaming this on the fan manufacturer. Surely Fisker has a fan specification, determined this part met those specs, and then perform an inspection on each shipment of those fans to ensure they continue to meet spec.

Right???

It's possible that the fans have a manufacturing defect (like some of the A123 batteries did), which maybe wouldn't be noticed since it's unlikely Fisker tests each fan to that extreme before installing.

Fisker's Official statement is here: http://media.fiskerautomotive.com/global/en-us/Media/PressRelease.aspx?mediaid=928&title=media-statement----woodside-ca-incident

The final conclusion was that this sealed component had an internal fault that caused it to fail, overheat and start a slow burning fire. ... At the same time an additional fuse will be installed for added protection.

It is interesting that Fisker now feels they need to install a fuse to prevent the fan from drawing too much current. It seems unlikely that the fans wouldn't be fused at all, but maybe the line they're on is shared and fused at a too high threshold.

 

[bonnie]

It's possible that the fans have a manufacturing defect (like some of the A123 batteries did), which maybe wouldn't be noticed since it's unlikely Fisker tests each fan to that extreme before installing.

Fair. I was coming from the perspective of the medical device industry. If a product fails, I don't get to say 'but it was the supplier'. We chose the supplier, we evaluated their internal controls, we set the product specs, and we inspect the component when it arrives for what we feel are critical safety and performance requirements. If a risk assessment had determined that an internal fault could cause a problem, we'd have controls in place for that (or backup mitigation, like the 2nd fuse mentioned).

And yes, the supplier could have a problem and we could miss it. But did they KNOW that the failure of this fan in this way could cause a problem? I'd hope they'd done that level of risk assessment. And at the end of the day, I don't get to blame the supplier because ultimately, I'm accountable for ensuring adequate controls.

But this isn't a medical device.

 

[raymond]

But this isn't a medical device.

A failing medical device would typically kill at most one person. A car catching fire in a home or a failing brake could easily kill more.

Should cars have stricter QA applied? Have we gone overboard in the medical world? Or is this just the way of the world and I shouldn't be asking these stupid questions?

 

[bonnie]

A failing medical device would typically kill at most one person. A car catching fire in a home or a failing brake could easily kill more.

Should cars have stricter QA applied? Have we gone overboard in the medical world? Or is this just the way of the world and I shouldn't be asking these stupid questions?

A single part failing the same way in a medical device, distributed to thousands of people, could kill thousands of people. Remember, it's risk-based. "If this fails, how does this impact the user?" A low-risk device (such as a IV pole) doesn't have many risks to consider (though a breaking caster could cause things to fall on a patient). But a ventilator for home use ... well, you better know all the ways it could fail and make sure you have taken appropriate mitigation.

But I agree regarding your point on quality activities for vehicles -- blaming the supplier for something within your control seems wrong. I'd at least want to know, for every part, how it might fail and if that failure would cause an unsafe condition. And if an unsafe condition existed, I'd look at my prevention and detection controls. And if I couldn't detect it during an inspection, then I'd do more prevention (which is always better anyway). And if I couldn't control the supplier, I'd have a design that could eliminate the problem.

 

[SoCalGuy]

Not to cut Fisker some undue slack, but remember their model is to outsource most of the components. The value-add of Fisker is on engineering/integrating these components into the "Karma application," that is, into a plug-in hybrid/extended range EV. I posted on FiskerBuzz earlier today a list of the major components that comprise the Karma - and their manufacturers. Nearly all of the key components come from third-parties - from the battery, electric motors to even the steering wheel, sound system and seats. Fisker relies on their suppliers to go through rigorous and adequate QA testing since as one poster pointed out, it is unlikely and unreasonable to expect Fisker to test every component to that level of QA (remember, the Chevy Volt cost upwards of $3B to develop, while the Karma has cost approx. $1B). They rely on their suppliers' ability to meet spec, and in the case of this fire, it would appear that the fan manufacturer had a design defect that is affecting literally every Karma being produced.

 

[vfx]

Tesla's only recall was blamed on Lotus (kinda).

 

[smorgasbord]

As you can see in this post: PEM-motor-gets-too-hot-fans-failed?p=137310, Tesla used off the shelf cooling fans in Roadster.
Here's a post with the spec sheets:PEM-motor-gets-too-hot-fans-failed?p=138758

If you're thinking Tesla is building its own electric fans for Model S you're almost certainly wrong. At most, Tesla specifies a unique size, power, configuration, etc., but there's another company building it. Just like Tesla isn't doing the chrome plating for the door handles. Just like Tesla isn't sewing the seat coverings and doesn't have a pasture out back raising the cows for the leather seats. At this point in time, there isn't anything Tesla could do to make electric fans better than anyone else.

So, having specified and tested some samples of fans, Fisker and Tesla, and probably every other company in the world, is then content to simply order them, unbox, and install. Fisker/Tesla probably pull one out of each batch at random and test it to be sure it's meeting specs, and it if were considered critical they'd test random samples in various failure modes. But, they obviously can't test each fan to failure because then they'd have nothing to actually install.

Back to this case here. If the fans have a design defect, then the blame goes to Fisker, as they should have detected that in both design reviews and failure mode testing. If, however, there was a manufacturing defect, then the blame goes to the manufacturer - although depending on how many fans are defective it's arguable that Fisker should have caught it in random sample testing.

That's my view, anyway.

 

[bonnie]

Not to cut Fisker some undue slack, but remember their model is to outsource most of the components. The value-add of Fisker is on engineering/integrating these components into the "Karma application," that is, into a plug-in hybrid/extended range EV. I posted on FiskerBuzz earlier today a list of the major components that comprise the Karma - and their manufacturers. Nearly all of the key components come from third-parties - from the battery, electric motors to even the steering wheel, sound system and seats. Fisker relies on their suppliers to go through rigorous and adequate QA testing since as one poster pointed out, it is unlikely and unreasonable to expect Fisker to test every component to that level of QA (remember, the Chevy Volt cost upwards of $3B to develop, while the Karma has cost approx. $1B). They rely on their suppliers' ability to meet spec, and in the case of this fire, it would appear that the fan manufacturer had a design defect that is affecting literally every Karma being produced.

But they picked that business model. Doesn't that make it their responsibility? Who you buy components from, how you use those components, the people you hire - those things are all within your control.

It isn't reasonable to expect them to test every component. It IS reasonable (imho) to expect them to understand how various components might fail & to take appropriate action to ensure the end result meets their quality standards. It's bs to not do your homework and then point the finger at other people.

 

[bonnie]

As you can see in this post: PEM-motor-gets-too-hot-fans-failed?p=137310, Tesla used off the shelf cooling fans in Roadster.
Here's a post with the spec sheets:PEM-motor-gets-too-hot-fans-failed?p=138758

If you're thinking Tesla is building its own electric fans for Model S you're almost certainly wrong. At most, Tesla specifies a unique size, power, configuration, etc., but there's another company building it. Just like Tesla isn't doing the chrome plating for the door handles. Just like Tesla isn't sewing the seat coverings and doesn't have a pasture out back raising the cows for the leather seats. At this point in time, there isn't anything Tesla could do to make electric fans better than anyone else.

So, having specified and tested some samples of fans, Fisker and Tesla, and probably every other company in the world, is then content to simply order them, unbox, and install. Fisker/Tesla probably pull one out of each batch at random and test it to be sure it's meeting specs, and it if were considered critical they'd test random samples in various failure modes. But, they obviously can't test each fan to failure because then they'd have nothing to actually install.

Back to this case here. If the fans have a design defect, then the blame goes to Fisker, as they should have detected that in both design reviews and failure mode testing. If, however, there was a manufacturing defect, then the blame goes to the manufacturer - although depending on how many fans are defective it's arguable that Fisker should have caught it in random sample testing.

That's my view, anyway.

So my big question is "Did Fisker know that if the fans failed, it could be a problem?" If they didn't do a full risk assessment on the design, then there is no excuse. OTS parts are common and acceptable. Not knowing the impact of a failure ... not acceptable.

 

[SoCalGuy]

But they picked that business model. Doesn't that make it their responsibility? Who you buy components from, how you use those components, the people you hire - those things are all within your control.

It isn't reasonable to expect them to test every component. It IS reasonable (imho) to expect them to understand how various components might fail & to take appropriate action to ensure the end result meets their quality standards. It's bs to not do your homework and then point the finger at other people.

I think as others have pointed out, they do sufficient testing as a consumer of OEM components - they do not test each unit produced, but a random sampling to ensure quality. I think its worth highlighting that even high end luxury manufacturers like BMW and Mercedes, with orders of magnitude more resources than Fisker, have had countless recalls for their vehicles due to design and manufacturing defects. This is not a unique problem isolated to a Fisker or Tesla - this is just a fact of life.

Frankly, I would be more concerned if the fundamental cause had to do with the unique drivetrain tech and/or control engineering. Thankfully, it is a minor component that failed.

 

[bonnie]

My POV comes from having to do exactly what I've said. If it were my business, regulated or not, I'd do the same thing. I'd want to understand where my potential liability was and decide if it was necessary to take more action, including designing in redundancy if I felt I couldn't rely on a component.

It's not 'just a fact of life' in my industry. But I understand that a car manufacturer might not have to meet the same standard. Just sharing my personal experience, along with a belief honed by over 20 years in this business.

Edit: Over 25 years in this business. Dammit.

 

[smorgasbord]

...over 20 years in this business.

Edit: Over 25 years in this business. Dammit.

Time flies when you're having fun?

You're right that Fisker needed to know what the consequences of fan failure were. As long as we're speculating here, though I give them the benefit of the doubt here. I'd guess that the fire didn't start because something wasn't being cooled, the fire started because the fan motor or wiring literally overheated.

The thing that caught my eye was the retro-fitting of a new fuse. That tells me that Fisker has now identified a design defect in their use of the fan. I would expect all electrical components to be fused or breakered to prevent runaway current draws. It's SOP to prevent too much current being drawn over any electrical line, so I wonder what Fisker missed that they're now correcting.

 

[stopcrazypp]

Rather than trying to figure out who deserves the blame (in reality, even if Fisker points the finger to the fan manufacturer, most people will still at the very least assign some blame to Fisker if not all, given Fisker is the one doing the recall), I just hope Fisker did a very thorough investigation and a similar fire will NOT happen again after this recall. That's the most important part, since these kind of defects hurt the reputation of the whole plug-in industry and calls into question the ATVM loans.

 

[bonnie]

Rather than trying to figure out who deserves the blame (in reality, even if Fisker points the finger to the fan manufacturer, most people will still at the very least assign some blame to Fisker if not all, given Fisker is the one doing the recall), I just hope Fisker did a very thorough investigation and a similar fire will NOT happen again after this recall. That's the most important part, since these kind of defects hurt the reputation of the whole plug-in industry and calls into question the ATVM loans.

I think it's more about figuring out what went wrong in the process - without understanding that, there is no reassurance there are not other lurking problems.

 

[Doug_G]

The thing that caught my eye was the retro-fitting of a new fuse. That tells me that Fisker has now identified a design defect in their use of the fan. I would expect all electrical components to be fused or breakered to prevent runaway current draws. It's SOP to prevent too much current being drawn over any electrical line, so I wonder what Fisker missed that they're now correcting.

It's not difficult for electronics to start a fire without blowing an appropriately sized fuse. I've seen it happen - years ago one of our engineers started an overnight test. When he returned in the morning the machine wasn't responding, so he opened the sealed NEMA enclosure and was rewarded for his efforts by being engulfed in smoke. Everything in the box had been destroyed - thousands of dollars worth of electronics. We later found that one circuit board had a manufacturing flaw - on an inner layer of the circuit board itself - and that had started a small fire. Everything was appropriately fused and the engineer's design was in no way responsible for the failure. We changed circuit board suppliers after that.

The point is, it doesn't take that much energy to ignite something very small, electronics are full of small things, and small fires can spread and become big fires.

Most likely the fan is part of some larger system that is fused; however, the fan when shorting out didn't draw enough power to blow that fuse. Adding a fuse in line with the fan allows them to put in a much smaller fuse that has a higher likelihood of blowing if the fan draws too much current. But you can't just litter a vehicle with thousands of little fuses.

It's can be difficult in a complex machine to understand every possible failure mode. The aviation industry does a pretty darn good job of that, actually, with detailed fault trees and calculations of probability of failure for all critical systems. Although nothing and nobody is ever perfect, engineers included, mechanical causes of aircraft accidents are pretty rare events. So it can be done.

So sure, Fisker has to take the blame, even if a supplier let them down. Could they have done better? Probably. But even if you do everything right, stuff still happens.

 

[stopcrazypp]

I think it's more about figuring out what went wrong in the process - without understanding that, there is no reassurance there are not other lurking problems.

A bulk of the previous conversation seems to involve trying to figure out who deserved the blame rather than what really went wrong. And there's no way to examine further without looking at a more detailed report, since the only thing we have to go by right now is Fisker's press release (which blames an "internal fault" in the fan). I don't mind if Fisker doesn't release the report publicly, I just hope they really did a thorough investigation (rather a rushed one to address media attention), have nailed down the root cause internally, and that this recall will solve the problem.

There's also the NHTSA report on the Texas fire, which will hopefully provide more information.

 

[bonnie]

I thought it was more in the flavor of 'who is accountable and what should they have known'. Blame by itself doesn't do much good.

Fixing internal processes so it cannot happen again is key - and when the reaction is 'not our fault, it was the supplier', it doesn't point to fixing internal process. Ultimately, the company that puts the product on the market is accountable. I would have felt better had Fisker said 'it was the supplier and we are going to improve our internal process so something like this cannot happen again'. But they didn't.