Universal or Constant Velocity Joint for Steering

[katou]

Anyone heard about using one of these on a bike for steering?

From the beginning, I've wanted to build a recumbent bike, then add an electric motor. I have most of the design worked out, but the steering is causing me some problems.

The design is very much like ones currently available, but the "tiller effect" found on many current recumbents, I can not accept. It's inelegance from a mechanical/ergonomic standpoint is inexcusable.

So, I thought, hey, why don't I turn the axis of the steering 90 degrees? A very free-running universal joint can go almost 45 deg before starting to catch. Two put together, gets me 90.

I know that this is true because I built a model full-size and it does indeed turn very nicely.

However, I spoke to a recumbent bike designer in California (got lucky, he actually picked up the phone himself) and he told me not to bother with the universal idea.

He mentioned the following problems:

1. backlash in the joints allow the wheel (even when going straight) to turn back and forth a minute amount, enough that the driving experience feels uncertain.
2. backlash gets worse over time.
3. universal joints are heavy
4. you can reduce tiller effect to tolerable levels by changing steering geometry (underseat, remote with tie-rod etc)

Now, I pride myself on taking advice of people smarter than me, and this guy has been there, and tried exactly what I'm proposing to do.

On the other hand, I HATE tiller effect, and I wonder if I could find a special "low backlash universal joint".

Anyone know of such a thing?

Katou

 

[rkosiorek]

only extremely esoteric designs like the Austrailan designed "Thompson Coupling" but those are likely extremely expensive as well.

rick

 

[John in CR]

For constant velocity you need a double u-joint, though I'm not sure constant velocity is a requirement on a bike. Let me know if you figure out a way to eliminate the play in the joint. The Ground Hugger XR2 has some kind of joint in the steering. http://www.rqriley.com/xr2.htm

 

[katou]

CV is the generic name for what we in NA call a double U-joint, or double universal. They call it that because when the axle is offset, the turning of the axle does not happen slow/fast, the turning is constant in velocity.

The Riley uses a single U-joint. I propose 2 to turn the axis 90 degrees (actually it will be a bit less, but that's the idea so far)

I just need one that is designed for low (ideally zero) backlash.

Maybe a lossy damper in the headtube assembly to reduce spurious movement would do it?

Katou

 

[spinningmagnets]

I don't have any advice, but you are not the only one interested in this...

 

[liveforphysics]

Go to an auto-wrecking yard. Start harvesting some steering u-joints. Plenty strong enough, and some models are very compact and light u-joints with no noticable backlash until you get them past 45deg or so. Once you cross that ~45deg point, the backlash skyrockets, if you can keep it under 40deg at each joint, I think you would be be happy with a pair of automotive steering u-joints. And if nothing else, they are only going to cost a few bucks at a wrecking yard to buy and test them.

 

[Miles]

An early example...

http://www.dutchbikes.nl/pagina/frame_recumbents.html

Recently:

http://www.sigmabike.com/

 

[AussieJester]

AUTOOOOOBOT!!!! Miles

I never thought of using the steering universal joint excellent idea
would work well on a trike with independent rear suspension wouldn't
it Miles (yes miles good luckfinding a trike with these on the drive train :: stickz out tongue:: :lol: )

KiM

p.s love that carbon fiber recumbent more each time i seez it.

 

[Miles]

Always happy to oblige, Kim

http://www.hellbentcycles.com/work_shop.htm

 

[liveforphysics]

I like the FWD recumbent drive-line concept. This one is unsuspended, but with slip axles to each front wheel and a u-joint on each side, no reason it couldn't have as much suspension as you would want. For sharp steering angles though, going to a CV joint would be a much better option, and by CV-joint, I don't just mean out-of-phase u-joints, I mean a real CV joint.

From one of the .PDF's on the site Mile's linked:

This is a real CV joint.

 

[liveforphysics]

Also, if anyone needs and Honda axle CV shafts... I've got about 20 laying around with snapped off splines various other problems, but an in-tact joint. If you wanted to do a production run of something using them, all the guys I race with also have similar piles of axles and drivelines in various states of twisted, broken, stripped, etc. I would say less than half the time the CV joint itself explodes, which means lots of in-tact CV joints laying around that could be re-purposed.

 

[John in CR]

Also, if anyone needs and Honda axle CV shafts... I've got about 20 laying around with snapped off splines various other problems, but an in-tact joint. If you wanted to do a production run of something using them, all the guys I race with also have similar piles of axles and drivelines in various states of twisted, broken, stripped, etc. I would say less than half the time the CV joint itself explodes, which means lots of in-tact CV joints laying around that could be re-purposed.

For a sense of scale, what's the OD of that CV joint, and guesstimate at weight after minimizing the shaft? Maximum angle? I take it those CV's would have no play for use as a bike steering joint.

 

[liveforphysics]

They do about 40 degrees, then are mechanically stopped from any additional angle.

The splines are 28 millimeters on the small ones in 36 millimeters on the large ones.

The CV part is about 3 inches in diameter, and about 4-5lbs if you cut the axles away from it.

Maybe better suited to a motorcycle than a bicycle application due to the weight, but definitely a cheap/free zero-slop durable option for any angular linkages on projects.

 

[John in CR]

Maybe a cv joint from an ATV is called for then. It looks like they come with 2 of the joints for more angle, and would be smaller and lighter. Might even be able to use the shafting too. I've been thinking about doing a roll bar/cage thing with the front cargo bike I'm building. I need to measure to see if the kids would still have plenty of head clearance with an inline tube. Running the steering through that has the potential to be slick and the tube would add so much rigidity to the bike that I can go really lightweight for construction of the canoe-like cargo/passenger area. It's a far better solution for steering than the boat steering cable I was planning. sweet!

John

 

[katou]

Perfect! Any solution that involves a scrap yard pleases me mightily.

Any chance that they come in aluminum?

Thanks a ton! This is exactly the information I was hoping for!

But seriously Miles, I looked for any bikes that used this linkage for steering, and all I found was the riley. HTF do you DO that?

Katou

 

[liveforphysics]

Right on John Good call with the ATV CV-joint, that would definitely be a more appropriate size for steering applications.

Any chance that they come in aluminum?

Nope, but a couple of my buddies run titanium ones, but it's not for the weight savings, it's to save money on gear sets, because the ti axles act like a little torque shock-load absorber. You're looking at $5,500 for them in Ti, or free to almost free from wrecking yards in steel, so you would really need to be well budgeted to take that approach. Johns idea with using the ATV axles sounds like a much better way to go lighter IMO.

 

[Miles]

But seriously Miles, I looked for any bikes that used this linkage for steering, and all I found was the riley. HTF do you DO that?

I already knew them both, as examples......

If you can keep within 40 deg. (single joint), I think it would be better.....

 

[rkosiorek]

in experiments i have found that even out of phase U-Joints have very limited angles. trouble is that after just a few degrees they will no longer track angles and one will deflect more than the other. at that point it is no longer a CV joint. any futher deflection only aggravates the situation and they will go even further out of whack. as a driveshaft you have to limit the deflection to only a few degrees. 15 or 20degrees maximum for any single pair of joints. even using a pair of out of phase U-joints is not a true CV. the shafts still go through acceleration/deceleration cycles. but they are much less pronounced than just a single joint. this is because the input and output shaft are not perfect and there will be differences in the deflection angle of eacg

Rzeppa or BALL type CV joints are better. but have a lot of friction. mostly because all of the power is transmitted by very small contact area on 6 steel balls. this will not be significant if used for a steering column. still they are the best compromise for drive shafts. commercial units are limited to 45degrees. though i have seen some 52 degree units in catalogs. far too expensive for mere mortals to afford. they do make them in small sizes for use in custom machinery and lab equipment. but you would not find these small size ones in your average scrap yard.

in a steering system for a car there is usually a rack and pinion or ball screw that isolates steering movement from the steering wheel. this usually allows for a bit of slop in any of the ujoints used in the steering column. it is also why include a steering damper.

i like the idea of the ATV CV joint. but you also may be able to use the rack and pinion used on some of these.

rick

 

[amberwolf]

I looked into this when I started ReCycle (the unfinished first 'bent) and later for CrazyBike2, and kept finding discussions and build pages that described the backlash problem, as well as the size/weight problem.

Given that my other choices at that time due to alignment issues and lack of rose/heim joints small enough to be useful on a bike were cable- or chain- operated steering, I gave it a shot. There should be a post from around mid-2008-ish on my blog with some experiments, IIRC.

Basically, I tried a few double-U joints I found in various items, including riding lawnmower steering, and I didn't much like the way any of them felt. But all of those were used parts, some very thoroughly so, so they all had wear on them that might've been the source of the "wiggle". rkosiorek's points about angles of alignment fs rotation were other issues I had, since I needed a fair amount of angle and could only use one joint; I needed about maybe 60-70 degrees I think. I forget what it was now.

I did not end up even trying them on the bike itself, as just mocking them up without a load on them had this problem; the load would've made it worse.

New parts might be much better, especially those from quality-made sources.

That's why in the end I went with the tie-rod steering I did on CrazyBike2, because once I worked out the sizes of bolts, bearings, etc, I reduced the lash to nil (though it was very bad at first!). Once I did that, I eliminated the whole tiller effect from the system, basically giving myself a steering wheel made of handlebars, as if it was directly attached to the fork.

Also, changing the distance of one end of the tie rod to the steering tube at it's end, relative to the other tie rod and steering tube, let me give myself a kind of power steering, so that turning the bars a little could give a lot of steering, making it easier to turn quickly, or turning them a lot could give a little steering, making it less susceptible to false steering inputs from me as I hit bumps and stuff. I preferred the latter after experimentation with both.

With a U-joint, AFAIK you'd not have the ability to do the above, just as you can't with direct feed to fork on a regular bike.


One other option you can look into is a geared approach, using bevelled gears to give a specific steering angle. I seriously considered this for CrazyBIke2, using the angled gearbox (around 60 degrees I think?) out of my Ryobi AC powered lawn trimmer, down at the end of it's tube. But the gearbox was old and worn and had too much lash. I think when it was new it probably did not have much lash, but I didnt' get it until it was already well-used.