EV gear reducer has black oil, ref my post on over-constrained couplings. 1

[KiwiME]

thread78-464172

So, adding information to my prior post from April 2020, at 19,000 km because of my prior concerns I changed the oil on my EV's splash-lubricated reduction gearbox and found it coloured like black coffee. The oil is a 70-75 weight GL-4 and the original colour is unknown.
There was also a light loading of fine ferrous particles, noting that this gearbox has no magnet on either drain or fill plug, nor is one clearly visible affixed inside the casing. The latter is based on a teardown video on YouTube.

My theory is that the oil discolouration is a result of outer raceway spin in the aluminium housing. A likely candidate is the input-side bearing on the primary pinion shaft. The fine ferrous debris on the other hand can only be a result of gear wear slivers circulating continuously until ground to dust.

I'm thinking that the outer raceway spin may be related to the over-constrained design condition found at the input spline, described in detail in the earlier thread. Should bearing friction be increased by either cold thick oil or particles in the oil, I'd imagine the potentially light radial loading due to the motor bearing also supporting the reaction from gear engagement, raceway spin could be initiated leading to aluminium housing wear.

I can't easily get an oil analysis done here in NZ here but I've asked owners in N.A. to consider doing this as I'd bet many examples have this condition. The one other oil change published on YouTube shows the oil in even worse condition. I will look for an after-market (or OEM) magnetic drain plug but I think once traces of aluminium are found the damage is done.

Any thoughts on this?

 

[BrianPetersen]

That's bad. The teardown video that you linked to, shows metallic paste collected on (presumably) the bottom of the gearbox. That's very bad.

I found an English-language video that shows the original colour of that fluid and the same blackness as yours ... and he was changing the gearbox because it had gone noisy.

https://www.youtube.com/watch?v=dMlngPWEbQ4

edit: It seems to be an endemic issue with this car.

edit 2: One of the comments on the video linked above suggests that there was a known bad batch of reduction gears, and there are owners with much more mileage that haven't had this problem.

 

[KiwiME]

Ok, thanks, I just read through all those comments on the video from The Off-roading Mechanic. I had posted on that video a few times recently but had not read the previous comments all the way through. It did give me some ideas.

There have been a few gearbox failures on this car globally and I'd prefer mine not to be one of them as it's now out of warranty. When oil from a failed box is found to be black (as in this video) no one thinks any more of it. But no one with a good gearbox has ever changed the oil to my knowledge. So I guess I'm the first to do that and I was shocked to see black oil. Interestingly, owners of Hyundai Kona ICE with a DCT (dual clutch tranny) also see blackened oil.

Today I tried leaving a magnet in the used oil for a few hours and the debris stuck to it was extremely fine and showed up as black when wiped on a tissue. That's actually a great relief to me because a spun bearing (producing aluminium) would be the worst outcome. I think the fine ferrous dust and subsequent oil blackening can be resolved simply by the addition of a magnetic drain plug, like every other splash-lubricated gearbox in the world has.

Thanks for your comments.

 

[romke]

the oil blackening is due to oxidation, oil components reacting with hot air. that leads to reaction products with a far higher viscosity as the fresh oil and possibly to foul smelling paste like substances. the reason is overheating - possibly caused by faulty assembly or incorrect clearances. also noteworthy in the first (Russian?)video is the darkening of the cage containing the rolling elements of the tapered roller bearing and some rolling elements showing quite a bit of wear.the most likely cause is faulty assembly - tapered rolling bearings need to be assembled with a suitable preloading to prevent excessive sliding.

the fact that the damage occurs after such a short service also indicates a far higher oil temperature then normally would be expected. something must be wrong with this gearbox. the oil drain intervals are in itself quite frequent compared with other automotive gearboxes, indicating that the manufacturer already has built in some extra maintenance. a standard oil change used to be three times the engine oil change interval for a ICE. many gearboxes nowadays do not need any oil change at all,meaning that the gears will be functioning more or less correctly over a period of say 300.000 km.

 

[KiwiME]

Thanks for your comments. I'd certainly agree about the tapered rollers - it seems a puzzling choice due to the added effort required to set up a suitable preload in a production environment with the vertically-split casing. The cast ribbing around the bearings does look rigid enough however.

The Nissan Leaf reducer is very similar in design except that large ball-bearings are used at the output (diff) shaft to handle thrust. The Leaf also has two magnetic plugs (both fill and drain) and oil change videos on YouTube show exactly what you'd expect in terms of collected debris and oil colour at the specified service intervals.

So, since I last posted a helpful owner in Australia has had an oil analysis done at a mere 3,700 km. I'm no expert at interpreting these but the ferrous results are what I expected, relatively low particles but relatively high iron. That's because, without a retention magnet particles continually circulate until they are ground down to dust. I'd wonder if that's the reason for the black colouration rather than oxidation? There are no external signs that the gearboxes are overheating and I checked the oil smell, nothing out of the ordinary for gear oil.

But the shocker to myself and the car's owner was the excessive silicon and aluminium. The car was not driven on dirt roads and is nearly brand new so we can't explain the silicon. Based on the gearbox design the aluminium could only result from bearing outer race spin. A potential location and reason for that is the over-constrained situation occurring at the input shaft when the motor is mounted in place - a deficiency I've mentioned a few times in prior posts.

I'd like to hear further thoughts on the oil analysis. In the meantime myself and that owner are trialing the addition of a magnetic drain plug.

 

[3DDave]

Finely divided aluminum is black. I don't know if it is from rapid oxidation or is just that way. I have found fretting wear in aluminum mechanisms that looks exactly like graphite powder. That may not be the only contributor, but it certainly serves the purpose.

 

[TugboatEng]

Something to consider when sending oil out for analysis, it all gets filtered to 10 micron prior to testing. The smallest particle you can see is 30 micron. I have sent in oil samples with visible wear particles and they came back zero because they were removed before the test.

Some aluminum alloys have silicon in them. Bearing race rotation is always a possibility.

 

[KiwiME]

Thanks for the comments. I have two Kona EV and two Leaf EV analyses to compare. The analysis posted above is the second row of numbers, the Leafs follow.
The reason a Leaf is of interest is that the gear reducer design is nearly identical to that of the Kona. However, the Leaf uses Nissan ATF and has 2 magnetic plugs while the Kona calls out 70W/75W GL-4 and has no magnetic drain plug.

Several people have suggested that silicon could come from the silicone sealer used to seal the housing.
The Blackstone comments for the two Leafs are pasted at the top left and right under the comments for the first Kona.
I've split the image up into two so that it's more readable. Under "meter on fluid" is kms driven.

At this point I think the short answer for the Kona is to "install a magnetic drain plug at the earliest possible convenience".

 

[KiwiME]

Well, the end result is this:
1. It's not incorrect to say that dirty oil is not good thing in the gearbox of an expensive EV, especially when that's due to steel wear particles being ground into dust by bearing and gears.
2. There have been numerous gearbox failures but in no instance had any forensic analysis been carried out.
3. In most cases the complaint is a "clunking" or "clicking" noise. In three cases the final drive tapered-rollers failed.
4. I posted about the lack of a magnetic drain plug on numerous global forums and social media targeting the Hyundai Kona and eNiro, advising owners to do a single oil change at the earliest convenience.
5. About 10 owners paid attention and changed their gearbox oil. In all cases they reported that the oil came out black. Some DIYed, some used the dealer.
6. In two cases the Hyundai/Kia dealers resisted doing the work citing that it "wasn't due". One relented and billed it as a "diff oil change".
7. Several other owners questioned my truthfulness or motives.
8. One said they expected Hyundai to cover any problems that transpired. Ironically that owner had already had to replace their gearbox out of warranty at 60,000 km and had to pay the labor cost of several thousand $$.
9. I've got two magnetic plugs on mine with sparkling clean oil. I'm happy.
10. You can't save everyone.

"None so deaf as those that will not hear. None so blind as those that will not see."
- Matthew Henry

 

[gruntguru]

9. I've got two magnetic plugs on mine with sparkling clean oil. I'm happy.

. . and "hairy" magnets?

je suis charlie

 

[KiwiME]

Just to update my post for the record, summarizing the issue and adding the recent findings:

This EV (Hyundai Kona) has a simple 2-stage parallel-shaft gear reducer with a cast aluminium housing including an open differential on the output shaft, splash-lubricated with 1 litre of a 70W GL4. Input speed can be up to 10,000 RPM and the overall ratio is roughly 8:1.

A recent finding is there is a small non-serviceable ceramic magnet captured between the housing halves. Another recent discovery is that there is a circular bronze pad running (spring-loaded?) against the input shaft tail-end ball bearing inner raceway outer face. I can only assume that it's a brush intended to ground small voltages that might appear on the shaft generated by the motor. This would prevent discharge across through the bearing, intending to avoid an apparently common form of bearing damage when variable-speed drives are used. The Nissan Leaf has a pair of carbon (or sintered bronze) brushes running against the intermediate shaft. The Kona motor also has a circular carbon-fibre brush on the output shaft.

I had carried out oil changes at 19,400 and 19,500 kms, taking the opportunity to substitute the drain and fill plugs with magnetic types. The original oil was pitch-black at the first change.
About 20 other owners globally have also carried out an oil change with four obtaining a used oil analysis. Distance driven on their original oil varied between 4,000 to 100,000 kms but mostly at the lower end. All reported black oil, many with considerable sparklies noted.
UOA results (above) were relatively uniform and not wildly dissimilar to those from two Nissan Leafs. The Leaf gearbox is a similar design.

I've recently carried out an oil R&R at 21,300 kms and found the oil very clean. I just filtered it and reinstalled it.

That gives me confidence that there is not an ongoing issue, more likely that break-in wear particles are not being adequately sequestered by the built-in magnet. My tests indicate that the very fine particles don't seem to be readily attracted to a magnet and as TugboatEng pointed out, that content won't appear in the UOA iron PPM value either.

One owner reported that he had rebuilt the gearbox because the bearings where "gritty", adding that the magnet was clear of particles. I'm wondering if oil turbulence related to OEM magnet location may be a factor?

In any case the easy solution is for owners to carry out early oil changes to clear the break-in particles. I'm thinking along the lines of recommending 2,000 and 10,000 kms.

 

[3DDave]

Thanks for the update.

If the particles were from the carbon or bronze brushes they certainly would not be trapped by the magnet. I don't believe that turbulence would be enough to keep iron particles from sticking - in my experience trying to remove tiny particles from a magnet requires a tremendous effort that merely rinsing doesn't manage.

 

[KiwiME]

Since this thread is still open I'll update my progress over the last 5 months. The subject is regarding dirty gear oil found in the 2019-2022 Niro and Kona EVs.

1. I now have (11) used-oil analyses for first-time oil changes from contributing owners covering drain intervals of about 4,000 to 46,000 km. I've graphed them as PPM vs km drain interval. As you can see, the numbers are all over the place. Most report notes describe the numbers as 'high' or 'very high'. There are also (2) reports from owners who have added a pair of strong magnetic plugs, as I have myself. They indicate lower iron levels as you'd expect.

2. I did a third oil change myself at 24,000 km after a drain interval of only 4,000 km. The old oil is completely opaque with a hint of glitter. The magnetic plugs had minimal particle loading. I will add that my Kona EV runs very quietly. The new oil is a 70W GL-4 and is a clear brown colour.

3. There's still no understanding of what's going on. The best theory I have now is a bit weird, hold onto your hats! Based on a photo one owner made of his internal magnet with a borescope, it appears that it's badly worn around the edges. Compare the photo below with that above. Could the ceramic magnet be rattling around in the 'cage' due to oil flow, especially noting how that flow would interact off centre?

4. I'll just add that owner reports of gearbox 'tapping' noises while decelerating are flowing in nearly every week on forums. Both manufacturers have TSBs on rectification of this (requiring both gearbox and motor are replaced) but I'll stress that I don't have any evidence that it's related to the dirty oil.

 

[SwinnyGG]

If that magnet appears to be doing no one any good, and may be the debris source, I'd probably try taking it out at the next change and running a few k without it. If this were my car.

 

[KiwiME]

That seems logical but I'd have to remove the gearbox from the car and disassemble it completely to remove the captive magnet.

 

[KiwiME]

I think I've finally identified the root cause of the black oil and ferrous wear particle retention in the Kona/Niro electric cars.

I used an inspection camera to confirm that the internal factory magnet is indeed captured loosely in its cast aluminium pocket.
Being directly in line with the oil splash off the final drive gear it's inevitable that it will spin and rattle, just like an object being squirted with a hose.
There was very little ferrous particle loading visible despite covering 24,000 km. I'd speculate that particles caught on the magnet are being ground against the aluminium surfaces as it spins, breaking those down and creating aluminium particles as well.

A short video here where I move the magnet around by moving a stack of magnets on the outside surface. While I had the camera in place I attached those magnets where they might hold the internal magnet away from the sharp feature at the left, what I'm calling a "nib". In 4,000 km I'll check the fresh oil and see if it's clear.

The moral of the story here is that a internal magnet needs to be held securely. Most likely one attached to the drain plug is the best solution.

https://youtu.be/YquYZPTkcmI?list=PL5W-HwNCNIAIpXFl43pqCH9scrrRIm4zC

 

[SwinnyGG]

Interesting update. Be curious to see if you see a difference at your next change.

I don't know much about ceramic magnets. Would ceramic particles abraded off of the ceramic magnet retain their magnetism (and thus be captured by drain plug or other magnets in the case)?

One thought, in response to filtering and re-using oil.... oil additives do not have infinite life. Just because the oil is clean doesn't mean it's chemically the same as when new. May want to R&R the fluid even if it's clear at your next checkup.

 

[3DDave]

Having fractured ceramic magnets - yes, the particles are magnetic. I don't recall them as being particularly strong so any fluid flow would likely dislodge them. Often there are steel plates to capture and "focus" the magnetic field to a useful strength.

Also - finely divided aluminum particles are black so it makes sense if that magnet is rubbing particles loose.

 

[KiwiME]

It will be 6 months before I accumulate a few thousand kms and check the oil again. Even if it still turns black I'm happy to change it once or twice a year as it's an easy job.

I won't be filtering the oil again as I've found even an N95 mask does not filter it sufficiently.

Certainly I expect ferrous wear particles to lose their ferrous properties when 'worked' in circulation by gears and bearings but I'm not sure about those already on the magnet being ground further down by the rotation. Wear on the magnet itself seems uncertain at this point as one photo (above, someone else's EV) seems to show wear but my own inspection doesn't. It seems the material is very hard.

Since posting last I've learned more. First is that an earlier EV model, the Kia Soul EV, has a gearbox that is not quite identical to the Kona/Niro but seems like an earlier generation. Amazingly it has a rectangular magnet with ridges fitting snugly in much the same shape of cast pocket. You'd almost think that going to a round magnet was a cost savings implemented without much thought.

Also, it seems that the motor in the earlier (2018-?) Kona/Niro EVs also can have a problem with outer race spin and subsequent bearing failure at the tail-end bearing. At a glance it appears to me that the axial wavy spring was not strong enough to hold the raceway still. Video here from a Ukrainian mechanic if interested. There is also a chance, IMO, that there was electrical discharge bearing damage involved.

Lastly, Kia has revised their 2023 Niro EV model, keeping the same name but completely changing the powertrain. The motor is now a 'wet' type with oil used directly for stator cooling instead of the coolant jacket used previously. The integration with the gearbox has been changed to eliminate the bearing 'overconstraint' deficiency that I had mentioned about a year ago in the post I've referenced at the top.