[Macgyver]

Something like this.

Amazon.com: NEW 1 GALLON ALUMINUM FUEL CELL,DRAG RACE STREET,STRIP GAS TANK,ROLLOVER VENT : Automotive

Buy NEW 1 GALLON ALUMINUM FUEL CELL,DRAG RACE STREET,STRIP GAS TANK,ROLLOVER VENT: Fuel Tanks - Amazon.com ✓ FREE DELIVERY possible on eligible purchases

www.amazon.com

Put this where the OEM Intercooler was thereabouts.

 

[pboglio]

Will do, might have use for it.

So here is the CVT back apart. Primary pulley shaft is exposed. This is the reverse clutch pack, still need to swap out to the improved Raybestos clutches but for now that is what it looks like. That splined shaft inside is where the forward clutch drum reinforcement in the 2nd pic has to reside but there is no room for it currently. The NeWay 80* carbide cutter has (5) blades which can be seen in the pictures. I'm using a 5.5mm dia. mandrel designed for an intake/exhaust valve guide as that is what this tool is designed for. I need to 3D print a stepped sleeve that will accept the 5.5mm ID mandrel and also a .670" OD/1.185" OD to fit in the spline bore. This'll center up the cutter and stabilize it as I cut. Adding grease or some tissue paper I can tighten up the fitment so it's dead centered. Tried a few light cuts without it and the (5) bladed carbide cutter cuts thru that steel with ease. It's somewhat self centering but for this operation I have little room for error with probably +/-.003" being allowable for concentricity. To save time I'm cutting it in place and using rags/tape to catch and shavings and a vacuum. It's all get cleaned with a wire brushed and solvent before final assembly. To get to this point was about 2 hrs to teardown to this level but probably I could rip the entire CVT down in about 3-4 hrs not counting the pulley dis-assembly which is time consuming. Anyway, this is that machining operation that I've been going on about for awhile now.

For now I have to wait to 3D print that cutter mandrel guide & then order up new internal parts from Nissan as I mentioned earlier. Also need to order up some AMSOIL CVT fluid so I can soak the clutch packs overnight to prep them for install. I'm thinking 2 weeks waiting on parts and I should have the transmission buttoned back up. The torque converter input o-ring in the JF011E Master rebuid kit is getting tossed and a high quality Viton 75 duro seal will replace it, this being a highly critical seal mating to the torque converter. If I ever rebuild another transmission I'll probably Vapor Hone the aluminum cases like I did with my engine block, for now I'm leaving it as is since it would add more time to this build.

 

[Aus Richard]

Hi Everyone,
We've got the pulleys back now all polished with 320 grit on a lathe.
I went to put the pulleys back together when I discovered the thread of the retainer nut on secondary pulley was damaged. After alot of ringing around I found Automatic Trans-wreck near Brisbane who are absolutely fantastic to deal with. Bruce not only found me the nut but also the original sealing ring in very good conditon.
I have been putting the pulleys back together again to find out the secondary pulley retainer bracket keeps cross threading the bolts and it looks damaged. I am going to replace it with the one from the other secondary pulley. I have tried my 3 jaw puller to remove the bearing so I can access the bracket but I can't get under it. The bearing splitters I have are also too big. Does anyone have any ideas how to get this seperated please ?
Also I have attached some pictures of the pulley housing, it has a lot of scratches/gouging where the bearings sit. I'm wondering if it is still ok to use like this ?
Thanks in advance and have a great weekend everyone.

 

[pboglio]

Hi Everyone,
We've got the pulleys back now all polished with 320 grit on a lathe.
I went to put the pulleys back together when I discovered the thread of the retainer nut on secondary pulley was damaged. After alot of ringing around I found Automatic Trans-wreck near Brisbane who are absolutely fantastic to deal with. Bruce not only found me the nut but also the original sealing ring in very good conditon.
I have been putting the pulleys back together again to find out the secondary pulley retainer bracket keeps cross threading the bolts and it looks damaged. I am going to replace it with the one from the other secondary pulley. I have tried my 3 jaw puller to remove the bearing so I can access the bracket but I can't get under it. The bearing splitters I have are also too big. Does anyone have any ideas how to get this seperated please ?
Also I have attached some pictures of the pulley housing, it has a lot of scratches/gouging where the bearings sit. I'm wondering if it is still ok to use like this ?
Thanks in advance and have a great weekend everyone.

Yeah, that sucks it was cross threaded. That stuff isn't too hard to find from a salvage yard, you got lucky though. The sealing ring(s) come in the rebuild kit but if you want to reuse that original ring it's not a big deal. If you are talking about the bearing facing the pulley case (ie. smaller case), then I would take a dremel and cut the bearing outer cage off. Then take the roller bearings out and you will have access to pull on the inner bearing race. This is how I did it but there are now some fancy pullers that can pull on the inner cage between the bearing balls. There isn't enough space in behind the bearing either for the TJ-1 or split bearing pullers. For most rebuilds these end bearings wouldn't have to be touched but a high mileage pulley set yeah it's a smart move to replace them. Again, Nissan have these bearings for $15 but aftermarket they tend to be pricey, these are all over the web though there are a lot of Chinese knockoffs. Sonnax now carry the (4) Primary/Secondary pulley bearings as a complete set so I recommend that to anyone. Lots of options here.

The pulley brackets you want to be careful you put them back in the right orientation as it's easy to flip them over and get them wrong. They will look wrong but just follow the rebuild manual. I put a paint dot on the top-side and double confirm with the install manual cause if you get that wrong the bearing has to come back off which typically is then a throw away.

You're showing the pulley case in those pics. The scratches aren't a big deal unless they obstruct the bearing from sliding back in, I would test a bearing out of the box and see if it slips in. The caged roller bearing should slip into it without catching/binding. Might take some scothbrite and lightly knock down any burrs but it should be a snug fit. This will sound counter-intuitive but sometimes you will want to remove any oil/lubrication as the fit is so tight it'll block the bearing from slipping in. This was the situation fitting the pulley assembly back into the center case, the bearing didn't want any lubrication on the outer race/pocket. You'll see when you get to that point. Any other time like press fitting bearings onto the shaft then yes I use lubricant as you want to feel the bearing bottom out and if it doesn't you will be in big trouble as the cases won't fit back together properly. This is why I measure the distance from the end of the pulley shaft to the top of the bearing before I disassemble it. Then after press fitting I make sure I hit that same dimension. The oil lip seals I also had to lightly lubricate for press fitment otherwise it tears up the outer diameter. Some things to consider.

 

[pboglio]

Got the machining of the primary pulley shaft completed and I'll toss up some pics of everything once final assembly starts.

Ordered from Nissan the final batch of new internal CVT components to finish the build. The planetary carrier/ring gear/snap rings/seal rings/bolts. Lot's of these parts were so cheap I just ordered stuff even if I wasn't sure. I typically don't reuse small diameter bolts and simply upgrade them or buy factory new bolts. Most of the stuff from the Altima 2.5 CVT will fit except the reverse clutch piston for some odd reason. The differential is not the same due to AWD and these are not available but typically don't wear if at all except the smaller bearings.

For the stuff that is highly stressed inside the transmission the Altima parts are a very inexpensive option & the latest factory improvements/updates are utilized. The other thing is that Nissan also offer "most" of the parts that would come in the rebuild kits which can run nearly $500 and are mostly seals, snap rings, gaskets, etc. What they don't offer those parts can be purchased seperately such as the SAP valvebody rebuild kit, pulley gap rings, etc. The most important parts are the Freudenberg Nok teflon piston gap rings and the oil lip seals......these are available thru Nissan. The pulley piston gap rings aren't available thru Nissan but Cobra Transmission have them for like $15 ea. The other smaller kit's from SAP include a valvebody rebuild kit, oil pump seal kit, piston gap rings, etc. The clutch packs I buy directly from Raybestos in a kit, also the sump filter & beehive filter. SAP also offer a nice forward clutch bonded piston set (2 piece) with a Hi-performance blue option that is an upgrade, the reverse clutch piston for the JF011E won't fit our CVT which comes in the (3 piece) set so money is saved buying the (2-piece) kit for the Juke's. The Altima can utilize the (3 piece kit).

Thus I could probably piece together a CVT rebuild kit without any wasted kit parts and still be significantly cheaper than what the Master/Banner rebuild kit's offer. I'm working on putting together a list of this stuff as a resource at some point once I'm finished with the project. But generally there is a lot of support/parts availability for our JF011E CVT transmission.

 

[Macgyver]

The Monster build.

I hope she can roll out of the garage after all the time spent.

 

[pboglio]

Definitely a monster build...lol. Should have been a 2 year build but will take over 4 years. Most guys probably would have parted out by now or would've just swapped in a replacement CVT and been done with it. The money I could have saved doing that instead. But yeah the goal is to see it running and have fun with it.

 

[Macgyver]

If I did all that work. It would blow up on the first 0-60.

 

[pboglio]

Definitely a possibility. Hoping some changes in driving habits prevent that plus some torque limiting in 1st gear & 6th /pulley ratio where the CVT is weak and the revised boost mapping. Going to take a holistic approach of internal upgrades/maintenance/cooling/driving habits/tuning to make it last.

 

[pboglio]

Here is the primary input spline machining completed for the forward drum reinforcement insert. It required a facing operation to cleanup as I didn't want stress risers on the face of the spline causing a crack to form up, 1st pic is messy. The other pics show it deburred and cleaned up. It also better shows how much material is left on the spline hub. Critical here was to leave proper wall thickness in the spline hub and the pictures show I've left sufficient material. The 2nd pic shows a cross section of how the spline hub tapers inwards for maximum strength and the machining has only "slightly" weakened the primary pulley input spline though it was massively over-designed from the factory. The machining accuracy/concentricity is critical and that was achieved.

From here onwards it will mostly be doing some final internal cleaning, re-installing new components, setting shaft-end play, clutch pack clearances, etc. I have a few modifications I'll walk thru later about the (4th) clutch disc upgrade I'm doing and how I go about making that work though I covered it much earlier in this thread. There are further custom upgrades I don't discuss here but I have in the works to support the clutch pack upgrades. The entire drum/clutch/hub pack assembly is massively stressed especially on the AWD. I'm on the fence about sending out for a few sets of these custom designed upgrades as it would be simple for me to retrofit them into the transmission while the forward clutch drum is still out and I'd also have a few extra sets available.

The Raybestos clutches actually have a smooth engagement characteristic so this'll help with the damper cone spring being removed which is required to fit a 4th clutch disc. The clutch damper cone spring is only really needed for moving from PARK to DRIVE. This might affect ratio or gear shifts in manual mode as the factory do attempt to modulate clutch apply pressure for smoother engagements. The Raybestos clutches are actually much grippier than the stock Exedy cellulose clutches from testing on the bench. Based on the literature something like +20% higher clamping torque capacity and the carbon/kevlar/graphite/ceramic hybrid material can take some serious abuse without degrading or wearing out. By themselves they are an upgrade (RCP96-256). My feeling is a stock/rebuilt Juke would hold some decent torque just with the Raybestos (3) clutch module kit (RCP96-256) drop-in ready to go. The newer JF011E master rebuild kits from WIT are now including these Raybestos clutches and I can see why as they totally outperform the cheaper cellulose clutches from the Raybestos testing results.

So I've got a short list of components to re-install. All the remaining new internal components are on order with Nissan. Case needs some mating flange cleanup to get the Loctite case sealant off & reapplied. Ordering up: new electrical harness to the valvebody, new primary speed sensor (05189840AB) and secondary speed sensor (05189840AB), new wiring harness zip tie anchors, new CVT oil dipstick and o-rings, and a few other random things to tidy up the rebuild & install.

 

[pboglio]

OK, so I'm waiting on my parts I have some time to do some analysis as this project winds down.

To recap, the Gen1 Juke CVT run the 901066 or newer 901083 pushbelt at 30mm wide, 10 bands and 229.4mm effective dia. This is rated by Bosch at 350 N-m maximum torque but nominally it's designed for vehicles with about 250 N-m (184 lb-ft). Problem here is most tunes are pushing past the 350 N-m and at this torque level this wears out the belt prematurely.

The later JF016E transmission (+2013 Altima, etc.) run the 901074 pushbelt with 28mm wide, 12 bands and 229.4mm effective dia. These pusbhelts are close to the same size with just a 2mm width difference. Bosch rate this belt at +400 N-m to 500 N-m. Nissan rate this belt on that 3.5L engine at about 430 N-m (312 lb-ft) max and it's the strongest pushbelt they ever made before switching to the chain drive. @Scougar actually discovered this belt years back doing some research and wondered if it could fit our transmission.

Basically this newer belt has a few advantages: First, it has 12 bands vs. 10 bands although slightly narrower it's stronger and the rating indicates this. Both belts are made of the latest F7 super alloy so equally both are durable at their torque ratings. Secondly, the newer belt element (Type "C") angle is inverse to our current belt elements (Type "B") so it doesn't pitchover at higher rpms and higher torque load capacities. It also has a higher CoG (i.e. center of gravity) in the upper belt element which acts as an energy booster. Because of this it can grip more tightly at higher rpms and has a +10% higher torque holding capacity in general.

Problem here is that this belt is 28mm wide vs. 30mm wide for our current belt. The pulleys can't stroke out much less than 29.5mm without the variator balls bottoming out and limiting the variator travel, I tested this to confirm it. The SAP pulley variator slider pins typically replace these factory 6mm dia steel balls as an upgrade and this provides an opportunity. They are 6mm dia. x 12.5mm long (3 ea.) on the Primary pulley and 6mm dia x 25mm long (3 ea.) on the Secondary Pulley. To get the pulley to actually clamp on the 28mm narrower belt I'd have to accurately grind 2mm off each slider pin so the pulley's can stroke out further to clamp the newer belt. The pulley essentially ride on these "pins" like rails. Cutting these slider pins down by 2mm then can allow the variators to stroke closed another 2mm to compensate for the narrower style belt, something the little ball bearings would not have been allowed to do otherwise. Thus these slider pin upgrades potentially allow a pushbelt swap. Ironically these slider pins are already a factory upgrade on the JF016E for different reasons.

Next problem here is the ratio control follower, looks like a shift fork on a manual transmission. This little thing tracks the position of the pulley sheave in the axial or in-out direction and works with the ratio control piston & linear solenoid on the valvebody to control the pulleys. It's job is to help control the pulley position so the ratios are correct. I believe this change of +2mm extra stroke travel would probably be accepted by the valvebody TCM and compensated for without throwing a code. That is a big guess but a calculated one. But if the TCM can target the correct ratios and the primary & secondary speeds are correct it should have nothing to complain about is the hope. The ratio control follower, ratio control pivot fork & valve body ratio control piston also have a limit on their physical stroke travel and that is something I'd also have to review so possibly this ratio control follower & pivot fork would need some modification to account for it. This is all physically doable stuff but the details of compensating for altered geometry would have to be confirmed and of course the TCM or DTC potential issues still exist.

Basically I'm going to finish my CVT build just like I planned to. This post is just illustrating the potential for really closing the loop on upgrading the Gen1 Juke CVT and more for documenting some ideas/thoughts. The pictures illustrate just how much beefier this newer pushbelt is over the current belt. I may purchase this upgraded belt and start doing a feasability if I end up building another CVT for myself at some point. Always room for improvement.

 

[pboglio]

Just purchased a reinforced/upgraded CVT pushbelt for the Juke. Should be good to about 400 N-m (294 lb-ft) based on the Bosch spec.Got a few valvebody upgrades (i.e. Sonnax control pistons) but that'll come mid-next year and not a big rush. Basically I've achieved most if not all the design targets I set for myself back in 2018 when this project started. If I run into issues I'll deal with it but planning for it to go smoothly. Early spring will be spent getting the vehicle running & tuned so the focus will be on that.

After the Juke is running I might get into more of the custom fabrication stuff I have planned. I got by on this build with some of the bare minimum poer tooling and equipment but built or purchased what I had to over the course of these 4 years. Will try and do some fun stuff with the CNC vert mill and the TIG welder will definitely get used for the custom intake and exhaust work I have planned.

 

[chidog]

In case this helps anyone, installing these transmissions, here is an idea I came up with. It may or may not work on the ones in the Jukes though.

 

[pboglio]

Thanks Chidog. This'll definitely take the anxiety of worrying about the torque converter slipping off the oil pump sprocket. I like how you shimmed the bolt with washers to snug that up. I'll see if I can fab something up when the time comes to install the CVT.

 

[pboglio]

CVT is waiting on some backorder parts still.

 

[pboglio]

Took a few pictures of a CVT pushbelt I bought on ebay. Was advertised as new, and more like used/refurbished but worth it just for analysis. Typically a new bosch pushbelt will be bathed in red VIS oil to protect it for storage and initial install and this was bone dry and the plastic sleeve was opened so I knew it was used. I've seen some China/Aliexpress pushbelt knockoffs that are a dead ringer for a Bosch pushbelt so be aware. This is precisely why I buy all my parts from Nissan/Mitsubishi/Mopar or reputable Transmission suppliers but in this case I wanted to see for myself what this older belt looked like as it was touted as a "reinforced" belt.

The 901047 pushbelt was used on the 2007 Altima but used earlier on other applications as well. It has 12 bands (.0075" thk ea.) and 30mm wide using older F6 material. The 901066 pushbelt (Juke +2011 thru 2014) has 10 bands and 30mm wide using newer F7 material. The 901066 is now end-of-life and superceded by the latest 901083 of similar design/material but some improvements made. There is an efficiency improvement running 10 bands vs 12 as they have less friction between them as the belt bends. This material change from F6 to the newer F7 is somewhat critical as the F7 is 10% stronger and about 500% more durable in terms of fatigue life. The material was reformulated to remove Titanium inclusions with the goal of reducing crack propagation zones and then increased the strength using cobalt. That means even when they reduced the belt thickness or number of bands from 12 to 10 they are still about 150-200% more durable than the old belt while being nearly equal in strength. Now Hitachi have licensed the belt technology from Bosch and produce it for certain pushbelts. The diagram illustrates the type of improvement in fatigue life on the S-N curve (Load stress vs Cycles) and a blurb about the material technology.

The older belt is typically listed as a "reinforced" belt on many websites but in fact it was superceded for a reason, they tended to snap well before the warranty period. So not all the information out there is technically correct unless it's from Jatco or Bosch sources. The 901047 I'm showing was refurbished by re-grinding the flank faces which essentially destroys the elements, they are supposed to have deep ridges similar to a synchronizer ring but with larger lands. Those ridges slice thru the oil layer and allow the belt to land and grip the pulley. When they get wiped they float like a bearing on the oil and that reduces their torque holding capacity. You can see my 901066 pushbelt in the pictures. The belt is actually stretched about .060" and there is a physical gap between the links which makes it look odd. The belt elements are wiped completely clean and heavily burned up from the brake torque launches and abuse/neglect. It's my belief that the intense heat is what stretched the belt and "annealed" it though it still held intact even after 80,000 miles running 210 lb-ft @ wheels. This annealing/softening of the belt can happen if the CVT overheats especially from brake torque launches or heavy slipping at WOT. This is precisely why I recommend AMSOIL as it protects against belt slippage.

The oil pump flow/pressure valve failure also accelerates the CVT wearing out prematurely. The variator ball bearings can typically fail and seize which can cause the belts to snap. This is very likely the cause of CVT pushbelt failures at higher mileage. A higher torque CVT engine should switch to variator roller pins that eliminate these types of failures. Rarely these days does a belt snap by itself unless it's a very high mileage transmission. A worn out valvebody or failing oil pump will greatly contribute to the belt getting wiped and so will overheating and overfilling the sump causing oil pump cavitation. Thus maintenance and driving style factor heavily in the pushbelt life.

Anyway, finally wanted to compare these two belts. Pushbelts will always be a weak-link which is why Nissan and Subaru have switched over to chain drive that can handle much higher torque loads.

 

[pboglio]

Ordered up (3) gallon jugs of AMSOIL CVT fluid. Used their 6-month membership ($10) to get their 25% discount and free shipping on orders over $100. Cost me $179.60 which isn't too bad. Needing the AMSOIL to soak the forward & reverse clutch packs overnight before I rebuild them as this is how it's done. Standard fill is about 8-9 quarts but I'm also going to be running the Setrab external cooler so I'll need an extra bit of fluid.

The last of the new Nissan OEM CVT parts came in last week except some internal bolts for the oil pump/converter reactor plate. Still need to purchase the primary and secondary speed sensor and the thermostat assembly.......then done. Those bolts are coming in on Jan 21st so I'm waiting on that. Typically I don't reuse bolts on a rebuild if they are highly stressed and for this CVT build I upgraded or replaced almost all the fasteners which'll be torqued to factory specs. Unfortunately the select thrust bearings for the JF011E transmission are now non-available. Sucks when you place an order thinking it'll ship out and they cancel it when they realize the part no. is now obsolete and not made anymore. This special bearing sets the input shaft end-play after final assembly. I tried Nissan and Mitsubishi and it was a no-go. I do have several sets of each size from when I purchased some 3 years ago just in case and can make them work with some precision shims if needed.

 

[pboglio]

So now I have the CVT oil pump to install shortly. As I've mentioned before I've upgraded the flow control valve on this pump with a valve from Sonnax essentially a pressure relief valve but hard anodized upgrade (Sonnax #33510-02). This upgraded valve may be the single most important upgrade in a CVT because it keeps the belt alive under heavy load. When this relief valve get's jammed due to poor maintenance......the pushbelt is a goner and a big reason why our CVT fail early. You can see the jammed flow valve about 1/3 opened, I guess better than full closed or full open. That pump was rebuilt but it's getting replaced with this new OEM pump with the upgraded relief valve.

Aside from dumping excess pressure/flow when not needed, the relief valve acts as a recirc booster into the pump suction inlet with the purpose of delaying the onset of suction cavitation. Table 1 shows performance characteristics of the pump. The max rotational rpms is 7,000 rpms before encountering cavitation. I've seen dyno charts of Maxima CVT running up to about 7500 rpms but using a gerotor style pump so a little different than this vane style pump. My Juke has seen right about 6,900-7,000 rpms with the ECUTek flash in CVT normal mode. With the upgraded engine I'm targeting 7500 rpms.

Pictures of the new style pump for the CVT7 below, similar in construction to the JF011E but updated. Attached diagram shows how this little guys works: the eccentric pump vanes push in/out and conform to the bore walls due to centrifugal forces inside the eccentric bore as the pump shaft rotates, thus the chambers draws the oil in, compresses/pressurizes the fluid and then fluid exits at very high pressures at the small outlet. Pump output is about 5.5 Mpa normally and 6 Mpa max upper spec limit.

Oil Pump modification:
The CVT7 new pump exploded diagram picture Fig 7. is below and it show's the pump suction port is fully ported without visible restrictions. The new pump illustration Fig. 1 shows how physically wide open the suction passages are as well to help lower cavitation and increase flow. The Juke pump or existing type has a very sharp/convoluted inlet as the attached pic shows. This new pump also eliminates the relief valve because of this highly optimized internal geometry. The Juke is restrictive due to the cross channel machining that creates the oil gallery passages. At full pump flow this would be a decent amount of suction loss and potential cavitation into the pump inlet. Since I plan on exceeding 7,000 rpms on my Juke build, this is a good place to optimize the pump. An old hot rodder's trick is to flow port the pump inlet on an engine oil pump. I show the smaller pump high pressure outlet and this I won't touch as I don't have a good feeling of the secondary effects and neither will I attempt to port the internals of the pump though it might be beneficial. I've rebuilt another pump previously but this one is already upgraded and I'll port it carefully.

The suction port has a step for the filter o-ring.....that goes untouched to avoid a vacuum leak. The 2nd step is the stop for the steel filter intake tube. I'll leave that shoulder but port some of the inner corner that is obstructing flow. Anyway, should be a 2-3 hour job then it's ready to drop-in into the CVT.

[Credit: Shimono Hiromi, "Flow Control Valve-Less Vane Pump for CVT"]

 

[pboglio]

Below is a picture of the Sonnax #33510-01 (over-size) oil pump relief valve meant for the rebuilt oil pumps requiring over-size reaming, this I bought for a pump rebuild I never ended up using. New pumps you'd use the Sonnax #33510-02 as a drop-in upgrade which is what I ultimately went with. Differences physically are subtle with the OEM except the Type III hard anodize coating is far superior wearing than the OEM Type II anodize. Sonnax maintain the factory relief spring as the goal isn't to change any dynamic function, just improving durability and reliability.

Relief valve control architecture:
The function of the relief valve is: Reduce cavitation at the pump suction port due to high inlet flows & foamed sump oil(< vapor pressure condition), reduce cavitation inside the pump vane assembly (deadhead condition), reduce oil pump power draw at low load/high rpm conditions. This is achieved with (2) control valves inside the valvebody (Pressure regulator valve & Line control valve), 1 solenoid (Line pressure control), and (1) relief valve in the oil pump assembly. Typically the CVT oil is highly foamed and agitated and this valve is meant to assist reducing these conditions.

The valve itself operates using a pressure delta across P2 and P1 of the pump outlet metering orifice in the previous diagram. That is the reduced pump outlet I said I wouldn't port.......that would have been a hugely bad idea if I did, the relief sensing port sits right after it and requires this pressure drop to function correctly. The pressure delta is the force that moves the relief valve along with the return spring. When the valve opens it boosts the pump inlet pressure raising it above the vapor pressure and prevents the oil from boiling off. Opening during dead-head (low demand/high pump speed) conditions prevents the oil from getting superheated in the pump vanes which also leads to cavitation. The reason these valves don't have seals is to reduce friction which would be unsustainable at these enormous pressure and of course the seals would be blown out or destroyed again due to cavitation on valve shutoff which is a common problem. The valvebody pistons look similar to the oil pump relief valve/piston and use high tolerance fitment to reduce pressure leakage thru the gaps. These can be damaged from abrasive metal debri and also when cavitation occurs these valves run dry which is the main failure mode. Sort of a chicken and the egg situation, which comes first?

At low load and high rpm conditions the valve likely would open to reduce oil pumping losses. High load and High rpms the valve would likely open to reduce pump cavitation in the inlet. Low load and low rpms it would tend to stay shut depending on conditions. High load and low rpms it would to stay shut. All controlled by the TCM and using the pressure delta as the hydraulic force to move the control piston. When the piston is damaged and jams then it could happen in any position but generally failures typically would reduce pump pressure causing the pushbelt to get destroyed over time.

Ultimately everything comes right back to the CVT oil pump. But again keep in mind it's the Valvebody that commands the relief valve situated inside of the oil pump so they work together and feed each other basically. Thus a good quality CVT oil, changed often, to the right fill level (not high, not low), cooled, & clean, never sucking air, upgrading the relief valve if possible, driving style, etc. is what will allow a CVT to survive longer. When that system breaks down the failures start happening though there are other structural limitations to consider.

 

[pboglio]

Below is a typical CVT pump flow demand requirement on a 250 N-m vehicle. Pump output is about 10.3 cc/rev or 10.3 liters/min per 1,000 rpms.

So the various flow demands are going to range from 9.5 lpm to about 19.5 lpm or higher depending on higher rpm oil flow requirements not listed here. Since the pump moves about 10 liters/min every 1,000 rpms it most times will be open but some of the time it'll be closed. Since mine was jammed about 1/3 cracked open I was probably coming up short on line pressure & pulley pressure at low rpms, launches, etc but generally I had enough pressure at higher rpms based on my CVTz50 datalogs. If it failed completely closed then I can see problems not so much with pressure but actual power loss, fuel economy, fluid over-heating, etc. If the valve is working correctly then the TCM controls everything automatically and efficiently for most/all conditions. The CVT is a fairly impressive piece of technology even with it's flaws.

One could argue that shimming this relief valve closed at least will guarantee the pump always puts out the highest pressure possible and eliminate the issue of valve failure but it would come with some significant downsides that might be difficult to predict. On my build I don't touch the control valves as the TCM will actively fight any changes using the solenoid trimming functions and generally it's a bad idea. When everything is working correctly like new then these problems don't exist but that always isn't the case, in that situation the upgraded valve is part of that potential solution.


[Credit: Bosch, "Fuel Consumption potential of a CVT"]