Repair issues and fixes

[KiwiME]

The disc magnet (x2 @ 3mm thick each) behaves better than the rectangular variety but oddly still sticks with either pole orientation. It feels like the internal magnet is shifting to accommodate the change, photos show alternate poles. As I dithered around in the quiet of the evening I could actually hear the internal magnet moving.

I'm becoming more determined to try and get the inspection camera working because it would answer questions that won't be resolved with the Hall sensor. I already know the magnet must be oscillating during operation; more importantly I'd like to know which pole orientation pulls the internal magnet down in the hope that it won't move around as much.

 

[electriceddy]

which pole orientation pulls the internal magnet down in the hope that it won't move around as much.

"Most neodymium disc magnets are axially magnetised. This means the poles are located above and below on the circular surfaces.
They are so-called raw magnets where the magnetic flux lines are dispersed in the air around the magnet (in contrast to pot magnets). This large areal distribution allows for attraction of metallic objects from a large distance."
So, direction should not matter, it will attract from any flat side

 

[hobbit]

I'm still following along here, but busy with other stuff that as a side effect has me racking up a lot of travel
miles, so I might get to my next oil-change point that much sooner. FTR, my inspection camera has a
right-angle view adapter to attach to the end and try to give more "borescope" functionality, but it's
worse than useless as it splashes so much of the camera module's own "ring light" back into the image
that it basically washes everything out. I might come up with something to mask off parts of that and
try to get better views of the magnet...

_H*

 

[KiwiME]

... it's worse than useless as it splashes so much of the camera module's own "ring light" back into the image
that it basically washes everything out.

Same with my setup. But today I purchased the brightest white 5mm LED that I could buy off the shelf locally and built this. The separate illumination should take care of the glare problem. Diffusing the narrow beam of the LED was the only issue. Will try it out tomorrow with luck.

 

[KiwiME]

Here's my video from today's experimenting. It took me all day to do the work then download and edit the huge files. I have cut out a lot of footage to keep it somewhat concise and to the point, but it's still a bit tedious. Sorry, there's no sound, although that's probably a good thing from what I normally see on YouTube.
I think it's self-explanatory, if not ask.

 

[electriceddy]

Here's my video from today's experimenting. It took me all day to do the work then download and edit the huge files. I have cut out a lot of footage to keep it somewhat concise and to the point, but it's still a bit tedious. Sorry, there's no sound, although that's probably a good thing from what I normally see on YouTube.
I think it's self-explanatory, if not ask.

You certainly gave it your best shot at releasing the monster, but its cage is still intact. This does indeed show if it was freed form captivity, how much damage could result. Well documented. It definitely looks as though the "stop" is located at one edge of one of the casts, and the "holding pen" is located in the other as you previously suggested, and would indeed cause rattling and upon escaping, breakup and damage to the gears themselves and their associated bearings.
It's starting to look like the only cure would be to remove the GRU assembly, disassemble and remove the magnet for good, or in the least- provide a sure-fire way to prevent movement using external holding magnets in the right location.
And the true story continues....

 

[KiwiME]

I've realised that by attaching the external magnets I've passed on the opportunity to prove that the internal magnet is indeed rattling during operation. But in the end I'm already certain that's what's happening. The external magnets overwhelm the range of the Hall sensor aside from influencing the original behaviour of the internal magnet. I wasn't looking forward to building the required kit to support the Hall sensor and I can't drive around with an oscilloscope on my lap.

Mitigation would be the next step anyway and I've just jumped ahead to that. In a few thousand kms I hope to find the oil is still clean. But even if the external magnets work for me I think that it's a complicated fix to deploy fleet-wide because the internal magnet could be installed in either orientation (north/south) and I found it hard to detect from feel which orientation of the external magnets produced the desired attraction without being able to actually verify it with the camera.

Notably the rattling internal magnet seemed to be mostly free of ferrous particles (except for a few visible in the centre, photo below) presumably because the wiping action removes any that happen to stick. That explains why a Votex or equivalent does help reduce iron levels.

I should note that the external magnet Ii attached will not stop the internal magnet from spinning, only encourage it to stay in one place translationally. The fields of both are toroidal and so relative rotation about the common axis doesn't produce any change in the combined field.

Hyundai could fix this at zero cost simply by deleting the internal magnet and using a magnetic plug. That they haven't done this leads me to believe they don't know about the problem.

For everyone else, especially those out of warranty, all I can suggest is that an oil change every service or two will go a long way to help.


The diagonal line near the lower-middle is the silicone gasket. Left of the magnet is the "nib" and the magnet's housing is to the right.

 

[R P]

You've done a lot of research and testing with this issue, and I commend you for that. But your complete focus has been on the gearbox and what eventually causes the tick tick issue. I can't argue that a loose magnet is not a bad thing. However, I am not sure that is the whole problem.

In my case, I had the gearbox replaced after the tick tick sound. Then within a few hundred kms it came back even with a new gearbox. The old gearbox lasted almost 20K kms. The dealer then replaced the motor, and the tick tick went away, for good, or at least as long as I had the car (another 45K kms). And this has happened with others as well. Unless you replace both, the tick tick comes back. So if the gearbox caused this issue, why did it not finally go away on my car until after the motor was replaced?

 

[KiwiME]

... But your complete focus has been on the gearbox and what eventually causes the tick tick issue.

My focus has been exclusively on what causes the blackened and contaminated oil, the only issue my Kona has.

I would not be able to investigate the motor rumble / tapping noise without having a failed gearbox/motor assembly on hand to analyse, and that's unlikely to happen. Without that all I can do is theorise the chain of events that could connect the two defects. However, I'm confident at this point that the loose magnet, inherent in this gearbox design and placed where it is, is responsible for the dirty oil and I probably won't study this much further. If my "fix" turns out to work I'll make everyone aware of that.

The apparent suspected cause of the tapping noise per Hyundai/Kia's TSBs is motor/gearbox misalignment, however it's worth noting that this sort of technology is bread-and-butter stuff for high-volume automakers. Most if not all automotive drivetrains have multiple locations where a splined shaft coupling is used. It's inconceivable to me that Hyundai could mess up such a basic design and production detail, and for such a long time. More likely, in my opinion the chain of events is more elusive and takes a long time to play out. That most examples have black oil yet very few suffer the tapping noise tell me that there is a probability component to this and that could be down to the progression of bearing raceway wear after being initiated by the dirty oil.

 

[R P]

The apparent suspected cause of the tapping noise per Hyundai/Kia's TSBs is motor/gearbox misalignment, however it's worth noting that this sort of technology is bread-and-butter stuff for high-volume automakers. Most if not all automotive drivetrains have multiple locations where a splined shaft coupling is used. It's inconceivable to me that Hyundai could mess up such a basic design and production detail, and for such a long time. More likely, in my opinion the chain of events is more elusive and takes a long time to play out. That most examples have black oil yet very few suffer the tapping noise tell me that there is a probability component to this and that could be down to the progression of bearing raceway wear after being initiated by the dirty oil.

I think too, that the tapping noise issue had to do with the motor/gearbox alignment/clamping. The new motor had a new part number and I believe the diff was longer (maybe stronger) clamping bolts and changes to the flanges.

It would give us more confidence if Hyundai would come out and explain better what the issue and fix was. While I liked how my Kona drove and pretty well everything about it, it did leave me some concerns about long term reliability with these unresolved/unexplained issues.

 

[Ferenc Jakab]

Here's my video from today's experimenting. It took me all day to do the work then download and edit the huge files. I have cut out a lot of footage to keep it somewhat concise and to the point, but it's still a bit tedious. Sorry, there's no sound, although that's probably a good thing from what I normally see on YouTube.
I think it's self-explanatory, if not ask.

It's definitely loose, moving around. Love your persistence. Wonder if it's still doing it in my 2022 model?

 

[rgthoma]

**Just wanted to provide an update. I took the car in to the dealer for the 5000 mile service and mentioned the "wheel of fortune" noise. The dealer seemingly went through the normal protocol of replacing the brackets first before diagnosing the traction motor. I got a call yesterday indicating the traction motor needs to be replaced but that particular part in on backorder with no ETA. So Hyundai is offering a rental car through a national chain but does not cover the entire cost. Apparently Hyundai covers $44/day and the daily cost through their authorized rental company is $45/day. Oh and a car isn't available until next week! Wow! It will be curious to see how long this takes to get repaired. I believe the lemon laws in my state say that repairs should be made within 30 days. Of course the current supply chain issues might come into play I suppose. I like the car but certainly don't like paying for the car loan, plus a very small fraction of the loaner car, plus the fuel while the EV sits waiting for parts.

Here is another update. I got my car back last night. 29 days in the shop! Here is a copy of the work that was done.

 

[KiwiME]

It's definitely loose, moving around. Love your persistence. Wonder if it's still doing it in my 2022 model?

I'm known for persistence and long ago my employer got a letter from one of our customers, Lotus Cars, commending me for how diligent I was while repaired and calibrating one of their engine development dynos. Engineering is all about getting every detail right. One mistake and you get blamed for everything that ever happens with anything remotely connected with the product. Get everything right and no one notices. Did we notice that no other EVs have gearbox issues?

I'd be confident that your 2022 has the same issue. The suggestion I made today on the global FB group are:

First oil change at 4,000 km (2,500 miles).
Subsequent oil changes at every second service.
​

If cost was an issue I'd skip the blue coolant change in favour of a few gear oil changes.

 

[electriceddy]

First oil change at 4,000 km (2,500 miles).
Subsequent oil changes at every second service.

Jiffy Lube should jump all over that one
At $160/HR for that service at the Hyundai dealership, one would be able to afford ~3 of them for the same cost, especially if you include coupons.

The Top Automotive Oil Change Franchises of 2022
1. Jiffy Lube International
Jiffy Lube International is the main player in the franchise segment of the quick-lube chain industry. Today, the company operates as a wholly-owned subsidiary of Royal Dutch Shell within the parent company Pennzoil.

 

[Ferenc Jakab]

yeah, I got onto this thread too late for the 4000. I'm at 13000 so I'll have the dealer do it plus the magnetic plug at the 15000. Too arthritic to get under the car myself unless absolutely unavoidable.

 

[KiwiME]

... the progression of bearing raceway wear after being initiated by the dirty oil.

As a related note, I placed a relevant post on the main Kona FB group which attracted the usual amount of temporary interest. As supporting documentation I posted a photo of the internal gears (from the Polish Ioniq at ioniqforum.com) and one person questioned some minor-looking damage in one area. I initially dismissed it as being irrelevant to the subject but after revisiting it, I think there is something interesting.

The area is the gearbox input shaft where it's supported by the gearbox bearing closest to where the motor connects via a splined coupling. The photo shows burnishing and heating from what is clearly the shaft turning inside the bearing inner raceway.

Normally turning here (technically "inner raceway spin") would not happen, because under radial loading the bearing rolling friction should be much less than that between the bearing inside diameter and the pinion shaft. However there are three conditions present that could upset this balance:

(a) Moderate bearing radial loading is not guaranteed to be present all the time because the motor bearing close by provides redundant radial support via the splined coupling when that coupling is under torque. Splined connections act like a solid shaft when loaded. This is the "geometric over-constraint" design error I've mentioned in the past. The Nissan Leaf and other older-gen EVs also use this sort of design feature and it's a recipe for problems. No new EVs do this now, not even the Ioniq 5. All have removed this bearing entirely!

(b) Entry of particles or other debris into the bearing ball tracks can momentarily increase rolling friction.

(c) Thick oil due to cold weather will increase rolling friction until it warms up. This shaft turns at 7.891 times wheel speed.

Whenever turning friction at the damaged area in the photo drops lower than that of the bearing itself, chances are turning will happen here instead of the bearing and the shaft is not well-enough lubricated to avoid the type of damage we see in the photo. Eventually a radial clearance can develop and that could result in a knock whenever the torque at the splined coupling is reduced to low level where the motor bearing won't also rigidly support the gearbox input shaft.

 

[Ferenc Jakab]

The legendary knock!

 

[rgthoma]

Do you think the support brackets installed per the TSB provides any relief to this problem?

 

[KiwiME]

The legendary knock!

Yes, and that movement could damage the spline on the motor shaft which is why that needs replacing as well.

Do you think the support brackets installed per the TSB provides any relief to this problem?

I can't see that it would despite that it's the first step in the TSB. Changes like this are typically focused on "NHV" (noise, harshness and vibration), generally meaning audio-frequency resonances in the structural elements. It could mitigate the high-pitched noises that a number of owners complained out years ago. Motor field vibrations could have excited that structure.

The vast majority of older Konas on the road manage just fine without the revised bracket and bolts and that says a lot.

 

[electriceddy]

The vast majority of older Konas on the road manage just fine without the revised bracket and bolts and that says a lot.

Newer models also, it seems not every car is created equal. I had "the knock" on my 2018 MY, no signs yet on the 21MY.
Changing the oil will probably help to contribute to less probability, in conjunction with that...I also reduced the regen to level 1, knowing that was also a contributor to additional stress on the drive system. It does bump to level 2 on occasion using the auto regen function, but only happens periodically.