(Note: we lost the last week of posts from TC, and I wanted to make sure that this one made it across. I've copied my posts; please feel free to copy yours over if you think it's valuable.)

-= POST 1 =-

I feel compelled to write something down publicly on the subject of Jim LaBreck's Cobra IRS cradle for the 3rd gen RX7. While I have great respect for Jim's work as a pioneer in the V8 swap area and am grateful for the time and energy he donated in the creation of his cradle as well as his selfless nature in creating a group buy around that cradle, I've come to the conclusion that the design is flawed and could potentially lead to catastrophic failure that would result in damage, injury, or death. This is why I am posting this--not to disparage Jim in any way, but rather to warn others who are using or may use his cradle.

The original JimLab cradle thread is at the rx7forum.com here:

http://www.rx7club.com/showthread.php?t=313112

Jim has unfortunately pulled down his pictures on that thread, so I'll try to supplement my own here.

Jim's original design had a crossmember with two tabs for connection to the Cobra pinion mount points. This crossmember was braced by two longitudinal mounts. The points of contact were the locations of the OEM stiffening braces that attach to the rear subframe--two for the longitudinal bars, and two for the two ends of the crossmember.

The pinion mount attaches to the crossmember using rubber, poly, or solid mounts.

You can see the pinon mount here:



And the general shape of the crossmember here:



(The forward tabs on the left side of the photo were not part of the original design. More on this later.)

Now, a few notes. First off, the original design had a cutout for the exhaust. The first person to actually use this cradle bent the crossmember almost immediately. Jim chalked this up to its use on a drift car, but provided a patch for those that had the cutout. A subsequent run of the cradle did not have the cutout. Here's my patched version:



After assembling the car I noticed that there seemed to be a lot of play in the rear of the car. I decided to hook up a webcam and zip tie it to the back of the car. Here's what I found:



What you're looking at (and HEARING!) is me doing some runs around the block in 2nd gear. I'm gently rolling on 30-40% throttle in a car that was untuned at the time. I'm guessing that I'm putting a maximum of 150-200 ft-lbs down. You can visibly see the crossmember bending under the stress. This is the unmodified cradle that still has the original exhaust cutout.

A quick digression: how much force is that on the crossmember? Well, the worst case scenario is in 1st gear. The 200 ft-lbs is multiplied by the first gear ratio (2.66:1) and the rear end ratio (3.73:1) which is about 10x. The distance from the axles to the pinion mount is about 9", so the upward force on the crossmember in this case is roughly 2700 pounds. It should be worth noting that the system is constrained by the grip of the rear wheels, so on street tires you will probably max out at 3500-4000 pounds. With drag radials, this could be significantly more.

So, we can clearly see that this design, especially with the exhaust cutout, is far to weak to handle the forces involved.

To address this, I added forward braces that tie into the transmission tunnel brace pickup points. I also welded on the cutout patch that Jim provided.



I'm a very skeptical guy when it comes to 'seat of the pants', but you could really feel the difference when these modifications were done.

Now, let's switch gears to the second half of Jim's kit, the rear cover bracket. The rear of the differential was supported with a tube that connects to the stock rear pickup points, but points backwards and requires removal of part of the subframe for clearance. All the mount points are solid connections to the cover and chassis.

This by itself is a problem. Besides noise issues, there is a lot of evidence that the rear cover of the Cobra IRS is a weak point. Although Ford does mount the cover solid to the subframe, the entire subframe is isolated from the chassis with rubber bushings:



So Jim's solid mount is a problem. But that wasn't the biggest issue. After modifying the crossmember I got to drive the car around for testing. My car had a serious drone about 75 MPH. I checked the usual suspects...wheels, tyres, exhaust, and so forth. Finally I measured the pinion angle.

The Hinson engine cradle puts the front of the motor up at about 2 degrees. Good driveline design has the diff at the same angle so that the input and output shafts are parallel. (In fact, many builders put the diff a bit lower to compensate for bushing and chassis deformation under load.)

Jim's kit had the diff up at about 7 degrees, or 5 degrees higher than normal. So I killed two birds with one stone by raising the rear of the diff by over an inch. I also added poly bushings:



This took care of the droning, reduced cabin noise substantially, and took some of the stress off the rear cover.

I was hoping that I was done with this saga, but I was sadly mistaken. After a track day at Laguna Seca, I noticed some grease on the bottom of the car:



The CV boot has been cut by contacting the subframe. After looking more closely, I saw that the pinion mount tabs were severely bent upward, especially on the passenger's side. After making the crosmember stiffer and tying it in to the chassis, the tabs were under increased strain. A problem with my CF card means I don't have a picture of this, but the tabs were bent up by 3/16" to 1/4" and were dished--the bolts were trying to tear themselves through the tabs.

Let me pause here to describe what would happen if these tabs failed. The nose of the diff would be violently thrust upward. If you were lucky, the pinon mount would smash into the bottom of the chassis and the rear bracket would hold, keeping the diff in generally the right location. If you're not lucky, the diff moves enough so that the driveshaft falls out of the transmission. It digs into the pavement...you get the idea. (This is why they make driveshaft loops...)

In retrospect, the idea that 1/8 plate steel in single shear (cantilever) could hold 3000 pounds of force was a very bad one. I'm not sure how many cars are running with this setup, but I'm sure that this issue will become obvious very quickly.

I was lucky. No failure, and the CV joint was okay. I designed a system that captures that mount point much more securely and will not suffer from tearing:



I ran this setup at Laguna on Monday and everything seemed to work well. I haven't inspected everything yet, but I'm going to do so this weekend.

I've spoken privately about these issues to many people, but after this latest incident I'm worried that someone is going to do some serious damage to their car or themselves. Again, I'm not trying to disparage Jim in this post, but the design of this cradle is flawed in many ways. Those who are considering it should be aware of these issues and be prepared to invest time and money to make it safe and reliable.

I should also say that my modifications are for discussion purposes only. I'm not saying that I've fixed the issues here, but that I'm trying. If you duplicate these designs for your own use you are doing so at your own risk.

-= POST 2 =-

Despite the clear desire for community to see a flame war here, I'm going to keep this civil. As I said before, I posted here because I feel that the design of the cradle--with or without the cutout--could lead to people hurting themselves or their cars which I know Jim would not want to see. So I'm going to clarify a few things, but I feel like I've done the right thing in sharing my concerns.

True. I had issues getting the cradle to line up with the subframe holes. I didn't want to enlarge them like Drew did, so instead I loosened the subframe to make things a bit more pliable. At the end of the day I had to use a ratchet strap to move one of the longitudinal bars into place, but it was probably something due to my car and not the design.

Jim did indeed offer to weld the patch on which was a nice gesture. I wanted to try to make it work without the patch which clearly failed. I just wanted everyone else to know that if they bought one of these without a patch, they should have it welded on immediately. But even with the cutout patched, the crossmember needs to be supported along its length to keep from moving vertically. The video shows only a fraction of the force that you would see under a 0-60 run, let alone a drag launch with radials. And since most people are running this setup to get gears that are better for the 1/4 mile, it's highly relevant.

But I should say that although Jim was clearly annoyed with me at the time, he definitely went out of his way to work with me which I appreciated.

I don't have any photos of the original cradle, as I stated in the post. I only have pictures since the first round of modifications which I clearly explain the the text of the article. Unfortunately the pictures in Jim's original thread are gone. I'd be happy to post some here if I could get them.

But for those playing along at home, here's a comprehensive timeline:

1) Receive cradle from Jim
2) Try to install, have issues, return and request that the pinion mounts be 'boxed' in from the front (which Jim does for free, and also throws in the exhaust patch and upper pinon brace--very nice!)
3) Install with some elbow grease, but without the patch. Diff has Torsen T-2R and Moroso 3.90 gears which whine.
4) Video under car reveals significant deflection of the cradle under moderate loads
5) Add patch and forward tabs that mount to tranny tunnel brake points. Rebuild diff with Ford Motorsport 3.73 gears. Whine goes away, but still lots of noise from the driveline in general.
6) Highway testing reveals drone at 75 MPH+
7) Try three sets of wheels and tyres. Have car re-aligned. Have new midpipe made with resonator. Finally have pinion measured to discover misalignment.
8) Fabricate new rear bracket, raising cover and adding poly bushings. Drone goes away, and driveline noise decrease substantially.

At this point, I haven't even gone WOT on the car because it hasn't been tuned.

9) Dyno car (421 RWHP / 408 RWTQ) and attend first road race
10) Discover cut CV boot and severely deformed pinion mounting tabs with clear signs of bolt tearing and dishing of the mount surface and washers
11) Fabricate new pinion mounting tabs and re-align diff
12) Go road racing again without incident, but full inspection pending

What I was trying to point out with the solid mount issue is that most road cars don't solid mount. This is for noise and longevity reasons. But I think Jim has a point, the solid mount is not so much a flaw as it is I think an area that could be improved. The Cobra rear cover is a known weak point, and I think Jim was planning at some point to make the rear bracket tie into the cover bolts to address this issue. Anything done here is an improvement.

The pinion angle, however, is a clear design flaw...

I had Moroso 3.90s for a while, but switched to FMS 3.73 gears. In both cases I was still running with the original solid rear bracket and Steeda poly bushings for the pinion mount. The FMS gears do not have the high-pitch whine that the Moroso gears had, but when I added the poly bushings the overall noise was reduced dramatically. It's still not as quiet as the Mazda rear, but it's much better.

Let's skip the rhetoric. Of course I tried muliple shim configurations, including no washers at all below the poly bushing (as shown in the first thread photo). Even at this height, the angle was way off. Each shim moves the angle perhaps 1/2 of a degree, maybe less. And since the diff was pointed up, it was not possible to move the nose down any further.

At least one other board member has verified this issue. But like I said above, I'm writing this so that people who decide to use this cradle are aware of issues that others have run into. They can measure the pinion angle themselves. If it's 5, 6, 7, 8...they're all unacceptable and will cause noise, poor joint lubrication, and potentially failure.

I'm not going to resort to ad hominem slights. There are enough issues that I've outlined here that make me feel that the cradle as designed is simply not safe or correct. If just one person avoids damaging themselves or their car by reading this post, then I feel like I've done the right thing.

And if I haven't belabored this enough, I'm not attacking Jim on a personal level. I don't think Jim offered any guarantees with his cradle, nor should anyone imply them. This is clearly caveat emptor territory. I'm trying to show some clear caveats. Getting this cradle to work correctly will take some work that is not trivial.

In fact, the diff ripping out of your car at WOT is not trivial, either.

-= POST 3 =-

I had a number of requests for an updated video from underneath the car. For reference, here's the original video which is a smooth roll on of 30-40% throttle:

Original Cradle, Second Gear, Partial Throtte


That was with the unmodified cradle. You can see that there is a lot of motion and that the cradle is clearly moving relative to the chassis. Here's a second video that I didn't post earlier that was first gear pulls, again with only partial throttle:

Original Cradle, First Gear, Partial Throtte


You can see even more motion. For the record I think welding the exhaust cutout would have helped here, but not enough. Also, I want to be clear that I never went past 50% throttle on the original cradle because the car was not tuned. By the time I had it dyno tuned I had already done several of the present modifications.

Okay, fast forward to the present. The following videos were taken today with the modified cradle. (Details of the mods are posted above.) You can see the brackets in the foreground that tie into the transmission tunnel brace, as well as the cutout patch and new pinion mounts.

The first video copies the conditions of the original one, namely a smooth application of 30-40% throttle in second gear:

Modified Cradle, Second Gear, Partial Throtte


I tried hard to put the camera in the same place as before. As you can see in the new video, the cradle barely moves, and in fact seems to have done a good job of reinforcing the entire rear of the car. Most of the motion that you see is bushing compression on the top of the pinion bracket (bottom of the video because the camera's upside down).

Since this video wasn't nearly as dramatic as the first one, I decided to do some near-full throttle 2nd gear runs. I started at about 2000 RPM and ran up to 5000, then back down, and repeated:

Modified Cradle, Second Gear, Near-Full Throtte


Again, this car is producing 408 ft-lbs of torque at the rear wheels.

You can see a bit more flexing of the chassis, but again the majority of the movement is due to the bushings compressing. Amazing how much those poly bushings move, isn't it? You can see why chassis designers usually put the pinion angle slightly lower than theory suggests because the flex of the bushings and chassis can move the pinion flange 1/8" or more.

Finally, I decided to really get on it, simulating the kind of hard acceleration that you might see at an autocross or road race. I did two full-throttle accelerations in first gear which both broke the tires loose pretty quickly. At the end I'm at about 5000 RPM and getting on/off the throttle very aggressively.

Modified Cradle, FirstGear, Full Throtte and Quick On/Off


Pretty amazing, hunh? You can see a significant amount of flex from the chassis, crossmember, and even the forward mount tabs. The differential is really moving around. This is with street tires--I don't drag race, but I'm guessing that good radials can put 25-50% more force to the ground before the wheels break loose.

So that's where I am now. I haven't seen much video from underneath cars, but I'd say I'm mostly comfortable with this setup. You can see why chassis designers try to tie the roll cage into the third member pickup points. I'm not sure how long those transmission tunnel brace mount points are going to last under that kind of compression. For those that are willing to weld things to the chassis, boxing and reinforcing that point might be a good idea. But in the end there's no substitute to tying the whole thing together with a nice cage.

This also shows why the torque arm or PPF is a elegant solution despite its drawbacks. You only have to deal with 4000 lb forces where the TA meets the pinion--but not against the chassis or subframes.

For those that would like to use video as a diagnostic tool, it's really not that hard. All you need is a laptop:



A cheap webcam that's probably in your desk drawer:



A USB extension cable to get from the cabin to the underside of the car (or engine bay, or wheel well):



Then zip tie the camera in place:





And route the USB cable into the car:



And away you go. Takes about 15 minutes to set up. I should have used some tape on the cable but I wasn't going very far and the cable couldn't really get anywhere dangerous.

What can't be shown with video--and what I think is the area most people should concentrate on--are the mounting tabs for the pinion. As I made the cradle stronger, the instantaneous force on the mounting tabs increased and began to cause serious deformation. I would recommend reinforcing the tabs by making the horizontal plate thicker--at least 3/16" if not 1/4"--and boxing in the sides completely. This should keep the tabs from bending upward or dishing due to the bolts trying to tear through the lower surface.

I accomplished this using 1.75" steel tube with 3/16" caps. The mounts were welded into notches on the crossmember. You can see this in the pictures above.

-= POST 4 =-

Shit!!!

Mike, when I read this I had one of those 'a-ha' moments that made me wish I had paid more attention during my freshman statics analysis class.

Mike is absolutely right, but it's not very intuitive. As mechanics, we usually talk about torque as force x lever arm. For instance, when you have a 1 foot torque wrench and you put 10 lbs of force on it, you generate 10 ft-lbs of torque at the bolt that you're tightening.

So when we think of how the driveline works, we are tricked into thinking of torque as happening at a single point--like the axles of the diff--and we calculate the force as a distance from that point of rotation.

This is wrong.

The idea of torque acting at a point is very misleading. The example of tightening a bolt is a special case where one end of the system is free to translate (the end where your hand is) while the bolt is fixed.

But in the case of a driveline, both ends are fixed. Let's look at the setup:



In this image, the vertical black bar represents the rear diff mount point. This could be the stock pickup or the JimLab bracket. The front black bar represents the transmission mount.

The torque of the wheels against the ground is counteracted by a torque in the driveline. The driver's side wheel wants to turn counter-clockwise, so the driveline wants to rotate the opposite direction.

Okay, so let's look at the stock setup with either the powerplant frame or a torque arm like Hinson's. The drawing is actually a bit inaccurate--because the engine/trans is supported by bushings, the actual level arm is longer (somewhere between the enigne and trans mounts) but this would only increase the difference between the two setups, as we'll see.



Now, let's remove the torque arm and put a mount similar to the JimLab (or what Kevin Doe is building):



The point here is that you shouldn't think of the torque acting at the output shaft location. If this were actually the case, you would see a different force upward at the pinion than you see at the rear cover. And from a statics perspective, all the forces have to balance out. Torques have to be canceled by normal forces.

Like I said, this is counter intuitive. So I tried to come up with an example that you could try to prove this to yourself.

Take a pencil and place one end of it underneath the edge of a table. Using your left hand, hold the other end. Now, take your right hand and grip the pencil about 1" away from the tip under the table. Twist the pencil so that the end under the table is forced up into the bottom of the table. Take a note of how hard you have to hold the other end of the pencil with your left hand.

Now move your right hand so that you're holding the pencil at the other end, about 1" from where your left hand is. Twist the pencil in the same direction. You'll notice that the force required to hold the pencil with your left hand doesn't change.

You can do the same experiment with a light broom handle--just don't use something so heavy that the weight of the object becomes and issue.

The point is that the normal forces don't care where along the solid body the torque acts. All that matters is that the body is under a torque which must be canceled by normal forces that balance the static system.

So, what does that mean? The right way to think about the rear end is that when you ditch the PPF or TA you reduce the length of the system that is subject to the output torque of the driveshafts. Regardless of where the output shafts are on the diff--towards the back, or the front--the torque acts on the solid body. So the forces at the front of the diff (in the upward direction) will be the exact same as those at the back. They will also be equal will balance out the torque on the system.

By making the system longer, you reduce the forces on the front AND the back. Shorten it, and you increase the forces at BOTH ends, just like Mike said.

Now things get really interesting. If the torque at the wheels is about 4000 ft-lbs, the torque on the driveline is the same. If the distance from the back of the diff to the end of torque arm is about 5 feet, the force on the transmission (up) and the rear bracket (down) will be 800 lbs. However, if you shorten that distance by mounting the pinion bracket to the chassis (like the JimLab setup) the system is now only about 1.5 feet long. This means that the same 4000 ft-lbs now puts about 2700 lbs of force on the pinion (up) AND the rear bracket (down).

Meaning that the rear bracket tabs and subframe are now being loaded roughly 3.5x what the factory intended. What does that look like?



That's first gear, two hard pulls (with wheelspin) and then on/off the throttle at high RPM. That's not the bushings deflecting! You can see that the poly bushings are not moving, but the bracket and subframe definitely are.

Mike, thanks for bringing this to my attention. Again, I haven't seen a lot of video from under cars, but the movement of the rear bracket gives me some concern. It has me thinking that the next move for me is to build a rear bracket system like Kevin's...

-= POST 5 =-

To be honest, if I were to do it again I would have stuck with the stock rear and just bought a KAAZ diff. I think the Cobra solution is ultimately superior, but if you get serious about road racing them you'll need to invest in a diff cooler as well. It gets pricey.

I think what might have worked better was to find another diff--if one exists--that is lightweight, strong, and designed for a torque arm or power plant frame. For instance, the Viper rear end:

http://www.turbobuicks.com/forums/sh...ad.php?t=61291



Which I think Mark Corbett is working on, IIRC.

-ch

__________________


1994 Mazda RX7 (FD3S) R2 with LS1/T56 and Cobra IRS
430 RWHP / 410 RWTQ @ 2939 lbs in full street trim

Haha! Jim and I had a round about this stuff a long while back. He was very intent on the inferiority of my TA design. I couldn't make him understand the reasoning that you have layed out here in great detail. I applaud your thoroughness in testing and verification of results. Simple use of physics and a calculator can yield great information.

Lane

Nicely done hyperion, thanks for the second effort required by the post lost in the TC transfer.

I agree that the PPF is an elegant solution. I hope mine works.

What's wrong with the Cadillac rear that nobody chooses it?
I'm guessing weight and few gear choices?

Wow, I love it when people, will stand next to their point and are willing to prove it. Rather then just BS their way through it. Now to really get some good info on the RX7 we would need to see this done on a stock rear I think this will give us a good idea of what is really acceptable as far as rear end movement goes. Now I know that most of us here (including me) don’t see the stock rear as acceptable is any way. But this will give us something to compare your info too.

__________________
Josh
I will finish your LS1 swap PM me!
74 MGB GT V8 12.83 (75 shot)
93 LS1/T56 Rule 3 No V8 RX7's

Threads like this make me want to put webcams under my cars and watch stuff under there work

__________________
Community Moderation System and other forum news

Hadn't thought of that, but I know a few guys locally who have big LS1's with the stock rear end. Hey guys--want to meet up and do some video runs? I'll bring the webcam!

It would be especially good to find someone with delrin or poly diff bushings so that we aren't looking at deformation of the OEM rubber bushings but of the subframe instead.

It really is easy to do, and shows you quite a bit about how things are actually working. I used this technique to see how bad my exhaust crashes were before I put in the poly engine mounts:



Easy, fun, and a great diagnostic tool.

-ch

__________________


1994 Mazda RX7 (FD3S) R2 with LS1/T56 and Cobra IRS
430 RWHP / 410 RWTQ @ 2939 lbs in full street trim

I have Delrin bushings in my car but I got to get the damn car on the road. I already have many test I will be doing for PS I can add this to mine as well I'm sure I got a webcam around some place if not i will go get one.

__________________
Josh
I will finish your LS1 swap PM me!
74 MGB GT V8 12.83 (75 shot)
93 LS1/T56 Rule 3 No V8 RX7's

I totally agree, we should even see if we can get a Rotary FD to play along with our testing. What is acceptable and what's not??? Are we engineer's? or are we just speculating on something that we don't know? Hyperion, I think your work in figuring this out is just amazing, I don't really know what to call it because I don't think I would of ever thought of it, but we need to go back to the begining, test a Rotary FD, then an LS1 FD with torque arm/stick diff, then even one with my rear setup.
I am not against what is being done here, I think it's great! I just don't have the knowledge or training to know what is acceptable and what's not? I really think we need to go about this from the begining like I stated previous, Rotary stock FD, and then take it from there.

I don't know if this helps anybody but here are a few pics of the Cobra cradle I copied while I was doing my research on the swap.

__________________

How about a corvette rear. That 3.42 would be perfect.