Thanks for the comments gents, it has made me re-think!
Originally I was a little nervous about mounting like this, and I did think twice about it at the time but then decided on gut feel it'd be alright. I had seen / read that MK mount their centre bearings directly to the tunnel panels. They also do this with it in the wrong orientation - rotated through 90 degrees (the bearing is slightly off-centre within the rubber to cope with sag over time) - and with the first part of the prop at around 8 degrees (!!). The prop manufacturer recommended to me that it was dead straight, certainly no more than 3 degrees. I've mounted it at under 0.5 degrees and I have a bit of adjustment available to square this up if needed. So there should be no / very minimal loading from vibration or phasing.
The prop weighs 8.3kg of which the majority is in the middle, however the two yokes and flanges at each end make up a reasonable weight, so say approx 6kg load at the centre bearing. There's 6 rivets in there all mounted in shear. Two are 3.2mm, four are 4mm. Suspension can be designed around a 3-2-1 principle (3g vertical loads, 2g longitudinal, 1g lateral - https://www.eng-tips.com/viewthread.cfm?qid=300205) - I'd double this to be on the safe side - so thought I could use this for the prop mounting too.
So the prop 'weighs' 36kg in bump, 24kg in braking / acceleration, and 12kg in cornering. Therefore in bump, each rivet takes 6kg shear force (way under the maximum load to allow for poor hole / material prep: products.afi.cc/contentonly.aspx?file=images/vendors/POP.pdf - Page 14). In braking / acceleration they take 4kg, in cornering they aren't loaded at all but the side panels are. Pressing with 6kg of force on each side panel yields almost no deflection.
In short, I think I'm covered - however - given the side effects of a floppy prop at 7000rpm I think I will add an extra diagonal brace to firm things up. I have a feeling the main thing that may contribute to any failure would be fatigue of the side panels in the area around the rivet heads, so reducing this loading should sort it.
A bit more work yesterday and this morning yielded progress on the first wishbone. I'd had a nosey around a load of different vehicles from single seaters to grass trackers (for when you drive the car @maurici) and liked the look of these: http://berrisford.co.uk/index.php/parts-components/wishbones/bottom-wishbone-5-8-unf.html and these: https://www.radicalonline.co.uk/products/SW0013-%2d-Wishbone-Lower.html
Reading some blogs written by an FSAE (Formula Student) judge was very interesting too, especially helping to consider load paths: https://www.formulastudent.de/pr/news/details/article/pats-column-rod-ends-in-bending/
Ludemann Engineering also have loads of photos of similar stuff they have built: https://ludemannengineering.com/2012/06/16/fabricating-the-lower-a-arms/p1100115/
I am aware the design I've gone with isn't 100% ideal, but I feel it's a good compromise given the tools I have and how overbuilt the wishbone / joints are. Anyway, onto manufacture...
Spherical bearings need a little bit of 'dimensional assistance' to fit into the housings...
Threaded sleeves turned down, hammered into the end of the tubing, welded round, ground back, then tubing bent:
Starting to machine the ends of the tubing to fit the spherical joint housings:
Still in two minds whether to use the 'classic' (but slightly bodge-y) upper wishbone method of sticking a rose joint in the end of an A arm, or doing it properly with another spherical joint and then having any caster / camber adjustment down at the chassis end. I figure I'll only set the car up once or twice ever, I know my camber and caster settings already to within half a degree, and I can always make more upper wishbones if necessary; so will probably go down the spherical route as long as the lower ones don't turn out to be a mission!
Back to some torque talk () earlier in the thread, regarding how 'BECs have no torque'. I've made some assumptions for the following charts, which are:
- BEC 500kg with driver, CEC 600kg with a Type 9 or 15kg less with a 4 speed box, therefore both fairly dedicated sprint / track cars.
- Both top speed of very close to 125mph, in 4th gear for the CEC because 5th is useless without shedloads of power, 6th for the BEC.
- The Fireblade engine makes more than standard power in a car because of exhaust limitations when installed in a bike. I've been conservative and gone with 185bhp / 110Nm flywheel power (even though some speed series cars a have this and more at the hubs...!). For the CEC, I've gone with 200bhp / 200Nm at around 8000rpm.
- These are calculated using peak torque, working from the engine back to the hubs.
- The Nova box is on there for comparison because the engine I have has one in apparently, rather than a standard one.
So here's the first graph. This shows raw torque at the prop vs gear. The bike has a ~1.72 drop gear built in, so the 2.28 1st gear (standard Honda box) becomes effectively a 3.92, and so on with all the other gears. Yes, over 450Nm at the prop in 1st gear with a 'puny' bike engine! This is what caught my attention in the first place and made me do these comparisons.
Second graph - this is now tweaked to take into account the diff ratio. The BEC loses out a bit here as it needs a taller diff to make the 'right' maximum road speed.
But then, when we add in a 'per ton' part to the calculation, the reduced weight of a BEC makes a big difference:
This shows that, at peak torque in 5th gear, the BEC has around the same torque at the hubs per ton as the CEC in 3rd. The closeness of gears also means that with each gear change, the bike drops back in closer to maximum power.
"He's being stupid!" I hear you shout. "This can't be right as bike engines have no torque, and they need revving to make their power!". Well yes they do, but this is in essence why they make power. BHP is a theoretical thing calculated from torque and rpm - rpm is effectively 'how many times per minute does the engine put out the torque it's making'.
If the shape of the power curve was wildly different (eg. the bike was really 'peaky') once a couple of dyno plots were stretched to fit the same axes then this may be right, but... Here's an overlay of standard CBR1000RR in yellow (tweaked up to 185bhp from 175) vs a 200bhp Zetec with a set of cams (black line):
Therefore, in fact, it kinda demonstrates you don't need to rev the BEC as much (relative to it's limiter), given that power tails off towards the top end. This may be different in a car install though, due to the exhaust differences mentioned above.
To put a practical case to it... 70mph cruising in top gear - just over half revs in both the CEC and BEC in this comparison, around where each engine will make maximum torque. Plant your foot and the BEC sends 761Nm to the rear hubs, the 4 speed CEC 825Nm and the 5 speed 718Nm. Add in the weight factor, and the BEC is putting a good chunk more torque per ton to the road.
I'm convinced the 'BECs have no torque' myth is debunked by this, and it certainly feels like it when you drive one!