Didumos69

Speaking of collars. Maybe a bit off-topic, but one thing I never understood is why the with its overly fat collar does not fully insert into a axle hole. This causing the ball to misalign with a one-stud-grid.

I noticed that too. I hope he got the inspiration for his limiter from this idea . Actually, Charbels 919 shifting mechanism appears to be a combination of this 90 degree limiter and piterx' shift drum.

@ImanolBB, @Rodzilla4, @brunojj1, thank you guys! The reason for not using knob gears at the bottom side of the unit is simply because I don't have enough of them. They do fit in perfectly though and I did use knob gears in the LXF-file. When braced well bevel gears are strong enough and won't slip in this kind of application. However, the advantage of using knob gears is that you can't make a mistake with the orientation of the axles involved. It's important that one of the axles has a 90 degree orientation and the other has a 45 degree orientation, so the mesh makes a 45 degree orientation difference. About the operation of the shifters: Each paddle pulls a pusher against a knob gear independently. When the knob gear gets pushed over the 45 degree point the 90 degree indexer takes over and completes the 90 degree step. This completion makes the knob gear push the pusher outwards, which allows the pusher to be pulled back into it's starting postion by the silicon band. While being pulled back, the 5x7 frames to the sides of the unit (with their smooth sides facing outwards!!!) make sure the pusher is pulled back inwards too. To repeat this process the paddle doesn't need to go all the way back. Half way back is enough to make the next shift. When a paddle sits in its default position the attached pusher does not block the knob gear. If it would it would obstruct the operation of the other paddle. For me this is convenient, because it allows the shifter axle to be operated from outside the unit, which in turn paves the road for HoG shifting, something I incorporated in my ultimately playable Porsche. I agree on the yellow bushes, but color vomit is a different topic. I left the half bushes yellow to get a reference to the original and because I try to reuse parts as much as possible when I make modifications. Modding the real Porsche, cool, huhu!!!

Scattered over the last few months I've been working on a proof of concept of an improved paddle shifter unit for the Porsche 911 GT3 RS (42056) that uses the stepping principle from Attika's stepper. After several redesigns I finally came up with something that I'm so content with - because of its reliability and sturdyness - that I thought it would not only be valuable as a 42056 MOD but could be of interest to other car builders too. Just like Attika's stepper this setup relies on the interaction with a 90 degree indexer. And just like the orginal 42056 unit it makes use of silicon bands to return the paddles. I tried to hide the silicon bands inside the unit, but that would give too much unbalanced stress to the whole shifting system. The operation is optimal when the bands are close to the paddles. About the operation of the shifters: Each paddle pulls a pusher against a central knob gear independently. When the central knob gear gets pushed over the 45 degree point the 90 degree indexer takes over and completes the 90 degree step. This completion makes the knob gear push the pusher outwards, which allows the pusher to be pulled back into it's starting postion by the silicon band. While being pulled back, the 5x7 frames to the sides of the unit (with their smooth sides facing outwards!!!) make sure the pusher is pulled back inwards too. The video shows a setup that mimics the situation in the Porsche, with the 90 degree indexer integrated in the gearbox. When combined with the indexer, the unit itself can be used in the original Porsche as is. Feel free to add comments or post (suggestions for) improvements. LXF-file here.

All replies make perfectly sense to me. And I also have the experience that the 1/2 pin should first be inserted with its pin side and then you can attach something to the stud-side, not the other way around. I prefer the 1/2 pin over the 3/4 pin when I want to avoid slack, but only when the attached element is locked-up someway to avoid it from getting disconnected when it gets stressed. Here's an example: I prefer the 3/4 pin over the 1/2 pin when I want to avoid a thin liftarm with axle holes from sliding off the axles. Here's an example:

That's a very useful investigation, thanks! I assume you replaced all 6 red clutch gears, including the two next to the H-frame. I was planning to do the same, but won't bother after your attempt. This brings me back to the conclusion that the real friction problem inside the Porsche gearbox has to do with the two clutch gears transferring torque on axles rotating at different speed. P.S. I suppose you al know about the Unofficial errata for the Porsche.

The requirement that it has to be done with parts from 42036 only, makes this a kind of puzzle. And I think it's very clever how Victor solved this puzzle, but I wouldn't regard it as a compact solution though. Thin 12t bevel gears take only 6 square studs (for rotating parts) to make a 90 degree mesh. Normal 20t bevel gears and knob gears would take up to 15 square studs to do the same. But this solution takes 51 square studs of space. If this is what it would take to make a C-model out of this set, I would suggest reconsidering all options. Maybe you can do without a 90 degree mesh after all.

Well, let me start with saying: Welcome back! Your story is one of many and we refer to this event as coming out of your dark ages, but maybe you already know about that. I would suggest more or less the same; try to buy cheap sets to get a lot and a great variety of parts. If you want to get into 4x4 and SBrick stuff immediately you could also consider starting with a MOC of one of the top-designers these days, for instance Madocca1977's Baja Trophy Truck with SBrick. Anyway, whatever you do, have fun!

I would like to add my 42056 front axle mod for my ultimately playable Porsche to this collection. It is tailored to relatively heavy builds with its 4 hard springs. Compared to the original 42056 front axles this setup includes Ackermann steering and a slider for the gear rack. It uses the full length of the suspension arms and adds a second (push rod) shock on each side. It eliminates bump stear, increases the front ground clearance to 1 stud and secures the suspension arms much better. In my Porsche this mod matches other mods, for instance the rear HoG steering mod, which connects to the axle with U-joint at the bottom of this setup. LXF-file here (note that I used placeholders for the new 42056 wheel hubs).

I was actually referring to some internet links (pointers) with background info about torsional bending of the chassis. But these are very nice images and make very clear what you mean with this torsional bending. Thanks!

The portal hubs look very clean and the whole setup looks very sturdy. When studying the LXF-file I noticed the pivot offset of the steering rods lies one stud outside the steering axis. Doesn't that give a negative Ackermann effect or is that negligible?

Looking realy cool so far! And indeed I think half a stud bending under 2.5kg load is not bad at all. About the torsional bending in the chassis. Do you have some background pointers on that? I would like to know more about it.

This makes perfect sense to me. All together the new red clutch gear only has disadvantages compared to the old dbg clutch gear. It comes closer to filling one stud and it misses a side with small radius to avoid the large disc brake effect. The only advantage you could see is that it can be used with two driving rings - one on each side - but I haven't seen any practical application of that property yet. As for my ultimately playable Porsche: With this in mind I will be replacing the 4 essential red clutch gears - the ones that are actually used to engage with a driving ring - by 4 dbg clutch gears anyway, even when it won't make any other mod unnecessary. Simply a matter of good practice to avoid unnecessary friction. What remains is the quest for a good explanation of Victor Imaginator's observation that the old clutch parts outperform any combination with newer parts.

Too bad. Thanks for clarifying! Sorry for the noise.

I mean something like this:

Your case - stating that the new parts, including the 3L driving rings and 3L connectors, are causing more friction all together than the old parts - is getting stronger and stronger. I still would like to know why. Could the length of the 3L connector as described by Charbel be the actual problem? Maybe the 3L connectors with ridges can each be replaced by the old 2L ones each combined with 2 half bushes. And maybe in the Porsche the new connectors without ridges can each be replaced by a 2L smooth connector combined with two half bushes.

No I didn't, but given this discussion I think it would be worth a try. Therefore I plan to undo the eliminate friction in the gearbox mod and replace the two red clutch gears just in front of the gearbox with dbg ones and see what happens... I'm also thinking about undoing the change-over catches extension mod and see whether replacing all other red clutch gears with dbg ones would solve the red gear getting clamped issue.

I was thinking of the same thing. I may sound convinced in my previous post - and I am - but would be even more convinced when I could verify that replacing the red clutch gears with dbg ones doesn't solve the friction problem by itself. To find out I will undo the eliminate friction in the gearbox mod and replace the red clutch gears with dbg ones.

I agree! So far I only see LDD-images or renders. I guess you already built a proof of concept for some parts, for instance for the front axles. Do you have pictures, or is it not ready to show?

Before concluding that it's better not to use any of the new clutch parts, I think it's certainly worth trying to combine the old clutch gears with the 3L driving rings. This might very well be part of the friction problem in the original 42056 gearbox, but I'm quite sure it's not the only cause. If it would be, the problem would be equally severe in all 4 gears when pushing along the car. However, my observations are that 2nd and 3rd gear suffered the most from friction, which are the gears that actually have the red clutch gears rotating at different RPM compared to their axles. Gear / axle RPM ratio for each gear: 1st gear: 1:1, 2nd gear: 1:3, 3rd gear: 9:25, 4th gear: 1:1. I also recall having tried to give the 5 16t gears in a row some extra play by agitating the bracing a little bit, but without success. One thing I couldn't understand though is why do clutch gears induce so much friction on their axles when rotating at different RPM? One would expect an axle rotating in a pin hole to have the same problem. My explanation for this is that gears and axles are made from a different plastic blend than liftarms and beams. They are slightly softer, more silicon-like. Liftarms are harder and have a much more smooth surface. My conclusion is that axles rotating in hard plastic pin holes suffer from less friction than axles rotating in softer gears with clutch. Victor Imaginator's findings with his lengthy 5+R gearbox and the issues with allanp's 5+R gearbox, together with my observations in the 42056 gearbox all point in the same direction: When you mesh a clutch gear with more than one gear and it needs to transfer torque while its axle rotates at different RPM, the clutch gear will induce severe friction to its axle. Btw, I had similar problems with Paul Boratko's AWD transmission, which is based on the same idea as allanp's 5+R gearbox. All together I think it would be a good practice to have clutch gears mesh with one gear only so they only transfer torque when they are engaged. I would even go as far as to no longer consider building a car when it doesn't oblige to this principle. Brings me back to Paul Boratko's 4-speed gearbox shown in the video mentioned by Jonfensu. I think that gearbox is as good as it can get with Lego. Well, luckily the images in this thread show which gears we are talking about, altough your image kind of ruins that . But I understand the confusion. I think 'gear with pin hole' would be better than 'gear with clutch'.

To get back to the original topic, I also compared the new and old clutch gears. My perception is that the holes of the gears are equal. I tried to wobble each of them with an axle through them and couldn't really notice a difference. What I did notice was that the newer red clutch gear practically takes one stud of space, whereas the old clutch gears clearly leave some space for play. Here's a comparison of (from left to right) the new clutch gear, the old clutch gear and the normal 16t gear when given exactly one stud space. The last one takes even less space than the old clutch gear: Here's a video showing the induced friction when all 3 gears are given exactly one stud space. I think the difference is clear:

Nice! This all starts to make sense. Btw, the problematic clutch gears (the ones to the left of the first image) in the 42056 gearbox could be replaced with normal gears relatively easily. All remaining clutch gears only mesh with their corresponding idler gear, just like your gearbox above, and the modified gearbox operates without noticeable friction, just like your gearbox above.

This might have to do with clutch gears transferring torque on axles rotating at different speed. A not-engaged clutch gear that needs to transfer torque from one gear to another (inbetween gears) actually works as a kind of lever and will push itself up or down against its axle. When at the same time the clutch gear's axle is rotating at different RPM, it's easy to understand that this will cause friction. This is also an issue in the 42056 (Porsche) gearbox.

That's a serious design! Didn't see it before, As of now I will be following the Unimog mods thread too. Another way to gear this setup down would be to use thin bevel gears instead of knob gears for the portal hubs. If the bracing is okay this should work just fine. I tested this and couldn't get the bevel gears to slip in my setup. LXF-file here.

What exactly do the old style parts do better? Less friction or something else?