Tesla 'open source' their patents

Seen something interesting in the news or on the intertubes? Discuss it here.

Moderators: Zamfir, Hawknc, Moderators General, Prelates

User avatar
Crissa
Posts: 294
Joined: Sun Mar 20, 2011 8:06 pm UTC

Re: Tesla 'open source' their patents

Postby Crissa » Tue Jun 24, 2014 6:12 am UTC

Just because you use off the shelf cells doesn't mean you don't also then tie those cells together in a revolutionary way.

Like I said, not insurmountable - but so far, unique. And kudos for them for making sure that doesn't stay unique for the twenty years of their patents. That way we all can build off of their experience.

-Crissa

KnightExemplar
Posts: 5494
Joined: Sun Dec 26, 2010 1:58 pm UTC

Re: Tesla 'open source' their patents

Postby KnightExemplar » Tue Jun 24, 2014 2:11 pm UTC

Crissa wrote:Just because you use off the shelf cells doesn't mean you don't also then tie those cells together in a revolutionary way.

Like I said, not insurmountable - but so far, unique. And kudos for them for making sure that doesn't stay unique for the twenty years of their patents. That way we all can build off of their experience.

-Crissa


Its been almost two weeks, and I still don't see any legal documentation that states what Elon Musk means by "uses a patent in good faith". So far, they HAVEN'T released their patents, they just released a blog post that is stirring up good press.

Again, its not a release of patents until they offer a license for it. I've already linked to the Microsoft example earlier.
First Strike +1/+1 and Indestructible.

stevey_frac
Posts: 947
Joined: Tue Oct 20, 2009 10:27 pm UTC

Re: Tesla 'open source' their patents

Postby stevey_frac » Tue Jun 24, 2014 3:15 pm UTC

Crissa wrote:Just because you use off the shelf cells doesn't mean you don't also then tie those cells together in a revolutionary way.

Like I said, not insurmountable - but so far, unique. And kudos for them for making sure that doesn't stay unique for the twenty years of their patents. That way we all can build off of their experience.

-Crissa


So far, the only evidence I've seen of Tesla's 'Novel means of connecting cells' involves skipping the voltage sense lead on every other cell. That's hardly earth shattering innovation, and definitely not an insurmountable lead.

Honestly, I think half the reason why Elon released these patents is to generate a pile of press and also because they were indefensible in court anyways. I can just see the court case now.. 'But your honour, it's every OTHER cell!!' *mind blown*

Heisenberg
Posts: 3789
Joined: Wed May 14, 2008 8:48 pm UTC
Location: Uncertain

Re: Tesla 'open source' their patents

Postby Heisenberg » Tue Jun 24, 2014 3:28 pm UTC

KnightExemplar wrote:Its been almost two weeks, and I still don't see any legal documentation that states what Elon Musk means by "uses a patent in good faith". So far, they HAVEN'T released their patents, they just released a blog post that is stirring up good press.

It may be on a case-by-case basis. For example, BMW met with Tesla and discussed sharing technology and charging stations.

Musk's promise is hedged, and I doubt he's going to just open the henhouse and let the Big Three take what they want. It's more of a "come to me with your idea, and if I think it'll further the electric movement, I'll give you free stuff."

moiraemachy
Posts: 190
Joined: Wed Jan 04, 2012 9:47 pm UTC

Re: Tesla 'open source' their patents

Postby moiraemachy » Tue Jun 24, 2014 4:28 pm UTC

@stevey_frac : I didn't claim that Tesla charges their cells individually, and would bet against it.

I don't have that much knowledge about batteries (I am loving to review some knowledge in this topic), but for now, I am calling it as I see it: Tesla rocks, everyone else drools. At least in the 2010-2012 period, everything suggests that they made a better, cheaper battery from off-the-shelf components, against competitors who used customized solutions in larger scales (and, considering scale, investment and weight per kWh, I would bet that their solution still is the best). If this is not better engineering, I don't know what is.

Now, I am not so sure about how they did it, but the patents and the buzz around it suggest me the following (may contain bullshit, proceed with care): the naivest approach (nobody does this, but I will start with it for simplicity) is to connect the engine (engine's inverter actually) to a 400V bus, and connect several 400V cell strings (I don't know if this is the technical term, but I like it, so I am now calling individual arrangements of cells in series "strings") to this bus, in parallel. You charge by putting current in the bus. The problem here is that a string's charge is limited by the crappiest cell, and all the strings must all remain the same. So after the shittiest cell is filled, you can't charge anyone else.

Having a weak link is not a problem in a laptop battery, with strings of 3 cells. In a EV, the problem scales. One solution is to use batteries with little variance in their capacities (cue car manufacturers doing research on battery technology and saying it is not mature yet). Other is to use your battery conservatively to avoid putting too much strain on the weak links (cue the Chevrolet Volt not letting you fully discharge its battery).

Or maybe you can rethink your topology. The simplest one is to add some circuitry that disconnects individual 400V strings from the main bus and simply connects its cells in parallel. It sounds like this is what the Smart Electric Drive does in the balance step: all cell voltages are monitored, and if a significant imbalance in the voltages of the cells of a string is detected, it is disconnected from the main bus and undergoes balancing. This does not fully address the problem: you might need to balance some strings when power is demanded, and a failed cell brings down its whole string.

But if you are going to rethink the topology, why stop there? You already have wires going to almost every cell to measure voltage, and you already paid for the transistors that gate all this current. By adding little additional electronics you can add mechanisms to bypass failed cells without compromising the whole string, you can do balancing and charging simultaneously, you can save money by outsourcing failure protection out from individual cells...

Which leads me to this patent that took me way more time than it should to find. I think this is the biggest one: US 20090023053 A1 (I couldn't get permalinks in this http://patft.uspto.gov website, so this link might die. In this case just search for the number)

Now, they don't give away the topology they are using, but I am guessing this: they make 400V "strings" by stacking modules of cells in series. Each of this modules is comprised of several cells in parallel and a charge balancing circuit. And in a string, a charge balancing circuit balances modules. So now, you balance charge at string and module level instead of only at string level, and a failed cell will make a module underperform, instead of ruining an entire string. But since you can charge-balance, not a big deal. And don't get me started with cross-string balancing...

What I am saying is: I bet there is a team of guys at Tesla whose job is to think about battery topology shenanigans, and they love to talk about it in computer science and network theory terms. Other car manufacturers have a much smaller amount of cells (the Leaf has around 200 cells and the Volt 300), so I bet their topologies are basically "do what laptops do, but make it 400V, 25kWh". But that, with current tech, just doesn't scale very well.

stevey_frac
Posts: 947
Joined: Tue Oct 20, 2009 10:27 pm UTC

Re: Tesla 'open source' their patents

Postby stevey_frac » Tue Jun 24, 2014 5:52 pm UTC

moiraemachy wrote:@stevey_frac : I didn't claim that Tesla charges their cells individually, and would bet against it.

I don't have that much knowledge about batteries (I am loving to review some knowledge in this topic), but for now, I am calling it as I see it: Tesla rocks, everyone else drools. At least in the 2010-2012 period, everything suggests that they made a better, cheaper battery from off-the-shelf components, against competitors who used customized solutions in larger scales (and, considering scale, investment and weight per kWh, I would bet that their solution still is the best). If this is not better engineering, I don't know what is.

Now, I am not so sure about how they did it, but the patents and the buzz around it suggest me the following (may contain bullshit, proceed with care): the naivest approach (nobody does this, but I will start with it for simplicity) is to connect the engine (engine's inverter actually) to a 400V bus, and connect several 400V cell strings (I don't know if this is the technical term, but I like it, so I am now calling individual arrangements of cells in series "strings") to this bus, in parallel. You charge by putting current in the bus. The problem here is that a string's charge is limited by the crappiest cell, and all the strings must all remain the same. So after the shittiest cell is filled, you can't charge anyone else.

Having a weak link is not a problem in a laptop battery, with strings of 3 cells. In a EV, the problem scales. One solution is to use batteries with little variance in their capacities (cue car manufacturers doing research on battery technology and saying it is not mature yet). Other is to use your battery conservatively to avoid putting too much strain on the weak links (cue the Chevrolet Volt not letting you fully discharge its battery).

Or maybe you can rethink your topology. The simplest one is to add some circuitry that disconnects individual 400V strings from the main bus and simply connects its cells in parallel. It sounds like this is what the Smart Electric Drive does in the balance step: all cell voltages are monitored, and if a significant imbalance in the voltages of the cells of a string is detected, it is disconnected from the main bus and undergoes balancing. This does not fully address the problem: you might need to balance some strings when power is demanded, and a failed cell brings down its whole string.

But if you are going to rethink the topology, why stop there? You already have wires going to almost every cell to measure voltage, and you already paid for the transistors that gate all this current. By adding little additional electronics you can add mechanisms to bypass failed cells without compromising the whole string, you can do balancing and charging simultaneously, you can save money by outsourcing failure protection out from individual cells...

Which leads me to this patent that took me way more time than it should to find. I think this is the biggest one: US 20090023053 A1 (I couldn't get permalinks in this http://patft.uspto.gov website, so this link might die. In this case just search for the number)

Now, they don't give away the topology they are using, but I am guessing this: they make 400V "strings" by stacking modules of cells in series. Each of this modules is comprised of several cells in parallel and a charge balancing circuit. And in a string, a charge balancing circuit balances modules. So now, you balance charge at string and module level instead of only at string level, and a failed cell will make a module underperform, instead of ruining an entire string. But since you can charge-balance, not a big deal. And don't get me started with cross-string balancing...

What I am saying is: I bet there is a team of guys at Tesla whose job is to think about battery topology shenanigans, and they love to talk about it in computer science and network theory terms. Other car manufacturers have a much smaller amount of cells (the Leaf has around 200 cells and the Volt 300), so I bet their topologies are basically "do what laptops do, but make it 400V, 25kWh". But that, with current tech, just doesn't scale very well.


First off, This was your statement:
Yeah, you're describing a charging cycle which touches the entire batter while Tesla is describing a charging cycle where each cell of the battery has its own cycle


How am I not to interpret that as 'Tesla charges each cell individually'? Unless you're talking about the balancing step? Which, we've already agreed, most modern packs perform?

Instead of discussing your hypothetical topology, let's discuss the actual topology.

It goes like this: 11 modules in series, each of which contains 9 bricks in series, each of which contains 69 cells in parallel. This means that when a single cell underperforms it's not that big of a deal. This is important when you're buying cheap cells from China with crappy quality control. It also means that you can pop a couple of cells and not have it be too big of a deal, other than it limiting performance. Pop enough cells in the same brick, and your performance will tank, since your current through the device is limited to the current through the any individual brick. This is, once again, pretty standard battery design. You replace the 69 cells with one or two large format cells of better quality, and bingo, you can do exactly the same things you did before. You are still charging by applying a current across the entire pack, and you still have a per cell balancing step, just like my Smart does. This isn't amazing new technology. It's a standard engineering practice. Interestingly, the Leaf has two such strings, instead of just one for the Tesla. Does that make it's battery pack more advanced? Surely not...

At best you can claim that the Tesla pack used to be cheaper. I'm definitely not seeing any advantage other than potentially the cost of cells, but as the cost of cells plummets world wide, I'd bet the cost of engineering and building that way more complicated pack will outstrip the cost of building the more resilient large format cells. The large format cells have the advantage of being inherently more safe, and not requiring as sophisticated a cooling system. This in turn means you can pack more electrode into less space for the entire pack. I realize that saying anything critical of the car company the internet is in love with will earn me more knee jerk replies without substance, but meh.

So, no. Tesla doesn't win any amazing awards for battery design. In fact, as far as I can tell, I can get nearly the same charging and discharging system if I do a bulk order from Panasonic for the same cells. I believe they'll even solder the cells to the charging circuitry for me, for a small fee. This isn't to take away from what they accomplished, and continue to accomplish.

What I'm trying to say is, Tesla recognized that battery technology had developed enough that they could take off the shelf tech, and make an awesome all electric car, and good for them! They've done just that. The Tesla is the best car in the world. But it's not there because of mythical advances in battery tech that they alone posses. It's because they built an awesome car, because they built an awesome network of fast chargers, and it's because the gadgetry in the car is cool, and finally, it's because they packed more battery into their car than competitive vehicles, which has more to do with the cost than available space.

User avatar
Crissa
Posts: 294
Joined: Sun Mar 20, 2011 8:06 pm UTC

Re: Tesla 'open source' their patents

Postby Crissa » Tue Jun 24, 2014 8:26 pm UTC

You quoted me and attributed it to someone else.

And even then, there's nothing not-unique just because you can tell a manufacturer to assemble something a specific way. That's how manufacturing works.

The best patents are the ones which are just taking what's out there and assembling it a way no one had yet to make work.

-Crissa

moiraemachy
Posts: 190
Joined: Wed Jan 04, 2012 9:47 pm UTC

Re: Tesla 'open source' their patents

Postby moiraemachy » Tue Jun 24, 2014 9:30 pm UTC

stevey_frac wrote:It goes like this: 11 modules in series, each of which contains 9 bricks in series, each of which contains 69 cells in parallel.
Nice, I was looking for that. So, from now on, I will use the terminology modules (36V, 621 cells) and bricks (4V, 69 cells).

stevey_frac wrote:You replace the 69 cells with one or two large format cells of better quality, and bingo, you can do exactly the same things you did before.
Reliability goes down. Say one in 100 cells from Tesla die in 10 years. Say 10 dead batteries will ruin your brick. This means that the odds of one brick failing in 10 years are around... 1 in 300 million. I doubt any individual cell would get close to that.

Right now, there is a large trade-off involving battery reliability, capacity, durability and cost. Why not use existing electronics that makes batteries more reliable, at little additional cost? Even if they do achieve some breakthrough in battery technology, Tesla's approach is still the way to go: why not use, say, 5 big reliable cells in parallel in a brick instead of one single cell, and have your battery still work if one of them fails?

stevey_frac wrote:You are still charging by applying a current across the entire pack, and you still have a per cell balancing step, just like my Smart does.
If you are willing to get the most of every cell, the balancing must be done while charging and while discharging. I don't know how the Smart does it, but if you want balancing to happen while batteries are providing energy, you will need some active electronics, which will be costly. But in Tesla's design, you could in theory reuse the electronics that turns off defective cells for this purpose. (using cells like the capacitors in an arrangement like this http://www.avdweb.nl/Article_files/Sola ... citors.jpg)


But anyway, you have being claiming that Tesla is basically doing textbook battery design. Again, I am not a specialist, but I find this highly unlikely. Can you prove that their approach is not innovative? Their approach being defined as: making high-voltage (300V and up) batteries by connecting several "bricks" in series, with these bricks looking like this: (fig 2 from US 20090023053 A1)

Image

That is, a big number of cells gated by some switching element, connected in parallel, with some monitoring electronics to detect and prevent faults.

stevey_frac
Posts: 947
Joined: Tue Oct 20, 2009 10:27 pm UTC

Re: Tesla 'open source' their patents

Postby stevey_frac » Tue Jun 24, 2014 9:51 pm UTC

Crissa wrote:You quoted me and attributed it to someone else.

And even then, there's nothing not-unique just because you can tell a manufacturer to assemble something a specific way. That's how manufacturing works.

The best patents are the ones which are just taking what's out there and assembling it a way no one had yet to make work.

-Crissa


Sorry on the misattribution.

And as I shall state again, as plainly as possible: This is not novel. Tesla isn't the first entity to use battery stacks. It's common, normal, standard battery construction.

Note how this is an example of battery stack management tech from 2008, pre-model S

And also note how it's just a minor shrinking of another piece, the reduces chip size, and part count, not a massive leap forward.

And no, the best patents are those that articulate something novel, and interesting, not those that provide marginal value to existing standard engineering practices. I doubt if a lot of these would even stand up in court.



But anyway, you have being claiming that Tesla is basically doing textbook battery design. Again, I am not a specialist, but I find this highly unlikely. Can you prove that their approach is not innovative? Their approach being defined as: making high-voltage (300V and up) batteries by connecting several "bricks" in series, with these bricks looking like this: (fig 2 from US 20090023053 A1)


Please see the above link that describes electronics to do cell monitoring and balancing, per cell, as a minor improvement, mainly in size over existing systems that already did the same thing. As I've stated before, this is standard battery building tech, and has been in use for a long time. It is not novel.

You claim that Tesla does balancing on discharge.. I'm not sure that's true, as how would you detect an unbalanced discharge? All the voltages of the discharging cells across a brick are identical, by definition. And seeing how the trigger to stop discharge of a cell is voltage based, you can only conclude when you should stop drawing power from the entire brick.

Reliability goes down. Say one in 100 cells from Tesla die in 10 years. Say 10 dead batteries will ruin your brick. This means that the odds of one brick failing in 10 years are around... 1 in 300 million. I doubt any individual cell would get close to that.


You are making the assumption that large format cells manufactured for automotive purposes, and small format single cells manufactured for laptops have similar reliability. This is a faulty assumption. Also, you are assuming that cell failure mode is destruction. This is generally not true, the failure mode is overwhelmingly decreased capacity, and in that, the cell format is somewhat irrelevant. The cells will wear fairly evenly between formats, with neither providing a distinct advantage, unless the chemistry is different.

The large format cells would be worse off in the event of a cell short, but then, most cars have multiple strings. You can simply disable the faulting string, and use the other string to get to a dealership, where you can replace the faulty module. If the large format cells are much more resilient to this type of behaviour, then it's less of an issue.

Edit:

The Chevrolet Volt pack is 3 cells in parallel, 96 cells in series. So, I think that meets your requirement of 'other people are doing this'... it's just using large format cells instead of tiny cells.

If you are going to argue the only novel thing about them is that they've adapted to use tiny cells, I'll concede that. No one else is doing specifically that with cars. But, I don't think that gives them any cost, reliability, and performance advantages, or at least not any more. It may have had cost advantages in 2012.

moiraemachy
Posts: 190
Joined: Wed Jan 04, 2012 9:47 pm UTC

Re: Tesla 'open source' their patents

Postby moiraemachy » Tue Jun 24, 2014 11:05 pm UTC

stevey_frac wrote:You claim that Tesla does balancing on discharge.. I'm not sure that's true, as how would you detect an unbalanced discharge? All the voltages of the discharging cells across a brick are identical, by definition. And seeing how the trigger to stop discharge of a cell is voltage based, you can only conclude when you should stop drawing power from the entire brick.
You are right, balancing is not done in cells connected in parallel. Balancing is only necessary on elements that are connected in series, that is: among bricks, and you have brick voltages. My proposal would be to have some floating cells in every brick that can be connected to the next brick when balancing is necessary. But I don't know how Tesla actually does it.

All I know is: if you use cells in series, the low capacity cells will be fully discharged while the high capacity cells still have energy. In order to use this energy, you need a mechanism that transfers energy from high capacity cells to low capacity cells. Something like this. If you don't do this, then your balancing step is only useful to avoid overcharging. It will not allow high capacity cells to compensate for low capacity cells.

The IC you linked to basically does the measurements for a system that does charge balancing. But this is not what I am claiming Tesla's innovation is. I am claiming that their innovation is replacing one cell, in a string of several cells in series, by a "brick" of several cells connected in parallel with some gating circuit in each cell, allowing faulty cells to be removed from the brick without making the brick inoperable (and allowing measurements of individual cell capacity by temporarily turning off a cell and measuring the change in the discharge rate). This is the stuff described in Tesla's patents. If a cell in a brick stops working, the brick's capacity is reduced, but a balancing circuit like the one I just linked could take care of this, allowing other bricks to compensate. All this with the specific goal of increasing reliability of unreliable cells.

If this is not new, then I will concede that they don't deserve the hype.

stevey_frac wrote:I don't think that gives them any cost, reliability, and performance advantages, or at least not any more. It may have had cost advantages in 2012.
Assuming same battery technology, Tesla's topology is fundamentally more reliable (unless cells get more reliable than the switching electronics, and that is just not happening) and more expensive (more electronics). The question is exactly what are the trade-offs. I think Tesla's approach is still worth it based on a vague impression that electronics are cheap and batteries expensive. (so making cells more reliable can't be that cheap)

stevey_frac
Posts: 947
Joined: Tue Oct 20, 2009 10:27 pm UTC

Re: Tesla 'open source' their patents

Postby stevey_frac » Wed Jun 25, 2014 1:18 am UTC

moiraemachy wrote:
If this is not new, then I will concede that they don't deserve the hype.

Assuming same battery technology, Tesla's topology is fundamentally more reliable (unless cells get more reliable than the switching electronics, and that is just not happening) and more expensive (more electronics). The question is exactly what are the trade-offs. I think Tesla's approach is still worth it based on a vague impression that electronics are cheap and batteries expensive. (so making cells more reliable can't be that cheap)


This Link discusses almost verbatim what you describe Tesla as innovating including:

"Instead of large cells, use parallel strings of smaller cells"
"The failure of an individual cell in a parallel configuration will not cause the failure of the whole battery which could possibly continue functioning at lower power."

The way back machine suggests this was written in 2012, slightly before the release of the model S. I've been really trying to avoid an appear to authority here, but this kind of thing was discussed in my engineering classes.

You suggest that Tesla's design is more reliable. In absolute terms, that may well be true, but cell failures, and by that I mean failures that we have to engage Tesla's failure mechanism, are incredibly rare. A cell failure rate of 1 in 200k was deemed unacceptably high, and a recall was issued for Dell laptops. Failure rates are now estimated to be in the 1 in 10 million cells range. That means, with 200 cells in a car, that car has a rate of 0.002% for it's batteries, over it's lifetime. That's acceptable, especially, if the car can engage some sort of limp home mode, which the Leaf and the Smart can do.

That's not a reliability problem. That's roughly --- 3 or 4 orders of magnitude more reliable than your average car powertrain.

KnightExemplar
Posts: 5494
Joined: Sun Dec 26, 2010 1:58 pm UTC

Re: Tesla 'open source' their patents

Postby KnightExemplar » Wed Jun 25, 2014 1:00 pm UTC

Heisenberg wrote:
KnightExemplar wrote:Its been almost two weeks, and I still don't see any legal documentation that states what Elon Musk means by "uses a patent in good faith". So far, they HAVEN'T released their patents, they just released a blog post that is stirring up good press.

It may be on a case-by-case basis. For example, BMW met with Tesla and discussed sharing technology and charging stations.

Musk's promise is hedged, and I doubt he's going to just open the henhouse and let the Big Three take what they want. It's more of a "come to me with your idea, and if I think it'll further the electric movement, I'll give you free stuff."


Then it isn't Open Source.

I don't need to ask Linus's permission to use or modify the Linux Kernel. I don't need to ask Microsoft to use or modify the C# compiler. "Open Source" means opening the henhouse and letting everyone out there get what they want (with minor exceptions: like Copyleft or Trademarks).

Perhaps Tesla is in "open negotiations" with other car companies, or something. But what they're doing right now is hardly "open source". Ford, Mazda, and Volvo all shared the "C1 Platform" for example, and all electric vehicles share a common charging platform already: the SAE J1772. If Tesla wants to impress me, they'll have to do more than just collaborate with others by sharing its proprietary charging platform... especially when Musk is promising "Open Source".

--------------

I'm being a stickler about this because companies of the past have screwed over the OSS community. AT&T made Unix de-facto open source. They simply never enforced the copyright... so when people were off Xeroxing the Unix source code, AT&T didn't really care. The underground pseudo-legal code was part of the primordial soup of code that eventually grew into the Open Source movement. (Stallman at this time believed it'd be important to rewrite a lot of the code from scratch anyway, to provide a strong legal basis for OSS)

Eventually, AT&T sold off its copyright in confusing ways, and the company SCO eventually was led to believe it held the UNIX copyright. Armed with the "confidence" that they owned UNIX, SCO then underwent a campaign to attempt to sue Linux users... starting with IBM's contribution to Red Hat. Code that was influenced by the Lion's book finally came under attack by SCO. AT&T may have done the right thing by treating their code as de-facto open source (they never really enforced the copyright), but without an explicit legal license, future companies (like SCO) were able to strong-arm a good portion of the OSS community.

Fortunately, Novell ended up being awarded the Unix copyright and the case settled as SCO went bankrupt. But the lesson here is that future IP holders can be dicks. Even if you have 100% trust behind Elon Musk and Tesla... you cannot just assume Tesla will always be in control of these patents: they may be sold off to someone else, and then _that_ person might be a dick about it. So yes, the legal documentation that precisely gives up patent rights (or copyright rights) is important.
First Strike +1/+1 and Indestructible.

moiraemachy
Posts: 190
Joined: Wed Jan 04, 2012 9:47 pm UTC

Re: Tesla 'open source' their patents

Postby moiraemachy » Wed Jun 25, 2014 2:59 pm UTC

stevey_frac wrote:This Link discusses almost verbatim what you describe Tesla as innovating including:

"Instead of large cells, use parallel strings of smaller cells"
"The failure of an individual cell in a parallel configuration will not cause the failure of the whole battery which could possibly continue functioning at lower power."

The way back machine suggests this was written in 2012, slightly before the release of the model S. I've been really trying to avoid an appear to authority here, but this kind of thing was discussed in my engineering classes.
First, Tesla filled the patent I mentioned in 2007. Second, the parts you quoted follow a discussion about EV batteries. In fact, that session explicitly contradicts your figure of 1 failure in 10 million cells, which was already strange to begin with (under what conditions? A plane battery that undergoes extensive testing, is replaced every few years, and operated with low depth of discharge and under very specific conditions? Yeah, that will improve reliability).

And even if the parts you quoted were written before 2007... listen, I am not claiming Tesla invented the parallel circuit, which is the most obvious way to increase reliability. I am claiming they innovated by doing the following, simultaneously:
1) Make a battery with several modules ("bricks") arranged in series to provide a high voltage
2) Build these modules with several cells in parallel, with one switching element in each cell, allowing individual cells to be disconnected.
3) Monitor the voltage in every module, and perform balancing among modules.
4) Periodically disconnect individual cells, allowing measurement of cell parameters without having one sensor per cell, simply by measuring the difference in rate of discharge.

Yes, this does not sound that complicated. Still, I don't think this constituted "pretty standard battery design". (this also does not touch "savings in safety components")

And finally, you have been focusing on failed cells in your last post, but this is missing the point: even if cells don't fail, their capacity decreases. And individual cells have their capacities decreased at different rates. When operated in series, these differences tend to increase due to positive feedback. By using modules ("bricks") with more cells in parallel, the variance in brick capacity is greatly reduced. This means the balancing circuit will have less work to do and weaker cells will not suffer so much.

User avatar
Crissa
Posts: 294
Joined: Sun Mar 20, 2011 8:06 pm UTC

Re: Tesla 'open source' their patents

Postby Crissa » Wed Jun 25, 2014 4:51 pm UTC

Post note: It wasn't really SCO that did the patent trolling, but instead someone who had bought SCO and fired all the developers.

Much like AT&T today isn't AT&T but instead SBC wearing the name of the company. (There is infact a rump company named 'AT&T Legacy' that holds many of AT&T's old assets that SBC didn't buy. Like their bank accounts and billing system.)

-Crissa

stevey_frac
Posts: 947
Joined: Tue Oct 20, 2009 10:27 pm UTC

Re: Tesla 'open source' their patents

Postby stevey_frac » Wed Jun 25, 2014 5:39 pm UTC

moiraemachy wrote:1) Make a battery with several modules ("bricks") arranged in series to provide a high voltage
2) Build these modules with several cells in parallel, with one switching element in each cell, allowing individual cells to be disconnected.
3) Monitor the voltage in every module, and perform balancing among modules.
4) Periodically disconnect individual cells, allowing measurement of cell parameters without having one sensor per cell, simply by measuring the difference in rate of discharge.



The Chevrolet Volt does *all* of those things with large format cells, with the possible exception of balancing on discharge, which I'm not sure makes sense. Your step 4 (which is really a condition of step 3) doesn't make sense in normal operation, since the if you command a 20 amp current through the stack, and then disconnect a cell in parallel with other cells, the other cells will simply increase current to make up the difference, this might result in a bit of a voltage drop, but it's going to be tiny, and possibly lost in the noise, and even then, you're now testing the remaining cells connected instead of the cell you disconnected. You could only perform that test if your battery was being effectively shorted, which, you really REALLY don't want to do.

I gave a link for the cell reliability. Feel free to argue with it, but it's not just something I made up. Personally, I'm surrounded by li-ion batteries. I've got a couple hundred in my car, my cell phone, the cell phones of all of my friends, their friends, laptops, etc... And the only catastrophic failures of the type Tesla can protect against that I've heard of were years ago, with the Dell laptop recall, and the failure rate on that was 1 in 200k. Have you heard of a lot of Li-ion cells dying?

As for battery wear... well... uhm. Who cares? These are pretty solid batteries. No seriously... They're really solid. That's why they come with massive warranties, and why car makers spent lots of time getting batteries *right*. And because the large format cells are trivial to replace, if you battery starts to degrade, they can use the BMS to find the worn cells, and replace *just* those cells, and your batter gets a nice big boost! Done! Any mechanic can replace cell 2A as instructed by his computer read-out. They are designed to be serviceable In fact my yearly service does *exactly* this. If there is an under performing cell, they'll replace it under warranty. Done.

In any case, like I've said previously, I think their pack is overall a pretty standard battery design, adapted to deal with the larger number of lower quality cells. It's pretty obvious I'm not going to convince you of that at this point, so I may as well stop trying. At the very least, the radical claims that were in this thread of been scaled down from 'Tesla rules, everyone else drools' to 'Better lifespan due to discharge balancing'. That may well be true, but I'd claim the difference is pretty marginal. I'll call that a win, and move on.


Return to “News & Articles”

Who is online

Users browsing this forum: No registered users and 12 guests