JRP3
Hyperactive Member
Center screen is setup to articulate, seems to only be missing the motors.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Investor Engineering Discussions
- Thread starter MC3OZ
- Start date
Very interesting video from Limiting Factor:-
I arrived at similar conclusions on LFP and LMNO, mainly because Limiting Factor is my primary source of battery information.
The useful additional point he made on LMNO is that good electrolytes are currently expensive.
It could just be confirmation bias, but that is the kind of R&D challenge I could see Tesla taking on.
In that context all, of Elon's comments now make sense to me, they are interested in LMNO, but it is not yet over the line.
There were additional interesting comments form Limiting Factor in the video comments:-
I arrived at similar conclusions on LFP and LMNO, mainly because Limiting Factor is my primary source of battery information.
The useful additional point he made on LMNO is that good electrolytes are currently expensive.
It could just be confirmation bias, but that is the kind of R&D challenge I could see Tesla taking on.
In that context all, of Elon's comments now make sense to me, they are interested in LMNO, but it is not yet over the line.
There were additional interesting comments form Limiting Factor in the video comments:-
JRP3
Hyperactive Member
And guess who's been working on high voltage electrolytes for years?
cleantechnica.com
Jeff Dahn's Research Team Improves High-Voltage Lithium-Ion Battery Cell Performance Through Ethylene-Carbonate-Free Electrolytes - CleanTechnica
Sign up for daily news updates from CleanTechnica on email. Or follow us on Google News! Jeff Dahn’s battery technology research team at Dalhousie University in Canada has developed a new means of improving high-voltage, lithium-ion battery cell performance through the use of cyclic carbonates...
cleantechnica.com
petit_bateau
Active Member
I had a chance to view the latest Munro stripdown video of the Model S Plaid motor/inverter today.
It was interesting and well worth watching although I disagree with Sandy on a few things (not for the first time) there was something he called out that also struck me as odd, i.e. this is not a skewed winding (see the 15:05 point). Now it has been several years since I last did a PMG/PMM but back in the day I worked with both skewed and unskewed. Ordinarily in a PMG (or M) one skews to reduce cogging and ripple, but this is not skewed. Sandy mentions that in driving the S it feels smooth, and presumably is also audibly quiet (human ears are pretty good at picking up harmonics, and I've heard no complaints), and so they must be managing this some other way.
Q1. How are they managing it ? (presumably with a more complex inverter waveform management ?)
Q2. Given that the inverter is common to the 3, the Y, the S, and the Plaid S, (and ?? presumably the X, and in due course the Semi and the CT and the Roadster 2) then is there skew in the windings of the others ?
By the way : The Tesla parts bin is going to be a core foundation for a lot of other industries and sectors .........
It was interesting and well worth watching although I disagree with Sandy on a few things (not for the first time) there was something he called out that also struck me as odd, i.e. this is not a skewed winding (see the 15:05 point). Now it has been several years since I last did a PMG/PMM but back in the day I worked with both skewed and unskewed. Ordinarily in a PMG (or M) one skews to reduce cogging and ripple, but this is not skewed. Sandy mentions that in driving the S it feels smooth, and presumably is also audibly quiet (human ears are pretty good at picking up harmonics, and I've heard no complaints), and so they must be managing this some other way.
Q1. How are they managing it ? (presumably with a more complex inverter waveform management ?)
Q2. Given that the inverter is common to the 3, the Y, the S, and the Plaid S, (and ?? presumably the X, and in due course the Semi and the CT and the Roadster 2) then is there skew in the windings of the others ?
By the way : The Tesla parts bin is going to be a core foundation for a lot of other industries and sectors .........
Last edited:
Not in the windings... But yes the 3/Y rotor has skew in the magnets.
petit_bateau
Active Member
Thanks, interesting to hear the 3/Y are skewed. Yes, either way works as is obvious you know.
So ........ supplementary question. Given that they have obviously cracked it in the S Plaid to have non-skewed by using clever software, and thereby achieving higher torque, can you see any reason why they don't just load that software into the Y and the 3 and get an equivalent performance improvement. And of course move to non-skewed which makes for cheaper manufacturer in the 3/Y ?
It could be related to the new multi-part laminate design and carbon fiber wrapping. So the same software might not work on the less expensive traditional rotor.
petit_bateau
Active Member
Not really imho. Separate issues.
The carbon sleeving is for very high rpm without airgap reduction. So in principle one could have the fast acceleration (or decelleration) from the high torque (i.e. non-skewed, which has a higher peak torque than skewed) without needing to go to high rpm (high max speed) that in turn rquires the wrapping. (btw sleeving is an old trick, not new).
And/or reduced build cost (non-skewed). Again no wrapping required. And fuller commonality.
The tweaked laminate comment you are making (presumably the T 'pole' laminations) aren't actually very different than ones I've used in the past so I am struggling to see what is new about them. I guess I'd have to put them into a magnetic CAD package to see what (if anything) in new. Maybe I am missing something.
Bottom line : I think there might be genuine free lunch in retrofitting this into the 3/Y in terms of a combination of reduced build cost and/or improved performance (acceleration). The clever is (I think) in the inverter software. Software has zero marginal cost.
But am I missing something ?
No, the carbon fiber wrap holds the small laminate pieces in, which isn't necessary with a one piece laminate. So the wrap isn't just for high RPM. It also reduces the necessary clearance which could be related.
@petit_bateau So you have used multi-part laminates on motors you have designed? (Where there are small pieces on the outside with nothing to firmly attach them to the rest of the rotor.) Because Sandy Munro says he has never seen anyone do that before.
petit_bateau
Active Member
There's a lot that Sandy hasn't seen before. Doesn't mean it hasn't been done before even going back into the dim distant past with ferrites long before neos came in.
There are other ways to hold pole pieces and/or magnets in place - pins, glue, end clamps, assemblies, stainless bands ......... this is all actually very old tech that we are seeing here, albeit with modern materials and modern design tools and modern methods that were not previously possible.
(Edit: I've been in the design team on some, and have inherited others, and seen yet more. But I would not want to claim credit as being 'the' designer, just one of the team. System architect giving the PMG designer a brief, yes; PMG designer no; PMG design team yes.)
(Edit: Dredging my memory as to how it was done on one design. Was with aluminium alloy rivet-pins through the laminated poles pieces from the sides, into clamping end shells in aluminium, and the end shells also clamped the magnets into position. So the pole pieces floated, just as you see in this Tesla design. That was a design introduced in approx 1980 and was magnetically equivalent to a Lucas design that probably dated from 1950 as it was a module in the British motorcycle and (??) auto industry back then. I'm not sure if the original Lucas design had the floating poles as I never stripped one of those to see - there were very few coming through our hands on refurbs. So, not much is truly new, most things are incremental improvements. There were other designs...)
Last edited:
I watched this hoping for clarification on Plaid vs LR differences and where the extra plaid power comes from. The carbon sleeving is great but he never said that is what gives extra power just that it is for very high rpm. Seems like same inverter, same battery pack. The whys were missing in this. This is ok if his target audience was a bunch of engineers that could make the non stated conclusions. I doubt that this is the bulk of his YouTube viewers.
Is it just having two rear motors that can rev to 20k rpm? And of course the software to do it. The LR MS is no slouch but I'm not sure what I paid for. Hell he spent a significant amount of time praising the cross utilization of M3 parts which is a great idea but not a differentiator. I do love the car and it is the best I've ever owned. Just wish I understood where the extra power/cost comes from. I'm ok if it's just carbon sleeving and additional rear motor.
Maybe I missed it. Your thoughts?
petit_bateau
Active Member
Plus being non-skewed means that more of the magnetism gets turned into torque, so both greater acceleration and higher top speed. (with software doing the smoothing)
Plus some indication that the neos may have a stronger magnetic field than previously seen, though I think jury is still out on that. But if so that is also extra acceleration and extra top speed.
Plus (if I recall correctly) the car is overall lighter than it used to be. And it seems to be more rigid. And better behaved (handling) due to the improved suspension.
It is possible that the old S/X inverter was not as good as the newer 3/Y/Plaid inverters, and if so that would also help. But since I've never paid any serious attention to the old S inverter that is only a guess.
Basically lots of small refinements cumulatively add up.
Two new Tesla Patents (paste number into this search: US Patent Full-Text Database Number Search )
11,260,898 Steering system for a vehicle . This may be the rear steering setup for Cybertruck
11,260,809 Wiring system architecture. Flat ribbon type power and data distribution system. With bolt on/ snap on control modules and sub harnesses that connect to end devices.
11,260,898 Steering system for a vehicle . This may be the rear steering setup for Cybertruck
11,260,809 Wiring system architecture. Flat ribbon type power and data distribution system. With bolt on/ snap on control modules and sub harnesses that connect to end devices.
Thanks for this link, cool thread.
About this cable thing. Not only this is idea for the robots in removing repetitive motions in Manufacturing, but this has a hidden use case maybe (haven't read the patent filing).
Forgive any ignorance, but what if those layers were stacked like a capacitor, about 100 layers, and each carried a signal. Noisy unless you isolated with alternating ground layers and had everything between films. A connection is made by designing the film with wider passage holes like they do with vias on a PCB, so you can't really "strip" this wire, it's predesigned for a specific build. Takes ribbon cable to a whole new dimension. Each layer a predetermined gauge to get fancy with power and signal combined. (OK, maybe I'll smoke some more, I think it's working and I like guessing what it is...)
About this cable thing. Not only this is idea for the robots in removing repetitive motions in Manufacturing, but this has a hidden use case maybe (haven't read the patent filing).
Forgive any ignorance, but what if those layers were stacked like a capacitor, about 100 layers, and each carried a signal. Noisy unless you isolated with alternating ground layers and had everything between films. A connection is made by designing the film with wider passage holes like they do with vias on a PCB, so you can't really "strip" this wire, it's predesigned for a specific build. Takes ribbon cable to a whole new dimension. Each layer a predetermined gauge to get fancy with power and signal combined. (OK, maybe I'll smoke some more, I think it's working and I like guessing what it is...)
The communication lines are differential so low emission. Power return is usually via the chassis, so no opportunity to leverage capacitance there.Thanks for this link, cool thread.
About this cable thing. Not only this is idea for the robots in removing repetitive motions in Manufacturing, but this has a hidden use case maybe (haven't read the patent filing).
Forgive any ignorance, but what if those layers were stacked like a capacitor, about 100 layers, and each carried a signal. Noisy unless you isolated with alternating ground layers and had everything between films. A connection is made by designing the film with wider passage holes like they do with vias on a PCB, so you can't really "strip" this wire, it's predesigned for a specific build. Takes ribbon cable to a whole new dimension. Each layer a predetermined gauge to get fancy with power and signal combined. (OK, maybe I'll smoke some more, I think it's working and I like guessing what it is...)
View attachment 775332
This setup cuts mass and volume down versus individual circular cross section wires of standard size, each with a layer of insulation (needed to chassis, but redundant to wire pairs). Can trade width for thickness so even high current signals don't intrude greatly into the cabin.
Manufacturing costs and reliability should be improved also vs hand routed and inserted harness construction.
Bet TSLA
Active Member
This is funny. The best coating is apparently no coating. Who would've thought?
cleantechnica.com
The performance of the cathode-electrolyte interface made using this method was “comparable to the best interface resistances we have seen in the literature,” but those were all achieved using the extra step of applying coatings. “We are finding that you can avoid that additional fabrication step, which is typically expensive,” Yildiz said.
New, New Solid-State Battery News From MIT
A solid-state battery has many technical challenges. Researchers at MIT say they have solved one of those hurdles.
cleantechnica.com
The performance of the cathode-electrolyte interface made using this method was “comparable to the best interface resistances we have seen in the literature,” but those were all achieved using the extra step of applying coatings. “We are finding that you can avoid that additional fabrication step, which is typically expensive,” Yildiz said.
JRP3
Hyperactive Member
CATL may be making progress with Sodium ion
Article: CATL new patent allows anode-free sodium-ion battery density to go above 200Wh/Kg
Video:
Article: CATL new patent allows anode-free sodium-ion battery density to go above 200Wh/Kg
Video:
Every phone since decades has similar camera-chips & a dedicated IR-Filter in Software. Some Apps allow you to circumvent the IR-Filter & you can "see" IR-Sources bright (TV-Remotes, modern "face-recognition"-stuff, etc.). A significant step of that "image processing" that is dropped is exactly to drop the IR-Range and do white-balance and many other things.
Karpathy talked about this in some presentation, but i am unable to find it with a quick google. I am not saying that it is FLIR-Level detail. But it is there & can be seen.
I tinkered with that stuff in my free time a lot (The wii-remotes have a small IR-Camera in them & you can do fun stuff if you setup some LEDs around your monitor
But we can agree to disagree. This is fine.
Visible CCDs cut off at 700nm or so. IR remotes are Near Infrared (NIR) aka Short Wave (SWIR) of around 840 or 940nm (<1um), which is outside the filter, but intensity still lets them register on the sensor. One method of extending this ability is using an IR pass filter to block visible and then increasing exposure time. IR-Photo.net | Infrared Photography
Some day/ night cameras with IR illumination have a selectable physical IR cut filter to provide both color quality and night sensitivity, but that is still NIR.
Contrast this range to standard FLIR which is Long Wave Infrared (LWIR) at 8-15um wavelength. That's 10x NIR. They also use 3-5um (MWIR) still far from the standard cut off filter. What is the typical spectral response for certain camera/lens combination?
Thermal Imaging Overview
Similar threads
- Marketplace listing
