Carbon Fibre pushrod, a closer look

[Brian.G]

that sorta makes sense not the way I would build that part I guess.

Indeed, there are other ways for sure.

Brian.

[gambler]

Looks like the outer jacket was formed from a woven "tube" of carbonfiber, and blown out with a bladder or mandrel, then machined on the ends afterward.

[riff_raff]

Brian.G-

Thanks for the fantastic pictures. It's too bad the pushrod had to be sacrificed.

The laminate looks quite good, and just what you would expect for a composite structure in compression. Lots of uni plies with an occasional 45 weave to keep the uni's aligned. The laminate also has excellent consolidation with no visible voids in the highly loaded areas.

The same can't be said for the insert bond lines though. Most importantly, the machined inside surface of the composite was cut back far too much (in one picture it appears to be almost 50% of the wall thickness). While the wall of the insert has some lengthwise thickness taper, the very inboard wall should be as thin as possible (1mm or less). Having a very thin, flexible wall at the edge of a bond line helps to prevent peel failures. The voids in the bond line could also have been avoided by putting vent holes in the insert for the excess adhesive and trapped air to squeeze out.

Otherwise, it look good.

slider

[Brian.G]

Brian.G-

Thanks for the fantastic pictures. It's too bad the pushrod had to be sacrificed.

The laminate looks quite good, and just what you would expect for a composite structure in compression. Lots of uni plies with an occasional 45 weave to keep the uni's aligned. The laminate also has excellent consolidation with no visible voids in the highly loaded areas.

The same can't be said for the insert bond lines though. Most importantly, the machined inside surface of the composite was cut back far too much (in one picture it appears to be almost 50% of the wall thickness). While the wall of the insert has some lengthwise thickness taper, the very inboard wall should be as thin as possible (1mm or less). Having a very thin, flexible wall at the edge of a bond line helps to prevent peel failures. The voids in the bond line could also have been avoided by putting vent holes in the insert for the excess adhesive and trapped air to squeeze out.

Otherwise, it look good.

slider

Yes, thats one thing i forgot to mention along with the square cut, a LOAD of wall was indeed cut away.
You mention vent holes, why so? Was the insert bolted up to a jig on inserting?
Because the insert has a large hole running through centre(where spacer is) The bolt hole tappings are also not blind, and open to the interior. You would think this would be more than enough venting, unless as mentioned insert was bolted to a jig.
Ive always considered vent holes to be a bit pointless, although I understand their theoretical use.
I always thought that interior pressure would be a good thing, In that it would pressurize surplus epoxy, (as in image) back into and out the bond line before the insert was pushed fully home. If anything, Id be pressurizing the rod to promote this. Obviously correct epoxy viscosity would be needed to both stay around the internal circumference of the insert, but yet be thin enough to be forced back out between the two.

Perhaps my theory would need testing, but from what I see here, ample venting makes no odds, or maybe the operator was just having a bad day.

Brian,

[humble sabot]

Well the viscosity is the constant in the equation. The solution is to adjust the fit or add a vent.

[marcush.]

I have to admit that I´m surprised with the detail workmanship flaws in these critical components.
I thought the process would not allow things that are obviously not best practise in cf fabrication.

was there a mention of the origin of this part?

[mep]

Regarding the 90° end mill,
it might be because of a practical reason because no ball mill long enough to reach the deep pocket might be available.

However milling into the fibres should clearly be prevented.

[marcush.]

you would lay up with peel off tape to create the bonding surface ,no?

[flynfrog]

you would lay up with peel off tape to create the bonding surface ,no?

generally peal ply makes a bad boding surface. abrading the surface gives a much higher level of surface activation and a better bond. You can also sand down to the fibers so that you bond to them instead of the resin.

[marcush.]

you would lay up with peel off tape to create the bonding surface ,no?

generally peal ply makes a bad boding surface. abrading the surface gives a much higher level of surface activation and a better bond. You can also sand down to the fibers so that you bond to them instead of the resin.

so why would someone even consider using peeloff ply?
I see the charme of using a perforated release film offering smooth surface finish and on the back side of the part you have in todays systems no need for release agents that could spoil your bonding surface..

[flynfrog]

They use it as a cheap short cut in non critical areas. Peel Ply does not give you a good bond on a molecular level. Abrading will energize the atoms on they outer layer surface providing a much better bond than using peel ply. Also peel ply can leave Teflon contamination in the bond. Look up some mil specs for bonding most of the ones I know require abrasion.

[marcush.]

They use it as a cheap short cut in non critical areas. Peel Ply does not give you a good bond on a molecular level. Abrading will energize the atoms on they outer layer surface providing a much better bond than using peel ply. Also peel ply can leave Teflon contamination in the bond. Look up some mil specs for bonding most of the ones I know require abrasion.

I´m aware of this .Attended a lecture by Brian o´Rourke(chief engineer composites at Williams GP ,don´t know if he is still in that position ) about structural bonding of carbon a long time ago when he basically teached exactly this.

[bill shoe]

Brian-

Somehow I didn't see this thread until today. I occasionally write posts, and when I do I give some opinion about such and such thing, pretty mild. Your original post here is a really impressive lump of substance. Thanks.

[xpensive]

I can't help from thinking about a comparison with an aluminium pushrod in terms of weight, strength and stiffness,
what's the cross-section area of a typical pushrod like that anyway?

One of the benefits with the pullrod is the minimal cross-section area needed, see RBR, when no Euler-case needs to be considered and the rod itself is made of metal, don't know which kind though, with a simple turnbuckle for height adjustments.

[hardingfv32]

what's the cross-section area of a typical pushrod like that anyway?

55 mm X 15 mm

Brian