Tesla Large Drive Unit (LDU) Motor Teardown and maintenance

[SuperV8]

Just thinking out loud!

1: How about an oil/ATF cooled rotor rather than coolant? I would think much less likely to leak - at the expense of lower heat transfer.

2: How does the model 3 front drive unit cool? This is also an induction motor - so will also get heated - and I can't see any coolant fittings on the case - so I presume either internally oil cooled? or not cooled at all?

[asavage]

I think the M3 front induction motor uses an oil-to-water heat exchanger (see exchanger on left side of the pic below). It's cooled by oil internally, and the oil is pumped to the exchanger that's bolted to the exterior of the DU. Additionally, the coolant loop cools the inverter electronics first (see the coolant fitting on the right of the pic below) before routing to the heat exchanger, so the more temperature-sensitive electronics get the cooler coolant.

Tesla Model 3 FDU, top view

My suggestion idea No. 7 above was to use a non-water-containing coolant (basically: glycol) instead of a 50/50 water mix, so any leakage would cause no corrosion nor electrical harm. I bought a gallon of a typical gycol coolant, and a spare Tesla coolant pump and PWM module, but it's all sitting on one of my workbenches right now, waiting for some testing time.

I don't know if it's practical to convert the LDU to oil cooling for the rotor, and an external heat exchanger, but it's possible. The external aluminum coolant tube that the LDU uses to move coolant from the inverter side to the rotor side, could be interrupted, and a small closed-loop oil circuit established for the rotor side, plus a heat exchanger.

This might be practical for recycled LDUs in EV conversions, but possibly not practical for Model S installations, which is what Howard is aiming to achieve.

My drawing tools are limited, but here's a mockup of a oil-to-water heat exchanger, plumbed in to an LDU:

Tesla Model S LDU, Inverter-to-Rotor coolant flow

Tesla Model S LDU, idea for oil-cooled rotor

[howardc64]

DELETED

[howardc64]

DELETED

[asavage]

Fortunately, LDU inverter-side coolant leaks are rare

[howardc64]

DELETED

[howardc64]

DELETED

[tesland]

We opened a drive unit for the first time.
This one is making the (in)famous milling noise when driving.
no leakage in the motor found.
Which bearings are the usual suspects. Is there a way to check the bearings
IMHO the bearings of the rotor shafts are noisy
Any suggestions are welcome.

https://youtube.com/shorts/LgdjdP0V-ok?feature=share

[asavage]

From memory (ie this is not Gospel):

• The far end rotor bearing (the one that gets the most coolant ingress from the leaky seal at the far end coolant manifold, by the rotor speed encoder)
• The near end rotor bearing (due to early production units not having enough rotor grounding (via Aegis ring brush) for stray AC induced rotor currents)
• And the pinion bearing at the rotor near end (due to the later use of "ceramic" ball bearings on the the previous bearings, moving the rotor ground path into the gearbox).

There are LOTS of threads and images available for these, including brg replacement part Nos. I'm not aware of any of the other gearbox brgs being an issue in typical use; racing or higher output use cases might have some other gearbox brg failures, but I don't read in those circles.

The coolant seal replacement is the primary focus of this thread, so reading it top to bottom will give you the current thoughts on replacement/modification. If you have no coolant leakage past the far end rotor seal, yet have milling noise, I'd guess you have an early production (2012-2013?) LDU with a failed Aegis ring and steel balls in the rotor bearings. Replacing the bearings (for noise) and the seal (because you don't really have a choice, once it's been disturbed) will likely get you back in service; the pinion bearing may or may not "need" replacement; if it was my budget, I'd replace it anyway, but my budget isn't yours, so it's in your lap on that one.

[howardc64]

DELETED

[asavage]

Unless I knew the brgs were installed <10k miles previously, I wouldn't even try to evaluate them, they'd be replaced. I am not that tight-fisted that USD$300 (or whatever) for two rotor brgs would seriously be a consideration; I do not want to take this apart more than once a year, these locations are known problem areas for this assembly, and trying to skimp here is for low-budget/lots-of-time-available/don't-need-high-availability situations.

That was preface to say: I do not test inexpensive bearings, I don't evaluate their condition or listen to them more than a few seconds; I am a parts-replacer when in this situation. I have the advantage of owning the tools needed to do that, so the raw materials cost is my only cost, other than time.

Of course, it's quite easy to damage bearings during installation, esp. the ceramic replacements. Use a press or find a way to minimize impact side loads on ball bearings during install.

[Roller brgs are much more tolerant of shock loads, and can carry much higher loads in general, but due to their increased contact area, they also have a lot more drag, so they're never used for high-speed applications.]

[howardc64]

DELETED

[hmejl]

Hey folks
I have now pulled my third ldu apart, all bearings changed and seals to.
But we get a bearing noise from 10 to 30 mph when acc slowly any idear of what could have happend ???

bonus info
we had a 0,6 mm groove on the rotor but found a local company there grinded the shaft down, and the put a new chrome face on the shaft ant then grinded it back to messurement and it is perfect surface for a shaft seal now

[howardc64]

DELETED

[hmejl]

- What bearings did you change? 2x rotor bearing, 2x primary shaft bearing? (6207/6208)
- How did you install these interference fit bearings?
- Did you close the gearbox horizontally or vertically? I closed mine upside down horizontally (to minimize damaging the inverter) I'd imagine vertical assembly is ideal but would probably need to build a box as the entire assembly is ~3 feet tall without flat ends.

My Tesla installed reman LDU had a 30mph decel whine and found 6207 spun in its bore during my rebuild. After my rebuild, has a speed based whine > 20mph. Have 3500 miles now, not sure its getting worse and sound seem to change with the warming weather.

Hey Howardc64
I have replaced all the bearings. and mountet all the bearing in a press. But i have just dismantle another one and meassured all the bearing gaps and such. And found not all bearings are all seatet on the shafts, and might think i also have a bearing there is spinning
So i will take it apart again.
Thanks for the answer

[howardc64]

DELETED

[asavage]

- Mine and other intermediate shaft's counter bore bearing (FAG) had about 1cm x 2mm of red retaining compound on the side of outer cage in the bore. @asavage suggested this would accomplish nothing . . .

To clarify: the retaining compound visible was at the bottom of the counterbore, not on the sides, and there is insufficient surface area at the bottom of the counterbore to accomplish anything when using retaining compound there. Anaerobic retaining compounds -- indeed, all retaining compounds -- are effective as a function of surface area, which is why they're applied to the OD of the outside brg race and/or inside of the bore to which the brg is fitted.

From what I saw, the visible compound in Howard's gearbox was either incorrectly/sloppily applied, or was excess that was wiped off the counterbore's sides during assembly, which is the most charitable interpretation. Certainly, "loose" retaining compound at the bottom of the counterbore serves no function, but it may be a side-effect of applying excess during assembly. I advised Howard to not place any there directly, but only to the counterbore's sides prior to fitting the brg.

I usually see this kind of assembly designed with the brg having a light interference fit, but properly fit -- looser -- plus retaining compound has been an alternative fitment I've seen. Tesla may have started building the LDU using one technique and later switched to another.

[howardc64]

DELETED

[spiff]

Hi Everyone,
I've been combing through all these pages, Howard's site and Johan's youtube videos (thanks to everyone who's contributed to this body of knowledge!) and need some tips. I've disassembled my LDU and am now in the process of re-assembly after waiting months to get new seals (Chengming) and one rotor bearing.

Does anyone have a picture or sequence of how the circlips, seals, bearings go in on either end of the rotor? I unfortunately didn't take pics or record it when I was disassembling and now realize there may be some variability in how to do it.

Johan,
I noticed you had a cad model you referred to in some of your videos. Do you have a clear picture you'd be able to provide? I seem to have the brush ring that provides electrical connection to the rotor....

Thanks in advance!

[howardc64]

DELETED

[asavage]

Had to close it up with bolting the 2 halves.

For others: this is never a good idea. There are probably thousands of motorcycle case halves cracked and in scrapyards from using this technique. I've done it myself (when I was 14 and working in a motorcycle repair shop!). Though I'm glad you got away with it with only a mystery-cause whine.

Never use the fixing bolts to draw assembly halves together.

[spiff]

Thank you both!! This is invaluable information and advice. I'll post everthing in my build log when I'm ready.

Thanks again to this community.

[howardc64]

DELETED

[spiff]

Okay. Got the rotor back in and end plate on but not the coolant seal and cover yet. That's enough for tonight. In the future, does anyone know what kind of grease goes on the spline end of the shaft?

[asavage]

I believe that no-one has information about what grease(s) Tesla has used over the years, and the subject comes up every so often.