DIY carbon fiber axles/driveshafts? Also: anyone have a broken 350z driveshaft I can have?

I went down a bit of a rabbit hole yesterday. Follow me for a minute:

1. I noticed that carbon fiber tubes are very cheap on ebay. I also found some american manufacturers that look like they are also making some tubes for REALLY reasonable rates. Like not scrap metal cheap but could be used on a challenge car reasonable. (for example, this $70 piece could make both axles for my fiat challenge project with tube to spare https://dragonplate.com/Braided-Carbon-Fiber-Round-Tubing-1-ID-x-48)

2. I got a bit engineer-y (watch out) and looked up some formulas for torsion of shafts: https://www.engineeringtoolbox.com/torsion-shafts-d_947.html

3. I looked up some carbon-fiber driveshafts specs from aftermarket companies like QA1. 

4. I looked at the 3M website and learned that they recommend DP420NS exopy adhesive for CF to metal joints, and it has approximate shear strength of 4500 PSI when cured. This stuff is also not going to break a challenge budget.

Here's where I am:

1. I'm considering buying a couple tubes and some epoxy and doing some torque tests. 

2. Engineering calculations using Modulus of Rigidity of steel (11 Mpa) vs CF (33 Mpa) shows that a CF shaft of 1.09" OD and 1.00" ID should have a very similar overall strength to that of a stock 240sx solid steel axle of 1.11" OD. 

3. Looks like CF driveshaft manufacturers are a little more conservative with their modulus of rigidity. If I take a 3.5" driveshaft in steel with .083 wall thickness and assume* it has about the same torque handling as the same companys' 3.25" driveshaft in CF with .120 in wall thickness, I get that the manufacturer is using a modulus of rigidity for CF to be about 80% of that of the steel. This 80% is based on a *big assumption, but it gives me a least a good feeling about being within the right ballpark. 

4. I'm much less sure of the calculations surrounding the adhesives. I guess the process is to basically glue the metal fitting into the end of the CF tube, but I don't know how to calculate about how much surface area I might need. I can calculate surface area of the metal-CF joint and multiply by the PSI of the glue, but I'm not sure how to figure out what that number needs to be. I assume that CF driveshafts have a metal fitting glued into each end and those fittings are probably about 3 inches deep by 3 inches diameter or so. So they would have about 30 square inches of glue area. This is where the 1 inch tube really hurts, since the diameter is small I'd be looking at maybe 10 square inches of glue area. 

So, my questions:

In a very lightweight (1500 lb) car with a motorcycle engine (low torque high HP), am I crazy to try and make CF axles by taking the CVs I currently have, cutting the shafts, and turning the metal down to just fit into the CF tube and then gluing them in? I have to convert nissan inners to fiat outers already so there will already be cutting and fab involved no matter the material. 

Does it sound possible to take a broken 350z driveshaft (did you know they are all CF from the factory?) cut the broken CF away, find a metal sleeve that fits just inside the ID of the CF, and re-glue it and re-use it in a shorter application? Not shooting for thousands of HP btw. 

I know this is not twisting, but this is a 6 lb truss holding 800 lbs at a huge cantilever. Made out of the same tubes I linked above. 

Nearly double your glue-wetted area by grooving the face of the metal flange and pressing glued CF end into the groove.
The overkill on the aftermarket shafts is to try and counter the shock loads and unknown temperature environments.

stafford1500 said:

Nearly double your glue-wetted area by grooving the face of the metal flange and pressing glued CF end into the groove.
The overkill on the aftermarket shafts is to try and counter the shock loads and unknown temperature environments.

You mean like this? (this would be a cross-section) leaving the hatched blue areas of metal on so you can glue to the inside and outside of the CF shaft?

This.... is interesting.

Is "rigidity" really a concern?  I don't care if it twists, I care if it yields.

Knurled. said:

This.... is interesting.

Is "rigidity" really a concern?  I don't care if it twists, I care if it yields.

Rigidity (modulus) is important because it changes the vibration characteristics of the driveshaft.

OP, I think you'd have trouble balancing it.  I was reading this article: 

https://bangshift.com/general-news/project-cars/project-buford-t-justice-update-our-caprice-9c1-gets-the-shaft-a-dynotech-drive-shaft-that-is/

And there seems to be a lot that goes into the design of a driveshaft so that it doesn't blow up.  

In reply to Knurled. :

Well, maybe I am misunderstanding (there are torsional calculations on that same page to measure how much twist in degrees there would be with a certain load on a certain shaft - but I didn't run those because like you i'm not so worried about "how much" it twists as long as it doesn't break).

My understanding was I was comparing the MAXIMUM stress that could be applied to each type of material before a failure occurred. I don't think these formulas are attempting to incorporate real-life things like the spring effect of an axle and shock loads. So I'm just trying to 'make a shaft of similar resistance to failure' as what is already there (which I'm pretty sure wouldn't break). 

Am I missing something here?

350Z's get parted out all the time thanks to people crashing them, but you won't find them in junkyards. I scored a LSD for my Challenge car by keeping an eye on part-outs. 

But the 350z driveshaft diameter is probably WAY bigger than whatever size of half-shaft you're using. If I were you, I would compare a hollow tube steel vs solid shaft vs CF tube. You might get 80% there with hollow steel tube using the same calculations. 

And since this is a spinny part, weight doesn't matter nearly as much as moment of inertia. Calculate the inertia of each shaft for comparison. Now compare that to the moment of inertia of an entire wheel. I bet the wheel + tire makes the inertia of the shaft look negligible, which makes all this effort somewhat futile. 

But if you need to make custom shafts anyway due the nut-so swap you have, might as well go carbon fiber!

In reply to Robbie :

Yes

In reply to maschinenbau :

The 350z shaft would be to try on the AMC.

so here's a question - how do I calculate the MAXIMUM torque that the shaft would see?

Simply take the engine max torque and multiply by the gear ratio? I think that means for the bike engine car i'd be around 220-250 ftlbs max. Maybe double or triple for shock loads and safety factor. (78 rated ftlbs times 10:1 in 1st times 12/43 sprocket teeth?)

I could easily rig up a way to hold one end of the shaft and apply 500-750 ftlbs to the other (hoist my chubby self on a few feet of bar), and see if it breaks. Is it that simple?

Edit: maybe I did the sprockets backward. soooo, 3000 ish ftlbs on the axle? does that sound right?

I have an RX-8 carbon fiber driveshaft in the garage. It's beefier than I thought it would be, and the RX-8 was not a torque monster.

This place is amazing. Home built ferraris and 3d printed velocity stacks and now CF drive shafts. Amazing.

Brett_Murphy said:

I have an RX-8 carbon fiber driveshaft in the garage. It's beefier than I thought it would be, and the RX-8 was not a torque monster.

Ooooooooh. Any idea how long it is center of u joint to center of u joint?

Are all rx8s carbon fiber shafted?

Edit: according to ebay, its 42.5 u-u and it doesn't look like all rx8 shafts are CF. 

You are unlikely to have balance problems with the bike powered car due to lower rotating speed of axles versus the drive shaft on a live axle car like your AMC or Bangshift's car.

Robbie said:

so here's a question - how do I calculate the MAXIMUM torque that the shaft would see?

Simply take the engine max torque and multiply by the gear ratio? 

NO!  This was actual a question a prof put to us in Sr design class....figure out why the school sae baja car broke its diff 3 times in 1 race.  Over the jump, break the diff even thought is was rated way over the engines peak torque output.  The solution was, much to the profs amazement the diff should explode.  The car goes over the jump and the engine free revs to well over operating rpms do the engine and most importantly the flywheel are spinning equivalent to about double what the vehicle speed is.  The wheels come down with huge tires sized for a water event and good for about 40-45 hp, 5 times the engine output and easily able to stock the energy stored in the flywheel.....diff explodes.

Point is you need to look at dynamic conditions...they are always much worse than simple statistic conditions.

....but that shaft seems plenty big for your application.  I say build it!

Max torque specs on those CF tubes don't tell the whole story. Because the thing with a driveshaft is it's spinning really, really fast, so the other big factor you have to deal with is harmonics. I'm not really smart enough to even begin to approach the harmonic and vibrational physics you're dealing with, but someone here probably is. It's definitely a problem to overcome, though. Looking forward to keeping up with this thread.

Question, how much benefit will this add? As compared to a cheaper oem type shaft, drive of half? I tend to look at cost benefit ratio. With my new ecotec engine  in my Alfa Romeo spider, my friends told me to turbo it. I can easily get up to 350 hp.. Without to much trouble, but my tires are limited to 205,50,16. I plan to stay about 210 to 225 up. With this I can get 0 to 60 times about 5.3 to 5.7 seconds

 I could add much more h.p , but this would not gain me to much more value. I applaud  your asking questions, they can be fun to think about, but will a carbon halfshaft or driveshaft  help a lot? Yes they are cool to have,and tell your friends about, but $$$$ ?

In reply to Alfaromeoguy :

I think part of the point was that he was thinking that he could do this cheaply. 

CF driveshafts actually improve driveability - I think they act like a shock absorber in the driveline. Our NA/NB V8 cars can have trouble with "trailer hitching", the car bouncing off the lash in the diff and an aggressive cam struggling to maintain even combustion at parking lot speeds. The CF driveshafts eliminate that. It was a big surprise to us the first time we experienced it.

But you do have to watch out. I shredded a CF driveshaft in our ND V8 in about three days. They have a very friendly failure mode as they just shred and don't cause any other damage, but it's still a busted driveshaft. It wasn't extreme use, just 3rd gear acceleration out of a moderate corner on the street.

The failure was either heat-related - packaging is tight near the shaft - or it was an incorrect weave. You have to watch for that with carbon, it's not omnidirectional in strength so a chunk of tube designed to be stiff in bending may be quite weak in torsion. Even though the shaft came from a well known Driveshaft Shop, the chunks didn't look like the right weave. They did get an aluminum replacement to us the next day, so I give them props for that.

I think that was a 3" tube, by the way. 

I'll see if I can dig up pictures of the chunks, but it weren't pretty.

grover said:

In reply to Alfaromeoguy :

I think part of the point was that he was thinking that he could do this cheaply.  That was my point,cost benefit ratio

Some parts can be homemade,but some should  not be. Engineers  are very smart people,they have very powerful computers and a lot of experience. Trying to glue one up, can be challenging. So be careful  on this. If you do this, I hope it works well

In reply to Alfaromeoguy :

A lot of engineers are not necessarily smart people, just well educated.  Education is more important for something like this anyway.   This thread is about exploring what we need to calculate for when designing something like this.

Keith Tanner said:

CF driveshafts actually improve driveability - I think they act like a shock absorber in the driveline. Our NA/NB V8 cars can have trouble with "trailer hitching", the car bouncing off the lash in the diff and an aggressive cam struggling to maintain even combustion at parking lot speeds. The CF driveshafts eliminate that. It was a big surprise to us the first time we experienced it.

 That is EXTREMELY interesting!!!  A bit late for me since I am almost done replacing the patron saint of trailer hitching with a more city drivable engine.  I honestly think a lot of my knee problems stem from shifting in and out of gear every few seconds when in traffic for the past 8-10 years.  Couldn't cruise or coast in gear whe under 45-50mph or so, it was accelerate and Neutral.  Accelerate and Neutral...

More brain dump to multiple people since I don't feel like replying individually to multiple more comments.

The goal as I initially understand it was to DIY adapter shafts, not to save weight.  Even with machining costs and the price of the tube stock and adhesive, you'd still be WAY cheaper than having a couple axle sticks made, which last time I priced out, were in the $200-300 each range for cheap materials.

Shock loads:  You'd want to account for a minimum of, say, 3:1 safety factor, lots more if you plan on getting aggressive.  I am thinking here of certain Audi transmissions that are tough as nails unless you hop a curb on slicks at 600hp.  Or all of the drivetrain problems I had with an engine that made half the horsepower and 1/3rd the torque of the big turbo and V8 swap guys who never had issues.  I think, but can't read Italian so can't verify, thay Lancia used titanium axle shafts on their Delta S4 rally car not just for weight, but to protect the drivetrain.  Titanium is a very flexible material relative to steel of the same dimensions or fatigue strength.

Driveshaft diameter:  No matter what  the faster it spins and/or the longer the shaft, the larger the diameter needs to be.  This is for basic harmonics reasons.  Nobody runs a 40" long 1" long steel shaft, it is going to be 2-3" tube.   One interesting thing was the "rope drive" in the rear transmission Tempest.  It was installed in a curving arc with four support bearings not just for a lower floor, but it also nuked harmonics.

Knurled. said:

In reply to Alfaromeoguy :

A lot of engineers are not necessarily smart people, just well educated.  

As an engineer who works with engineers all day every day, I couldn't agree more. 

Robbie is taking the right path.  What forces are at work here and what equations attempt to analyze for them?  Then it comes back to if your assumptions are close enough.  Is this really all the forces the part will see?  Is "strong enough" really all we need to be looking at?

I think Robbie could easily calculate if a steel drive shaft is going to be strong enough.  Calculating the torsional strength of a tube is pretty straight forward.  The question becomes how does cf act differently than steel when used as a drive shaft?  

As someone with very limited cf experience I think Keith is onto something.  A cf tube made to resist bending will not be great in torsional loading.  Weave pattern is going to be very important.