Investor Engineering Discussions

[BioSehnsucht]

While there may be less friction loss (none between the gears, but still whatever they rotate on, unless also magneticly supported) with magnetic gears, somehow I doubt there's a total absence of energy loss.

Remember kids:
1) There is a game
2) You can't win
3) You can't break even
4) You can't get out of the game

 

[mongo]

Ugh, only 108 miles range (that's w/o the EZ-bake oven turned on), 115 mph charge rate on L3, and 630 Wh/mi rated energy consumption, I think I'll just wait for the Tesla van2rv conversion, thank-you...



North America's FIRST All-Electric, Zero Emissions RV? Winnebago eRV2 Facts & Features | Munro Live

Due to being a conversion, the user loads are a completely separate system from the chassis drive system (which is for final mile deliveries, thus the short range).

 

[Artful Dodger]

Due to being a conversion, the user loads are a completely separate system from the chassis drive system (which is for final mile deliveries, thus the short range).

Yeah, the Ford Transit van only has a rated range of 126 miles. Imma weight for da TESLA...

 

[Optimeer]

CATL has just announced a new battery with an energy density of 500 Wh/kg suitable for passenger aircrafts (cand cars)

https://twitter.com/x/status/1648630741801246720

What do you guys think about this battery?

One picture in the tweet shows the A380 aircraft which imho is a bad choice imho.

 

[JRP3]

Cycle life, costs, volumetric density? I'll believe it when 3rd parties have tested it.

 

[mongo]

Cycle life, costs, volumetric density? I'll believe it when 3rd parties have tested it.

Don't forget a pack, BMS, and thermal system.

 

[JRP3]

Don't forget a pack, BMS, and thermal system.

That's the relatively easy stuff, cell chemistry/construction is where the magic happens.

 

[mongo]

That's the relatively easy stuff, cell chemistry/construction is where the magic happens.

True, but Wh/kg for a large pouch is as optimistic as it gets.

 

[Buckminster]

https://twitter.com/x/status/1649164598602534917

 

[MC3OZ]

The video is about "Grid Forming Inverters", but the term "Grid Forming Batteries" is used fairly interchangeably.

The most interesting part of the talk is the presentation from the Tesla engineer starting around 13:22, the presenter immediately after her is the next most interesting/relevant presentation. However all of the presentations have some merit and are informative, if you don't know much about the topic.

All 3 batteries sites featured in the presentation have Tesla Megapacks. Megapacks running in "Virtual Machine Mode" provide grid forming (system strength) services in a very flexible way.

One of the presentations also covers the difference between "Grid Forming" and "Grid Following Inverters' and why that is important.

If you have a conversation with a Fossil Fuel loving power engineer and they tell you the grid can't run without synchronous generation, they are probably talking about "voltage system strength". Tesla batteries have been above to provide "frequency system strength" since soon after the commissioning of Hornsdale. The addition of "Virtual Machine Mode" now allows Tesla megapacks to provide voltage system strength", this also improves the ability of some sections of the grid to island and keep running in the event of broken transmission links.

Grid system strength is an issue in Australia, the strength varies in different parts of the grid, and the presentations cover how this affects the commissioning of solar farms and wind farms and how system strength must be maintained as coal and gas generators exit the grid. Even grid is different, in fact every part of the grid is different, but sooner or later most countries will hit the same issues. And this is another reason why demand for Megapacks will increase.

Batteries battle with spinning machines for key role in 100 pct renewables grid

Tesla says big batteries can deliver system strength services around one third cheaper than spinning machines currently favoured by network companies.

reneweconomy.com.au

If you really want to dive into the topic the next video goes deeper.

There is a product from Hitachi called an "Enhanced STATCOM" which provides a more heavy duty solution to some problems and "Synchronous Condensers" which are more expensive traditional solution. These are just additional reasons why Fossil Fuel loving power engineers are wrong.

But overall Megapacks running in VMM are probably the most economic and efficient solution.

The "Enhanced STATCOM" presentation around 40 minutes into the video is the most interesting part.

 

[mongo]

I don't know why you keep getting this wrong. Again, tabless can only mitigate what would otherwise be a slower charge rate in the larger cell.

According to the Battery Day presentation, the tabless design enables to increase the diameter of the cell from 21 mm to 48 mm without impacting supercharging times. This enables cost savings for the cells of approximately 18 % compared to 2170 cells and a range increase of approximately 7 % (if all other parameters are considered as equal).

View attachment 931740

View attachment 931743

The tabless design is patented by Tesla. So competitors and suppliers can only get these benefits by cooperating with Tesla. I think it is also clear that the cylindrical design is superior compared to pouch cells or cubic design, with one major exception: CATL LFP (since BYD had battery fires).

I think we should generally not assume sandbagging by Tesla unless there are clear indications.

I can post the data I have later

But lowering the resistance allows for more current for the same voltage rise, which is one of the limiting factor here

Yes, the diffusion is the limitation, what do you do to increase the diffusion? You increase the force against it, which is more voltage which results in higher current, and thus more ion diffusion

Look at any of Bjorn videos where he shows current and voltage during charging

At 50% the pack is pretty much at its maximum voltage, so you can’t charge faster without pushing the whole pack voltage over the maximum, which makes a speed run of degradation

On Tesla packs, as soon as the power starts going down, it’s because the pack is at close to or at full pack voltage, so even if Tesla wanted faster charging, they barely have any room to spare

The tables electrode will reduce the voltage rise on the jelly roll by changing the average distance of any given point in the electrode from 250 mm to 40 mm

The papers I posted quite a few pages ago agree on that, not taking this from nowhere

No you don't because that causes undesirable side reactions which cause electrode plating. Again, this has been covered. Also Tesla themselves does not show faster charging speeds on their own presentations of the 4680.

4680 form factor does not influence chemistry limitations.
The supercharging size vs time graph is normalized to 21mm diameter, a tabless 2170 would be able to charge faster than a tabbed 2170.

Tabless reduces resistance, and reduces resistance uniformity. This results in less voltage delta across the cell which allows higher currents without resulting in overpotential of any area. Having anode and cathode current paths in opposite directions helps also (though slightly offset by copper vs aluminum conductivity).

Reduced resistance and an improved thermal path improves heat generation and rejection allowing higher currents vs overheating (limited by chemistry and maximum potential).

Other notes:
Panasonic is making multitab 4680s.

 

[JRP3]

Obviously they don’t

Imagine if they public told 4680 vehicle will charge faster with the limited supply of 4680s they have

And I suggest you take a good look on both papers again

They go in depth into lithium plating and that it will still be avoided with faster charging times due to the tables electrode

ShieldSquare Captcha


ShieldSquare Captcha

Larger format cells provide benefits for energy density and power output but potentially exacerbate the internal current and temperature heterogeneity by having a longer jelly-roll. Increasing the cylinder diameter may also prove problematic for effective thermal management due to decreased surface-area-to-volume ratio of the cell. To address these problems, Tesla proposes a "tabless" current collection method 14 by using the current collector foil itself with a contiguous array of current collectors extending from the edge of the foil. This should mean that the current distribution inside the cell is much more uniform with the majority of the edge of each current collector foil being held at the same potential. In theory, this design reduces much of the ohmic loss inside the cell and with it much of the heat produced.

With the larger cell format, the heterogeneity in current collection resulting from high ohmic losses along the length of the jelly-roll are exacerbated and must be addressed by the tabless design. When using the traditional end-tab design, the ohmic losses from the current collectors results in five times more energy being lost as waste heat, compared with the tabless case.

What this means is that the tabless design reduces the negative impacts of the larger cell size not that it increases the C rate over existing 2170 cells.

 

[mongo]

Another point, Tesla said 50% of vehicles delivered were LFP

LFP is many times worse in respect to BMS calibration since it has such a flat voltage profile, so if a user don't fully charge almost daily and doesn't drain the pack often, the BMS is clueless where the bottom is

And due to that it's even more conservative than before to avoid a bad situation

Moreso where in the middle it is but, yeah, if it doesn't get to see one end or the other, it'll drift.
Roughtly:
99%-20% is a 0.5V drop
20%-14% is a 0.4V drop

The BMS can do coulomb counting, and Vdrop vs current, not just voltage so it has some other methods.

 

[GhostSkater]

Moreso where in the middle it is but, yeah, if it doesn't get to see one end or the other, it'll drift.
Roughtly:
99%-20% is a 0.5V drop
20%-14% is a 0.4V drop

The BMS can do coulomb counting, and Vdrop vs current, not just voltage so it has some other methods.

Yes, been a while, but I did a BMS design specialization some time ago, at least started, it got on a part that dove way too deep into the math itself that wasn't what I was looking for

You need way more than voltage and coulomb counting for an accurate SOC estimation.

Voltage is the worst in the SOC estimation unless you pack is at an even and stable temperature, had no use for a long period of time (24h+) and you did a good characterization of the cells used beforehand, without it, your SOC estimation can de 30% or more off, even more with LFP

Coulomb counting works only if you are doing a single cycle from 100% SOC because every current measurement has an error, and as time goes on the integration error gets out of hand, in a few cycles you might completely lost track of it if they aren't from 100% SOC

What you need is an algorithm that takes voltage, coulomb counting, temperature, current, SOH and the history of all those values to be able to compute SOC and to keep track of it

There is no way to directly measure SOC.

Ok. there is, but involves having a probe measuring lithium concentration inside the jelly roll, and even that isn't accurate always because at heavy usage, not all points of the cell are at the same SOC at the same time (something the tabless electrode helps a lot in evening out and thus my comments on being one the ways it helps with faster charging)

 

[MC3OZ]

Tesla has faced a number of manufacturing hurdles with 4680 battery manufacturing and progress has been slower than expected. Beyond manufacturing issues though, as I discuss in this video, Tesla also discovered and fixed a performance issue that was caused by a crucial component of the dry electrode manufacturing process, the polymer binder, which was breaking down at higher voltage, leading to a possible reduction in cell performance or outright cell failure.

 

[MC3OZ]

IMO the stainless steel skin will turn out to be an important part of the body strength for towing and hauling.

 

[mongo]

IMO the stainless steel skin will turn out to be an important part of the body strength for towing and hauling.

Please explain further, aren't those loads primarily concentrated in the rear axle assembly?
Center structure handles pitch torque and center loads. Towing should be mainly inline other than tongue weight. Payload in bed should be mainly vertical.

 

[JRP3]

He didn’t say much about it, but he shut down Bill Maher’s claim that we will run out of water and insisted that desalination is “absurdly cheap” and will solve it.

I still have concerns about unintended consequences of that on a large scale. There is no free lunch.

 

[Gigapress]

I still have concerns about unintended consequences of that on a large scale. There is no free lunch.

I do too, but what alternative is there? That’s a genuine question. What else would we do?

Humans have been geoengineering the hydrological cycle at massive scale since the dawn of civilization thousands of years ago. Irrigation is an older technology than the wheel. Currently we get our water mainly from sucking rivers, lakes and aquifers dry. For one of the most extreme examples, today 80% of the Colorado River’s flow is used for agriculture and distributed across the Southwest via a giant system of aqueducts, and even more is used for other human purposes. Barely a trickle makes it to the Gulf of California anymore. Elsewhere in the world, we have nearly killed several major lakes including the Aral Sea and Lake Chad by overconsumption of incoming river flow for agriculture.

We also tend to put big dams on our rivers for hydropower and for storage reservoirs. For example, here in the Pacific Northwest we have no less than 274 dams scattered throughout the Columbia River watershed. This has had a lot of negative unintended consequences as well. It’s disruptive to salmon migration which has had catastrophic consequences due to salmon being a keystone for the entire regional ecosystem. Dams also impact sediment deposition patterns, nutrient cycling, oxygenation, temperature and more.

There is no solution I’m aware of that can supply adequate water for 10B humans that doesn’t involve large scale geoengineering. I don’t have deep expertise on this but I think that manipulating a tiny fraction of the world’s seawater and finding a more responsible way to dispose of it than typical current practices is the much lesser evil when compared to what we’re doing now to the comparatively limited and precious freshwater resources naturally available on land.

 

[JRP3]

I do too, but what alternative is there? That’s a genuine question. What else would we do?

What we do is not allow Elon's fake "population crash" narrative to go unchallenged or hand waved away with "just desalinate seawater". Yes we will likely have to use desalination as the least bad choice to overpopulation, along with many other bad choices we'll have to make to deal with it.

There is no solution I’m aware of that can supply adequate water for 10B humans that doesn’t involve large scale geoengineering.