Rob T's 818R Build Thread

[Hobby Racer]

I would definitely replace the bearings on both sides. The other spindle is kind of a shot in the dark. It could be fine if the the one that failed was due to metal fatigue because of being loose. There is not an easy way to tell if the good side is really good. I guess if the spindle is inexpensive, just replace it since you will have it apart to replace the bearings.

[J R Jones]

Rob, Good advice from John. I would also find a way to check the run-out on the axle shaft. Not easy with the CV joint. Maybe in a lathe with the end in a tailstock center. Perhaps a driveshaft shop can help.
jim

[C.Plavan]

I'd like your opinions, please. As I already have the "patient on the table". What are your thoughts about replacing the bearings and spindle on the right rear as well? Having the failure at speed was not an experience I'd like to repeat. Looking back, I think I had "warning" with the brakes rubbing slightly after sessions on the Left Rear. I do not have any of that on the right. However, based on the forensic evidence, the crack seems to have been around for a while....I don't know how long it took to manifest itself in either rubbing or a total failure.

Replace everything. Safety first. Watch and replace bearings every couple of years. Those bearings were new in 2013 when I built car. I bet the bearing was wearing everything (the noise you heard) and led to complete failure. Also repack the bearings yearly. Just my .02 cents.

[C.Plavan]

Hi Chad! I've been following your posts on the NP01 FB page. I take it you like the "Gen 2" upgrades?

It's taken a long time but I think we've about figured out how to race these 818Rs reliably, despite all of the ones at VIR failing after 2 days!
But yes, I keep up closely with what is happening with the NP01. Unfortunately to have anyone to race against I'd have to do some long travels, there is no NASA in the NW. Anytime you'd like to experience ORP you're welcome to come up here as my guest. I'd love to see your car on track here. We've had 3 members buy Radicals this year, and not the SR3, the SR8 and the one with the Ecoboost twin turbo V6. Crazy fast but they are just as unreliable as any other race car.

Not to hijack- but yes I updated to Gen 2. $5000 for new complete motor with 25 more HP and 400 RPM limit (new cams and ecu flash). I have my orginal motor that a may turbo charge for fun (glutton for punishment :P ) Also all the new aero and tunnel diffuser. Such a great race car and cheaper to run than a Miata. I get tired thinking of everything you guys have to do to the 818R. Don't worry about racing other NP01's, race it in a class it falls into where you are at. They are doing great in the WRL endurance races.

[Rob T]

The hub shaft broke at the end of the bearing. I got everything else apart yesterday. Now on to finding parts to replace.

IMG_6693.jpgIMG_6694.jpg

1) Hub (both wheels)
2) Seals (both wheels)
3) Bearings (both wheels)
4) Rotor (at least one wheel) and Bolts
5) Brake Mounting Plate (Subaru part)
6) Brake Plate to Caliper mounting plate (bent)

The odd thing to me is that was absolutely no bearing play evident after the last track day. I checked with the car in the air. The only clue was the dragging brake after a track session, that went away after the car cooled down.....

[Bob_n_Cincy]

Rob,
Did you find your axle nut? I'm curious if it spun off or did it get its threads ripped out of it?
Bob

[J R Jones]

Rob, There was an anomaly stressing the broken bearing carrier. Not all the parts can move with a rhythm that creates fatigue that causes the outer part to break-off.
Logic suggests the upright is stable and does not move with a rhythm.
That leaves the rotating parts.
bearings
axle/CV
wheel lug hub
the wheel
brake rotor
I suggest checking the run-out of all the rotating parts.
jim

[Rob T]

Bob: The track was "hot" the rest of the time I was there, so no opportunity to go look for it. That, and not knowing where exactly it came off, leaves over 3 miles of "haystack" to find a "needle" in. That said, I didn't notice any residual material in the threads of the axle. My money is on it spinning off completely after the break when the wheel could move differently than the axle when the hub was broken. Direction of travel would unscrew the nut.

[J R Jones]

Rob, In an effort to be as objective as possible here, I have a problem with your concept.
Before the break the splined portion of the hub was intact held together by the nut on the axle shaft. Because of the splines, the hub and axle turned together. No torsional load on the nut. (assuming the splines were OK)

After the outer portion of the splined hub broke free, the splines kept the hub segment turning with the axle. (unless I am missing something)

If we look at the timeline, there does not seem to be enough time after the the car became unstable, to your stop (10 seconds?) for the nut to turn off.
The ET of the last session when the hub broke was ~4.00 minutes. Was that enough time for the nut to turn off?
Have you summarized the total time the wheels were turning at the event?

It seems to me a tolerance opened up somewhere and/or something was running "out" causing the hub to fatigue and seperate.
jim

[Sgt.Gator]

I just checked the axle nuts on my STI. Two of them are starting to loosen:

Driver Side Rear:



Passenger Side Front:

[Rob T]

Maybe we are on to something here....If my failure helps others from not having the issue or something really bad happening, then it wasn't just a bad day at the track....

Can anybody confirm what the nut torque is supposed to be? I have several documents with different numbers and have heard different numbers as well.

Factory Five mentioned to me that they always checked the torque before their track sessions. I am trying to reconcile that with the dimple on the nut that is supposed to "lock" the nut in place.

[Hobby Racer]

I use 140 ft lbs. of torque as stated in the 2006 Subaru WRX Factory Service Manual.

I also have one side that is starting to loosen, looks like Gator's pics. Not sure how to best lock these if the dimple does not do the job.

[Sgt.Gator]

Yeah the FSMs jump around a bit.
Hobby, I think you may be looking at the rear axle spec, not the front axle spec.

2003 Imprezza FSM, 2005 Imprezza FSM:
Front Axle: 162 Ft-Lb.
Rear Axle: 140 Ft-Lb.

2009 Legacy GT FSM:
Front Axle: 162 Ft-Lbs
Rear Axle: 177 Ft-Lbs

2019 WRX/STI:
Front Axle: 162.3 Ft-Lbs
Rear Axle: 162.3 Ft-Lbs.

Since we have Front Axles now at the back of the 818 that means 162 Ft-Lbs.

I put the yellow paint across them once they are tightened and staked down just to make it pretty obvious they are loosening if I remember to even glance at them when changing wheels and adjusting air pressures. That's my way of checking the torque. I can't imagine actually putting a TQ wrench on them at the track, or anytime if they haven't moved.

I've been at the track 3 times when wheels have come off from axle nuts....luckily never my own.

[Hobby Racer]

Yeah the FSMs jump around a bit.
Hobby, I think you may be looking at the rear axle spec, not the front axle spec.

2003 Imprezza FSM, 2005 Imprezza FSM:
Front Axle: 162 Ft-Lb.
Rear Axle: 140 Ft-Lb.

That is correct. The 818 uses the front inner CV joint in the rear at the trans input but uses the rear hub at the wheel and therefor you need to use the rear hub torque spec for the the axle nut.

140 Ft-lbs is the correct spec

Rob - how well do the splines on the axle shaft mate up with the new hub? Any apparent damage?

[J R Jones]

A torque specification is relevant to the nut and the shaft threads. The other bits being clamped-up are non-deformable and don't care what the clamp load is.
The Subaru design intentions' here are violated, front/rear, steer/non steer, side loads.
Despite the threads and matching torque, if there is a choice, given the non-production loads involved, I would go with the higher torque.
BTW the clamp load of a 20mm thread @ 140 lb ft is ~15,000 lbs. The clamp load @ 162 lb ft is ~18,000 lbs.
I'm with the Sgt on this.

John is appropriately concerned. It is unlikely that (lateral) thrust loads are loosening the nut, it is more likely twisting movement at the splines.
jim

[Sgt.Gator]

I looked up the specs in the 2005 Legacy GT FSM. It also uses the 5 x 100 hubs and R-160 rear diff.
Front: 162 Ft-Lbs
Rear: 177 Ft-lbs.
I personally will use the 162 ft-lbs on the 818.

[Ajzride]

Drill a hole and use a castle nut

[Hobby Racer]

It's not a matter of using more torque. The tapered roller bearing cartridges in the hubs are designed for a certain amount of clamp force provided by the axles nut being torqued down. Each of the examples quoted above use different tapered roller bearing cartridges in their hubs and therefor need different axle nut torque settings. The hub bearings for the different vehicles/models are not interchangeable and neither are their torque settings.

Will the bearings fail if torqued to a higher value, no, but they will certainly wear quicker.

Drill a hole and use a castle nut

I wish it were that easy. The splined sections on the axles are induction hardened. I don't think they can be drilled without very expensive tooling.

[Sgt.Gator]

Learn something new everyday!

I talked with another Subaru racer. He says the rear bearings being the press in replaceable type aren't effected by the axle nut TQ. The shoulder of the CV and the nut pinch the inner race and don't load the bearing.
YMMV!

[Hobby Racer]

I talked with another Subaru racer. He says the rear bearings being the press in replaceable type aren't effected by the axle nut TQ. The shoulder of the CV and the nut pinch the inner race and don't load the bearing.
YMMV!

That is exactly what the nut does. It pinches the inner races together and puts a thrust load on the bearings. Look at this cutaway of a Subaru style rear bearing cartridge and it becomes clear.

UNIPAC_340x257.jpg

[jforand]

Hobby has this one nailed. The nuts and axles shafts are not going to care what torque is used. They will be able to handle either easily. What matters is the bearing preload exactly as Hobby described. You need to determine what bearings are in your rear knuckle and go with that OEM torque value or whatever an after market bearing recommends should you be using something special.

My fronts are peened. There is no rotational torque up front on the 818 so I don’t think they will loosen. As long as the peen stays put I think it is safe to call them good. My back nuts (from the driveshaft shop, came with axles) do not peen. There is no locking mechanism. There are a couple options…..torque them very frequent as part of a checklist or locktite them torque and paint. If they ever move you might need to get creative with the retorque depending on the locktite you used (red for instance). Blue you could just retorque.

I will provide one note of caution with the red locktite. I could not keep the axle nuts tight on the Mustang. I ended up grabbing the red locktite. It worked very well. When I did have to take the nut off later I used a dewalt electric impact gun and it actually tore up the threads on the axle!!! The secret would be heat and working the nut back and forth until if fully loosens. If you power it off in a single brute force session it gets worse and worse. I got to drive all over town trying to find the correct die to re-chase the outer couple threads. Of course this was the day before a track event

[Frank818]

The hub shaft broke at the end of the bearing. I got everything else apart yesterday. Now on to finding parts to replace.

IMG_6693.jpgIMG_6694.jpg

1) Hub (both wheels)
2) Seals (both wheels)
3) Bearings (both wheels)
4) Rotor (at least one wheel) and Bolts
5) Brake Mounting Plate (Subaru part)
6) Brake Plate to Caliper mounting plate (bent)

The odd thing to me is that was absolutely no bearing play evident after the last track day. I checked with the car in the air. The only clue was the dragging brake after a track session, that went away after the car cooled down.....

Rob, were your drive shaft rods Subaru standard size and steel strength, or you were running slightly larger rods (I think it's possible but can't recall if mines are larger) but most of all with greater steel strength (chromoly or something like that)?

[Frank818]

Maybe we are on to something here....If my failure helps others from not having the issue or something really bad happening, then it wasn't just a bad day at the track....

Can anybody confirm what the nut torque is supposed to be? I have several documents with different numbers and have heard different numbers as well.

Factory Five mentioned to me that they always checked the torque before their track sessions. I am trying to reconcile that with the dimple on the nut that is supposed to "lock" the nut in place.

Oh absolutely! I will do like Gator and mark mines too, keeping an eye couple of times a year.

I use 140lbs in the back, like Hobby, even though I have more weight. Never was able to determine if I should go higher with 200-250lbs more weight in the back.

[Rob T]

Update: Waiting on parts....have everything I need but the backing plate for the left rear. First problem was I ordered the wrong one. I guess there were two brake packages for 2006 WRX's. The caliper bolt positions are different. I needed the one that fit the 2004 WRX. Second problem was the right part arrived damaged. Now I am waiting on a replacement. I am ready to put it all back together when the parts get here....Next track day is August 20th at CMP.

[Rob T]

Quick update:

1) Put the left, rear side back together yesterday. Had to machine the new backing plate for the Wilwood bracket clearance.
2) Started on the right rear. Disassembled, pressed out old bearing parts, cleaned everything, put in new bearing, pressed studs into hub and assembled everything. Today I mount the brake caliper, check the alignment with the rotor and re-install.
3) Neither of the rear axles had any part of the two inner seals. This may have contributed to the bearing/spindle failure on the left side. The right side (unbroken one) was pretty dirty.
4) I have not had a chance to pull the outer bearing and race off the right hub yet. I want to see if there is any evidence of cracking....

[Rob T]

I pulled the race off the right side hub yesterday. No evidence of cracking or anything abnormal. Got everything back together. Today I bleed the brakes and get the car down from the lift. The new bearings seem good. Just a little more friction than before with all three seals brand new and making good connection.

[J R Jones]

Rob, did you identify any damage to the rollers or races on the left side? I sure would feel better if you could find an anomaly.
Did you research hub failures as the initiating problem? Warranty failure data is never made public.
jim

[Frank818]

Rob,

It took me way too much time to log back in and tell you this, but if you ever want to make the axle/hub/CV assembly stronger, you could follow what I did.

All my parts came from DSS (DriveShaftShop. Now try to say that super quickly without screwing it up). Back in 2015 I didn't trust the subie axles would handle my torque in a 2WD and the guy at DSS told me they have seen way too many of these setups to fail. Rare are those that reliably handle more than 450hp (or 450tq).

What DSS proposed was to bore/broach the hubs to 1 inch using bigger splines and fitting a 1in rod (chromoly I think). CVs are also beefier. That upgraded setup should handle 700-800, he said.


Here are my hubs and iterations (it took a few CVs to find the right ones fitting properly, the ones grey could not push in enough and I had to get the black ones. Also the blue boots got trashed cuz the clamp had a too thick buckle and was hitting on the spindle right above the loooooooooooooong lateral link bolt. My inner CV boot also got changed for a Hypervelocity boot due to my constant high angle, but on the pic it's the kind of boot you'd get for the subie. All in all the thickness of the CVs, splines, etc., are correct):


2015-11-17 05.46.23.jpg2015-10-20 16.12.53.jpg2015-11-17 05.44.52.jpg2016-03-04 05.51.04.jpg2015-11-18 16.40.44.jpg

On the above 1st pic, the right CV is the OEM subie and left is DSS.

[Rob T]

Frank: Thanks for the update. I am running DSS axles. Splines go almost all the way through. The axle survived this whole mess with no issues. I agree that the axles are awesome.

The new hubs look to be a different pitch. Do they fit with everything else? I have the standard (modified) backing plate with the Wilwood brakes. Also, a new pitch would obsolete two very nice sets of wheels. If my failure was indeed caused by a loose axle nut (my car is only about 300hp/300ft-lb), my new marking and checking procedure for the nuts should help. The right rear was fine. Thankfully, I did get a warning when the hub was cracked, I just was not able to diagnose it. Now I know.

[Frank818]

The hubs fit everything else, they just need a specific CV for the added diameter and apparently stronger splines. I think I couldn't install the inner-inner seal so I have 2 instead of 3 seals, but wheels, bolts, offset, Wilwood, that all fits normal.

It's true if your loose nut was the culprit than you're probably fine with OEM specs, I forgot about that!

[J R Jones]

Rob, I do not understand the specific differences in the alternative parts you and Frank have discussed. There is a basic relationship between splined parts that is relevant to axle nut loosening.
As we have discussed the axle nut clamp load is huge, way more than dynamic side loading. It does not allow the axle to thrust in and out of the hub, causing loosening.
Accelerations and decelerations try to rotate the axle in the hub. The class of fit is important to minimize clearance and rotation. Even a micro-rotation counterclockwise tries to turn the nut off.
With clean splines, male and female, test the spline fit. Slide the axle into the hub without the nut and bend the joint to evaluate fit. Then rotate the axle in the hub, again looking for spline clearance.
Obviously no clearance is preferred, maybe a little interference is better. I do not know if a clearance specification exists. New parts would be the benchmark.
You mentioned "Splines go almost all the way through", that is not ideal. More spline engagement is better than less spline engagement.
jim
Does your hub fracture align with the "Splines go almost all the way through" end of the splines????

[Rob T]

A successful track day at Carolina Motorsports Park Yesterday.

1) No Issues with the new bearings, hubs, left rear rotor or backing plate!
2) No Fuel pump issues now that I went to a direct wired (circuit breaker, separate switch) fuel pump.
3) Only 6 cars in my run group.
4) Ran 4 sessions.
5) In the last session, I saw the DAM number with a low of 0.835 and end at 0.935. The fine knock was at 0.33. They have always been 1 and 0 before. I did see an IAT at 124F when I was pushing hard. Previous sessions had been no higher than about 118. It was 88F outside and probably hotter over the track.
6) Water temp hit 200. Oil temp was 200, even when running hard, so I feel very good about the third stage on the dry sump. Oil pressure was 80psi when at speed on the track. 40+ in the paddock.

Video and analysis to follow....

[Frank818]

Oh yes that is really good news!!! You must be so relieved now, I'm happy to hear that!

[Rob T]

As this is the first time I have seen knock/timing adjustments while on the track, do I need to be concerned about this? I am not a Cobb expert.

5) In the last session, I saw the DAM number with a low of 0.835 and end at 0.935. The fine knock was at 0.33. Both are displayed real time on the Cobb Unit. I was not logging at the time.
They have always been 1 and 0 before. I did see an IAT at 124F when I was pushing hard. Previous sessions had been no higher than about 118 IAT. It was 88F outside and probably hotter over the track.

[blomb11]

Given the high IAT temperatures you must have seen some knocking which resulted in the DAM going down as you mention. As long as you do not see anymore knocking the DAM should correct itself and go back to 1. Per Cobb as long as you are between 0.6-1.0 you are okay and it should resolve itself if the knock has stopped and gone back to 0. The DAM can also be triggered by bad fuel sometimes. I have had high knock around -2 and saw my DAM come down around 0.8XX as well, but it corrected itself later completing the next driving cycle not under high load.

Something to watch, but it does not sound like you need to be concerned unless it continues. You will see knocking at high IATs...

[Sgt.Gator]

I'm emailing you the Cobb Tuning Datalogging Subaru How To Guide. But in short:

Monitor: DAM
What it does: DAM, or Dynamic Advance Multiplier, is a learned correction that adjusts overall timing.
What the numbers mean: For the 02-05 WRX, this number ranges from 0 to 16. For all other turbo model Subarus, this number ranges from 0 to 1 in tenths of a decimal (ex: 0.8). The number advances upwards when no knock is detected and decreases for extreme knock situations. You always want this number to be at it’s maximum value.
Why you monitor it: If you ever see this number decreasing, you have a potential severe knock situation.
What to look for: The starting value will depend on the vehicle and tune, but it should always learn up to the maximum value. This value will reset to it’s starting point after a map Reflash, an ECU reset, or a battery disconnect. If you see this value begin to drop under any other circumstance, you most likely have a severe knock situation. You should immediately analyze the tune and mechanical condition of your vehicle.

Monitor: Fine Knock Learning
What is does: Fine Knock Learning is a learned correction that makes small corrections to timing once the DAM has settled.
What the numbers mean: The values are degrees of timing being added or subtracted. An initial correction is typically -1.4, but is learned away in increments of 0.35.
Why you monitor it: When Fine Knock Learning hits extreme values, the ECU may start to adjust the DAM.
What to look for: Occasionally under low loads you might see an initial knock value followed by several values of that number decreasing in the incremental value (ex: -1.4, -1.05, -0.7, etc). This would mean that the ECU is already learning the correction away, and would be nothing to worry about. These values would also immediately go away after an ECU reset. If the ECU is making consistent and multiple knock corrections under load (such as full throttle and full boost) all in a row, you could have a potential knock situation.

Hope this helps!

[Rob T]

Thanks for the help Gator. I appreciate it.

[Rob T]

8/20 Track Day Summary at CMP

1) I drove 4 sessions, about 33 laps total, 75 miles, and used 11.25 gallons of fuel or about 6.7 mpg.
2) Highest transmission temp I saw was 210 using and IR gauge
3) AWIC supply (to the cooler) went from 90 - 100 - 104 - 110 as the day heated up. It was over 90 for the last session. IAT went from 102 to 115 to 118 to 124. I did see some DAM adjustment and Fine Knock Learning in the last session at the 124 IAT.
4) Oil temps were 200 in the last session. No issues with pressure, ever.
5) Water temp was 200 in the last session, but 190 for the others.
6) My best lap was 1:55.304
7) Best acceleration was 0.72G
8) Best braking was 0.92G and there is more there, I am sure.
9) Right Turn G's max 1.53
10) Left turn G's max 1.45 (100 treadwear Federal tires)
11) Max speed 108.29
12) Max calculated HP - 244
13) Max calculated TQ - 335
14) No issues with fuel delivery.

I am still working on driving skills and consistency. It's great to have the car running well, and be able to focus on driving.
Next track day scheduled for 9/10. I am hoping for a little cooler weather. 90+, high humidity and a heat soaked car, seem about the limit for what I have.

[Frank818]

Really happy everything got finally sorted! You deserved it very well, it's time to clock in lots o' miles and smiles!

[Rob T]

Track Day Scheduled for 9/10. Weather looks good so far with highs about 85F. Carolina Motorsports Park