DIY Tesla Drive Unit Repair

[howardc64]

Has any consideration been given for the temperature differential to the squirrel cage style rotor? I believe the rotor is all metal, so it can take extra heat, however you will lose some performance once you pass 100 Celsius due to I2R effects. I suspect these features will be most pronounced during extended freeway drives. I can get mine running I will work on data for performance versus temp effects on the rotor

google "induction motor broken rotor bar" and "induction motor laminate failure" Lots of studies illustrate temperature effect on rotor thermal stress. Rotor is the most interior and naturally the hottest and hardest part of an induction motor (IM) to cool.

I do read service life of (IM) goes down drastically with temperature increase (some say 10x shorter service life with just 50C increase). Furthermore, highest current flow is starting the motor thus industrial AC IM applications advise to keep it running all the time. Usage in a car is constant start/stop/regen condition if not highway. The power meter on the instrument panel provide a pretty clear visualization how much current is going through the IM rotor. Even gentlest acceleration and regen hits like 50-60kW while highway cruising is like < 20kW. So just normal start stop driving is likely the most thermally stressful condition.

It could very well be a 20+ year life cycle got reduced to even 2-4 years by coolant delete which is still far better than Tesla's reman LDU leakage failure rate.

It'd be wise to start researching robust cooled rotor solutions as well as rotor sourcing if they do start failing in a couple of years time. I've been in touch with leading edge coolant seal + water pump design companies to get them interested. Difficult to create momentum without rotor failure appearing. If rotor failures appear, sourcing that will be very challenging and expensive.

 

[philip295]

Does the regen have a big heating effect on the rotor? Is it worth setting it to low after a coolant delete?

 

[howardc64]

Does the regen have a big heating effect on the rotor? Is it worth setting it to low after a coolant delete?

I don’t know but I’d assume so. Magnetic field is generated by current flow in the IM’s rotor in drive and regen. Google induction generator for info. But it’s beyond me.

Don’t know how much it helps to reduce heat with less regen. Judging by the instrument panel, regen current is not insignificant. It’s an interesting thought. Maybe better to just bleed the heat at the friction brakes.

 

[geordi]

Regen seems to max out at 50 kW, while drive power can be up to 6 times that. I don't think it will contribute much to heating based on the time-of-use and the lower power level... Unless you are regularly coasting downhill from Donner Summit - 47 miles all downhill at a significant grade.

 

[brainhouston]

I'm installing a LDU rotor temperature monitor (wiring connectors on order, raining today, etc. delays), but I'll only be getting "after" data (after "coolant delete"), not "before".
...
Pictures here show a blued rotor on an LDU having had "coolant delete" applied, implying significant heating:
..
I expect we'll hear tales shortly about exploded rotors after "coolant delete". In my RAV4 EV, with rare towing and a lighter load than a MS/X, I hope to avoid that. On Tesla Factory Reman units, I assume they push some firmware update that limits total 100% power output algorithmically. Or, perhaps, they alter the stator temp sensor to monitor stator cavity temp. It'll be interesting to see one of the Rev U LDUs torn down and see if there are other hardware mods than just that new coolant manifold casting.

seeing how most heat is concentrated in the middle of the rotor, measuring temp on the side might not be that useful...
but even with rotor cooling intact, it only runs the coolant in the center and not the surface, so i wouldn't expect much heat taken away from surface of the rotor
this might be some corner case rotor... need more data...

Regen seems to max out at 50 kW, ..

On my X, it gets up to 70kW, but yeah not that much compared to drive power

 

[asavage]

seeing how most heat is concentrated in the middle of the rotor, measuring temp on the side might not be that useful...

If you've got an idea on how to better measure LDU rotor temperature, I'm all ears.

 

[brainhouston]

If you've got an idea on how to better measure LDU rotor temperature, I'm all ears.

i don't, just saying we might not catch the data from hottest part of rotor...

 

[asavage]

I could place a probe on the end of a rod that protrudes into the center of the hollow rotor shaft -- in place of the tapered alum. tube I'd cut off. Maybe 6" long, so it's centered along the axis.

But it would be a lot more work than where I placed the IR sensor, and honestly I think that the rotor's temperature is going to be pretty homogenous after a few minutes. It'll be hotter in some places, but I'm not looking for a thermal analysis of hot spots, only looking for some base data.

If I was engineering to prevent failures, I'd approach this very differently.

 

[howardc64]

Did some additional research on IM thermal analysis. This is definitely a lost art. Not the common digital computer HW/SW design knowledge in US these days.

Expert Knowledge Source and Book

Found a couple of authors of a key text book for motor designs through their online articles. US prof is the CUSP dept in Univ of Minnesota (I've reached out). Scottish prof is Univ Glasgow. Both are running a dept with commercial industry involvement. Here is the book if seeking punishment

https://www.amazon.com/Design-Studies-Electric-Machines-Miller-ebook/dp/B0B82XXQQ

Univ of Minn Prof

Here is his lecture on electrical motor thermals

One particular interesting section is his discussion in insulation service life that have long been accepted at 50% reduction per 10C increase. My understanding is the insulation is mostly the around the stator windings. I don't know if the laminates between rotor disk plates suffers this similar degradation. If it is only the stator, LDU stator is probably fine with continued stator liquid cooling.

Another interesting section is the discussion of poor thermal conductivity of insulators around the stator windings. Heat is actually generated in the wires and need to get them out. Sounds like they change the the insulator formulation and even encase the whole thing in ceramic in attempt to increase thermal conductivity. Again, LDU likely don't have to worry about the stator. Not much has changed with rotor coolant delete.

The standing non convection air between the rotor and stator in LDU is a significant insulator due to air being a poor thermal conductor. In a finned air cooled design, there is high volume of cooled air flowing which compensates for the poor thermal conductivity.

Here is the whole course in video and power point if seeking slightly less punishment than reading the book ( link )

Univ of Glasgow Prof

He regularly pens some articles. Found this one showing basic thermal density of an IM. Note the highest thermal is concentrated on the end rings @ 150C+ (figure 6, do note the whole scale from cold to hot is < 1C in this figure) This is where we can find articles studying welding junction failures between the bar and ring (broken rotor bar)

[JAC228] Induction Motor Thermal Analysis | Simulation Technology for Electromechanical Design : JMAG

In this example, evaluating induction motor temperature distribution.

www.jmag-international.com

Empirical Data

Best empirical data for rotor coolant delete is the Tesla Roadster. Basically the same LDU design but 1/3 power and pushing a car weighing 1/2 as much. Roadster rotor is neither air or liquid cooled and seems fine for long term. Searched but couldn't find any used Roadster rotors pulled for thermal discoloring inspection. Would be interesting to see one but really only a couple of Roadster shops exist and don't hear them messing with the DU.

 

[mr_hyde]

Kudos for installing the temp sensor but it would have been REALLY helpful to do some before/after comparisons. When your readings report XXX*c we still won't know what that means. Did it go up XX*c or did it go up 50%? I'm personally not willing to do the extra work to R&R the motor an extra time. Maybe it is easier to do in the Toyota or Mercedes application and someone on that side of the fence is curious?

Also, a consideration that isn't discussed much is the relative temperature of the coolant reaching the gearbox. In the original configuration, very hot coolant is likely heating the gearbox oil or at least pre-heating it from the cold start and then having a marginal thermal delta once everythings has been moving for a while. After the delete, the majority of coolant will be cool - the coolant loop is generally closer to ambient where the battery is happiest. It will still get the hot coolant from the stater and inverter, but that hot coolant enters and exits near the bottom of the heat exchanger with the majority of the coolant spending time in the heat exchanger being from the flyover tube. I wonder if that will have any impact on anything related to the gearbox?

 

[asavage]

Kudos for installing the temp sensor but it would have been REALLY helpful to do some before/after comparisons. When your readings report XXX*c we still won't know what that means. Did it go up XX*c or did it go up 50%?

Agreed, and that's what I said three weeks ago. But, we don't have that pre-mod baseline info. Somebody will have to make the sacrifice and get it. I'm not going to reverse my coolant delete to obtain it; neither are you, so the numbers I obtain will be data without a baseline reference. It's still data, it's still usefule, and it's still $400 (with tools) and 12 hours I've invested so far to get it.

 

[philip295]

Can we assume that Tesla will have all this data? They must have tested before and after their revision U.

 

[asavage]

Can we assume that Tesla will have all this data? They must have tested before and after their revision U.

Sure, we can assume that, but it makes no never-mind to us: Tesla doesn't sell parts, doesn't share info, has no PR department, fired (then later rehired some of) their SuperCharger development team . . . what Tesla has or does with that knowledge is irrelevant to we customers; we'll never see it.

Additionally, any rotor temperature data gathered will be specific to whatever temperature measurement method is used, so unless someone duplicates my sensor placement and uses an MLX90614 IR sensor on their own install, any "before" temp data couldn't be used as a reference/baseline anyway. IOW, it's unlikely that data I gather could be used to accurately compare to anybody else's "before" data, or even "after" data . . . unless they duplicate my design choices, which again isn't likely.

But . . . I figure that it really is "better than nothing".

 

[mr_hyde]

Can we assume that Tesla will have all this data? They must have tested before and after their revision U.

They do and we will never see it. All we know for sure is they believe this will fail 'less bad' than the leaky seals. They just need to get them out of warranty so it is our problem and not theirs. If you put on your tinfoil hat, you see the distinct possibility they know this will melt all the rotors in 5-10 years which is a great way to get a bunch of old cars off the road. NLA the LDU and sue the pants off anyone that tries to make replacement rotors which we might all need in a few years. I hope I'm wrong, but would anyone put it past them?

 

[brainhouston]

.. After the delete, the majority of coolant will be cool - the coolant loop is generally closer to ambient where the battery is happiest. ...

not sure i fully understand this statement..
from my long trips experience, batteries heat up to 30-35C (range mode on, so just from DUs heat)
pretty sure Tesla does it for faster supercharging

here's a pic after an hour drive at the start of SC:

 

[howardc64]

not sure i fully understand this statement..
from my long trips experience, batteries heat up to 30-35C (range mode on, so just from DUs heat)
pretty sure Tesla does it for faster supercharging

here's a pic after an hour drive at the start of SC:
View attachment 1049818

Have seen this menu previously (Tesla activated diag mode or I think rooted cars only) Was always curious about the stator temp of ~170C (prior link) The other temps are much cooler (inlet to the motor, 2 on the inverter) 2 on inverter are in coolant flow passage and inlet is surely coolant somewhere before the inlet tube. 120C on inverter might be near the IGBTs (high power transistors) on the phase board itself.

This stator sensor is buried in the varnish at the gearbox end of the stator windings so not sure it's measuring coolant or embedded near the windings itself. This sensor is the difficult wire to reassemble the 2 halves of the gearbox and need to pull it through tiny gaps with a catchy rubbery protective outer tubing. I used a sleeve to make it much easier to reassemble ( link )

Came across this stator sensor manufacturer ( link ). There seems to be 3 different potential installations. Not sure what LDU uses (4 wires) Any rebuilder that has disassembled the stator windings probably have a better idea how its mounted.

Anyway, my take is the 165C and 120C from above are not measuring coolant. Rather near components that has a lot of current flow.

Regarding the coolant temp, looks like entry is slightly above battery temp (probably heated slightly through the charger) and exit is close to 80C. My inverter disassembled pic ( link ) shows 2 coolant passage temp sensors. One from the stator cooling channel exit (top of pic), the other inverter exit to the gearbox heat exchanger (bottom of pic)

 

[brainhouston]

Have seen this menu previously (Tesla activated diag mode or I think rooted cars only)

yes its enabled with root
every car has it but the pass it dynamic so only Tesla has it
with root, u lock out tesla so pass becomes static

Anyway, my take is the 165C and 120C from above are not measuring coolant. Rather near components that has a lot of current flow.

the way i interpret this (also matches smt) is red number are limits
the actual measured temp is the white numbers on the left

not super positive if its coolant or components temps but..
when my MS coolant wasn't flushed and car was going in limp due to front heat sink temp, it was about 170F (76C)
which does seem to match the 78C max for inverter, although inverter has its own line...

 

[Decibelle]

Sorry if this needs to be in a different thread. I have been following the LDU conversation for a long time and always hoped my drive unit in my 2015 P85D was immune from the coolant ingress. I have on several occasions taken out and inspected the speed sensor and the first time, 2 years ago, I took it out it had 2-3 drops of something looking like grease. I have since then removed the sensor every 6 months and last couple of times I noticed something that looked more like a liquid. Maybe a drop or two. My issue now is that I’m starting to hear the high pitched sound when putting light pressure on the “gas” pedal. The noise is not there during coasting or generative driving. It sounds more electric that bearings, kind of like a mig welder. According to a former Tesla technician, it definitely are the bearings. Am I correct to assume that coolant has entered and is now causing the bearings to start going bad? Or are these two different issues? I assume a coolant delete will not help in my case as I’m guessing the bearing noise will only get worse over time. Picture attached of last time I checked the sensor with 6 months worth of buildup

 

[howardc64]

Sorry if this needs to be in a different thread. I have been following the LDU conversation for a long time and always hoped my drive unit in my 2015 P85D was immune from the coolant ingress. I have on several occasions taken out and inspected the speed sensor and the first time, 2 years ago, I took it out it had 2-3 drops of something looking like grease. I have since then removed the sensor every 6 months and last couple of times I noticed something that looked more like a liquid. Maybe a drop or two. My issue now is that I’m starting to hear the high pitched sound when putting light pressure on the “gas” pedal. The noise is not there during coasting or generative driving. It sounds more electric that bearings, kind of like a mig welder. According to a former Tesla technician, it definitely are the bearings. Am I correct to assume that coolant has entered and is now causing the bearings to start going bad? Or are these two different issues? I assume a coolant delete will not help in my case as I’m guessing the bearing noise will only get worse over time. Picture attached of last time I checked the sensor with 6 months worth of buildup View attachment 1049855

I think Tesla LDU factory/reman packs seal lips+shaft with grease on install (counter what every PTFE seal install manual says... must be pristine clean) Thus every 1st pull of speed sensor from Tesla have grease on it. After cleaning it off, subsequence pulls with no longer have grease and only shows coolant if you have a leak.

@mr_hyde also has "not too many drops on speed sensor on 6+mo 10k miles ago pull" but hearing high pitch whines on acceleration now (mig welder sound is a good description). He has original factory LDU from 2014. I can hear the high pitch from 2 houses away when he comes by. We'll probably pull the speed sensor and do the coolant delete if present. Curious which bearing is sounding so bad.

 

[howardc64]

the way i interpret this (also matches smt) is red number are limits
the actual measured temp is the white numbers on the left

not super positive if its coolant or components temps but..
when my MS coolant wasn't flushed and car was going in limp due to front heat sink temp, it was about 170F (76C)
which does seem to match the 78C max for inverter, although inverter has its own line...

View attachment 1049841

Interesting thought on red # can be limits. I see the prior post + follow up post with @vanR on what the temps sources might be ( link ) Interesting in that thread there are MCU coolant pics with red #s in white and some white # don't have temps.