Tesla Small Drive Unit (SDU) Gear Box Fluid leaked

[alexbeatle]

Folks,
Sometime ago I resealed my SDU due to gearbox oil leak. I then refilled it with fresh oil.

I did not have the setup to spin the drive unit at the time.
Today, I was ready to spin it. Wanted to do a basic motor test before applying HV. Pulled the below orange rubber plug (phases connections cover) and lot's of fresh gearbox oil poured out. I quickly plugged it back, but quite a bit came out.
Does it mean I damaged something in my drive unit when I was resealing?? I did not expect the gearbox oil to be where the motor connections are.
Please let me know.

The + and - on the input tested fine:
+ to - precharged and then showed open circuit
- to + settled on 0.555 ohms

[alexbeatle]

Ok. I'm prepping myself to pull the unit out and so refreshing my memory how I took things apart last time.
In doing so, looks like the gearbox and the motor (rotor and stator) section are suppose to have gear fluid. Before I was thinking that only the gearbox section should be in the fluid.


Inverter section should not be having any gear fluid., which is why these wires from the inverter to the motor windings have o-rings and a cover.

I'll start by removing the inverter side and see if any obvious damages can be spotted.

If any of the o-rings or grommets are damaged, can I simply seal these wires with a gasket maker? Like this one...

I'm prepared for the worse, but lets see if any gods old or new are on my side....I was communicating with the drive unit and changing parameters, just didn't apply any bus voltage yet.

[alexbeatle]

Ok. I couldn't wait until the weekend to do everything at once. Doing bit by bit when I get a minute..
Drain fluids...oil and coolant ..

Some was found in the HV ports too

I think this is where it started the leak..

A little pre-story. I received the DU bone dry. I swapped the board and bench tested. Put in the vehicle. Filled up and moved on to doing other tasks of conversion. After a while started noticing some drips from the gearbox as shown previously. Drained, pulled out the DU, opened it up, resealed, refilled. No leaks from the gearbox on the outside after that. Maybe it's been damaged during me disassembling, but chould be that it's been leaky from day 1...

At this point I should assume all electronics parts have been touched by the gearbox fluid.

Please let me know if you have any suggestions on cleaning and testing?

[alexbeatle]

Planning to clean the LV drop in board with just a contact cleaner.
Not sure what to do with the HV IGBT section. Doesn't seem like it's dismantable.
https://youtu.be/1Hdr3wlHCAY?si=GncX2P1 ... Hf&t=3m38s

[jrbe]

O-ring + make a gasket = almost a guaranteed leak. Best to make sure all the o-ring grooves and bores are clean and not scratched. Use new o-rings.

Isopropyl alcohol should work well to clean all of this. Make sure it's completely dry before you assemble it.

[alexbeatle]

O-ring + make a gasket = almost a guaranteed leak. Best to make sure all the o-ring grooves and bores are clean and not scratched. Use new o-rings.

Isopropyl alcohol should work well to clean all of this. Make sure it's completely dry before you assemble it.

Thank you.

Cannot clean inside of the IGBT section, can I? Doesn't look disasembable.

[jrbe]

At the worst you should be able to pour isopropyl alcohol into the power stage section and swirl it around. Doing this a few times should get it clean. Making sure it's dry might be challenging though.

[alexbeatle]

At the worst you should be able to pour isopropyl alcohol into the power stage section and swirl it around. Doing this a few times should get it clean. Making sure it's dry might be challenging though.

Yeah. If it was LV I'd do it without thinking.

[alexbeatle]

Opened the motor leads cover. All o-rings look clean and undamaged. The cover looks clean too.

Planning to make a gasket on the cover using the gasket maker I used to fix the other leak - permatex optimum black (sensor safe). Though that was metal-to-metal, this will be plastic-to-metal. Should work just as good, right?

These folks use some kind of rubber lube, anybody familiar with this product?

[modellfan]

Interestingly I had the same issue and already started a topic on this: viewtopic.php?t=4143. Therefore I am very interested in your outcome. My SDU Rear unit is leaking just around one drop per day into the inverter housing. When I bought it used it hat just some dozens of ml in the inverter. I am pretty sure that this is the temperature sensor leaking. It is loose (see the video in my post) and I rarely can't feel a press fit of a o-ring.

However it is looking different then yours not having a rubber gromet on it, I am looking forward to see your solution to remove the leak. If you have some pictures frome the other side of the temperature sensor I am realy interested about it.

[jrbe]

These folks use some kind of rubber lube, anybody familiar with this product?

Your best bet is to use a dab of transmission oil as the lube. Mixing different lubes / oils / greases can have bad results.

For the loose temp sensor o-ring, you could likely find a thicker o-ring that fits properly that's compatible with Tesla transmission oil. Sounds like they aimed too low on the o-ring squish % or ended up with a temp sensor with a smaller o-ring OD area.

[alexbeatle]

Used the optimum gasket maker.
Put a thin layer on the inside of the holes of the plastic cover to seal any gaps between plastic and wires. Made a thin gasket between the plastic cover and the metal of the of the frame. Thin layer inside of the sensor hole. Hand tightened.
Torqued after 1h of curing.
Put a generous layer of gasket maker around the edges.
Waited 24h.
Filled with new ATF.
Spun both wheels at the same time every once in a while so that the oil pump would push the fluid all around.
24h later, so far no leaks - see attached
Sure it looks ugly, but if it works it works...
This gasket maker is super flexible, should allow for any movements without compromising the seal.

[alexbeatle]

Well this issue has returned.
Not sure what to do at this point other then buying new unit, which cost like $2k in my part of the world.
Any suggestions are appreciated.

[barracuda816]

can you pressurise the gearbox side with some kind of bike pump, (would recommend less than 10psi or so) and squirt some soapy water around look for bubbles?

[alexbeatle]

can you pressurise the gearbox side with some kind of bike pump, (would recommend less than 10psi or so) and squirt some soapy water around look for bubbles?

Good thought. Question, anyway that I can pressurise with the fluid still in the gearbox? SDU only has a drain hole - no fill hole - fill via shaft insert. Cannot think of a way to plug the shaft insert with the pump connection.
Trying to do all the investigations with the fluid still inside. It's red in colour so very visible when leaking.

How I discovered it. I had the car in the garage - checked via 3ph HV connection plug - no leak. Yesterday, I pushed the car out to the driveway (with some minor maneuvering) in prep for my first test drive. Decided to check the 3ph HV connection plug again - and the fluid poured out.

Today, I raised the rear wheels off the ground, drained the coolant from the inverter and removed the inverter side. Gearbox and motor side is still in the vehicle. No obvious spots for leakage. Wiped everything clean.

I spun both wheels by hand several times - no leaks. As I understand, need to spin both wheels simultaneously to engage the gears (not just the differential) and promote fluid swirling. Put one rear wheel on the ground, leaving the other suspended (did each side) - no leak. Both wheel on the ground - no leak.

Right now I left the car tilting to the passenger side (inverter side) to see if this will add enough pressure to the inverter connections and if it shows red drops within 30mins or so.

[jrbe]

I'd suggest way lower than 10psi if you decide to pressurize it.
The oil will thin out a bunch when it gets warm and will leak out easier then.
My suggestion is to drive it and check. Then if it seems good, check the oil level at 100, 1000, and 3000 miles.

[alexbeatle]

I'd suggest way lower than 10psi if you decide to pressurize it.
The oil will thin out a bunch when it gets warm and will leak out easier then.
My suggestion is to drive it and check. Then if it seems good, check the oil level at 100, 1000, and 3000 miles.

My biggest fear is that the gearbox fluid is flammable, is it not? So it will cause fire if gets onto the inverter under HV current load.

Incidentally. I put all wheels on the ground and tried to push the car around - not much different from how I did when I pushed it out of the garage yesterday. With the inverter still off, checked again - no leaks.

[jrbe]

Oil is difficult to ignite. It needs to be thin and well atomized to light. Or it needs very high heat.
There's also a fixed amount of oxygen available inside the transmission, so even if it did manage to light it would be very short lived.

[alexbeatle]

Well, I was about to attach the inverter back to run the motor for a bit, but decided to check under the encoder cover again. And...a little more fluid came out. I suspect this is where the leak is. Because before doing the aforementioned bus motions to troubleshoot I wiped everything dry.
Last time I assembled it, it was clean and torqued.

Is this a replacable seal or there's a lubricant that can be added to the existing one to make it better?

[jrbe]

Used the optimum gasket maker.
Put a thin layer on the inside of the holes of the plastic cover to seal any gaps between plastic and wires. Made a thin gasket between the plastic cover and the metal of the of the frame. Thin layer inside of the sensor hole. Hand tightened.
Torqued after 1h of curing.

Was the plastic cover fully seated using the bolts before waiting the hour for it to cure? If they weren't bottomed you may have caused an issue crushing / popping / distorting the make a gasket when tightened. This make a gasket style should be gooped, torqued, done. The gap, wait, torque technique can cause leaks.
I mentioned it a few posts back, using make a gasket with o-rings is pretty much a guaranteed leak. If you did, you can expect leaks in those places. I'd give an o-ring with make a gasket a 20% chance of sealing.
The encoder cover just has a shaft seal under it to leak oil, no? Having that o-ring seal better would just mean an oil puddle in the cover, no?
The o-ring looks proud of the cover, it should work fine as long as it's not nicked or dirty.

[alexbeatle]

The encoder cover sits over the 3ph leads. So it's possible that it could leak over them or enough that the fluid puddle will reach the leads. I had some pictures here.

...

The O-Ring looks undamaged and clean. But! It does looked more squeezed into the groove at the bottom than in other places. Though probably it doesn't matter when torqued to 25Nm. Wiped again and torqued the cover, will push the car around again tomorrow. With some thing to catch the leak from under the cover, if any.

Was the plastic cover fully seated using the bolts before waiting the hour for it to cure? If they weren't bottomed you may have caused an issue crushing / popping / distorting the make a gasket when tightened. This make a gasket style should be gooped, torqued, done. The gap, wait, torque technique can cause leaks.
I mentioned it a few posts back, using make a gasket with o-rings is pretty much a guaranteed leak. If you did, you can expect leaks in those places. I'd give an o-ring with make a gasket a 20% chance of sealing.

I followed the sealant instructions - bolt, wait an hour then torque. I looked at the way ZeroEV put sealant on the plastic cover - looks like they used sealed over the O-Rings - https://www.youtube.com/watch?v=NBFSS82ZLY4&t=340).
I still haven't concluded that it is my previous seal leaking or the encoder cover, but I see your point. Thank you for helping me out with this!

[jrbe]

The encoder would very likely be completely fine running submerged in oil. The shaped o-ring is likely meant as a backup to keep oil out of the inverter area if the oil seal leaks. Seems like the shaft seal has a leak there if there's oil inside.
The o-ring swelling sounds like they may have used the wrong material for the o-ring to be compatible with the transmission oil long term.
Or, more likely someone added stop leak to the transmission to "fix" a leak. It's possible all the seals are shot if someone did this. They swell and wear, so if the stop leak is removed (transmission oil change)they can shrink some and leak again. Based on some of your struggles it sounds like this may have happened.

In the video they used rubber lube on the o-rings "the red stuff". So they could not have used make a gasket on the o-rings as well, at least in any useful way. They could have used make a gasket away from the o-rings, on the flat face., but the rubber lube would very likely prevent the make a gasket from adhering well or at all.

Did you use some strong cleaner like acetone, brake clean, etc. to make sure the oil was completely off whatever surfaces you used the make a gasket on?

[Steve16]

Had a similar leak in my SDU. Found that the temperature sensor wire plug was not fully seated. You can see the photos and description here: viewtopic.php?t=4143