ifeelbeta.de

  • Increase font size
  • Default font size
  • Decrease font size
iFeel beta

Beta prototyping Robots

Some time ago, we used a special process to create our casted 3D printer parts in larger quantities. In this process, we printed master models with our 3D printers, very similar to what you would do for creating printed 3D printer parts.  Then we put those masters into fitting boxes and pour silicon material into it. This way we got a negative form when we carefully remove the master again. Finally we used the created negatives to cast new copys of the master model by pouring PU casting material into it.

During this process, we decided to find out how much time it would really need to get from a digital model to multiple casted copys of this digital model. We were interested in making other parts than 3D printers, so we started a small experiment to get from digital design to small scale series as fast as possible.

To make the project more fun, we started with designing a small robot model. It is split into body, legs and weapon parts with different variations of each so we can combine many different looking robots once they are casted.

The digital design of all those robot parts looks like this:

And digitally assembled, they can look like this:

After we designed those digital models overnight, we started the next morning with printing those parts. As all those parts are pretty small, it took us only a few hours on our mendel 3D printer to make master parts of these digital objects. The whole morning was spent with waiting and feeding the 3D printer so we got all parts ready. The chaingun with an added plate for casting can be seen here:

It was around midday, and we had the master parts for this robot models ready. So after a lunch break, we started with putting the masters into boxes and securing them so they don't float up while casting. Then we mixed the two part silicone mold material and filled those boxes carefully to create our negative molds. We made a foto during this process:

Again we had to wait, as the silicone mold material needs some time to cure and to become solid. Depending on roomtemperature between 2 and 4 hours. So later, near the evening, we could take the silicone out of the boxes, carefully remove the printed masters. After this we only needed to mix some two part PU casting material, add the desired color and cast robot parts one after the other. The PU cures really fast, so you can make new casts around every half an hour.

Unfortunately some parts of our models were too detailed and thus these features didn't look perfect when casted. But the robots still look very nice after assembly. We made them in different colors:

In conclusion, it is possible with this technique to get from a digital model to multiple casted realworld copys of it, in one day, from moring to evening. And most of the time is spent on waiting for materials to cure and for the 3D printer to finish its jobs. But there are also limits to this technique, as some details, like the rockets on the green one, are very hard to cast and it's very likely for them to have casting errors. So the process might be better for more simple parts than ours. But all together it's a great way to get a small scale production of digital objects into the real world without too much effort.

If you want to recreate some of this, you can find the stl and openscad files for this robots on thingiverse as object nr 23788

We also made a video about this whole project. Take a look here:

 

 

 

 

2Printbeta feat. Richrap

For all who don't want to read the whole blogpost: Watch this video (link).



----------------------------------------------

At the beginning of this year we made experiments with paste extrusion.
First we tried it with a big syringe extruder connected to a bowden cable.
But the bowden cable was not the right decision.
Most power gets losts inside this bowden cable and though the paste inside has to be very fluid to get through.

Even bigger transmisson doesn't solve the problem: The syringe also has it's limits and may burst by the pressure.

So time passed and we dropped this project because the new 3D printer wallace arised.

A while ago Richrap posted it's genius solution for paste extrusion: The universal Paste Extruder for 3D printers



This approach seemd to work, so we made some research ourselves.
The mechanism is very simple. A timing belt is used to squeeze a syringe. The power is boosted by a transmission.

The problem: Richraps model is non-parametric. It only fits to 10 mm syringes.

So we decided to improve this idea by creating a parametric version of a syringe extruder.
We've got plenty of 20ml syringes in stock, so we made it suitable for those syringes.

2Printbeta proudly presents: A heavy duty universal syringe extruder



Because 20ml syringes are much harder to extrude than 10ml syringes, we had to change the tramsission of the gears.
We added another gear to get enough torque.



The prototype is very gigantic but works perfect.

Insted of a metal pulley we tried to use a printed one.
Therefore I modified greg frosts script for pulleys suitable to T5 belts by iteration.



Against all assumtions: The pulley is strong enough to told the torque of the gears.

The additional gear forced us to place the idler on the other side, right above the motor.
It doesn't need to resist much pressure. It is only meant to force the timing belt in place.
If it would be screwed on tight, the pulley would move very hard.

A zip tie on opposite side of the idler pushes the timing belt away from the extruder.
This prevents jamming with the incoming timing belt.
The other side of the belt is clamped by the syringe itself.


Like the richrap model it fits to a standard Prusa x-carriage. It has the same mounting holes as the universal wades extruder

The nozzle is something special:
First tries with a 0.5mm cannula failed because always some larger sugar crystals remain in the sugar powder and block the hollow needle.

The used nozzle is a M5 cap nut with a hole through the cap. For rough prints it's enough just to drill a hole through.
In our case we drilled a 2.5mm hole and cut a M3 thread into it.
To get a small hole without 0.5mm drill we used 8mm hexagon socket M3 grubscrews. The hexagon socket is conical shaped and quite deep.



Simply cut the top with an abrasive wheel and remove material until you can see a little hole.
By removing more you can widen the hole.
Being very careful, you can make tiny holes with a diameter <0.5mm

The screw is now only as long as a M3 nut.

To secure the short M3 screw we used a M3 nut cut into halves.
This way you can screw it into the M5 cap nut and still secure it from dropping out.



The pastruder works just fine. It passed heavy duty tests with frost, hazelnut spread and marzipan without any problems. We used the same motor as for the wades extruder and the torque was high enough.

Here you can see a pictures of our very first extrusion test:

The second test with hazelnut spread was successful using a hollow needle as nozzle.



Since you can not print with high speed we tried to print on something else than our printbed - fixed only by mass inertia:

To get better results, we printed on a butter raft:

The third test was again heavy duty:
Driven by the good printing properties of our very viscous frost, we tried to print marzipan.

The cannula blocked, but the 1.5mm and also 0.8mm nozzle didn't.

The results are awesome:



I hope you enjoyed the perspective of printing you own thread candy :-)

Greetings from the lake constance

BonsaiBrain

 


Page 6 of 52