Jump to content
LegacyGT.com

Comparison of USDM, JDM & COBB bars


vel0mac

Recommended Posts

Introduction and Method:

 

This test compares three different anti-roll bar configurations on a completely stock 2005 Subaru Outback XT (including stock RE92A tires). The XT is a family driven car. My goal is to improve the handling of the car by increasing the stability and sharpness of its cornering and also reducing or eliminating the ‘wobbly’ effect through the corners. Typically, I use this car for street and daily driving on asphalt roads. I do not race or autocross, and if the car is used on dirt roads they are generally smooth forest service roads. I am not interested in race-type modifications. I do not use the car for difficult off-road driving.

 

The first tested setup was as the car had been driven for the last year plus: showroom stock as sold in the USA (USDM). The second setup tested was to replace the USDM rear bar with a Japanese (JDM) rear bar, keeping the stock USDM front bar. The third and final setup was to replace both front and rear bars with a matched pair of COBB bars. In all three cases, stock end links were used.

 

One of the concerns/problems when fitting stiffer rear bars to these cars (and the Legacy) is that the rear mounting tabs may deflect unacceptably, if not break off entirely. In theory, if the geometry of the bar and its end links is correct, all loading should only be linear columnar loading on these tabs. However, theory often departs from reality, especially if any adjustments are made or different end links are used that introduce non-normal loading vectors. Still, even if all loading is columnar, the tabs can buckle if sufficient load is applied. The solution that COBB and Perrin have implemented for this problem is to install a stiffening bracket on this mounting tab that reduces the flex and strengthens the mount.

 

Installation and Fitting:

 

Installation of the rear bars was done with the car on jack stands with the rear wheels removed. The exhaust was not disconnected or lowered in any way as suggested in the COBB instructions. The bars were removed and reinstalled by threading them out through the wheel wells and down in between the exhaust’s y-pipe.

 

The COBB rear bar was set to its softest setting.

 

The front bars were removed and installed according to the COBB provided instructions.

 

In both front and rear cases, there were no immediate fitment problems and both installations were very easy. There are visible differences with the front COBB bar that caused some problems later on in the driving tests. See the “Problems” section for more detail on this.

 

Below are some photographs of the different bars.

P1040023.JPG.2f8ed1a6e75c277a3471570873e38adb.JPG

P1040024.JPG.5b5e695be830ca26ab0b8b55abf758bf.JPG

P1040026.JPG.7ca3f1604b11fd1afbf954b47e264f07.JPG

P1040027.JPG.5e88ae66acd94bbb2c52d3327c453bf8.JPG

Link to comment
Share on other sites

  • Replies 126
  • Created
  • Last Reply

Top Posters In This Topic

Static Measurements:

 

Since enough people are interested in numerical data, I took numerous dimensions for comparative reasons. A dimensional table is found at the end of this section.

 

It should be noted, however, that simply measuring a bar’s diameter is insufficient to determine its stiffness characteristics. The USDM bars are the only bars in this comparison that are solid construction. The JDM and COBB bars are tubular and as such wall thickness, in addition to outer diameter, is also an important consideration to the bar’s stiffness. The bar’s wall thickness was estimated by measuring the swaged end of the bar and dividing this value by two, the assumption here is that no additional material flow or extrusion occurred during the swaging process.

 

Finally, since anti-roll bars are torsion springs, the overall bar material composition, material length and moment arm are contributors to the bar’s characteristic. I do not have the means of determining the material composition, but I did measure the other dimensions.

 

Resultant static measurements were based on lifting one rear wheel to a known height and measuring the rear bar mounting tab’s deflection by mounting a dial indicator to the sub-frame via a magnetic base. This measurement was taken under four conditions: a.) stock USDM bar and no stiffener; b.) JDM bar and no stiffener; c.) COBB bar and no stiffener; d.) COBB bar with COBB stiffener.

 

After reviewing the measurements, I’m not sure that the method for measuring tab deflection that I used actually gave me the data I was attempting to collect. I believe that the sub-frame that the magnetic base of the indicator was mounted to was also deflecting, which skewed the value of the indicator tip. I believe the measurement method was in error because the ‘deflection’ values for the COBB bar, with and without the stiffener, were almost identical. And yet when I physically try to deflect the tab by grabbing it with my hand and moving it fore and aft, there is a very significant difference in deflection with and without the stiffener. The stiffener bracket clearly reduces deflection and after seeing how easy it is to deflect the tab by hand I would not mount a stiff rear bar without the COBB stiffener add-on.

Finally, I jacked up one rear wheel until the other rear wheel was lifted off the ground. This was in attampts to measure the overall 'system' stiffness as measured across a diagonal loading.

Static Measurements Table.pdf

P1040019.JPG.f6116a5422ed46549935fa2321930a31.JPG

P1040021.JPG.fb41fdf6e7fd12fcddc9a0bb98317292.JPG

P1040022.JPG.b56c880601eee9bf75961be9f5ccc851.JPG

Link to comment
Share on other sites

Driving Impressions:

 

The driving conditions that we have here are a nice mix of high mountain passes, tight twisty canyons and wide-open freeways. There are also many forest service roads. Most of the secondary roads are chip-sealed, and in the neighborhoods there are many drainage dips. The freeways are a combination of asphalt and concrete (both new grooved concrete and old cracked and rutted concrete). The freeway speed limit is 75 MPH, which means that often the actual speed required to keep up to traffic is 85+ MPH. We do get snow here, but given that I just installed the different bars, I have not yet driven them in the snow.

 

In all cases, there were three drivers that tested the car: myself (the gear head), my wife (who is not a gear head) and my son who has only been driving for a year (the newbie).

 

Stock conditions:

Well, this is a well-known condition to all (or at least most) of us reading this. The stock XT has great power, but the suspension cannot handle it effectively. The car wobbles side to side and drifts erratically in high crosswinds, especially at highway speeds. Cornering is OK, but not the greatest. And, the dreaded pogo-stick ride is nearly nauseating – but that is a different problem.

 

 

JDM Rear Bar:

Adding the JDM rear bar definitely improved the handling of the car. The understeer was noticeably reduced, and overall the car felt a lot more stable under all conditions. Steering was more precise and confidence inspiring. In fact, the first comment that my wife made after driving the car was that she felt far more confident in corners, and she could ‘go faster than before but feel more in control’.

 

The ride did not change as much as the handling did. The ride was slightly firmer when hitting unevenness in the road, but it was barely different than the stock bar.

 

Overall, I think the JDM bar is the minimum these cars should have. They clearly should have been shipped off the showroom with them.

 

There were no fitment or noise problems with the JDM bar. The setup was as quiet as the stock car, even though it was mounted using the stiffer AVO urethane bushings.

 

 

COBB Front and Rear:

The best way I can describe the feel of the COBB setup is that the F&R combination made at least twice the difference that the JDM bar made. That is, if the incremental change between stock and JDM was worth one ‘unit’ then the difference between stock and COBB is three ‘units’ of handling improvement.

 

No additional road noise or clunking was noticed. The car is very, very flat through corners. Steering is far more precise. My wife described it by saying, "the car goes exactly where you tell it to go now." Cornering corrections due to suspension settling are almost gone. Tire squeal happens at considerably higher cornering rates. The little skipping over bumps and manhole covers is nearly gone. The high-speed stability and tracking are much improved – especially in strong crosswinds (this is a huge bonus around here). The overall feel is far more ‘lively’ and fun. It is really, really fun to ‘stick’ the car on the exit of a hard corner with full throttle 2nd gear accelerations!

 

As would be expected, the independence of the suspension is reduced and the ride is stiffer when traveling over undulating terrain. If the dips/bumps hit both wheels on a given axle at the same time, there is no difference over stock. It still bobbles up and down. However, if the dips/bumps hit one wheel on a given axle first, I think the ride is better and less bouncy -- even though it is a stiffer ride. The under-damped pogo-stick characteristic of the car has been slightly reduced, but much of the wobbly-bobble-head-doll cornering has been reduced. Yes, the car still needs new dampers – badly – but it is better than before, especially when cornering or going over bumps even on a slightly diagonal bias.

 

On a side note, I got new rear shocks installed at the dealer under warrantee. The new shocks seem to be slightly better than the old shocks (only 12,000 miles on the ‘old’ shocks!), and the car feels like both axles are now equally under-damped rather than the rear being more under-damped than the front. The car still needs better dampers on both front and rear, that’s for sure.

 

If you drive your Outback on rough dirt roads, the COBB matched set may not be the best solution. The smoother ride of the highly independent suspension will be gone, that is for sure. Plus, I’d imagine that a lot of pounding on rough dirt roads would put its strain on component parts like end links now. But, I seriously doubt that these bars are designed as off-road accessories…

 

Nonetheless, I did drive down some dirt roads that were not too rutted or rocky, they had some harsh washboard, and the ride was not that bad. It was notably stiffer than before, but still tolerable. And, there seems to be a slight increase in road noise, possibly due to the urethane bushings. There is also an increase in the steering gear ‘rattle’ that is well-known to exist in Subaru cars. This rattle is only noticed on full-lock turns over rough terrain at parking lot speeds (broken pavement, dirt lots, etc…). The car made this noise a bit before I mounted the stiffer COBB front bar, however the noise has increased notably. I do not think this is a function of the COBB bar, per se, my guess is that any bar this stiff would cause more steering gear rattle. Certainly, on smoother and more ‘quiet’ roads and parking lots there is no noticeable increase in noise due to any binding or interfering parts.

Link to comment
Share on other sites

Problems & Solutions:

 

But, alas, not all is perfect. On day 2 of having the COBB bars installed I snapped one of the front end-links. And it happened in a very innocuous low speed situation, fortunately. I pulled off to the side of the road to fiddle with something under the dash. The asphalt dropped off a couple of inches and the washboard clay shoulder slanted down away from the road. When I stepped the car back up on to the asphalt the car was heavily diagonally loaded. Just as the front left wheel came up on to the pavement the right front wheel’s end link popped out of its socket, rendering it completely separated and useless. I drove home pretty slowly and pretty wobbly.

 

When I got home I looked under the car, made sure all was installed correctly (it’s hard to screw this up) and looked more closely at the photos of the two bars. And then I reviewed the measurements that I took.

 

Looking at the annotated photo below, you can see that the COBB bar is shorter than the USDM stock bar by 22mm (1086mm versus 1064mm), and the distance between the outermost bends is greater (unfortunately, I didn’t measure this parameter before installing, I do not have a specific value for this dimension, but it seems to be an inch or more). This combination will allow misalignment of the end links -- any end links -- not just the stock Subaru end links, especially if the bar ‘walks’ off center. Misaligned end links will greatly increase the stress on the links.

 

And that is exactly what happened. The COBB bar ‘walked’ side to side over to the left side of the car about an inch, which caused a severe misalignment of the right end link and it popped out of its socket. I replaced the end link and repositioned the bar to its centered position and made sure the bar was parallel to the axle since it is mounted with slotted u-brackets (I measured from the bar to specific points on control arms and sub-frame parts to ensure parallel mounting). After driving for a couple of miles, the bar had walked off center again and the end links were again at their extreme angular misalignment position. This time, I caught it before anything broke.

 

All spherical joints have an acceptable angular misalignment. If the angular misalignment is too great, there is an abrupt increase in axial loading of the joint. Most spherical joints are not designed for very much axial (thrust) loading. If the front bar walks off center an inch, which is roughly the distance it moved in my case, this will cause a misalignment of 23 degrees on the front 60mm link. Most commercial spherical rod ends only have a full range of less than 65 degrees (plus/minus 32.5 degrees), so a misalignment of 23 degrees is quite severe. In fact, a 23 degree off axis position of the stock Subaru end link stud is the very extreme that it will move before it hits its socket lip. No wonder the link blew its head off!

 

As a means of correction to this problem, I machined 4 spacers for the end links. Each spacer is 5mm long, so that the total amount of correction is 20 of the 22 mm. This will move each of the links 5mm closer to the center of the car and, effectively, widen the sway bar’s width by 10mm. Spacer length of 5mm was limited by the amount of available stud length. Washers would probably work here also. But, I had access to some machined stainless spacers...

 

In addition, I also added some stainless steel shaft collars (McMaster-Carr p/n 6436K38) to the bar so that it cannot walk off center. These collars are ½ inch wide each, so this would suggest that the bends are at least an inch too far apart (total) as the stock bushings fit right against these bends.

 

After making these modifications, the bars have worked perfectly. No noise, no clunking, no rubbing, nobroken links. And I even went back to the scene of the crime and did the same thing that broke the link in the first place (dynamic diagonal loading of the wheels) and there were no problems afterwards. It has now been three weeks since I made the 'corrections' to the initial install and still no problems.

 

Summary:

 

The handling of the car is greatly improved with the COBB bars and at this point in time, after the modifications I have performed, I am happy with the COBB bars -- although it is unfortunate that the front does not fit better out of the box. After communicating with Trey Cobb, it appears that the fitment between the XT and a LGT is in fact different. From looking at the photo that Trey sent me, the LGT that COBB has does not exhibit the front bar fitment problems that I experienced. However, the required additions to make the COBB front bar fit an XT are relatively minor, and I think they are worth the effort.

The rear fitment is fine -- no problems at all, no noise. I do not have the clunking that others have talked about with their lowered Legacy cars.

 

The COBB bars are probably not the best solution if you are going to drive the Outback on rough dirt roads. If this is the type of driving that you do, I think the JDM rear bar would be a better way to go. But, if you stay primarily on asphalt or smooth dirt roads, and you want a far more lively and stable feel, then use the stiffer COBB setup, it really transforms the car nicely. It is a LOT more fun to drive now!

 

Now all I need to find are better dampers for this car and it will be all set! (Koni, are you listening??)

161769909_Front_Dimensional_Differences-small.JPG.0d5e9010533f393e2b6dda1004796edc.JPG

P1040038.JPG.c152a5242cf57b9f76cf5a34616f4261.JPG

P1040039.JPG.0f9439b3e4fa15299b693096318e533c.JPG

End_Link_Spacers.jpg.57c46f1b1f4a404b92d65fb9705d5533.jpg

Centering_Collars.jpg.4a60d23b31b200ae67c82137c8719fe0.jpg

Broken_Link-small.JPG.8ebb4a9a51839e94f284a0919ee1d7d1.JPG

Link to comment
Share on other sites

Someone make sure Cobb sees this.

I sent them the original *.pdf over a week ago.

 

I think (I hope!) they are taking all this into account when they look at redesigning the rear bar (for the numerous clunk issues that many of you have had) and the front bar for fitment. Also, one person mentioned that his front mounts had bent and he had told COBB about it. All these issues together are probably why they have delayed any shipments for 6-8 weeks. Most likely they are back at the bend layout phase.

Link to comment
Share on other sites

Does anyone know if the fitment on the front might be due to his car being an XT? I dont remember hearing fitment issues on other similar threads.

One way to possibly tell is to measure the original Legacy bar as I have done and compare it to the measurements in my table. Or, check part numbers between the two OEM bars. I do not have access to stock Subaru part numbers. If the bar is the same between the two cars, then I'd imagine that the fitment would be the same.

Link to comment
Share on other sites

I sent a link to this thread to Josh at COBB and he said the problems were due to it being an OBXT.

 

COBB does not recommend these bars be installed on anything other then a Legacy!

 

It seems as though you have discovered a nice way of making the bars work, all the more power to you!

Link to comment
Share on other sites

:eek:

 

Wow! Mad Props! That was a fantastic write-up!

 

I read all of it, and I don't even own an Outback! :D

<-- I love Winky, my "periwinkle" (ABP) LGT! - Allen / Usual Suspect "DumboRAT" / One of the Three Stooges

'16 Outback, '16 WRX, 7th Subaru Family

Link to comment
Share on other sites

I sent a link to this thread to Josh at COBB and he said the problems were due to it being an OBXT.

 

COBB does not recommend these bars be installed on anything other then a Legacy!

This sounds like it may be a cop-out (or should I say COBB-out?) to me. I just called the local Subaru dealer and they told me that the p/n between the Legacy and the OB are the same ... sometimes. The p/n variance depends on production date and equipment, but there is definitely overlap between the two models within the same model year. And it obviously follows that if the p/n is the same, then fitment should be the same.

 

Also, when I talked to COBB on this BEFORE I ordered them, they told me that they would work for an XT, so I ordered them. Then after my order had been placed, and independently, sebberry emailed them asking about his OB and they told him that they had not yet tested them on an OBXT. So, I called them back ... They back-pedaled by telling me that they would really like to advertise and sell them as working for the OBXT, but in fact they actually hadn't tested them on that car yet.

 

I offered to drive to SLC from Denver to have them mounted on my car so they could be tested on an XT. As a compromise, and to avoid a 15 hour round trip drive for me, they offered a full refund if I was unhappy if I would give them feedback -- which is why I wrote this thing up the way I did.

 

So, it seems like their right hand may not be speaking with their left hand. And, it seems like we, on this forum, may have a more consistent 'collective memory' of any events and communication.

 

I am not annoyed with them over the fitment issues. I think that there is a certain amount of experimenting that will need to take place. They are a small outfit and they cannot afford to test their stuff on all possible combinations -- that would be very expensive. I work for a big multi-national tech company and we even have a hard time affording all the testing required for our stuff. So I believe that there is some amount of risk that the buyer may need to assume when doing something different (like I did). However, I do wish that they had their story straight in the beginning, and they should be able to keep it straight given that they are a small outfit...

 

Just my $0.02, and probably worth that much.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...

Important Information

Terms of Use