Видео

hahaha, it hurts a little but I didn't apply too much force. I need to improve the tolerances

on grease

thinks, I need to improve the teeth profile to reduce the gap because can you can lose torque. The most efficent it is the cycloidal gear teeth profile but then you lose "precision". The long teeth help you to mantain the gear in place and to avoid backlash.

@@urielalbertodiazreynoso6309 Interesting, thank you

yeah, I know, I wanted to know how much would last this thing.

sorry, I will search for the file then I comment you back

@@urielalbertodiazreynoso6309 not a problem thank you for taking the time to look for it

github.com/udiazr/OpenSCAD

ruclips.net/user/shorts7A8bf8zxnrA

I didn't invented it either.

These small 5V motors don't spin too fast and they spin with small torque. I haven't had any visible wearing. The same design but with 12V, 3A johnson motor friction and heat can be an serious issue.

Hola Marco. En la liga de github github.com/udiazr/OpenSCAD/ puedes encontrar el código fuente de este motor (motor25.scad) también el modelo STL para un motor de 25mm de diámetro. La verdad, aprendí sobre estos engranes primero viendo un videos en youtube hace un par de años. Luego leí la patente sobre el reductor cicloidal. No sé cuál sea tu formación, pero el funcionamiento de este tipo de engranes es como el de una palanca tipo 2. Este tipo de engranes en realidad son dos etapas que están "soldadas" (por así decirlo) en un mismo engrane, todo esto para reducir el número de piezas necesarias (aunque con ciertas desventajas). Mi reto fue hacer el reductor de engranes funcional más pequeño posible para ver el límite de una impresora 3d y de este tipo de diseño. Saludos desde Morelia.

It is about 1:n where n is the number of teeth. The ring teeth should be elastic similar to a belt, maybe can be stiffer, try with PETG

It sounds correct, but I haven't calculated yet.

A better approximation it is 2.5:1, the idea is simple, a normal cycloidal gearbox without the inner pinion would be 12:1, but the pinion "multiplies" the number of excentric movements by 5.

thanks

@@urielalbertodiazreynoso6309 liar

Thank you. The idea came to my mind when I figured that the elastic part of harmonic or belt driven gearboxes is what gives you precision. I don't like the inner mechanism and complexity of harmonic gearbox so I designed a simpler one. The problem with a elastic part is that you only have zero backlash at zero torque. I'm not sure about torque increase, you can stop the outer ring with your hand. The longer the teeth the lesser efficient in torque transmission but the lesser backlash. I'm thinking what could be the optimal teeth profile. I will try more teeth profiles then I will make a proper torque test.

@@urielalbertodiazreynoso6309 great work , i think less backlash is better than increasing torque because you can always stack 2 gearboxes with low backlash

Efectivamente, hice una cama caliente para la impresora MPMD utilizando una placa de FR4. El circuito espiral funciona como una resistencia que calienta la cama al pasar corriente. Además, FR4 tiene una excelente adherencia en primera capa. Lo único quizá que no me agradó tanto, es que me quedó un poco grande el valor de la resistencia pero tengo trabajando a esta MPMD con 18V y por lo tanto no hay problema. Pero se podría hacer un espiral con menor resistencia y por lo tanto generar más calor a costa de mayor corriente.

I've uploaded it to github.com/udiazr/OpenSCAD it is called pump03.scad

the backlash is extremely low, more formally, the backlash tends to 0 when the torque tends to 0 (it is very similar what you get with belt based reduction). You can get some backlash applying a very high torque and deforming the tpu ring but you will get at most 1/4 tooth displacement (in this case approx 2°)

I'm using unipolar PSU (a bunch supply at 19V aproximately). also I use half-bridge for right and left channels and full bridge for the subwoofer to get a 6db gain in the bass section.

@@urielalbertodiazreynoso6309Sir, I thank you for the quick answer! I am in no way advanced in this field but how does your setup allows to use single half bridge per Left&Right speaker without switching the voltage polarity to the speaker?

@@kristsm half bridge configuration allows you to push and pull a load as in full bridge configuration or with a bipolar psu, the only caveat it is that load will see only half the voltage.

With 5 pieces you can make a 100:1 two stage cycloidal reducer, with this pinion you can make a 20:1 with 6 pieces with the same design (putting the pinion inside)

Yes, the only thing to do is to adapt it to a stepper motor. Check this 2nd iteration ruclips.net/user/shorts-OH2spgqlLM

The idea is simple, a normal cycloidal gearbox without the inner pinion would be 12:1, the cycloidal gear "advances" one tooth each "revolution" (eccentric movement). The pinion "multiplies" the number of excentric movements by 5. so a more approximated total reduction is 2.5:1

XD thanks

In theory yes, but sadly in the practice no. The very high ratio and the poor tolerances of a cheap 3d printer make most of these gearboxes nonbackdrivable.

Here you can see a torque test: ruclips.net/user/shortskAjC7EDokVM Thanks

Thanks but I don't know those gear types. I only mixed the ideas of the cycloidal (eccentric movement) and harmonic (flexible teeth thightly fitted) gearboxes

Si, para el tamaño y la facilidad de imprimir me parece muy bien.

You can download the source code and STL files at github.com/udiazr/OpenSCAD

I haven't tried to do a gearbox for heavy loads, but it is interesting. I'll try it next.

I use sinousodal gears ( the tooth profile is sin(x) to keep it as simple as possible). The top ring has 2 teeth less that the bottom ring. The bottom ring, the sun and the planets have the same tooth size. Top ring has a bit bigger teeth to make the top and bottom ring the same radius tooth_top=tooth_bottom*n/(n-2). I haven't analysed how the torque is transmitted, is not good but is enough for this application. The tooth size is so similar that makes little difference when n/(n-2) approaches to 1. When teeth size is very different it stuck easily. The idea is similar to the vernier. For example, If we have 20 teeth (t1) in bottom and 18 in top ring (t2). In angular dimensions, 9 * t2= 10 * t1. We have 'positive interference' (tooth coincidence) every 10 t1 (two times in these rings). Here, we have two planets so teeth number must be even, with 3 planets we must have 3 teeth less and teeth number must be divisible by 3.

@@urielalbertodiazreynoso6309 ah i see, that makes sense, you’re sacrificing reduction ratio for the ability to use the simpler planets. In this case it’s very similar to a standard harmonic drive too.

@@johnfelt1089 yes, it is almost the same as the harmonic drive. Also it is the same as the vernier but in reverse. I don't like 'normal' teeth, cycloidal teeth have much better properties overall, but sometimes it is very easy to figure out some designs with 'normal' teeth. I try to do everything as simple as posible. I'm working in a reducer for a 1000KV brushless motor. The key is to use TPU to reduce noise. I've reduced heat and I will reduce vibration too. Brushless motors are much more efficient. I think we can achieve better results using them.

@@urielalbertodiazreynoso6309 Agreed, for my use case brushless motors are necessary due to their power density. Noise has been a major problem in my gearbox, and I’ll be trying a tpu rotor in the lower stage soon.

@@johnfelt1089 TPU will help a lot.

the sun and ring gears are made of PLU, but planets are made of TPU. It is very difficult to break TPU. Thanks.

Plastic is a lot stronger than people give it credit for