A python trike is born...

With all of these reports, it sounds like all that the Python needs to be perfectly stable at high speed is a fast single board computer, a dozen or so actuators and linear sensors, and some good AI code to assist the pilot. Since a lot of machine learning code is done in the language Python, it might work out quite well.
 
Hi folks thanks for all the interest suggestions and humour, work temporarily distracted me from essential diy activities so I’ll try to answer now .
 
With all of these reports, it sounds like all that the Python needs to be perfectly stable at high speed is a fast single board computer, a dozen or so actuators and linear sensors, and some good AI code to assist the pilot. Since a lot of machine learning code is done in the language Python, it might work out quite well.
Yes but as a friend told me once « it could be worse « « you could be a nuclear engineer « or say there could always be a complex engineering solution to a problem that should not exist, like building an unstable bike after more than 100 years of common sense
 
My Python is currently "rusting resting" in my shed.
It was an interesting experiment. I was able to ride it "hands-free" in empty car parks and do figure of eights etc. with no trouble.
On the road it was a different matter. The smoothness of the pivot I made interacted all too well with the road's camber and I often found it trying to cross the road as the trike bent in the middle.

Going downhill at speed was "scary" to say the least. ;)

I shall probably recover the useful parts and scrap the frame to make room for....... ?
« resting » that reminds me of the « monthly python dead parrot » scene. Resting let’s hope you can revive it.. I felt the same it s fun on flat car park deadly downhill in our mountain roads.
 
So idea prop up the front wheel so it is vertical and can't turn.
Then add packing under the rear wheels and measure the pivot angle.
When you have something near mine make some rear drop out extensions [ they could just be bolt on ] and see what happens the the handling ?

Paul
Ah interesting experiment I ll try I should have built it with rod end bearings that can modify the angle, but altering the wheel height is a test alternative thanks
 
Thanks stormbird for the comparison side by side and analyzing behavior of the trike.

I will try to summarize my understanding, I was probably wrong in the first place when reading the python archives I should have reached out to ask you before welding…

As for bottom bracket vs wheel seat and steering pivot position…
I tried to replicate the position on my tadpole, and the theory on the python pages..

So the seat to bottom bracket distance is fixed by leg length or inseam,
I wanted 20 inch wheel so there was room and the question was where the wheel should be.

Placing the pivot and wheel close to the body seemed to be logical as steering movement is at the hip joint and that minimized negative trail as well I remember reading some theory on the python page saying that would reduce pedal induced steer.

So it seems that in practice your geometry works better with the wheel more forward close to the bottom bracket and seat more back with room for the pivot in between.

As for rear brakes thanks I was going to couple the two back brakes, but it makes sense that unbalanced braking would induce steering in the python ( but again what doesn’t ? What keeps it straight is yet to be found… ) so keep them separate…

I agree that turning is very short and driving with hands off looks cool, but randomly steering the wrong way is a serious concern… I should give it more time to learn…


Thanks a lot for the python free list I ll try to see if survivors have suggestions…

Thanks for for pointing out the virtual pivot steering bike from Germany.
It looks elegant and vps is impressive but may have caveats one finds out by trying..
I have the impression the contact point of the wheel and instant pivot point move a lot so although trail is positive dynamics may be different from a real pivot.
It looks nice and vps seems to give the ability to set caster angle and trail freely without a big head tube between the legs, but then the guy still discarded his prototype in the end, not sure to understand why, still a very interesting idea.
Considering the trouble I have with one pivot, I ll avoid the complexity of four for now…

So thanks again for suggestions and sharing experience I ll sleep over it and try improving the wild snake …
 
I think the main problem is I can't put my finger on why mine is OK and others are not !

Your experience sounds more like DannyC and Pegasus than mine , but why ?

I did lengthen mine to 48" w/b simply by moving the rear wheels backwards and it behaved just the same ?

By no means is mine perfect ALL Pythons show instability at speed , and I would certainly not ride one without steering handle bars.
[ as a lot of the 2 wheelers are ]

The rear brake steering is useful when free wheeling down hill and you just want a slight turn to follow the kerb , the turn is very slight and
easily held once you expect it.

Sadly my trials and tribulations learning to ride one has been lost in the intervening years I do remember thinking the steering was very heavy
that turned out to be my legs resisting the movement of the front , once leg steering is mastered that goes away.

I do remember my first excursion out of the estate I live on was about 1.5 miles to a friends on a nice wide pavement and had no difficulty on the way there however on the way back I realised I had removed my hands from the handle bars and was cruising alongside a road with 40mph traffic on it and my hands on my lap !

One thing you could try is somewhere safe with a gentle slope once it is moving take your feet off the pedals and nudge it of course ? does it
correct itself like a normal bike would i.e it is stable or does it continue off course i.e unstable.

The virtual pivot from what I understand worked well however his experiment with wood did not stand the stresses very well and was beginning to
fall apart not helped by the fact his front end pivoted forward and was held by some wire that was beginning to fray.

Paul
 
Ah interesting experiment I ll try I should have built it with rod end bearings that can modify the angle, but altering the wheel height is a test alternative thanks
I think I have this wrong ? should you not be lowering the back end to reduce the pivot angle ?
Kids bikes come with 18" [ rare ] 16" [ common ] and 12" wheels so plenty of choice to change the pivot angle ?

Paul
 
Hello thanks for your encouragements and suggestions, sorry I had no update this week I had no progress to share,
work and also the "normal" tadpole trike broke on me that was a bit dangerous - there was an unique M8 screw to serve as an axle for chain pulley and support the seat, so when it broke the seat sank and wobbled, and the pulley and guide tubes blocked everything.
I had fun extracting the broken screw from inside the frame and cleaning up the threads ... 🥵

So back to the pyhton trike, I welded on the posts for the rear V brakes, and painted the frame blue
I thought the color was an essential element to improve steering stability ... at least it is sort of reassuring to finish something.

I also fit an ugly but removable carrying box, the rear frame was sized to accept a standard euro box 60x40x42cm for loading.
weight in the back did not change much the steering but it was kind of expected, the CoG is hardly moving by putting 10kg in the back.

On a positive note I noticed I am getting used to pedalling without causing too much pedal steer, by pushing slightly towards the outside.
As you said also getting used to steer with the legs, they are way stronger than the arms and determine steering, arm restance is futile.

Uphill here in the mountains is fine the trike is like a tractor, with the MTB gear ratio on a 20 inch wheel it climbs well but slowly
Uphill is also very stable in direction, I think because the weight is behind the front whell traction, so whatever trail / angle it works.

Downhill (very gentle like car park 2% slope) I tried "free caster front wheel" as you suggested, not reassuring
The way I found to stabilize steering downhill is stop pedalling and hold one leg straight against the frame, acting like a damper and steering.
I have to try a steering damper next as suggested earlier.

Thanks for the wheel size idea I didn't get it at the beginning that makes sense I can effectively change pivot angle and by putting in small whells, I ll try that I have made a quick drawing and your suggestion would allow to change angle from 62° to 58° with 16 inch wheels, nicely it does not entail modifying the frame.





54113772519_a4632e38b1_c.jpg


54112577547_4269b56436_c.jpg


The steering comes apart to take out the front, bearing cups need to be tighter than on normal bike because there are axial and radial forces here, those bearing cups may wear out quickly but wanted to reuse bike components / critical welds of a MTB frame, and the head tube is not easy to weld.
In retrospect I think your "rod end bearing" solution is more efficient & modular , easier to take apart and modify if need be.
I also supect that the slight friction of those bearings ( they have a bushing and are not roller bearings if I am correct ) may help with steering "nervousness"

54113703463_a6e8bf5275_z.jpg


Finally one thing that is sturdy and practical is the seat mount, made with dropouts :
I think I will try to lower the seat by another 5cm or so.
54113772639_f1c90e82cd_c.jpg


Thanks again for interest and suggestions, the kids bike rear wheels are worth trying out and I am way beyound ridiculous at this stage anyway ;)
 
The way I see a Python being ridden is it sort of has 2 modes ?
a) Mode 1 you are starting off [ or have not reached the maximum speed of the gear you are in ]
In this mode the weight you are accelerating if on level ground or could be the hill you are climbing is absorbing all the power you apply to the pedals and so almost no power goes into steering ? On mine I can see the front wheel hunts gently left and ride however because it is equal and rhythmic the resultant direction is a straight line.
b) Mode 2 we are now approaching the maximum speed we can achieve either because we need to change up or because we are picking up speed down hill and the trike is overtaking your pedaling input. For what ever reason the load [ that make the trike stable in Mode 1 ] has been removed from the equation. IF you continue to pedal that force now goes 100% into steering and causes the trike to become unstable. As you found leg steering is far stronger than arm steering.

Hope you a find a solution to your problem , they are great trikes however they will bite the unwary ! [ even mine ;) :) ]

Paul
 
Forgot to say yes my Python does come ' interesting ' at speed mainly the steering becoming very very light ? however I did ask before on this forum why my other delta a Hase Kettweisel also did this.

I have not seen a plausible explanation as to why , nor do I know whether tadpole trikes also do this ?

Paul
 
On a delta not only do you have just one wheel doing the steering but you are sat mainly between the non steering rear wheels so weight over that single wheel is low. On a taddy you sit between the steering wheels so grip starts off high.

Delta
One wheel + low weight = low grip

Taddy
Two wheels + high weight + weight transfer into a corner = high grip

It's down to loosing the self centring effect because grip is being lost. As you turn you should feel a resistance to that turn through the bars. You only get this if the tyre(s) has grip. You notice this loss on even the most minor inputs.

The light steering effect at speed is because there is less resistance to turning the wheel at speed because it is not fully following the direction it is being pointed in but is skidding to some degree. At lower speeds the wheel has sufficient grip to make itself and thus the trike follow the input. The wheel transmits this feeling of grip to the handlebars. At higher speeds the grip required to turn without skidding is higher because of inertia wanting to push it straight on and the single, low weight tyre can not overcome that inertia so it skids in a turn to some degree. This lack of grip is also transmitted through to the bars.

It's just like a car hitting ice in a turn. Suddenly the steering goes light because there is not enough grip and the steering wheel doesn't offer any self centring effect. Effectively the steering behaves as if the wheels were pointing straight and doesn't self centre.

On top of this a python may also have other issues because of it's unique geometry, but regardless that geometry doesn't negate the low grip scenario above.
 
Last edited:
The only way I can see to add more grip is to use a tyre with more road contact, ie wider, grippier and lower pressure plus, if you can find one, a more square profile. The only other way is to add more weight over it and that means getting nearer the edge of the stability triangle. Nobody ever claimed the Reliant Robin was a good track day car.
 
Thanks popshot that makes it clearer now and relates to the feelings on the road

The tadpole i have has low seat and quite a lot of trail and angle, it is very stable at speed and the resistance and feedback in the handles greatly increases when cornering so i guess i am still in the safe zone, really feels like you are grounded and on a track.

For the python the slightest steering input seems to be amplified, i dont know wether it is due to delta configuration or python geometry, or my diy implementation. There is a lot of weight (me) on the front wheel so I guess my design is not great altough i cant point out an obvious flaw wrt other python trikes. it feels like the opposite of the tadpole like having to focus to keep it straight.

Thanks Stormbird at least your experience is more positive that is encouraging or possibly you adapted faster to controlling the python, let s persevere.

I also relate to and agree with your explaination, uphill or accelerating the front wheel is dragging the rest as a trailer, hence aligned and stable. Downhill or decelerating the back is trying to go faster than the front, and you notice the " inverted caster wheel" is not such a brilliant idea after all.

I should train at the local supermarket trying to go straight with trolley wheels reverted, maybe one day i ll tame the python
 
Last edited:
Popshot

At last an explanation I can be happy with , so also I expect small wheels like 20" or less are not helping this situation. On a Python I am actually sitting where I would be on a tadpole ? just with the wheels between my legs instead of outside my legs. I have spent several hours musing on how to make a Python quad ? a pair of wheels with identical pivots could be connected like a parallelogram however we than have the thorny problem of getting drive to one or better still both of them ?

Something like this but for one rider :-

There is a video somewhere of him riding this solo with just a length of wood controlling the other empty side !

If I could get a pair of pedals like a pedalo between the 2 bb's however I don't think there is anyway of connecting them together with the extreme angles the front end goes through...

Oh well keep dreaming..

Paul
 
You just need three pivots all in a line. The central one runs to the bb. You then link all three with tie rods and rod ends. The only issue is placing the tie rods where they don't want the space your feet do. Ideally the tie rods need to go as far forward as possible. The best solution would be to get your feet high and the tie rods under your feet.

The knock on implications of having a pair of tie rods becomes one of getting on the quad as the rods intrude on where you want to stand as you mount it. Not a problem for young pups but the more experienced hands may curse them.

There is the issue of Ackerman and this can be introduced by a gap in the central tie rod mounts to the bb only pivot.

I can cobble a poor mans cad if you can't picture what I mean.
 
The idea of a python with two heads scares me even more than my contraption :eek:

Fair play to you if you can get it to work as a quad

How would you transmit traction to the two wheels without a small scale replica of a car drivetrain though ?
 
1STXUxx.jpg


Hopefully that explains it. The tie rods ideally need to go fairly far forward or their accuracy in controlling matters is working against the leverage that can be exerted against them. Unfortunately to put them ahead of the wheels means them sticking a long way out so they need to go behind the wheels. This will require fairly beefy tie rods I think to handle the leverage stresses. I'd go M10 male with 2mm wall tube and not the cheapest rod ends you can buy either. They will take some forces in that location. To help matters I've moved the wheels forward in their frames but no further forward than the far pedal so overall no longer. Whether this move helps or hinders all the foibles of pythons I have zero idea and whether two wheels halves or doubles python issues - same again. Guinea Pig required.

The Ackerman works like this...

VPINtBn.jpg
 
Last edited:
Those rod ends will have to be low enough mounted to clear the inner frames as it tilts as it turns and you will almost certainly need rod end misalignment spacers because of that tilt to get the maximum articulation from the joints.

Because of Ackerman moving the wheels differently the front drivetrain becomes ridiculously complicated. There are two solutions. The least appealing from an engineering perspective is to bring the drive down to the pivot and take it across to the outer pivots via sliding shafts with universal joints then back up via chain. Heavy and complex. The more sensible option is to take it to the rear. This needs a twisted chain solution like a fwd swb recumbent two wheeler without a MBB. The principle is the same but with less twist in it and the only reason a single front wheel python trike is fwd is because it is easier to do. There is no good reason not to take the drive to the rear.

The benefit over a conventional quad is it can be narrower as the legs turn with the front wheels so no space has to be reserved for wheels to turn into. If you aren't going to take advantage of that narrowness then I can't see any point in doing it. Going rwd means you can make the wheel frames narrower as you don't have any cassettes on the wheels saving a bit more width.

Simples.
 
Last edited:
Sorry Trickytrike I did not intend to hijack your thread with my ramblings.

Yes I was really expecting a miracle where someone spots a FWD solution I had not though of.

Is it worth building , I doubt it I am currently more interested in the quad I am currently building

and maybe a very light Python based on my racer with some weather protection [ sort of back burner should the quad disappoint ]

Paul
 
Click for DIY Plans!
Back
Top