TESLA high performance tuning discusion

[Arlo1]

http://www.autoblog.com/2015/06/23/evtv-hacks-tesla-drivetrain-video/

This is cool but I'm more interested in using my own stuff as I'm aiming for bigger power. It looks like Tesla uses 7-8 TO264 packages in parallel. I'm aiming for more then that

 

[MrDude_1]

Oh I see what you are saying.. If you want to go that route you can also use what the guys at EVTV have done to crack the can on the Tesla...

I havent followed it since the first "we got it spinning!" moment.. did they ever figure out the CAN messages? or are they still stuck at only playing back the recorded stream into the CAN network?

ideally, I would want a solution that plugged into the Tesla unit from its external connections without ever cracking the sub-assembly open.... then it becomes more of a mechanical install than anything else. I have a welder, lets do this. lol.

 

[Nuts&Volts]

I would suspect Garlits would rank near the very top when it comes to funding !

Haha I can tell you first hand that this is not the case. The name does have draw, but cant say its brought much money in

..but really, just saying, I doubt $$$'s were a big part of his technical choices !
..Just a matter of definition...
Some race series may be cost controlled ( where the rules dictate budget/cost limits on items..EG, a "claimer" series),
But top level drag racing is not.
It may be "limited" by some internal team financial priorities, but that would not stop guys burning up a few sets of brushes or a motor , if they thought there was a record to be nailed ! :lol:

Your quite right that a set of brushes is small potatos, but most of the time your not damaging brushes. The commutator can be easily damaged, the springs always melt and the brush rigging melts down too.

I'd love to stuff two model S motors in a dragster. Just need some better controllers to push them to their limit.

 

[okashira]

Oh I see what you are saying.. If you want to go that route you can also use what the guys at EVTV have done to crack the can on the Tesla...

I havent followed it since the first "we got it spinning!" moment.. did they ever figure out the CAN messages? or are they still stuck at only playing back the recorded stream into the CAN network?

ideally, I would want a solution that plugged into the Tesla unit from its external connections without ever cracking the sub-assembly open.... then it becomes more of a mechanical install than anything else. I have a welder, lets do this. lol.

Yes, looks like they have . considering buying their setup to play with and sell with working kit (with MS inverter/motor)

 

[Hillhater]

Your quite right that a set of brushes is small potatos, but most of the time your not damaging brushes. The commutator can be easily damaged, the springs always melt and the brush rigging melts down too......

Compared to the ICE guys engine budgets, replacing/rebuilding a complete motor after every pass is small beer I suspect !

 

[jonescg]

Interesting how far you can push a motor, and what the potential power might be.

I mean, Voltron surprised me with its 167.5 kW peak-rated motor. We dynoed the bike at 172 kW at the rear wheel. And we weren't even running this thing at the full welly - if we could run a 720 V battery with no sag we'd be pushing nearly 190 kW.

Consider that the AFM240 exists, and a pair of these 80 kg beasts in your EV would deliver well over 700 kW, you have a very fast machine available for $80k in parts.

So I wouldn't be surprised if you could push similarly high values out of a Tesla.

 

[Hillhater]

Isn't this what Drayson set out to show , ..multi motor AC drive....but in a circuit race car layout rather than simple drag racer ?
http://www.draysonracingtechnologies.com/products.html

 

[liveforphysics]

Interesting how far you can push a motor, and what the potential power might be.

I mean, Voltron surprised me with its 167.5 kW peak-rated motor. We dynoed the bike at 172 kW at the rear wheel. And we weren't even running this thing at the full welly - if we could run a 720 V battery with no sag we'd be pushing nearly 190 kW.

Consider that the AFM240 exists, and a pair of these 80 kg beasts in your EV would deliver well over 700 kW, you have a very fast machine available for $80k in parts.

So I wouldn't be surprised if you could push similarly high values out of a Tesla.

Maybe 95% of dynos sadly don't enable the user to disable the barometric pressure and temp compensation factor. This is why magizines dyno'ing a Zero typically show substantially higher output than the bikes spec sheet says it makes (which is the correct hp).

Just saw it happen recently on a friend dynoing his 2015 DS, dyno claimed 61hp at the rear tire, the true power the bike makes is 54hp (measured from a dyno that had the dyno mfg send a tech out to disable the compensation that doesn't apply go EVs).

 

[okashira]

Tesla has published their actual hp numbers. I was only about 8% off or less in my estimates from 6 months ago.
The numbers are even lower then I estimated.

532 hp (Ludicrous)
463 hp (Insane)

We know L is 1500A or 20.2A per cell

NCR18650GA would sag to about 3.34V per cell under a quick load at near full charge when loaded to 20A.
If hot, it might sag to 3.45V.
That's 67.47 W per cell or 479.3kW cell power in the car.
532hp is 396.7 kW.

So at full load the rest of the system efficiency is 396.7/479.3 = 82.8% This includes losses from: all wiring, fuse wires, connections, inverter, motor.
That is really good considering how small of an induction motor it is for the power.

Note that the battery cell losses amount to an additional 115kW of losses. (4.15V - 3.34V) * 20.2A * 7104

That brings total efficiency at peak load to 396.7/(479.3+115) = 66.8%

 

[riba2233]

Really interesting info, thanks for sharing!

 

[liveforphysics]

Tesla has published their actual hp numbers. I was only about 8% off or less in my estimates from 6 months ago.
The numbers are even lower then I estimated.

532 hp (Ludicrous)
463 hp (Insane)

We know L is 1500A or 20.2A per cell

NCR18650GA would sag to about 3.34V per cell under a quick load at near full charge when loaded to 20A.
If hot, it might sag to 3.45V.
That's 67.47 W per cell or 479.3kW cell power in the car.
532hp is 396.7 kW.

So at full load the rest of the system efficiency is 396.7/479.3 = 82.8% This includes losses from: all wiring, fuse wires, connections, inverter, motor.
That is really good considering how small of an induction motor it is for the power.

Note that the battery cell losses amount to an additional 115kW of losses. (4.15V - 3.34V) * 20.2A * 7104

That brings total efficiency at peak load to 396.7/(479.3+115) = 66.8%

The P90D I drove a couple days ago had a special setting for pre-heating the battery for maximum discharge performance. Those power numbers align pretty well with the physics of getting that mass of car to that speed in that amount of time.

 

[flat tire]

So, has anyone actually tuned or tried tuning a Tesla? I don't mean removing the drivetrain or installing a lighter battery pack either. Aside from heat management, which seems straightforward to improve with larger radiator / more flow volume, the car's stock performance is entirely limited by software. I take it nobody who's trying to figure out the controller SW has yet posted their efforts online? I can't imagine NOBODY is trying to hack these things for increased performance. I have experience in tuning / breaking ICE ECUs and the lengths people go to to break in / reverse engineer the engine control software are pretty remarkable. Someone with the required skillset could certainly get a leg up on the Tesla if there was interest, and I can't imagine there isn't interest.

 

[Punx0r]

AFAIK performance is already limited by the main pack fuse. I guess you could defeat that and up the current limit somehow, but pack failure may well soon follow.

I think these cars will have to get a lot cheaper, more prevalent and out of warranty (or nearly) before anyone seriously tries messing with them. They're currently a relatively rare, expensive premium car already in a high state of tune. How many people mod their new Ferrari?

 

[flat tire]

AFAIK performance is already limited by the main pack fuse. I guess you could defeat that and up the current limit somehow, but pack failure may well soon follow.

I think these cars will have to get a lot cheaper, more prevalent and out of warranty (or nearly) before anyone seriously tries messing with them. They're currently a relatively rare, expensive premium car already in a high state of tune. How many people mod their new Ferrari?

Actually, plenty of people mod their new Ferraris. Furthermore, I doubt maximum power is so close to the fuse threshold / pack capacity that you couldn't increase it a bit, but the main issue is continuous output on a track. The cars as configured from the factory could stand a huge improvement in that regard, which will likely require software mods as well as increased cooling.

 

[liveforphysics]

I've got not doubt there are tons of owners who would like them hacked.

Allow me to extend a public offer. Give me a p90d as a development car to tinker with, and after I've cracked how to hack it (likely with a hardware based current sensor spoof), and I will happily share how it's done.

Many folks spend more than a p90d costs on fuel for there boat for a single day of cruising around.

 

[amberwolf]

Many folks spend more than a p90d costs on fuel for there boat for a single day of cruising around.

I think you must mean "Many folks that own small ships or really big boats", cuz I can't imagine that simply "Many folks" even *have* boats with fuel-powered engines or motors on them, much less ones that use that much fuel.

 

[Hillhater]

There are already several folks working to hack the Tesla software...and hardware...with some success.
But this is a little more complex than an ECU, and it has actually taken a lot of time and resource just to be able to control the motor and inverter without the Tesla inputs.

 

[Punx0r]

Furthermore, I doubt maximum power is so close to the fuse threshold / pack capacity that you couldn't increase it a bit, but the main issue is continuous output on a track. The cars as configured from the factory could stand a huge improvement in that regard, which will likely require software mods as well as increased cooling.

Sure, you could run closer to the fuse threshold at an increased risk of nuisance tripping. Tesla recently replaced the fuse with an electronic circuit breaker with increased sensitivity which narrowed the gap between normal use current and fault current, allowing them to run closer to the pack limit, which resulted in increased wheel-HP and measured acceleration. I'd say they're running as close as they dare and the only scope for improvement is the traditional mass-production Vs. tuner acceptable risk.

Continuous power on Model S seems to be a big problem which I doubt can be meaningfully improved via software.

 

[okashira]

Tesla is already pushing their motor to more power in kW/kg out of any EV.. And it's not because their motor or inverter technology is any better then anyone else's

 

[Chalo]

Many folks spend more than a p90d costs on fuel for there boat for a single day of cruising around.

That depends on what you mean by "many", doesn't it?

I like to divide available resources by the number of people who use those resources to determine whether certain activities make sense.

Personal cars (any of them) only make sense if you disregard those people yet to enter the world. We're basically spending their birthright on ourselves when we drive cars. We can recycle the steel, but we can't recover the petroleum that is gone, nor the clean air and water and land that won't be back until after we are all gone.

The Sultan of Brunei practicing touch-and-gos in his personal Boeing 747 just for kicks, because he can, doesn't make sense in any rational analysis of the world we live in, in any time scale. It's like performance art to demonstrate in absurd terms just how little it matters to you that others are suffering and dying and laboring their lives away for lack of the resources you just wasted for no reason, because nobody was there to stop you.

 

[liveforphysics]

There are already several folks working to hack the Tesla software...and hardware...with some success.
But this is a little more complex than an ECU, and it has actually taken a lot of time and resource just to be able to control the motor and inverter without the Tesla inputs.

If the software has too advanced of fault protection, this is no problem. I will happily dremel the trace to the logic level gate driver input signals to the IGBT stage, as well as tapping the wires to the main pack contactor and make my own precharge circuit. Them feed those signals from a lebowski controller setup for induction or whatever works best, so many options.

If the pack turns out to be completely tapped for maximum practical power, I am capable and willing to sacrifice a little of its massive cargo area and add a few hundred pounds to an already heavy car to add an adequate amount of high-c-rate cells to get the power levels up to fully magnetically saturated on the front and rear motors iron for the whole quarter mile.

Im pretty confident its possible to drop one into the low 9s and maybe 8s by adding enough of the right cells in a supplemental pack to parallel directly with the main pack (a trivial modification as far as hotrodding goes).

Chalo- I agree with you. However, super hotrod Teslas shift perceptions of EVs for millions if not billions of people to lust for EVs over poison sprayers. This is beneficial for all living things. I do agree with you that bicycles are more sustainable than cars, but I disagree that just saying it to the world is a more effective use of time than making bad ass EVs that show the new era of supercar performance that leaves the world forgetting gas vehicles used to also try at it.

ATB,
-Luke

 

[drebikes]

Tesla is already pushing their motor to more power in kW/kg out of any EV.. And it's not because their motor or inverter technology is any better then anyone else's

I reckon it must be related to the thermal limits of their system. Here's a few cents bearing in mind I design IPM motors, I'm no more than an interested observer on all other aspects:
* in my EV designs the electronics must be sized for peak power, not average power because the Si dies get to max temp right after the max current passes trough them (Cth=0); electronics have no thermal inertia
* Tesla have an Induction motor. My own PMSM is 14kW/kg and its biggest issue is rotor temperature because you can demagnetise (200°C limit more or less for Samarium). Tesla doesn't have this problem because they have a squirrel cage induction motor - I have seen these at 350°C rotor temp, while an IPM on Neodymium like the Prius can't really go over 140°C.
*** ==> on the other hand one can't keep the rotor at 300°C for a long time as it will radiate heat to the stator where the usual limit is around 180°C (insulation usually); as such I can see the Tesla going nuts briefly, after a while you'd do some power derating as the rotor heats up.

Another limit I know of in motorsport batteries is inner-cell temp for batteries ==> again you can go nuts briefly, but heat dissipates from a cell only thru the tabs and to the water sleeve, you rarely find fully liquid-droned batteries due to weight. In this short period the inner cell temp rises from maybe a normal 60°C to 80-90°C (I know of A123 heavily power-modded LiFe); it's only the thermal capacitance of the cell that allows the burst as thermal dissipation is limited through something as small as the tabs.

So my 2 cents would be that we could tune the Tesla easily if:
*the cells aren't at the temp limit with a simple bypass of the DC current sensor and/or fuze
*or if we'd cool the rotor; that's tougher because bigger rads to cool water more would mainly cool the stator, the rotor would heat up just as fast. I only know of 2 ways to cool a rotor: lubricated bearings and forced air cooling in the airgap. Forced air should be easier - during a a drag race one could add some liquid nitrogen in the air mixture to help with the cooling too

 

[jmz]

Tesla is already pushing their motor to more power in kW/kg out of any EV.. And it's not because their motor or inverter technology is any better then anyone else's

I reckon it must be related to the thermal limits of their system. Here's a few cents bearing in mind I design IPM motors...

Ooh, a real live one! I'm not a motor guy, but I am a power electronics engineer, and over the years I've been slowly trying to increase my motor knowledge (particularly PMSM and SRM/hybrids).

Is anyone building non-hub outrunner PMSMs for traction applications? Seems like it might be easier to cool both the rotor and stator.

 

[flat tire]

The Sultan of Brunei practicing touch-and-gos in his personal Boeing 747 just for kicks, because he can, doesn't make sense in any rational analysis of the world we live in, in any time scale. It's like performance art to demonstrate in absurd terms just how little it matters to you that others are suffering and dying and laboring their lives away for lack of the resources you just wasted for no reason, because nobody was there to stop you.

Oh, come on, like you wouldn't also do touch-and-gos for kicks in your own 747 if you were a billionaire.

I only know of 2 ways to cool a rotor: lubricated bearings and forced air cooling in the airgap. Forced air should be easier - during a a drag race one could add some liquid nitrogen in the air mixture to help with the cooling too

Looking at Model S efficiency curves, there is a SHITLOAD of heat being generated in the motor under WOT (speaking of, is there an EV term equivalent to this?). You really would need some kind of active cooling like LN2 or water for track use to be practical with these motors and maximum torque output. Since the cooling system will be depletive (unless someone can engineer a phase change system to cool the motors, could be doable with ICE engine to drive the phase change, better battery technology, and and probably some coolant loss thru the bearings. Another potential phase change solution, probably still quite impractical to engineer, would be the same setup but with water as the refrigerant, to be condensed and recycled as it boils off, keeping the motor steadily at 100C. This would have a weight and energy penalty, but be really sick in action and allow the car to driven full-out on a track with the battery and controller as the bottlenecks.

More powerful motors would help too, so they can give the ludicrous torque we want but stay in the sweet spot of the efficiency curve.

 

[okashira]

Tesla is already pushing their motor to more power in kW/kg out of any EV.. And it's not because their motor or inverter technology is any better then anyone else's

I reckon it must be related to the thermal limits of their system. Here's a few cents bearing in mind I design IPM motors, I'm no more than an interested observer on all other aspects:
* in my EV designs the electronics must be sized for peak power, not average power because the Si dies get to max temp right after the max current passes trough them (Cth=0); electronics have no thermal inertia
* Tesla have an Induction motor. My own PMSM is 14kW/kg and its biggest issue is rotor temperature because you can demagnetise (200°C limit more or less for Samarium). Tesla doesn't have this problem because they have a squirrel cage induction motor - I have seen these at 350°C rotor temp, while an IPM on Neodymium like the Prius can't really go over 140°C.
*** ==> on the other hand one can't keep the rotor at 300°C for a long time as it will radiate heat to the stator where the usual limit is around 180°C (insulation usually); as such I can see the Tesla going nuts briefly, after a while you'd do some power derating as the rotor heats up.

Another limit I know of in motorsport batteries is inner-cell temp for batteries ==> again you can go nuts briefly, but heat dissipates from a cell only thru the tabs and to the water sleeve, you rarely find fully liquid-droned batteries due to weight. In this short period the inner cell temp rises from maybe a normal 60°C to 80-90°C (I know of A123 heavily power-modded LiFe); it's only the thermal capacitance of the cell that allows the burst as thermal dissipation is limited through something as small as the tabs.

So my 2 cents would be that we could tune the Tesla easily if:
*the cells aren't at the temp limit with a simple bypass of the DC current sensor and/or fuze
*or if we'd cool the rotor; that's tougher because bigger rads to cool water more would mainly cool the stator, the rotor would heat up just as fast. I only know of 2 ways to cool a rotor: lubricated bearings and forced air cooling in the airgap. Forced air should be easier - during a a drag race one could add some liquid nitrogen in the air mixture to help with the cooling too

Actually I was thinking that Tesla was accepting much less safety factor then other EV's. They push their motor harder because of all the testing they have done.
But you are exactly right and make an important point, the induction motor can handle much higher temperatures in the rotor with no magnets to kill.
The bearings are made by SKF are ceramic, good enough for oil & gas applications