DIY Tesla Drive Unit Repair

[asavage]

More:

A ReVolt representative has said:
We do whatever is needed. You can send the stator half only or the whole car. The more you send, the more we can check and fix, but obviously a stator only repair will be cheaper.

So, if a DIY-er with a low-iso stator issue wants to, they can ship the stator half of the LDU to Arizona for possible repair.

 

[brainhouston]

maybe also try a diff meter for sanity check
seems weird to get such low resistance due to coolant only...

 

[asavage]

maybe also try a diff meter for sanity check
seems weird to get such low resistance due to coolant only...

Well, the stators do short.

 

[Hank42]

You don't have to split the gearcase to pull the inverter stack, or the rotor. It's only when you really need to get to gearbox bearings, or you have a rotor that has to be pressed out from the pinion gear side, that you have to open the gearcase.

Exception: if you're going to have your stator soaked in magic dissolver or rewound or something like that. Yeah, it's got to be completely disassembled for those.

I was able to bend a 90° on the end of some thin plastic strips I cut from deli trays or something, and carefully slip them behind the stator wires when the bolts were removed, isolating the stator from the inverter's IGBTs. But I didn't need to, I just wanted an A-B comparison of the iso test results, which were minimal. You might try that before pulling the inverter stack, as Howard notes the problems with the fixing bolts he had (too short, not enough thread engagement, threads pulled after reman; o-ring seating difficult to verify/easy to damage). If you can test the IGBT outputs while isolated from the stator and without pulling the inverter stack . . . might glean valuable info to inform your next move, without diving off the deep end immediately.

Thanks again, was able to follow your suggestion and managed to test the resistance. Inverter Checks out fine, but the stator - oh boy is that shot. Pulled it apart and it was flooded

 

[blodg1]

Well, the stators do short.

Yep. Currently have three with dead shorts.

 

[philip295]

I fitted my coolant delete manifold at the weekend. The speed sensor had one drop of fluid on it and it felt slightly damp inside. I didn't disconnect the axles: I pivoted it on the front mounting after unbolting the rear and side mountings. It's rebuilt and seems to drive normally .

 

[Hank42]

I fitted my coolant delete manifold at the weekend. The speed sensor had one drop of fluid on it and it felt slightly damp inside. I didn't disconnect the axles: I pivoted it on the front mounting after unbolting the rear and side mountings. It's rebuilt and seems to drive normally .



View attachment 1048921View attachment 1048922

Good luck! I dropped the whole motor before performing the coolant delete and pulled it apart and a surprising amount of brown mucky liquid drained out.

 

[blodg1]

I fitted my coolant delete manifold at the weekend. The speed sensor had one drop of fluid on it and it felt slightly damp inside. I didn't disconnect the axles: I pivoted it on the front mounting after unbolting the rear and side mountings. It's rebuilt and seems to drive normally .



View attachment 1048921

Nice work! Just based on this long distance shot, the rotor end looks very clean. What revision letter is that unit?

 

[asavage]

Just based on this long distance shot, the rotor end looks very clean.

Mine (with low iso) had a spotless seal surface on the rotor shaft, but low enough iso to cause the BMS to open the contactors once, and a bunch of muck/rust on the stator's exposed laminations.

Best Practice is If coolant is showing on the speed sensor, it should be disassembled; at least the rotor cap removed to inspect the exposed bit of stator, and remove the rotor for stator cleanup, so the corrosion (damage) doesn't continue.

Mine had no coolant beyond 2/3 of the stator's length, so I pulled the rotor, tilted the LDU with the open end down, and cleaned away.

 

[geordi]

Well I'm officially scared off of ever attempting this in my own home shop.....

 

[asavage]

If you can get the subframe out from under the MS/X -- floor jacks and 4x4s -- have or can build a suitable workbench for the 300 lb. unit, and can buy/borrow a std. engine "cherry picker" (or do as I did, and use a come-along to a garage joist above), it isn't terrible. Yeah, I bought an expensive and versatile lift to both R&R the LDU in my RAV4 EV, and double as a workbench, but that's not necessary, and I actually bought it used in order to help drop the battery pack one day in future -- it'll do 2,200 lbs, and stores more compactly than you'd think. I'm working out of a very well packed 2-car garage.

The main problems I had were having to build three separate puller setups to remove the reluctor wheel, the rotor end cap, and the rotor (which was unusually difficult to remove in my case, but isn't always). I don't consider beating on the parts to be an acceptable way to remove things, but when needs must . . . you can.

I should set up an LDU tool loaner program.

Yes, it's work. But it can be done.

 

[philip295]

Nice work! Just based on this long distance shot, the rotor end looks very clean. What revision letter is that unit?

View attachment 1048975

It's a revision Q. It has a new motor, not remanufactured , at 43,000 miles with the first owner . I've done 56,000 and this was the first time I found any moisture. The speed sensor was dry 6 months ago. So I thought it was worth risking not dismantling any further.
I hope this was the right thing to do: I wiped lathe slideway oil in the shaft and on the reluctor ring to prevent rusting.

 

[Gtech]

IDK about in the Model S, but the LDU in the RAV4 EV has access to the three HV stator bolts without dropping the LDU. IOW, the orange plastic cap can be removed and an insulation resistance tester (applies 500V to the stator windings; this is not a function of a standard DMM -- I have a Fluke 1507) can be used to check stator isolation resistance without removing the LDU.

That cap is also a coolant puddling place for coolant making its way from the rotor to the inverter, so if coolant is seen there . . . it's probably in the inverter as well.

I assume such a test will become a standard part of a pre-purchase inspection for old Model S/X in future.

View attachment 1045803View attachment 1045802

Thank you for all your valuable information, think that pulling speed sensor + orange plastic cap from stator windings to check moisture but also checking isolation resistance up front gives a good idea if it's worth doing the coolant mod yes or no.

1. Maybe I overlooked this topic but do you know from your experience what a normal resistance for a healthy LDU is with an stator that is not soaked with coolant? TSM is talking about a value of > 3.5 MOhm but that is minimum not 'normal' because I think when you measure 3.6MOhm it's 'good' according to TSM but I think not for long.

2. If you measure B- or B+ to GND on the outside of LDU does the measuring voltage of insulation resistance tester also reach the stator windings or do you suggest measuring U V W to GND also? If so, what resistance do you think is healthy?

Thanks.

 

[asavage]

I don't have the breadth of experience with many LDU stators to give good advice on what is "good" HV isolation resistance. I'm seeing numbers quoted from 4Mohm to 120Mohm, which is quite a large range. I think that some numbers are coming from repair facilities that have a strong incentive to disqualify a stator that isn't comfortably above a minimum. Sorry in advance: lots of words follow . . .

Backing out a bit and looking at the entire HV system, the Tesla Isolation Faults guide says (pg. 6) that during various test procedures using the diagnostic menu or TDS (which utilizes the BMS for measurements, and therefore only measures at the battery side of the contactors), HV contactors open, that part of the system must remain above 1.8Mohm. Put another way, with the contactors open, that's measuring the isolation of the battery only.

[My understanding is that on the RAV4 EV (which is basically a Model S) the BMS will refuse to close contactors if it sees all isolation <1Mohm; if the contactors are allowed to close, and the isolation resistance drops under 1Mohm, the BMS opens the contactors and faults.]

On pg. 10, it states that "healthy" iso resistance is >3Mohm with contactors open, >2Mohm with contactors closed :

Later in the document, various components being tested independently by Fluke 1507/1587 have to pass these thresholds (pg. 11):

Note that the LDU is not listed above.

But none of those values in that last chart are for the bare stator windings: there is no test from Tesla for the stator. If the inverter is not in driving mode (powering the stator), then the inverter's IGBTs don't have the stator in circuit, and I don't see how the stator's iso can be tested independently (ie via an insulation resistance tester/megger/Fluke15x7) without accessing the stator windings bolts that I mentioned above, under the orange cover mid-gearbox front (rear, on RAV4 EV).

The first chart above says the entire system must be >2Mohm when "in Drive" or driving or inverter supplying power to the stator, but that's the entire system; each HV component "leaks" a bit, so you'd want any one component to be well above that.

---

I had one HV contactor opening incident, in Feb, when I demanded a bit more power than usual. More power = the inverter sends higher voltage & current to the stator, and like a clutch that slips only when you step on the gas, the stator's isolation resistance breaks down when faced with higher voltage. That's why you can't use a DMM's ohmmeter setting to validly check for insulation resistance: it doesn't challenge the insulation with HV! A "megger" (megohmmeter) or Insulation Resistance Tester (like the Fluke 15x7) put 500v on the windings and stress tests the insulation.

A DMM puts almost no voltage on the windings, and is pretty much useless for testing for isolation resistance of HV components, unless they are so badly damaged that they are completely shorted: well below test threshold. Passing a DMM ohmmeter test means nearly nothing.

After resetting the car, it drove fine for another 50 miles, whence I parked it until removal of the LDU.

My somewhat-dry stator, which had suffered coolant intrusion, measured 4.7Mohm after water/soap washing on an incline as shown previously and having had a small heater blowing air into it for about ten days: nothing very vigorous, and certainly not as much effort as I'd planned to do, because I considered 4.7Mohm to be "good enough" for my purposes. I had started with ~3Mohm before disassembly. I estimate that about two tablespoons of coolant were distributed around the reluctor wheel, the rotor's outer bearing, and the stator cavity. There was NO moisture on the inboard end of the stator.


Ugly on the stator outboard end:

Lots of corrosion on the stator outboard end, and visible damage of the (awful, waxy, non-epoxy) stator potting compound:

Moderate corrosion to 1/2 the stator laminations:

Dry, but with some evidence of previous coolant on the stator inboard end:

Measuring with the stator/inverter bolts installed, or removed and the windings' leads isolated (by using thin bent plastic strips behind the stators' leads) made almost no difference to that 4.7Mohm reading. If I'd had a leaky IGBT or coolant in the power section, this would have been different, but my coolant never made it to the center "tunnel" much less the inverter side, lucky me.

After reassembly and installation, the BMS reports total HV system iso as 3.3Mohm.

I would have preferred that the stator windings' iso be above 6Mohm, but in practice I'm content that I have no (or miniscule: difficult to remove inside the windings) residual water. I may live to regret not stripping the gearcase and soaking the stator half of the assy. in 99% IPA to dilute and "wash away" the remainder, but I'm taking a calculated risk by not doing so. I left no visible coolant, and removed all the corrosion I could reach without completely stripping the gearcase. Coolant could continue to degrade the iso, and there's little chance that remaining water will "evaporate" because there is almost no air exchange in the stator cavity. Time will tell if I get away with not having gone further.

 

[Gtech]

I don't have the breadth of experience with many LDU stators to give good advice on what is "good" HV isolation resistance. I'm seeing numbers quoted from 4Mohm to 120Mohm, which is quite a large range. I think that some numbers are coming from repair facilities that have a strong incentive to disqualify a stator that isn't comfortably above a minimum. Sorry in advance: lots of words follow . . .

Backing out a bit and looking at the entire HV system, the Tesla Isolation Faults guide says (pg. 6) that during various test procedures using the diagnostic menu or TDS (which utilizes the BMS for measurements, and therefore only measures at the battery side of the contactors), HV contactors open, that part of the system must remain above 1.8Mohm. Put another way, with the contactors open, that's measuring the isolation of the battery only.

[My understanding is that on the RAV4 EV (which is basically a Model S) the BMS will refuse to close contactors if it sees all isolation <1Mohm; if the contactors are allowed to close, and the isolation resistance drops under 1Mohm, the BMS opens the contactors and faults.]

On pg. 10, it states that "healthy" iso resistance is >3Mohm with contactors open, >2Mohm with contactors closed :

View attachment 1049208

Later in the document, various components being tested independently by Fluke 1507/1587 have to pass these thresholds (pg. 11):

View attachment 1049209

Note that the LDU is not listed above.

But none of those values in that last chart are for the bare stator windings: there is no test from Tesla for the stator. If the inverter is not in driving mode (powering the stator), then the inverter's IGBTs don't have the stator in circuit, and I don't see how the stator's iso can be tested independently (ie via an insulation resistance tester/megger/Fluke15x7) without accessing the stator windings bolts that I mentioned above, under the orange cover mid-gearbox front (rear, on RAV4 EV).

The first chart above says the entire system must be >2Mohm when "in Drive" or driving or inverter supplying power to the stator, but that's the entire system; each HV component "leaks" a bit, so you'd want any one component to be well above that.

---

I had one HV contactor opening incident, in Feb, when I demanded a bit more power than usual. More power = the inverter sends higher voltage & current to the stator, and like a clutch that slips only when you step on the gas, the stator's isolation resistance breaks down when faced with higher voltage. That's why you can't use a DMM's ohmmeter setting to validly check for insulation resistance: it doesn't challenge the insulation with HV! A "megger" (megohmmeter) or Insulation Resistance Tester (like the Fluke 15x7) put 500v on the windings and stress tests the insulation.

A DMM puts almost no voltage on the windings, and is pretty much useless for testing for isolation resistance of HV components, unless they are so badly damaged that they are completely shorted: well below test threshold. Passing a DMM ohmmeter test means nearly nothing.

After resetting the car, it drove fine for another 50 miles, whence I parked it until removal of the LDU.

My somewhat-dry stator, which had suffered coolant intrusion, measured 4.7Mohm after water/soap washing on an incline as shown previously and having had a small heater blowing air into it for about ten days: nothing very vigorous, and certainly not as much effort as I'd planned to do, because I considered 4.7Mohm to be "good enough" for my purposes. I had started with ~3Mohm before disassembly. I estimate that about two tablespoons of coolant were distributed around the reluctor wheel, the rotor's outer bearing, and the stator cavity. There was NO moisture on the inboard end of the stator.


Ugly on the stator outboard end:

View attachment 1049223


Lots of corrosion on the stator outboard end, and visible damage of the (awful, waxy, non-epoxy) stator potting compound:
View attachment 1049224

Moderate corrosion to 1/2 the stator laminations:
View attachment 1049225

Dry, but with some evidence of previous coolant on the stator inboard end:
View attachment 1049226

Measuring with the stator/inverter bolts installed, or removed and the windings' leads isolated (by using thin bent plastic strips behind the stators' leads) made almost no difference to that 4.7Mohm reading. If I'd had a leaky IGBT or coolant in the power section, this would have been different, but my coolant never made it to the center "tunnel" much less the inverter side, lucky me.

After reassembly and installation, the BMS reports total HV system iso as 3.3Mohm.

I would have preferred that the stator windings' iso be above 6Mohm, but in practice I'm content that I have no (or miniscule: difficult to remove inside the windings) residual water. I may live to regret not stripping the gearcase and soaking the stator half of the assy. in 99% IPA to dilute and "wash away" the remainder, but I'm taking a calculated risk by not doing so. I left no visible coolant, and removed all the corrosion I could reach without completely stripping the gearcase. Coolant could continue to degrade the iso, and there's little chance that remaining water will "evaporate" because there is almost no air exchange in the stator cavity. Time will tell if I get away with not having gone further.

Thanks, maybe if more people can post their iso we would have more reference of what is pretty good ( >5+ MOhm?) and what is not so good (probably something 3.5-4 MOhm)

Not sure why LDU is not listed in the overview but they mention it on page 19:

Also found this a few minutes ago:

So apparently the ISO can drop up to 0.2 MOhm when putting in drive

 

[Testurs]

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

 

[asavage]

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.

 

[Testurs]

That is interesting to see, especially with where the thermal distribution actually is on the rotor. I was thinking of putting a simple air pump that will blow air into the cavity on mine, or even just closing the far end of the metal tube on the the coolant tube/install a wider tube with a closed end. That would be a way to still channel coolant and remove at least some of that heat, because that just looks terrible for the rotor. I would think the stator insulation will rapidly degrade too.

 

[asavage]

Google says plain-ish steel blues in the high 500s°F (~300°C). The MLX90614 IR sensor I've installed can read up to 380°C (720°F), and I sure hope that the rotor isn't going to get much hotter than that.

But I wouldn't bet $100 on that.

 

[Testurs]

I am not an engineer, so someone will have to check the work here, but from a quick lookup I came to this:

"According to the Tesla Model S 2012 specifications, the motor has an efficiency of around 90% at 25°C (77°F). At higher temperatures, the efficiency of the motor decreases due to the increased resistance of the copper windings and other components.

Assuming a linear decrease in efficiency with temperature, we can estimate the efficiency of the motor at 300°C (572°F) as follows:

• Efficiency at 25°C (77°F) = 90%
• Temperature coefficient of efficiency (α) = -0.5%/°C (a typical value for induction motors)

Efficiency at 300°C = 90% - (α * ΔT)= 90% - (-0.5%/\°C * 275°C)= 90% - 13.75%= 76.25%

So, approximately 76.25% of the electrical energy from the battery would be converted to mechanical energy by the motor at 300°C (572°F).



76.25% is pretty bad, perhaps someone can comment how this feels real world in terms of performance after a long drive on one of these "updated" units? Kind of sad Tesla turned to this instead of actually fully fixing the problem