opensource 3d printable electronic shifting?

[geeksville]

Hi,

I'm forking this thread to discuss ideas raised by casainho and mittkonto: A small electronic shifting device would be a great project for someone who has time and interest. Using the board that casainho pointed to and a linear servo (like this: https://www.aliexpress.com/item/32916297734.html?spm=a2g0o.productlist.0.0.67e238c4HGEhkm&algo_pvid=fe9b80e1-c9c2-4b52-b57c-5d6e015e9a22&algo_expid=fe9b80e1-c9c2-4b52-b57c-5d6e015e9a22-2&btsid=c05607e9-5d02-4fa0-9520-264bb5d4fcd7&ws_ab_test=searchweb0_0,searchweb201602_1,searchweb201603_60 ), I bet someone could make a nice electronic shifter.

A few questions:

Anyone have an interest?
Anyone enough of an engineer to do the math and figure out if the servo I linked to has enough force to nicely pull a shifter cable?

I'd like to work on this someday but my recreational programming time is booked on SW102/850C stuff for the next couple of months.

The idea would be a small/open device that you could mount 3" to 12" away from your derailleur and it would pull the shifter cable to do nice shifting. This would be nice because it could talk to the SW102 ebike controller and coordinate shifts with motor changes, and even do nice things like slightly depower during shifts, and auto trimming after a switch to ensure the chain never rubs.

 

[amberwolf]

Have you looked at the recent (in progress) thread specifically about doing this?

https://endless-sphere.com/forums/viewtopic.php?f=2&t=102122

 

[geeksville]

Have you looked at the recent (in progress) thread specifically about doing this?

https://endless-sphere.com/forums/viewtopic.php?f=2&t=102122

No! Thanks for the pointer!

 

[geeksville]

ooh e3s's comment led me to google this:

https://www.instructables.com/id/DIY-Electronic-Derailleur/ and https://github.com/nabilt/diy_eshift

It looks like using a regular rotating servo has enough go for the task! And a little BTLE devboard connected to a servo would do everything much more nicely than this 7 yr old proof of concept. Though a linear servo would be nicer because the bike would stay in gear even if the motor was turned off .

 

[geeksville]

I went ahead and ordered one of these little linear servos. In a couple of months I'll do a small experiment with a derailleur on my desk and if it looks promising see about coding something real up (with a little plastic housing and minimum soldering required)

 

[e3s]

I see this calculator: https://daycounter.com/Calculators/Lead-Screw-Force-Torque-Calculator.phtml. maybe it's ok that servo. I think that the for safety reason is better to let the original shifter and the cable and put between the derailleur and the shifter a linear servo. cutting a piece of cable housing, and putting an homemade linear servo with a screw that pull the housing cable, consequently the same cable, you let the original config, you have the security that in a absence of energy you have a mechanical shifter and you have a simple solution that can be adapted to every derailleur.

 

[geeksville]

I see the appeal of keeping the original shifter, but I also like the idea of just making a little plastic box that can be mounted (using a regular derailleur) cable a short distance from the derailleur. If the electronic breaks you just end up in whatever gear you were in. a little plastic pod like the (kickstarter disaster IMO) xshifter: https://www.sportcrafters.com/products/xshifter

Ideally a printable housing + some screws + an off the shelf linear servo + a small devboard. Then I bet there is a fair chance one of the chinese clone houses would be interested in spinning a real PCB and just use the open source sw.

 

[e3s]

Take in consideration that sram pulls more cable, shimano less. Minimum 2 cm maximum 5cm depends from the number of the gears and if it's a shimano or sram derailleur. Or a less known mark.

 

[geeksville]

Really good point!

 

[e3s]

If you search stepper motor slider you find something cheap and almost ready to use

 

[geeksville]

If you search stepper motor slider you find something cheap and almost ready to use

Oh that's great! So many good options: https://m.aliexpress.com/wholesale/Stepper+motor+slider.html?channel=direct&keywords=Stepper+motor+slider

Now hopefully I can find a suitable tiny existing CPU board with BTLE and a stepper driver.

 

[guancio]

If you search stepper motor slider you find something cheap and almost ready to use

I don't think just a stepper slider would work. If you connect the slider to the shift cable, you must keep the motor pulling to counter the spring of the derailleur. The existing DIY proposals attach the stepper directly on the derailleur and remove the derailleur spring. You still have to keep the motor on to avoid chain movement, but the torque required will be smaller.

 

[e3s]

I'm curious to see a stepper motor directly in a sram derailleur
Anyway a nema stepper motor has enough torque. I read that some linear slide has only a force of 500gr at 360rpm. Maybe is not enough

 

[geeksville]

I don't think just a stepper slider would work. If you connect the slider to the shift cable, you must keep the motor pulling to counter the spring of the derailleur. The existing DIY proposals attach the stepper directly on the derailleur and remove the derailleur spring. You still have to keep the motor on to avoid chain movement, but the torque required will be smaller.

Though I think the existence and positive reviews of the xshifter product proves a stepper slider works nicely.

 

[guancio]

I agree, but I guess it uses internally some sort of mechanical means to counter the spring force, probably similar to the internal of a grip shifter.

On the other side, the reduction created by the screw of a linear motor may be enough to reduce the force needed to counter the spring. I guess I ll have to order a motor and check

 

[XShifter]

I saw this thread about electronic shifting. I can't give away any confidential info, but I can tell you its not as easy as it seems. There is no off the shelf solution that will do what you want. Every component on XShifter is custom designed and manufactured for this purpose. Fairly easy to make an electronic shifter, but not easy to make it small, light, cheap and marketable. Anyone interested in working on the XShifter project feel free to contact me privately. We are developing some very exciting new products.

 

[XShifter]

I agree, but I guess it uses internally some sort of mechanical means to counter the spring force, probably similar to the internal of a grip shifter.

On the other side, the reduction created by the screw of a linear motor may be enough to reduce the force needed to counter the spring. I guess I ll have to order a motor and check

In fact, we experimented with all of this. But there is 2 big problems. 1) the counter balance spring works inversely from the derailleur spring. They are pulling in opposite directions. So yes you can reduce the force a little, but not much. 2) Any spring large enough to counteract the force is larger and heavier than just increasing the battery and motor size, so its really pointless.

 

[XShifter]

Anyone enough of an engineer to do the math and figure out if the servo I linked to has enough force to nicely pull a shifter cable?

Not even close.

 

[mittkonto]

...There is no off the shelf solution that will do what you want....

Thanks for making that even more apparent!

..marketable. Anyone interested in working on the XShifter project feel free to contact me privately.

Personally; Not interested in a steep $400 cul-de-sac.

However... I would be interested chipping in $??? towards an open road though!
Doing my part injecting even more joy in pedelecs adventures!

...so its really pointless.

Not even close.

Hm... See, speaking for myself.. "open source", methodology of science, heck sharing cooking recipes or whatever. There is a lot of parts that goes into these concepts.

Could be, you (have) know.... We are all kids on the shore, with curiosity and enjoyment of that moment, seeing the day unfolding, picking up pebbles here and there. Marvel at cone shells, struck by awe, realising the nature of a shell is only because of the knowledge from, those shoulders of giants, one choose to stand on. Looking out over the ocean and just "maybe" enthusiastically think; there are ways to go. Vision.

Please, don't curb that wading.
"That's bad mkay."

 

[XShifter]

We can't make it open source, but hopefully in the near future, after its all stable, we will open the API so its possible for related products to join the network, make custom mobile apps, etc.

 

[guancio]

I agree, but I guess it uses internally some sort of mechanical means to counter the spring force, probably similar to the internal of a grip shifter.

On the other side, the reduction created by the screw of a linear motor may be enough to reduce the force needed to counter the spring. I guess I ll have to order a motor and check

In fact, we experimented with all of this. But there is 2 big problems. 1) the counter balance spring works inversely from the derailleur spring. They are pulling in opposite directions. So yes you can reduce the force a little, but not much. 2) Any spring large enough to counteract the force is larger and heavier than just increasing the battery and motor size, so its really pointless.

I personally really appreciate your feedback. I think it is really nice that someone involved in the business (don't seem to me that Shimano is posting often in this forum ) spends his time to give feedback. I personally think xshifter isn't too expensive in case of single gearing system. Of course it is 400 Euro, but if we take into account small number of sales, time for development etc, it seems OK to me.
ON the other hand, we are all nerds here and we like to rebuild things. My motivations in experimenting with bikes is because it is fun.

 

[XShifter]

I agree, but I guess it uses internally some sort of mechanical means to counter the spring force, probably similar to the internal of a grip shifter.

On the other side, the reduction created by the screw of a linear motor may be enough to reduce the force needed to counter the spring. I guess I ll have to order a motor and check

In fact, we experimented with all of this. But there is 2 big problems. 1) the counter balance spring works inversely from the derailleur spring. They are pulling in opposite directions. So yes you can reduce the force a little, but not much. 2) Any spring large enough to counteract the force is larger and heavier than just increasing the battery and motor size, so its really pointless.

I personally really appreciate your feedback. I think it is really nice that someone involved in the business (don't seem to me that Shimano is posting often in this forum ) spends his time to give feedback. I personally think xshifter isn't too expensive in case of single gearing system. Of course it is 400 Euro, but if we take into account small number of sales, time for development etc, it seems OK to me.
ON the other hand, we are all nerds here and we like to rebuild things. My motivations in experimenting with bikes is because it is fun.

Yes its expensive, but the cost of development was very high. Literally every component on this device is custom engineered and made. Still losing money on it. But we are working very hard to improve the design and reduce the cost so everyone can use it. All new technology starts expensive.

 

[guancio]

I saw this thread about electronic shifting. I can't give away any confidential info, but I can tell you its not as easy as it seems. There is no off the shelf solution that will do what you want. Every component on XShifter is custom designed and manufactured for this purpose. Fairly easy to make an electronic shifter, but not easy to make it small, light, cheap and marketable. Anyone interested in working on the XShifter project feel free to contact me privately. We are developing some very exciting new products.

I totally agree that make it small, light, cheap and marketable is challenging. However, as ebikers we may be OK to compromise a lot. For instance, we all have at least 500 Wh of energy with us and we recharge the batteries quite frequently. So something that consumes 0.5/1% of my battery, that I recharge every 3 days, would be acceptable (and not acceptable in general).

For similar reasons, we don't care too much about weight. Our bikes are already above 20 Kgs and we have plenty of watts to counter the weight of the shiftier.

I would like to experiment with a really cheap and dirt design. A trivial solution would be to take an existing grip shifter, mount it somewhere on the frame, and rotate it via a servo motor. The grip shifter will work as usual, balancing the spring using mechanical means. The servo should have enough torque to rotate the shifter, I think few kgs would be OK, and could be off between shifts.

 

[e3s]

I practically thought of two projects: a linear servo as we have already seen or a motor with a screw that turns a round gear and you can also print this in 3d. It's more compact but fragile. If you think about the compactness you have to say that there are also disadvantages. I would prefer a long screw that is more solid. I'm looking for pieces An idea about the size about could be 8 x5x3-4cm. Too big? It has to be said that the cable to be pulled is about 5 cm. The motor must have more than 2Ncm considering the friction of the bearing I thought to insert. I'll list the parts and then give me some advice on the dimensions.

 

[geeksville]

I practically thought of two projects: a linear servo as we have already seen or a motor with a screw that turns a round gear and you can also print this in 3d. It's more compact but fragile. If you think about the compactness you have to say that there are also disadvantages. I would prefer a long screw that is more solid. I'm looking for pieces An idea about the size about could be 8 x5x3-4cm. Too big? It has to be said that the cable to be pulled is about 5 cm. The motor must have more than 2Ncm considering the friction of the bearing I thought to insert. I'll list the parts and then give me some advice on the dimensions.

That sounds totally awesome. I'm not a ME so can't help with the math. But I can 3d print and even do the cad if you want to farm that off to someone (once you select a suitable motor and gearing)