Skip to content

Un Blog

Cositas para compartir.

On Lathes and Microwaves

The situation

I like making tools, but I'm always missing parts for them. Why?

  • I live closer to Antartica than most of you.
  • Things are expensive there, if available.

there

I am also a somewhat unreasonable, open-source philosophy maniac, obsessed with distributed manufacturing and technological freedom.

To me, these constraints are opportunities for empowerment.

Flowering Elbow explains the implications of this view in part 3 of his video on an "Idea to make ridiculous shelves from trash (& challenge the way you live!)". Limitations can make you strong.

The objective is to make parts for the micropipettes on my liquid handling robot.

I could have them CNC'ed by PCBWay, JLC CNC, a local shop, or the lathe at my friend's workplace, as I have sometimes done. Plastic injection is reasonable for large volumes, bot not for this niche case. These options are either not cheap, generally available, or adequate enough.

So.

Tree of Requirements

To make lab automation (fun) ubiquitous, the following is required:

  1. Develop a low-cost, hackable pipetting robot. ☑️
  2. Manufacture reliable micropipette tools, with their machined tip holders. ☑️
  3. Reliable access to a CNC lathe. 🏁

Work on the first two is not complete but rather done.

However, for a long time, getting a decent CNC lathe seeemed insane: too big, too expensive, too heavy. I really only needed a small CNC lathe to turn plastic into tip holders, I could not justify it.

Stuck, I spent many hours enjoying Tony's and Brandon's videos on machining, and dreaming of lathes and microwaves.

Then, a few months ago, I came across "The First Tool" lathe on Aliexpress. A cute little machine resembling a watchmakers lathe in size, costing 90 USD + shipping. It would need to become CNC to fit my purposes, which someone had already done of course: the Awesome CNC Freak, paving the way for the rest of us.

The tiny lathe now sits on my desktop, and I'm loving it.

The project

In brief, the objective is to make the lathe CNC and turn tip holders for micropipettes. I have some experience with stepper motors, am also interested in learning closed-loop motor control.

Excited by my ample ignorance in every other aspect of this project, I got to work.

Step 1: select components.

  • LinuxCNC on a Raspberry Pi. ☑️
  • Remora firmware running on a Pi Pico board. ☑️
  • "First Tool" micro lathe. ☑️

All of this I already had. So far so good.

Step 2: define milestones.

  • Spindle speed control and indexing. ☑️
  • Motorization of lathe axes. 🏁
  • Software and CAM setup.
  • Test cut.

I had been playing with this desktop lathe for a few days, figuring out its limitations. I concluded that it can turn aluminium, but rigidity issues and vibrations easily appear. Spindle speed control was also missing, as it could only turn at max speed (2400 RPM approx.)

This was solved with Pi Pico, a rotary encoder (AS5600), and a DC motor driver (VNH2SP30).

alt text

New pulleys for the axes were still due, all of them uncommon. There is no GT2 "stock" I can get my hands on, and McMaster might as well be myth over here. This brings us back to the situation thing I described before: How could I get them?

Or perhaps I could make them instead.

What do micowaves have to do with a lathe?

The Microwave

It turns out you can melt metals in the microwave, use that molten metal to cast pulleys, and finally use those pulleys to motorize the lathe's axes.

All thanks to Denny from Shake The Future. I highly recommend chekcing out his channel out, and joining his patreon.

Dennys videos are great, so I'll spare you the details and share results:

alt text

There are four steps in this process:

  1. Make a mold of the GT2 pulley, using a pulley 3D printed in PLA and plaster.
  2. Lose the PLA from the plaster in a DYI microwave kiln.
  3. Melt the metal in a DIY SiC microwave crucible.
  4. Cast the pulley using a rudimentary vacuum machine system.
  5. Cleanup.

The cleanup takes as much time as the previous steps, but its very rewarding:

  1. Cleanup the pulley's faces and bore in the desktop lathe (might as well use it!).
  2. Tap threads for the setscrews.
  3. Cleanup the teeth by hand.

Results

A perfectly improvable 70T GT2 pulley, 10 mm bore.

alt text

It fits the lathe's spindle nice and tight.

alt text alt text

Thanks Denny, a lot.

Reparación de Lavarropas Inverter LG

foto del manual, error LE

Si tu lavarropas de carga frontal LG F1380TD falla con error "LE" (sobrecarga de motor) es posible que el arreglo sea muy sencillo.

Todo gracias a David Lightman, quien publicó un video explicando el problema, y la solución paso a paso: reemplazar o reparar el sensor de posición del rotor.

sensor abierto

Este arreglo es relativamente fácil, rápido, y 24 veces más barato que llamar a un técnico. De última, algún amigue les sabrá dar una mano, o pueden pedirle a un técnico que lo haga por el precio de un trabajito simple.

Síntomas

  • El lavarropas se enciende normalmente.
  • Al iniciar un programa, el motor no se mueve, o se mueve erráticamente.

Video explicativo

Vayan al canal de David y déjenle un like y comentarios. Un video "sin vueltas ni endulzamiento".

Diagnóstico

  • Verificar que el rotor gira sin impedimento, girándolo a mano.
  • Extraer el sensor de posisición del rotor (efecto Hall) de la parte trasera del lavarropas, y usar un multímetro para medir la resistencia entre los pines 1-4 y 1-5.
  • Si alguna de las mediciones da algo diferente a 10 KOhm, el sensor está roto.

Solución

  • Para la persona promedio, comprar un repuesto e instalarlo uno mismo. Es bastante fácil (ver video).
  • Mi número de serie era 6501KW2001B CNM 140614.
  • En mercadolibre se encuentran con el nombre 6501KW2001B, y son relativamente económicos.
  • Si se dan maña para soldar electrónica, pueden abrir el sensor y remover las resistencias dañadas, e instalar nuevas.
  • Son dos resistencias SMD de 600-700 Ohm en paralelo (ver video abajo).
  • Lo más complicado es sacar la silicona sin dañar el circuito, para poder acceder a los componentes dañados.

sensor abierto

Resistencias dañadas (recuadro amarillo). Lugares para insertar algo, hacer palanca, y remover la silicona (flechas celestes) sin dañar el circuito.

Background

El lavarropas de mi viejo estuvo parado por años. Práctivamente nuevo, pero no pudimos encontrar los papeles de la garantía. Hace poco compramos otro, y decidí intentar arreglarlo antes de tirar el anterior, ya en estado de abandono...

lavarropas sucio

Buscando en internet me encontré con la solución de David Lightman.

Los motores tipo "inverter" en un lavarropas necesitan un sensor. Los motores universales que se usaban antes no lo requieren, pero no tienen las ventajas de un motor BLDC: mayor control via FOC.

Es decir, este motor no anda sin el sensor de efecto Hall, por más impecable que esté mecánicamente. Es un fallo electrónico.

Costos

  • Cambio de sensor: 9.500 pesos (7 USD), precio del repuesto, nuevo y original.
  • Presupuesto del técnico de lavarropas: 280.000 pesos (193 USD), reemplazo total del motor. "No tiene arreglo".

En mercadolibre pueden encontrar reseñas que relatan experiencias similares.

A gripper tool for OLA & Jubilee - Part 2

This is a follow-up post to A gripper tool for OLA & Jubilee - Part 1.

Background and requirements are listed there.

homing

Tracker: https://gitlab.com/groups/open-la/-/epics/20

Homing Implementation

The wripper sits on a freely-rotating wrist, which makes wiring sensors on the tool-side complicated.

To get around this, the homing sequence proceeds in two ordered steps:

  1. Home the wrist axis.
  2. Align the gripping axis to the second endstop.
  3. Home the gripper.

gripper-cad.png

The most reasonable endstop options, given what i had around, were either optical or magnetic. I went with magnets because they are far more forgiving than opto-endstops in terms of alignment.

I used electromechanical Reed switches, soldered to Makerbot-style endstop boards, each one fixed to the "stator" part of the gripper.

Control

This time I just stuck the control board onto the tool, and routed only 24V and USB to it.

I really didn't want to do any more cable management. Reducing the cable count to just 3 is a real achievement.

Behind this is a firmware capable of syncronizing motion between multiple control boards. At the time of writing, I used my klipper for CNC fork.

With it I get all of the klipper goodies, and two important bits:

  • Kinematics for an additional axis set; A for gripping and B for turning in this case.
  • Homing through generalized probing (i.e. G38 commands) and basic macros.

Klipper is slowly working towards its own multi-axis features, so hopefully I won't have to maintain the fork for too long.

If you know about other firmware projects that might replace Klipper, please let me know. There is now Prunt, but it's not what I am looking for.

Proof of concept

Here is a video of it homing clumsily on a pair of stepper motors: https://www.youtube.com/watch?v=3veRTsW12Zw

Up next

Get involved. Introduce yourself and browse to the OLA tools forum.

Get in touch, join the chat! https://discord.gg/GmCeXTHpM4

A roadmap:

  • Replacing the servos with NEMA8 motors.
  • Print a Jubilee-style tool mount.
  • Add endstops, probably optical, probably coupled to each other (or with just one).
  • Review or redesign the mechanics to fox the jankiness in the motion.
  • Figure out kinematics for rotation & gripping.
  • Have a little fin with it.
  • Integrate to the OLA lab automation stack as a new tool: https://gitlab.com/open-la

There will be a "Part 3", no ETAs though. :)

A gripper tool for OLA & Jubilee - Part 1

After a couple years, this project has finally received my attention.

alt text

Tracker: https://gitlab.com/groups/open-la/-/epics/20

Background

I want a gripper tool for this lab robot: https://docs.openlabautomata.xyz/

Notable previous works include:

Some requirements

  1. Gripping action: for tiny (e.g. seeds) and medium-sized stuff (well-plates and petri dishes). Nothing too heavy.
  2. Wrist action: with unbound spinning angle.
  3. Affordable, ubiquitous components.
  4. Mostly 3D-printed parts.
  5. Upgradable.
  6. Open-source hardware.
  7. Synchronizable motion.

Nice to haves:

  • Eccentric grip.

Implementation

  • Hollow-shaft concept: a pair of concentric shafts transmit force to different parts of the actuator.
    • The broader shaft (8 mm) rotates the wrist.
    • The narrow shaft (5 mm) passes through it, and actuates the gripper claws.
    • Because of this arrangement, the wrist is coupled to the grip action. Both motors will need to turn in sync for the claws to rotate and hold a certain grip.

alt text alt text alt text alt text

Proof of concept

Here is a video of it running clumsily on a pair of servos: https://www.youtube.com/watch?v=_W5kAROLW5M

Up next

Get involved. Introduce yourself and browse to the OLA tools forum.

Get in touch, join the chat! https://discord.gg/GmCeXTHpM4

A roadmap:

  • Replacing the servos with NEMA8 motors.
  • Print a Jubilee-style tool mount.
  • Add endstops, probably optical, probably coupled to each other (or with just one).
  • Figure out kinematics for rotation & gripping.
  • Have a little fin with it.
  • Integrate to the OLA lab automation stack as a new tool: https://gitlab.com/open-la

Check out "Part 2" of this project.

Traveling Prusa

So I decided to bring my Prusa MK3S+ with me to Bariloche by plane.

Preparations

Naturally it required partial disassembly to fit in the case. So I took it apart:

  1. Carefully cut the zip ties that secure the cables coming from the PSU, Z-axis motors, and LCD panel.
  2. Disconnect the heated bed cables, and remove the entire Y-axis (rods and all).
  3. Disconnect the LCD cables from the mainboard, and unscrew the entire LCD panel.
  4. Detach the entire X axis (extruder included):
  5. Remove the top-side stops of the Z axis
  6. Carefully moving it upwards by turning both leadscrews until the nuts disengage.
  7. Unmount the Z-axis motors, including their 3D-printed holders.
  8. Unscrew the "long" aluminium extrusions from the middle (i.e. not the screws on the front).
  9. Don't lose any part, nut or screw. Some of the tiny square nuts will fall out without you knowing.
  10. Carefully move the parts to your suitcase, tucking them in with abundant clothing (or bubble wrap and such things).

Here is the end result:

alt text

Putting it Back Together

Some notes:

  • Try not to swap the Y-axis rods with the Z-rods. You'll feel crazy thinking that the Z-axis rods "grew" on the plane.
  • There is not much else to this really. Use the official Prusa assembly guides when in doubt.

alt text