Tesla Large Drive Unit (LDU) Motor Teardown and maintenance

[johu]

No he must have replaced it with the DELETED text himself.
When I delete spammers I can delete their posts along with the user. Sometimes I delete users that have no posts and haven't been active for more than a year. But never ones that posted something

[howardc64]

Does anyone know what happened to howardc64?

hello, I deleted the contents myself, Johan is developing the website and info. Hope that helps. Thanks

[asavage]

Bit of crossposting: coolant manifold plug of 6061:
https://www.myrav4ev.com/threads/anothe ... post-30941

Teaser:

Rotor temperature sensor/display:
https://teslamotorsclub.com/tmc/posts/8230849/

Teaser:

[Boxster EV]

I’ve bitten the coolant delete bullet too. Just fitted this from Revolt Systems in California. My car is used as a Daily, so didn’t have the time to have off the road whilst the existing manifold was modified. Took about 2 hours to fit.

[asavage]

Some notes & pics on my journey:

LDU rotor cooling bypass ("Coolant Delete") modification

LDU rotor coolant bypass plug

LDU special disassembly tools

LDU Rotor End Cap removal tool

LDU rotor removal puller tool

LDU rotor temperature monitor (display)

LDU rotor temperature sensor: installed

LDU rotor temperature display: early prototyping with bad alignment.

[sirpewe]

What do you use for seal material (brownish on the maifold)?
And do you have any drawing or cad file for the plug you like to share?
Or possibly if you can sell a plug?

[asavage]

What do you use for seal material (brownish on the maifold)?

That picture is misleading; if you follow the link, it makes more sense. The red gasket you see above is for the pressure test jig I set up to provide proof positive that the plug would not leak.

The red gasket is not used for mounting to the LDU.

And do you have any drawing or cad file for the plug you like to share?

Follow the link, please.

Or possibly if you can sell a plug?

Howard & Mr. Hyde have a batch of SS formed sheet metal plugs coming, that will be cheaper than machined aluminum ones like mine. If you follow that link, you'll read that my two plugs were machined locally by a fellow who doesn't want to get into production; he does only one-off prototype stuff like this.

If there was sufficient demand, I could have a run made by eMachineShop.com . If I bought around (50), I could sell them for about $30 + shipping, but I doubt there will be sufficient demand, and the SS sheet metal plugs will be cheaper; most people will probably want to go that way instead.

I prefer the machined aluminum plug, myself.

[jrbe]

JLCPCB does cnc work now. You can setup a commission based on your design there if you want.

myrav4ev wants you to login to look at pictures. Not everyone wants an account.

The design could be modified to have an o-ring on the OD of the plug. If threading that bore in the housing is an option / easy to do, an o-ring plug / bolt can be used to hold the plug in place.

[asavage]

JLCPCB does cnc work now. You can setup a commission based on your design there if you want.

Lots of places do machining. I got one easy quote from one place, got a ballpark figure, and I'm not interested in chasing lower prices or lower minimum quantities. I'm retired and don't need the hassle of retail sales again. I already "sell" pre-assembled ethernet cables for the Tesla Powertrain Diagnostics program used in the RAV4 EV, and that's been hassle enough. Plus the inevitable people who receive something and then don't like it/claim it doesn't fit/claim it ruined their LDU/etc. Not gonna be me.

myrav4ev wants you to login to look at pictures. Not everyone wants an account.

You know what? I don't care. I am active in three communities around the LDU, and I'm only going to post this info once; it took HOURS -- almost two man-days -- to put the pics and descriptions together, and if someone can't be bothered to run down the pics, I just don't care anymore. Don't want to create an account? I get that. Go do your own measurements on your own LDU and make your own mistakes.

I can't make information barrierless for everyone.

It doesn't matter if I posted on Instagram, X, TikTok, or FB, somebody would complain that "it's too hard" to find the info. I wrote an entire G'damn Wiki on the Tesla GEN1 OBC and still had complaints about this, that, and that other thing. If I put it on YT, I'd get compaints. If I offered to mail detailed specs via snail mail, I'd get complaints. Github? Complaints -- got one just last week from there.

And then there's the people on phones who complain that the pictures are too large and their phone won't render them properly. I've run a hobby website forum for DECADES and have heard it all before, and I'm done trying to accommodate every last person's issue with FREE content.

I spent $50 on Toytota online subscritions and an entire Labor Day weekend a couple of years ago, pulling service content for the RAV4 EV, collected as much as I could and put it in a ZIP file . . . and got complaints that it was "too hard to find things".

In the past, I tried to maintain info in multiple locations, and it just multiplies the work. If you miss an update on one site, you field questions/corrections well into the future, and you just can't maintain multiple data stores. It has to be one place, I chose my one place (the RAV EV is the community I'm not active in), and it'll have to do.

The design could be modified to have an o-ring on the OD of the plug. If threading that bore in the housing is an option / easy to do, an o-ring plug / bolt can be used to hold the plug in place.

Yup, could be done with an o-ring seal instead of Loctite 640 . . . but I discarded that idea fairly early on a year ago. A mechanical retainer would be required, as you noted. And there's no need to ever remove the plug for any reason, so why bother with installing a failure point (o-ring)? Gluing and pressing it in is adequate and has no downsides over an o-ring seal.

[sirpewe]

Thank you for the fast feedback, much appreciated.

That picture is misleading; if you follow the link, it makes more sense. The red gasket you see above is for the pressure test jig I set up to provide proof positive that the plug would not leak.

[asavage]

The second plug, .005" oversize + Loctite 640, passed 30 PSI for two hours with no leakage. 30 PSI pressure differential is at least ten times what that plug should ever see, so I was satisfied with the result.

[sirpewe]

I found a drawing (yours?):

Are these the dimensions you used? (you mentioned .005 oversize which would equal 0.127mm so perhaps reduce 55.2 to ~55.1mm)
I have a contact that can help me mill and lathe so I will go and try this.
Plan to seal the cap with loctite 638 or 648.

Are there any risk of leakage on the other side (where the "tube holder" is pressed in) or does this has a tight fit from factory?

BR
// Per

[asavage]

That's not mine, though roughly the right idea. Some of us batted around several sets of dimensions. You'll notice that there's no counterbore dimension on that drawing, and it's 2mm too shallow IMO. And I wouldn't normally specify outside radii such as that; from an engineering standpoint, I prefer chamfers on outside 90° edges, and radiused corners on inside ones.

This is the drawing I handed to my lathe man. mm in blue ink, inches in pencil. Scale is correct.

Tesla LDU Coolant Manifold Plug, 20240408

I believe that the SS formed sheet metal plugs coming are 55.2mm (.008" OS) and are designed to deform & crush into place, similar to a seal cage. I am skeptical that that much interference fit is needed/desired in a more-solid 6061 plug, so I started with .003" OS, and while that would have worked fine, I liked the fit of .005" OS better. I have a "feel" for this, but while I worked in a mechanical engineering department with real engineers, I'm not one myself, and am more of a cut-and-try man when it comes to this kind of thing.

Also, the lathe man agreed that .003" would be "enough", so I let him convince me And, with the lightly pressurized system in the LDU, and an appropriate sleeve/bearing mounting compound, it would be.

[asavage]

I’ve bitten the coolant delete bullet too. Just fitted this from Revolt Systems in California.

While machining a part such as this is relatively straightforward for those with the skills to draw it and with CNC, what will impress me is when somebody will start offering the various formed o-rings for the LDU. AFAIK, these are currently unobtainable, and having bespoke formed o-rings made isn't as simple as handing a drawing and a slug of metal to your local machinist . . . or is it?

I was verrrry careful handling mine, knowing that if I damaged or lost one I'd be unable to replace it.

[Boxster EV]

While machining a part such as this is relatively straightforward for those with the skills to draw it and with CNC, what will impress me is when somebody will start offering the various formed o-rings for the LDU. AFAIK, these are currently unobtainable, and having bespoke formed o-rings made isn't as simple as handing a drawing and a slug of metal to your local machinist . . . or is it?

I was very careful handling mine, knowing that if I damaged or lost one I'd be unable to replace it.

Yes, they are seemingly unobtainium. I have a couple of spares that I managed to source from broken drive units. I also ran a very light bead of high performance RTV over contact surface to make double sure it sealed.

[asavage]

Because the cooling system is unpressurized, the seals don't have a lot of forces working against them, but still . . . elastomers don't remain pliable forever, and best practice is to replace them when they are disturbed. The fact that replacements aren't readily available is likely the only reason that we're re-using them.

RTV can reasonably replace/supplant them, but vastly complicates future R&R work. You'll see in my other thread RE special tools for the LDU that I could not reasonably remove the glued-on rotor end cap without building a puller adapter, because of the large surface area that Tesla glued (the edge of the ~12" dia cap), and that they used SO MUCH of the stuff, even though it doesn't have to seal anything but perhaps rainwater splash.

Beating on things to disassemble them isn't professional because it can change their dimensions beyond what glue and bearing tolerances can compensate for.

[Boxster EV]

What alternative do Tesla have for the end cap? Gaskets are not infallible either.

[asavage]

My point isn't that a gasket would be a more effective sealing solution for rainwater, it's that it makes disassembly -- even by professionals -- several times more difficult, for no good reason.

Some alternatives to RTV: the rotor end cap could have had:

• a formed elastomeric seal, such as is used on the inverter cover. I don't particularly like the inverter's seal, but it seems to work OK;
• an o-ring, which would have been ONE more machining operation on the existing casting. I really like this idea, as non-formed o-rings are readily sourced (or assembled on-site/in the field) and do not affect fit-up tolerances;
• a flat gasket; a mature technology for centuries. A downside is the variability of flat gaskets, which on an assembly such as the LDU affects axial play of the rotor. I would probably not spec a flat gasket here, were I asked.

Instead, they went the Subaru way and used robotic RTV squirting, putting serviceability way down their priority list. But I bet that assembly costs were pennies less.

Other advantages to RTV sealing is that it conforms to many surface irregularities, so machining tolerances can be lessened, and it seals on a relatively larger surface area than do o-ring or most engineered formed elastomer solutions. It's very susceptible to surface contamination upon assembly, though. It's big advantage is cost reduction during assembly -- more than negated by how difficult it is to remove the stuff to service it. Seriously, try out the various silicone-based RTV removing agents, and you'll find that until you get into the big $$$ solutions, none of them truly work. Not acetone, not MEK, not WD-40 or gasoline or diesel or anything you can buy off-the-shelf for less than hundreds of dollars, and until you get into really dangerous temperatures that would ruin the stator, not heat either. Even if you have a suitable RTV-dissolving compound at hand, you can only treat the edges of the rotor cap, so you can't break the bond because Tesla applied so much that it wraps around to the inside of the stator housing.

I've worked in the service industry through almost all my careers, and I see this kind of bonehead engineering everywhere: make it as cheaply as possible, assume it only has to make it out of warranty, and "we can keep burning things and burying things forever (until Mars)" mentality.

[AlphaTango11]

Hi there,

I've owned a 2014 P85D since 2019 that recently had coolant ingress at around 70k miles.

LDU P/N: 1025598-00-P
LDU S/N: T13E0015996R2

About a year ago, I used the Ceimin yellow triple lip seal - 30x55x8 apparently renamed to OD55 CTLBDYW-3965 and this 6007-2RZTN9/HC5C3WT SKF bearing for ~$110 each from Acorn. I also drilled an M6 size drain hole (tapped and then fitted a barb) and ran some clear tubing to drain/contain coolant. There was no coolant or residue present on the inverter side, only the rotor/stator side.

The good news: After reassembly and driving a few hundred miles over the past year (haven't driven it much due to the bad news coming up), there hasn't been any coolant leakage out of the drain hole. Good news as long as drain hole works as intended.

The bad news: The vehicle still sometimes throws external isolation alerts when it is still (gets down to about .5MOhm, otherwise stays around 1-2MOhm). I'm an ex-Tesla service technician, and the only low isolation is at the LDU, about 1MOhm instead of >4.

My main question is, how can I resolve this low isolation if the stator is epoxied? I did not remove the rotor (which I admittedly should have done). Is it likely that the rotor needs to be removed and fully cleaned?

I didn't see a way to attach images, so I uploaded to imgur.





Thanks!

[asavage]

The vehicle still sometimes throws external isolation alerts when it is still (gets down to about .5MOhm, otherwise stays around 1-2MOhm). I'm an ex-Tesla service technician, and the only low isolation is at the LDU, about 1MOhm instead of >4.

Interesting that the isolation resistance is variable. Why are you certain that it's the LDU whose resistance is changing? Given the history of the Model S I would have thought that other components would be more likely to yield variable isolation (external cables, especially: a little rainwater or through the car wash with worn connector seals, and Poof! The iso goes down.).

Are you relying upon the Tesla's diagnostics for those numbers? Or, is the variability of the LDU isolation resistance being measured by, say, a Fluke 1507 Insulation Tester or equivalent? While a DMM will spot the worst of low iso stators, it isn't designed to test the isolation resistance while under HV, and that's why Tesla specs testing using a "megger". I own a 1507. It applies 500v to the stator (or whatever I'm testing) and measures the leakage/decay time with voltage applied; a DMM doesn't do that.

(ref. Tesla's Isolation Testing document SB-10052460-6095. Not for you, but for others who will stumble across this thread later.)


Do you disconnect the LDU's HV cables and test the iso resistance of the LDU itself?

how can I resolve this low isolation if the stator is epoxied? I did not remove the rotor (which I admittedly should have done). Is it likely that the rotor needs to be removed and fully cleaned?

A couple of things here.
Cleaning the rotor isn't going to change the measured isolation resistance, because there's no electrical connection (at 500v or less) to the HV portions of the LDU. I mean, we clean the rotors, of course, when we have them out, but do they need to be cleaned? No.

Removing the rotor wasn't a walk in the park for my recent LDU repair; I fabricated a fixture setup to allow enough pulling force to remove the rotor with the inside bearing remaining in the case.

LDU rotor puller split shaft collar

LDU rotor puller

LDU rotor puller setup

It wasn't simple or easy, but other than replacing the inside rotor brg., it allowed me access to as much of the stator as is possible. I was able to use a water-based cleaner and surfactant to remove all the coolant and corrosion byproducts from the exposed stator laminations, "wash" the exposed surfaces, blow warm air on the stator for a day, then douse everything I could with IPA (99%! not 70%!) to further dilute any remaining water that stubbornly wouldn't evaporate.

Tilting LDU assy. for stator cleaning

Tilting LDU assy. for stator cleaning

Spraying IPA on a warm stator isn't exactly the safest of operations, and ventilation is really, really necessary. But IPA will merge with and dilute any water that can be accessed, and then it evaporates with a minimum of airflow and warmth.

Having said all that, a user reports having three unsalvageable LDUs with shorted (low iso) stators right now; not all low iso stators can be "dried out". There are shops that are removing the stator from the case and rewinding/repotting them, but I have almost no info on this myself; you can readily find more on YT, etc. but I don't think of this as an economical path in the US at this time; it's certainly less expensive to find a less-soaked/higher iso stator/gearcase, than to try to rebuild a wet stator here.

---

A point that others I've discussed this with seem to overlook is that any water inside the LDU, stays in the LDU upon reassembly. It won't magically evaporate and migrate outside. For purposes of imagining water in an LDU, it's sealed. Yes, it has a couple of vents (one on the ATF/gears case; one on the stator/inverter case) but aside from thermal expansion/contraction of a couple of liters of air, no air exchange occurs, so ANY moisture left in the stator cavity etc. will remain there, further corroding things and degrading the iso resistance.

You have to dry it out, to the best of your ability, before reassembly. It will not dry itself out later. I suspect that by no pulling your rotor and cleaning/drying as much as you could access, you may have been recycling old moisture in your LDU.

I didn't see a way to attach images . . .

It's possible that, as a member with no posts, the forum software may restrict your ability to use the Attachments feature. I know from personal experience that you can't add your Location to your profile until after you have posted at least once.

[AlphaTango11]

Why are you certain that it's the LDU whose resistance is changing?
...
Do you disconnect the LDU's HV cables and test the iso resistance of the LDU itself?

All good points. I have ruled out all other HV components through various means (primarily direct measurements using an isolation tester at the rear HVJB, but also keeping drive rails off (but contactors closed) and checking A/C compressor, PTC, coolant heater and watching isolation, which stays stable and healthy).

Of course I should have checked the drive unit directly, but by the time I had purchased a proper tester instead of using a DMM, I had already put it all back in.

Note that in these pics, the HVB was not installed in the vehicle (rapid mate had a crumbling seal, so I fixed that at the same time as the drive unit).

A couple of things here.
Cleaning the rotor isn't going to change the measured isolation resistance, because there's no electrical connection (at 500v or less) to the HV portions of the LDU. I mean, we clean the rotors, of course, when we have them out, but do they need to be cleaned? No.

Right. I guess I was afraid of dried coolant on the rotor somehow coming into contact with the stator windings/laminate, but it would quickly be deposited elsewhere anyways once in motion, I guess. As techs we were told that extreme contamination can cause low isolation (of course anything can conduct at sufficiently high voltage), but obviously the rotor is never supposed to be energized anyway.

At this point I suppose I need to check the stator windings resistance to ground directly (good access available at the orange center cover), remove the rotor to actually clean anything on the stator laminates, and then recheck after significant drying. No point in reinstalling the motor if the windings have unrecoverable low isolation.

I will say that the car drives fine, but I'm sure it's somehow bad for the inverter or windings (or my safety?). The performance is unaffected and the car hasn't ever refused to close the contactors for things like 12V support or vehicle entry.

All I know is that the low isolation resistance is hard to duplicate in motion, but can be duplicated occasionally when the car is stopped or nearly stopped (1-2mph) while the vehicle is in Drive.

I'm thinking that some of the bottom windings only have low isolation resistance whenever sufficiently high voltage is present to conduct across any contamination, but my assumptions may be flawed (Maybe all phases of the windings are ~0V when vehicle is stationary unless throttle is applied?)

I've also never seen low isolation alerts triggered while drive rails are off even with contactors closed, which I assume is because the windings are not energized until the drive rail is on or throttle is applied, but I don't know if that is true.

I monitor isolation while in drive through the vehicle's diagnostic reports over Ethernet (soon I will have a good CAN adapter), and the isolation measurements themselves typically have a ~30sec refresh.

I almost wish I could disable the rear drive unit and only use the front.

Removing the rotor wasn't a walk in the park for my recent LDU repair
...
Having said all that, a user reports having three unsalvageable LDUs with shorted (low iso) stators right now; not all low iso stators can be "dried out". There are shops that are removing the stator from the case and rewinding/repotting them, but I have almost no info on this myself; you can readily find more on YT, etc. but I don't think of this as an economical path in the US at this time; it's certainly less expensive to find a less-soaked/higher iso stator/gearcase, than to try to rebuild a wet stator here.

You have to dry it out, to the best of your ability, before reassembly. It will not dry itself out later. I suspect that by no pulling your rotor and cleaning/drying as much as you could access, you may have been recycling old moisture in your LDU.

Yes, removing the rotor seems to be the most challenging part, hence why I attempted to avoid it (and likely have created more work for myself). Good example of one way to do it, thank you. Seems better than snatching/dropping the drive unit.

It is entirely possible that there is too much residual moisture, as I only air dried it for a few weeks as I waited on all of the parts and got side tracked with other projects, but it seems some stators are capable of becoming unrecoverable.

So to summarize, the vehicle is reporting constant low isolation (1MOhm) that sometimes (unsure exactly why) gets even lower and triggers alerts (250-500kOhm). I will plan on checking all phases properly and going from there.

Thanks for the informative response!

[asavage]

IDK the spec on your MS (you have a GEN2 Rear HVJB and OBC, so it's Jan2014 or newer) but the GEN1 spec (during testing) is 1.8Mohm min. for the entire HV system (with the contactors open ie the HV battery not being under test) -- see pg. 6 of the Isolation doc. I linked above (again, this is for others following along). So, testing just one part and getting that low reading . . . is bad.

Thanks for showing your test methodology. This tests the HV leads to the LDU but not the stator because, as you noted, the stator is not actually in circuit when not powered by the IGBTs. That 0.9Mohm is quite troubling. I venture to guess that there's an issue with those leads to the LDU themselves, which is a failure point I've seen mentioned elsewhere sometime in the past. Corrosion at the plug-in EMC shields grounds/glands? This is good news, yes?

Yes, you can test the stator windings at the 3-phase HV wiring at the center of the gearcase, under that orange cover with the "do not disturb" iconography. For others (sorry for the poor lighting; these are found images, not mine):

LDU inverter/stator internal HV junction access

LDU inverter/stator internal HV junction access

My experience on mine was that testing could be accomplished without removing those bolts and isolating the stator leads from the inverter leads; almost no change in test results occurred when I tried both methods. The access to these isn't horrible on the RAV4 EV, but IDK what access is like on the MS with the LDU in the car. As yours is on the bench, this'll be easy for you.

[later]
If your inverter side had had coolant exposure, it creeps up the small harnesses and usually causes errors that the customer sees, but I could theorize that it could also reduce iso enough to give the test result you're seeing. But I see that you have the inverter cover off, and you'd have said if you had any blue drops of death sitting in the cover's bottom.