2021 - Latest models CVT issues?

FelineFreddie said:

@GraemeHarrison …
…another tractor manufacturer using CVTs.

Dealer Locator

www.caseih.com

Click to expand...

Yes, it's great to see that Case International Harvester has managed to add a CVT to its tractor lineup. The link you provided notes that it is a "mechanical and hydrostatic" system. In other words, CaseIH were smart in deciding to stay completely away from use of chains! As the video on that link shows, the CaseIH system uses fixed gear ratios for each gear, a sun/planetary gear arrangement and drives a hydrostatic system that 'boosts' the system 'between gears'. That gearbox may be called a CVT, but it uses entirely different technology to Subaru's approach. The main friction surfaces are the teeth of the gears. Your 'evidence' supports my claim that a CVT like Subaru's design cannot handle large torque and needs its gearbox oil changed far more regularly than Subaru advises... and the more torque x hours-used (eg towing or hills) the more frequently the user ought change that gearbox oil.

So what is the basis for why Subaru CVTs won't last anywhere near as long as the excellent 4EAT automatic gearbox they replaced? Just do the mind experiment of pulling apart that thick chain assembly inside the Subaru CVT. It is like six large motorbike chains, bonded side-by-side to make a stronger chain to 'pull' a car during heavy acceleration, hill climb or towing. Each of the hundreds of spindles of a chain 'links' are a cylinder of metal being 'dragged' around a central pin, as the chain flexes. Then each of the side-plates of each link rub against the side-plate of the next link, for the considerable portion where the plates overlap. Then one needs to consider the 'teeth' that protrude outwards from the hydraulically-expanded cylinders. Each time the teeth come up into the chain, they must self-locate into a gap between the pins of the chain. Because each of the 'drive' and 'driven' cylinders are constantly expanding or contracting radius, there is no 'natural' recurring fit of chain to teeth (as in a fixed-gear bicycle) but the micro-location of teeth to chain-link is always on the move - so fresh micro-surfaces abrade against each other, with every small adjustment in overall effective gear-ratio. Plus I suspect there is even more metal-to-metal contact with torsional forces. I suspect that with the expanding/contracting cylinders, that not every chain gap receives a tooth. In other words, of the six or so side-by-side chains, on any row across the wide chain, only some gaps receive teeth at a given frozen point in time, for any particular gearing ratio. This means that if some links are being pulled, the ones to its side are being twisted diagonally, as they are not being directly pulled along, but are being pulled by the link to their side. This means the side-plates are trying to twist 'away from the force' that is applied to the portion of the chain under load. This must increase micro-abrasion of side-plate to side-plate (metal-to-metal) scraping. Then once one has added up the total area of metal-to-metal contact, of chain pins, sideplates-to-pins, teeth-to-pins, and sideplates-to-sideplates, one gets an idea of the amount of total metal-to-metal scraping involved in a chain-based CVT. Then compare this to the areas of wear in a traditional fixed-gear gearbox. In a traditional gearbox, at any one point in time, only two sets of gears are enmeshed, with the others spinning disconnected. Of those two enmeshed gears, only a 1-3mm band across the width of the teeth are physically rubbing. The total area of metal-to-metal contact with fixed gears is FAR FAR LESS with fixed gears than with a chain (or six chains side-by-side). There are also bearings fixing the gear-shafts in place in each type of gearbox, so bearings play only a minuscule difference between the two gearbox technologies. Besides, the ball bearings are small tungsten steel balls that 'roll' (not drag) over other metal surfaces, so bearings can be rated for up to 1b revolutions under load, if well-lubricated. As I noted previously, the bearings in CVT fail (mine have) as a result of the chain shedding a lot of microscopic steel filings into the oil. On the issue of 'friction surfaces', a clutch is typically located outside the gearbox enclosure, so that the material 'shed' from the friction surfaces does not get circulated in the lubricating oil. Some motorbikes have wet-clutch systems, but require more frequent oil changes as a consequence. With automatic gearboxes, the extent to which friction plates are used determines how frequently the gearbox oil needs changing. The Subaru CVT contains a slippage component that provides the hill-hold capability, and which (once worn) gives rise to the unexpected 'creep' problem that is well-documented. That component may well contribute to the contamination of the gearbox oil. That element does not exist in the manual gearbox, by comparison. So, in summary, it is almost impossible to estimate total area of metal-to-metal micro-abrasion in the CVT. But just considering the hundreds of chain pins, hundreds of chain side-plates, tens of gear teeth, it is at least one order of magnitude larger than for fixed-gear gearbox. That is why I estimated it at 10-40x greater wear surfaces in earlier post. But once you have seen metal-heavy oil that came out of a CVT, as an engineer, you 'know' there is a problem with metal-to-metal friction in such gearbox design. I stand by my views on the CVT design shortcomings, as expressed earlier.

But interesting questions I'd love a Subaru engineer to answer are:
1. Can one easily swap-in a 4EATS (earlier fixed-gear 4-speed automatic) or manual gearbox to replace a CVT, with the overall electronics of car not objecting so strongly as to make car undriveable?
2. Would a teflon-based lubricant (Nulon N100 gearbox friction-reduction additive) be able to dramatically cut wear in CVT gearbox to prolong life. This product adds sub-micron PTFE (teflon) balls that get pressed into micro-cavities of metal surface, and roll around between metal surfaces, thereby preventing metal-to-metal abrasion. [I have no connection to Nulon, but have used this product in everything from older tractors to 4EATS gearbox with great extension of gearbox life.]
3. In light of now-documented high-metal content in CVT oil as these cars age, what is fresh recommendation for how frequently the CVT oil ought be changed?

Subaru should not be allowed to remain silent on these issues. There are clearly quite competent engineers within Subaru who are aware of the problems, and are being muffled - prevented from speaking out as to work-arounds, or improved service approach to dealing with CVT wear.

As to 'playing the man', if you wish to continue to challenge my qualifications, rather than the validity of what I have presented, you can email me directly on 'PROF at-symbol POST dot HARVARD dot EDU' citing 'Forester' in subject, and I will reply promptly, confirming that this email address is indeed mine. The email goes to a Harvard server, and is simply a life-long courtesy email address for me as ex-Harvard.

GraemeHarrison said:

So what is the basis for why Subaru CVTs won't last anywhere near as long as the excellent 4EAT automatic gearbox they replaced? Just do the mind experiment of pulling apart that thick chain assembly inside the Subaru CVT... The total area of metal-to-metal contact with fixed gears is FAR FAR LESS with fixed gears than with a chain (or six chains side-by-side)... as a result of the chain shedding a lot of microscopic steel filings into the oil.

Click to expand...

I forgot to mention as part of the 'proof' that chains are not good operating inside oil used for lubricating other surfaces, that Subaru ABANDONED using 'timing chains' - metal chains from main crank to operate valves, By the time of my 2.5-litre Forester petrol engine, Subaru had decided to use synthetic belts for that task. Now people can argue as to why Subaru would switch away from having a chain within the engine oil system, but I say that the reason is that the chain sheds too many micro-filings into the oil that is then lubricating the main bearings etc. At least with a synthetic belt, what micro-plastics are worn off its surface over time, such plastics will not 'cause wear' on the metal surfaces being lubricated.

So that is a decision, clearly made by Subaru against continued use of an existing chain. And remember that the engine oil is being replaced 10x more frequently than the gearbox oil, so the issue of chains releasing micro-filings is far more important in the CVT, than in the engine.

When reading about the stronger metal belt used by Subaru (as opposed to the flawed rubber belt used by Nisson), in my mind I always assumed it was also just a very tight metal belt utilizing friction to do the power transfer. I knew the metal belt would have to technically be more chain like with some sort of linking in order to bend and work as a belt replacement. But it never occurred to me the mechanism included many grabbing spikes on the "gears" of the CVT.
It makes sense this would work better to avoid slipping of only using friction. But it also makes sense this would create a lot more metal on metal interaction.
So the intuitive idea of this causing more metal shavings is logical. However, real world results from data (not just anecdotal) do not yet seem to support a firm conclusion...yet. Real world results do occasionally have a tendency to not match logical assumptions based on lab and design work. Which that uncertainty is why every auto company had huge research divisions collecting data from vehicles in service.

YoGeorge said:

Your ignorance completely disqualifies you from any intelligent discussion here.

Click to expand...

YoGeorge, you are unfairly seeking to shut down discussion on CVT issue. I did foolishly reference chain-v-belt issue of motor, but immediately recanted, noting it was a red-herring, as not handling the whole torque of the motor. You can't exclude someone from a discussion for one point they raised and promptly recanted. It seems Fozzy & you are 'Subaru apologists', seeking to shut down discussion of Subaru problems. I notice that when another improperly said "What about Case having done a CVT on a tractor?", you did not insist that they be excluded from making comments, when I showed that the Case CVT did not use any chain, but was based primarily on fixed gears. You need to be measured and impartial in your claims of who should enjoy freedom of speech.

The issue of CVT failures and CVT oil contamination remains an issue. The apologists claim that there is nothing to see here, and no CVT problems exist. But I now see that a humble Australian home vlogger has 85k 'views' of his piece about how Subaru CVTs have failed in Australia, and how Subaru Australia have done apparently nothing to help the affected customers, including failing to extend the warranty as Subaru of America did. This is a large number of views in a country of only 25m, given the title of the YouTube means one would need to be looking for 'CVT problems' to find the video:


You guys really need to move on from who you like, who ought to be able to comment... and move on to real CVT issue discussion.

Firstly, I'd like FozzyX3, YoGeorge, Kean & FelineFreddy to each disclose 'ANY' association (financial or otherwise) with Subaru, its dealers and its partners/contractors, as there seems more than an end-user devotion to defending Subaru, even in technical discussion of CVT wear/life. I made a complete declaration earlier that I have nil association with any competitor of Subaru, and confirm that for two decades I was a strong advocate of Subaru, but that I feel Subaru has gone against its USP (unique selling point) of high reliability and good resale value in AWD, but forcibly switching customers from Automatic gearboxes to CVT gearboxes with their chain systems.

As for FelineFreddy's comment "I would have thought someone making technical-based arguments re: belts .vs. chains wouldn’t have made such a glaring mistake in knowing what Subaru uses for valve timing…especially when criticizing or critiquing a device’s suitability for application", let me state that Subaru to this day uses a mixture of chains and belts for timing in its engines, based on model, so the matter is not as clear as one might think. But more importantly, as soon as I made a reference to timing belt/chain, I dismissed it as a red-herring, as the full torque of the engine is not applied, whereas in the CVT, the full torque of the engine is applied to the chain used there.

If one is to exclude someone for a 'glaring mistake' it is more appropriate to recall that I said the chain was not used by heavy vehicle manufacturers for high-torque power situations, to which an item of farm machinery was brought to my attention. But that 'CVT' was not driven by any chain or belt, but rather used 'permanently enmeshed' gearing in a sun/planetary arrangement. All it requires is that one have a source of power to drive the planetary gears. The Mk1 Toyota Prius used that system for its CVT, where ICE engine drove the sun-gear, and a separate electric motor span the enmeshed planetary gear, with the drives to powered wheels connected to outer ring-gear that enmeshed with only the planetary gear. So if planetary gear was powered in same direction as ICE, both combined to give higher-gearing output to wheels, but if planetary gear was powered in opposite direction to ICE, the opposing forces gave a low-gear effect, slow progression of wheels-on-ground without stalling ICE. One problem with that design was that reverse power was limited to only what the electric motors could deliver (ie no way to use ICE to power in reverse).

I have now talked to a wide variety of Subaru wreckers. All of them are aware of the failures in Subaru CVTs. One said he had a whole row of the earlier automatic transmissions ('4EATS' that were displaced by CVTs) as those were super-reliable and 'no-one needs replacements'. But they charge 4x as much for the CVT gearbox, as they are in demand. Now, one could argue that CVT boxes are 'in demand' for some reason beside them failing... but I can't imagine what that other reason could be. And when I spoke to the transmission expert who swapped out my CVT for a low-mileage one from a wrecker, he noted that one needs to change the transmission fluid in ALL transmissions, for them to last - particularly CVT transmissions. He noted that when a manufacturer says that something does not need to be serviced "for the life of the car", they actually mean "for the life of the warranty" as the car manufacturer is quite happy for users to come back and buy a newer model, if car got to end of warranty. Another keen mechanic with many classic cars in his garage said, "Some years back Mercedes extended their distance between oil changes from <10,000km to 20,000km and do you know what technology they employed to give that longer service life? Nothing! They had simply learned that their typical new car buyer did not want to be hassled to bring their car in for service as often. So the manufacturer just extended distances between services, knowing those first owners of the cars will have traded them in, before the shorter life of the engine will have impacted them." So the issue is not limited to Subaru, but I do believe that Subaru ought act on the knowledge it does have about CVT failure rates, and state that the CVT oil MUST be changed every 30,000-40,000km. They extended the warranty (only in the USA), and should go further and update the advice as to service regimen to match. I'm sure part of their concern is that the prior gearboxes had dip-sticks for easy checking of transmission fluid, whereas CVT has no dipstick, so oil must be dumped and hand-pumped back in... and the specialty oil is very expensive at circa 10x the cost per unit of engine oil.

I also spoke to Nulon, who make a wide range of friction-reducing additives for oils, mainly PTFE (teflon) in sub-micron balls. I've used their gearbox products in gearboxes including a 1962 Massey-Ferguson tractor, to great effect. But they do not have any product for CVT, as adjusting the friction could impact on how the device works. Nulon noted that the manufacturer would need to do the testing to see if Nulon's auto-gearbox teflon additive (G60) might lessen metal-to-metal wear in the CVT gearbox. I think pressure should be brought to bear on Subaru to do such tests as are necessary to find a friction-reduction solution for the CVT, as the problem will not remedy itself. And just like those TEPco engineers who knew it was 'crazy' to put the Fukushima stand-by generators on the outside of the tsunami seawall, no good can come from failing to speak up, when you know there is a problem not being addressed!