[pboglio]

I'm about to start on my CVT replacement/upgrade and figured I'd document the process in this thread. I'll layout my build plan which'll include a brand new CVT transmission replacement but also various bolt-on power mod upgrades. The list will change since I have to inspect the car. If items are worn, they'll get replaced at this point or just upgraded. This will also include a BOM for parts purchases and as many pics as possible. The CVT transmission teardown is going to direct me on where I "may" be upgrading the new CVT as there are several internal improvements to be made to increase the strength and reliability.

Goal:

Basic goal is to get the Juke AWD CVT to withstand more torque from the additional modifications I have planned. I'm realistic about the Juke CVT, it'll never be as reliable at torque levels that the 6spd RS cars are running, so the goals are going to be conservative. I could trade up to an RS 6spd, but I won't be doing that. I considered it but I'm already in another car payment and the time isn't right, plus I need an automatic for the daily. A stock CVT trans can hold 210 lb-ft @ wheels no problem for up to 75k miles and potentially longer if meticulously maintained and tuned correctly. Typically but not always a stock CVT will start slipping at the 245 lb-ft @ wheel level which isn't a good thing for CVT belt life. I suspect those with RS models may have higher torque slippage values due to differences in TCM tuning from the factory. The ECUTek flash tunes may also have some part to play but I can't verify that. Mine never slipped until the belt wear went critical due to mileage. The goal is a reliable 230- 245 lb-ft wheel torque level and it must hold trouble free for +50k miles. That is more than plenty for a daily driver.

I could pay Level10 to provide an upgraded CVT transmission, but that's not my plan. The (1) member (i.e. ARF) I know of who has completely upgraded the CVT seems to have found some reliability at elevated torque levels so it's possible. I'm going to attempt to tie-in as many of the CVT improvements as possible in a DIY project. ANY marginal improvement in reliability and torque capacity I'll take. I'll be utilizing as much of these "off-shelf" upgrades as possible. The valve-body may eventually get sent out for upgrade as well but I can always do that at a later date since it's a simple removal/install procedure on the car.

The Mamba 19T turbo, 2J Racing downpipe, High flow cat/Injen midpipe, 2J 255 LPH fuel pump, and an FMIC will get installed as well if I'm not pushed for time. There are so many parts getting removed for the CVT trans that the extra labor isn't much different. I'll also be replacing the upper & lower radiator support, & bumper crash beam due to light corrosion.

Backstory:

Back in March I started having drive-ability issues with my AWD CVT at 75k miles in terms of belt slippage and rpm hunting while driving. I had the basic bolt-on mods on my signature and was running these mods since about 2012. I'm not easy on my transmission but I also never launch it. After alot of diagnostic testing I ended up taking the risk and replacing just the CVT valve-body and switched to AMSOIL and this temporarily fixed most of these mechanical issues but a few remained. When I did it I almost knew right away I should have bought a new CVT for just another $1200 more, but it was done. Eventually the plan was to swap a new CVT anyway next year but this would buy me some time. Then at 80k miles I started getting bored at stock power levels and I decided the trans could take some upgrades, which probably wasn't the smartest idea. The additional torque going from a stock (the car was detuned at this point) 236 N-m to 320 N-m ended up being too much for my worn out transmission. It was running great for a couple days then it gave (1) massive slip during an awesome 3rd gear pull near 6000 rpms, at which point I decided to crank the boost back down. Well, 2 days later the CVT catastrophically failed while heading to work. There was no indication or slippage, just a roll into the throttle and then a "free" revving engine in all gears. I have had many transmission failures on modified cars and this is the nature of the beast. I kept my momentum up and maneuvered around a couple of cars and picked a nice spot to coast the Juke and it finally came to rest. I ended up calling a tow truck and had to make a quick decision on the side of the road to divert the tow truck to my house instead of the dealership. I had already priced out the CVT parts vs. Dealer installed and it was looking like $2500 vs. approx $4500. At my house the driver backed the flatbed up and perfectly rolled her into position neatly inside my (2) car garage. The Juke is now sitting about 20" up in the air on my QuickJack car lift with the front clip removed and the teardown process has begun.

Progress:

To start off, here is the Juke raised in the air 20" on the Quick Jack 7000SLX with wheel ramps as backup supports. The QuickJack is sitting on it's "safety-lockout bars" and not hydraulic pressure. The Lift is rated for 7,000 lbs so the 3,160 lb curb weight is hardly stressing the lift. The rating mostly pertains to the hydraulic unit as the frame itself can lift 21,000 lbs. Jack stands cannot be used as safety backups on this type of lift unless they are pre-loaded against the chassis, neither can crib blocks under the wheels. The risk of vehicle collapse is low but still there. I initially had 6" steel pedestal risers for the lift but it proved much too unstable to trust while wrenching on the car. I also had medium durometer "rubber" riser blocks but they were collapsing on the inside under the pressure load from the vehicle weight. I replaced them with solid pine wood 2"x4" to raise the high durometer "uni-body pinch-weld" polyurethane blocks enough to get correct vehicle & lift clearances. This solution is working great and the mounts are not flexing whatsoever. When the front wheels get removed the car will have wheels lying flat under the unibody to catch a potential collapse.

The other pics is my FMIC design with the Setrab CVT cooler/integrated fan. This'll go in place of the stock side mounted FMIC. I have the Nismo RS bumper grill insert illustrated to improve radiator cooling I'm going to lose with the FMIC. I may switch to the treadstone 22"x 6"x 3.5" core as an alternate as well to avoid having to cut/weld the Garrett core endtanks.

I've pulled the front clip and today I'm starting on the driveshaft/propeller shaft removal from the transfer case. I'll also be pulling the battery/intake/etc. to gain access to the (4) topside CVT bell housing bolts. The car has some corrosion so the "bottom-side" will be challenging. Next after that will be removing the right wheel, partial suspension, then drivers axle and transfer case.

 

[JuklearWinter]

Its come full circle. I hate that you're having to replace it after all your research and experimenting but I am really looking forward to watching you catalog the process. Hope everything goes smoothly!

 

[pboglio]

Thanks. It was depressing for about 1 day, but the tranny was on it's last legs and it was a matter of time and I was pushing it anyway. I'll have some good links to post with sourcing, pricing, time schedule and all that. Luckily I have the equipment and tools to handle this no problem. I'm planning for (3) months or better. I'll then follow up with some good dyno runs but I might be conservative on the torque since I'd like to dyno what I drive everyday.

I actually like driving a CVT car so that's why I'm keeping it and starting the build.

 

[squirtbrnr]

I will be watching this thread for sure. I like to know how things work and your analysis of the CVT will be quite intriguing to me.

On a side note, depending on how corroded your front crash bar is, since you’re replacing it, can I buy it from you?

 

[Macgyver]

Ditto. Its too bad that all this happened but now you know and we know since you are so kind in sharing your rebuild.

Will you use Amsoil from day one or the NS2 ?

 

[pboglio]

Squirtbnr, I'll let you know if I'll put it up for sale.

Macgyver,

I'll be running Amsoil from the get go. The AMSOIL has proven on my car to hold torque very well, even on my completely worn out belt. I'm looking at the Bosch 901047 belt rated at 350-400 N-m (257 - 294 lb-ft). I need to confirm the OEM belt is the Bosch 901066 before buying anything, the teardown should help here.

I just dropped the Injen midpipe/testpipe. Doesn't sound like much but it was 2 hours of cutting off the bolts. The Injen exhaust gasket gave up pretty good causing a big exhaust leak. Injen says they used 304 S.S. tube but it looks like 400 series to me based on the extensive surface corrosion. I now have a straight shot at disconnecting the driveshaft with alot more room to work as well. The catback and driveshaft will not be dropped as it's unnecessary. This is a huge time saver as well.

However, the front subframe is going to be hell. I need to disconnect the front sway bar end links, disconnect the steering rack and tie rods, disconnect the transverse link from the front knuckles, pull the axles, swing the knuckles out of the way, etc. Then finally remove like maybe 9 bolts to drop the front subframe onto a jack. If the bolts aren't corroded in place, this'll go good. The transfer case and CVT should then be fairly straightforward. I'm figuring maybe 5-7 working days if nothing is frozen in place, then the trans is on the workbench for teardown. I usually take my time on these tranny swaps since I'm not in a rush. I'll take detailed pictures of this part since it's kinda tricky.

 

[Macgyver]

I saw the rust on my 2012. So I took matters on and my present Juke. I treated the underside immediately. It is not gonna rust out thats for sure. If its not painted and rust treated. Its got Fluid film on it every year.

 

[pboglio]

Smart move on the rust preventation. My Unibody....looks really good actually. I had to cut the mid pipe exhaust bolts, swaybar endlink, knuckle/endlink ball joint, wheel nut completly off. That was just on the drivers suspension. All the front subframe..... bolts are hanging loose at this point.....they were not rusted on and easy to break loose. The drivers axle slid right out of the hub, usually a problem area. Waiting to pop it from the transmission tomorrow. Alot worse then I expected but the dremmel... with a cutoff wheel is a huge time saver. My 500 lb-ft impact was no match for the wheel nut. I had to cut 2 slices to weaken the nut......then impact it off. Works everytime. Subframe should drop by this weekend along with the CVT trans and transfer case. I'll be getting new wheel hubs, new axles, new front end links, new tie rods, new sway bar end links and new struts/springs, new brake rotors/pads. I'll have to refurbish the front knuckles......had an end link ball joint seize up.....gotta cut that out too. I'm gonna repaint the subframe as well, moderate rust but it's thick sheetmetal and costs nearly $900 new. Overall, progress is moving along after 1 week. I need to start cleaning up the work area when the right side suspension is completed. Then I'll setup a teardown station and a separate rebuild station for the cvt tranny using some folding tables. Finally I'll put the parts order list together and start spending some money.

 

[Macgyver]

Dayum. You got a ton to do. But at least you are doing it right.

Are RS parts any better ? I know some are beefier and from Nissan they may be close to the same price? Just a hint.

I am doing my Spoon Sub frame busings soon. Should be interesting.

Oh btw on your Quote. I like a target rich environment. Haha.

 

[pboglio]

I'm down for any RS parts for sure. If you have any ideas please let me know.

There is a bunch to do but I also have 3 months. I'm looking at some suspension upgrades as well, maybe the Bilstein B14 coilovers. My EVO X runs Bilsteins and they are nice, handles tight but still has some compliance. I'd like to reduce some bodyroll but can't find any front sway bar upgrades for the Juke, do you know of any?
I like that quote, it's one of my favorite lines.

 

[Macgyver]

The pics of your lower part of the radiator. Why does it look funky >?

 

[JuklearWinter]

I one-million percent cosign with Mac on the Spoon Rigid Collars. You can easily find my detailed review on them because that topic only has a handful of comments, but they are my favorite nin-vroom mod so far (probably my third favorite behind the tune and manny mounts). Zero body roll, torque steer is gone, the whole car just does what you tell it to do. I'm really not that interested in doing coilovers or bars/braces anymore.

I know Keoke put some Bilstens in recently on the topic of shocks.

So it sounds like you're going to stick to spec as much as makes sense, but will continue to find opportunities to make improvements to the OEM setup?

I'm encouraged to see your layout for the convection parts, because this is exactly how I planned to set my CVT cooler and FMIC up if the 2J ever materializes.

 

[pboglio]

MacGyver,

Condensor core has some bug smash goin on. Ugly but not enough for me to discharge the A/C system to replace.

Juklear Winter/MacGyver,
Interesting, I'll put the Spoon Rigid Subframe mounts on my list then if they are that good. Actually, I'm planning on upgrading where I can on the suspension if the cost is reasonable. For the amount of labor required, I'd do as much as possible if the subframe is getting dropped anyway.

The CVT cooler I had in mind is the Setrab 9-20 Row Dual Fan Pack setup (FP920M22I), very nice engineering. It's pricey at $380 but comes with the fans too. I'll have to run thru the calculations again for the BTU heat rejection and all that but the core is 6" x 14.5" so it's quite massive but fits nicely in the stock FMIC location. We know the CVT get hot as hell and this setup is primo. The airflow should be pretty good, I designed an inlet skirt to force air thru the core.

The FMIC I'd like to run are either the Garrett 24x7.9" (10-row core) or the Treadstone 22"x7.9" (TR-82) core for $271. Not too big but they have enough cooling flow for 500 h.p. I ran the Forge FMIC and the power and response are great but the cooling isn't fantastic on those long highway runs. For the Mamba 19T flow rates the Garrett or Treadstone would work pretty well while not being too laggy.

 

[lne937s]

As you are looking at treadstone, have you considered switching to vertical flow? Would reduce pressure drop and allow for a slightly smaller package overall.

TRV185 Series Intercooler 720HP- TREADSTONE PERFORMANCE

TRV125 Series Intercooler 500HP- TREADSTONE PERFORMANCE

MacGyver,

Condensor core has some bug smash goin on. Ugly but not enough for me to discharge the A/C system to replace.

Juklear Winter/MacGyver,
Interesting, I'll put the Spoon Rigid Subframe mounts on my list then if they are that good. Actually, I'm planning on upgrading where I can on the suspension if the cost is reasonable. For the amount of labor required, I'd do as much as possible if the subframe is getting dropped anyway.

The CVT cooler I had in mind is the Setrab 9-20 Row Dual Fan Pack setup (FP920M22I), very nice engineering. It's pricey at $380 but comes with the fans too. I'll have to run thru the calculations again for the BTU heat rejection and all that but the core is 6" x 14.5" so it's quite massive but fits nicely in the stock FMIC location. We know the CVT get hot as hell and this setup is primo. The airflow should be pretty good, I designed an inlet skirt to force air thru the core.

The FMIC I'd like to run are either the Garrett 24x7.9" (10-row core) or the Treadstone 22"x7.9" (TR-82) core for $271. Not too big but they have enough cooling flow for 500 h.p. I ran the Forge FMIC and the power and response are great but the cooling isn't fantastic on those long highway runs. For the Mamba 19T flow rates the Garrett or Treadstone would work pretty well while not being too laggy.

 

[pboglio]

Core area is too small. Trv185 is 18.5x6. The TRP82.....Is 22x7.9. That's 110 sq.in vs 173 sq. In. Great pressure drop but efficiency not so good. The forge is like 15.5 x 7.0 which is 108 sq.in., internal fin dedign is different too so not an exact match. That TRV185 is exactly the size of my STI intercooler and isn't too bad except for the heatsoak. BUT the side by side intercooler are better IMHO. Gotta try and see I guess.

 

[pboglio]

I've got as far as pulling both suspensions off. Game plan change: Engine, trans, transfer case are getting pulled as a single unit. Was gonna be tricky to support the engine while simultaneously pulling the trans on another jack. I needed to pull the stock turbo anyway. The timing chain, crank pulley, drive belts, water pump, oil pump, oil pan, and head gasket can now get swapped. Engine needs a little TLC...to keep it alive with the upgraded turbo. Debated on forged pistons and rods but I'm not going that deep into the engine just yet. On the stand I can easily swap everything while pulling the trans off first on the engine hoist. Sounds like more work but In fact once the electrical, power steering, ac, etc are unhooked it's practically out anyway. The upper and lower rad supports, radiator, fmic, crash beam, ac condensor core get yanked and it just gets pulled forward and out.

 

[Macgyver]

Yeah. It will be much easier when its out of the engine bay thats for sure.

Me personally. Internals are fine for the CVT power limitations. I am going to take the internals up to 350whp and test the waters some day. My Clutch. That will have to be upgraded.

 

[pboglio]

MacGyver,

Engine wasn't part of the initial plan. But what happened is that I'm at 80k miles and I would have rebuilt at 120k miles anyway. At this power level I think the motor is done at 140k miles but I gotta admit the MR16DDT is a tough motor. The plan is to make the motor bullet proof. The CVT will not be as capable as the motor, but I'd prefer to have a motor & trans at the same mileage. The torque will be checked at something like 350-375 N-m (257 - 275 lb-ft) and I'll use rpms to make the horsepower. I have to get with an ECUTek tuner that can up the shift point in manual mode from 6000 rpms to 7000 rpms.

Engine Parts will be 2j Forged Pistons, 2j Forged Rods, Cometic headgasket, new OEM crankshaft, RS bearings, etc. I'll rebuild the cylinder head but I have to teardown the power train to see what I can reuse, and what I will keep as a backup. I have to put together a spreadsheet and start budgeting, been busy with work and doing the actual teardown.

Anyway, my decision to pull the motor from the front is looking good. There is SO much room to service the entire engine once you remove the front clip/radiator supports, crash beam, FMIC/Radiator/Condensor core. I did a little work today and I have to drain the transfer case/CVT, pop the axles, disconnect the CVT shift linkage, disconnect fuel lines, heater hose, and disconnect the electrical and the power train should be completely free. I'm really loving the way the factory made front engine access possible. At this point I would do all my turbo work, timing belt, etc. by pulling the motor. This also makes inspection so much easier, I never noticed how bad the oil pan was leaking until I took the front end off.

Finally, I've been doing some research on the CVT upgrades and came up with an idea. I then did a Patent search and it looks like someone came to the same conclusion. Based on what I've already seen for wear on the pulley's I have a pretty good game plan to deal with that. The pulley axial ball bearings will be replaced with tool steel dowel pins, there are a few kits that already provide this upgrade. The balls tend to shatter and take out everything in the transmission, a very common failure mode. I had bearing chatter for the last 5k miles and I'll also be upgrading the main shaft roller bearings as well, there are numerous kits out there for this. The belt situation I have somewhat confirmed the Bosch 901047 belt had 12 rings of F6 material. The newer Bosch 901066 runs 10 rings but with F7 material having better fatigue resistance. It looks like the torque rating is the same 250-350 N-m, but that is the distributor estimate. Bosch maintains the 30mm x 12 ring steel bands are good for 350-400 N-m, but the F7 materials improves fatigue stress. I'm going to have both belts for inspection. The goal if it works is to do 10 bands of the F7 material from the 901066 belt on put them on the 901047 belt and keep the other (2) F6 bands if they even fit. This'll produce (10) F7 reinforced bands + (2) F6 bands on the older 901047 belt. This is a super long shot since everything has to be identical fitment but I'm willing to try. If it works, I should be able to get the full 400 N-m torque rating with the improved fatigue life of the newer F7 material. The pulley modification will be a nitrocarburizing process on the pulley sheave surfaces as well as the axial pulley shaft slot where the ball bearings run. The dimensions are going to grow about .0005" so I'm going to have to account for it and mask any close fit interfaces. I'll have the pulley surfaces checked for hardness before they go out and if they can be hardened, then it's a go. These surfaces I've confirmed on my transmission are hugely prone to wear and the hardening will greatly extend the life of the CVT. The belt can then be replaced periodically using a master rebuild kit, new bearings, etc. without having to replace the entire transmission.

Next post will probably be the engine/trans on the hoist. Hopefully mid-week or this coming weekend.
Cheers.

 

[JuklearWinter]

It has been really fun to watch your understanding of the CVT grow and grow over the last few months. There's probably not a lot guys out there without first hand design/build experience who are as dialed in to this transmission as you are.

I'm sure you have very good reasons for 140k being your estimated drop-dead mileage for the motor without rebuilding it, but can you share a little bit about how you came to that number?

Is the salt bath hardening something that you can do yourself or do you have to send it off to an outfit that specializes in case hardening?

For the axial bearing solution, did you find a pending patent for an idea you already had, which confirmed that you were on the right track? That is a super clever idea, to the extent that I understand it haha.

Good luck!

 

[pboglio]

Thanks. Yep, learning a bit for sure. Hopefully I can get some positive results that'll maybe push the limit a little bit higher on the CVT. Right now I am in the fact finding mode. Then I'll get to the build and validation phase. I usually push my designs to failure. But in this case I can't afford that......don't have the R&D budget......lol. Nitride process will be sent out. Pretty common for case hardening gears and crankshafts. The mileage for rebuild is my estimate. A boosted 4 cylinder that is modded doesn't last long. Eventually fatigue will set in. My 465 h.p. eclipse gsx went 140k. Then it crank walked. Compression and leakdown were decent and there was no oil burn.....like zero. But eventually it had enough. The MR16DDT is nowhere near as tough being an open deck aluminum block. I was looking for a good excuse anyway to build the motor......the trans failure gave it to me. The pulley ball bearing the to dowel pin conversion was offered by TTK.....and a few others, totally not my idea. Makes perfect sense though. It is probably the single biggest weak point from my understanding. The nitride idea was mine but someone already patented it and produced positive results. Once I tear the trans down I can document everything and share the results. Than I can adjust my strategy on the rebuild.