Reprapsource Blog

New ABS Plastic in stock!

2010-08-16T01:19:00.005+02:00

We have new ABS plastic in stock in the colors Natural/White and Black in stock. The plastic is made in Germany and comes on spools that are easier to unroll than the coils of Natural/White ABS we have.

Testing Wade's extruder drive [Update]

2010-07-10T17:06:00.010+02:00

I'm currently testing Wade's extruder with Nophead's Gear and Idler bracket on my Shapercube. Thanks to Prusajr's design, I was able to easily mount it. The mount with cable binders is not bullet proof but it gets the job done.

The dark stuff is "WEICON Flex+bond" glue, which I had only in grey colour. A combination of production error and too much force caused delamination. The glue holds it together nicely.

The main thing I dislike on Wade's extruder until now is the serviceability. When something is blocking the extruder, the drive teeth eat the filament quite fast. For example, a failed part after this happens looks like this:

And the drive looks like this:
Wade's design could be improved if this drive would be accessible from the top or side, without having to disassemble half of the parts to get there. I'll try implementing this in the next days.

I haven't found the perfect way to clean this up, yet. I was thinking about a screw on the other side that rotates with the same speed of the gear, but that would add a lot of complexity to it, because a second gearing would be required.
Another idea I had is adding a small wire brush, but that would not be effective on the plastic that got inside the teeth.

[Update]
I'm having success with drive gears made out of a M6 tap. The wider teeth won't eat the filament on normal operation. Even if, a short spray out of ice spray does the job.

PCB Heatbed Prototype

2010-06-15T01:12:00.002+02:00

Thanks to Prusajr I've got a prototype PCB Heatbed running on my ShaperCube.

This heatbed is an etched, single sided 220x250mm big PCB, ~1,5mm thick in total with 35ym coppper thickness.

Mounting it

The PCB itself is not flat, but can be bent slightly. Similar to Prusajr's design, I utilized magnets to keep the sides down.

The Magnets I use are 9mm in diameter and 5mm thick NdFeB ones. They have a catch: The ones I use can only withstand temperatures up to 80°C. Because of that, I didn't glue them onto the PCB, but I glued some strips of steel tape onto it with temperature resistant glue.

I made a pile of 2 magnets each on four sides of an MDF board hoping to keep the temperature down a bit. The magnets are glued to the MDF using some expoy glue. So far, this works nicely.

After a couple hours of operation, the glue on the middle piece failed. The ones on the side don't get that hot and are still working.

As it turns out, magnets are not the most elegant way to mount it but probably the easiest non-permanent "do-it-yourself" way.

Electrical issues

The heatbed needs a lot of power. It draws 11A @ 12V when cold. Using this kind of power from a ATX PSU requires to use two plugs and probably a very good PSU. I had problems that the voltage on the +12V rail dropped too much on my 400W PSU (which is appearantly rated for 16A on the +12V rail). Currently I'm using an external power supply for the heatbed.

Heating up & Polyimide Tape

To stick ABS onto the heatbed, I use Polyimide tape. Those can be ordered in the same size as the PCB is. Large strips of Polyimide tape are a bit tricky to get on to the PCB without air bubbles. The factory actually suggested a machine to roll it onto, but from $350 a bit pricey.

Anyone else use large tape and have different experience?

Printing quality

Apart from some air bubbles I had, the parts come out perfectly flat. A bit of warping still appears on the sides of the heatbed. Nophead wrote about this and has a nice solution for his Mendel.

The bigger problem I still see is in adjusting the heatbed. If the nozzle height is not set correctly, it can easily snap traces on the PCB. Once snapped, they can be repaired with solder and some sanding, but the results are far from perfect.

I was thinking of putting a thin sheet of aluminium on top of the heatbed to protect it. But this would need a better way of mounting it.

Shapercube batch 2 preparations and Batch 3+ stepper driver questions

2010-04-08T19:55:00.007+02:00

Batch 1 is all sold out except our now almost completed photo model.

The prototype of the slightly revisited version 1.1 is on the way and will be produced in the next days.
Batch 2 will be shipped in week 19 and can be preordered already!

Future releases

After batch 2 is gone, we will run into the problem that the stepper motor driver 1.2 will be sold out. We don't want to support it any longer because it has too many components and takes too long to solder. On the other hand, we don't want to ship out unassembled SMT components to our customers.

A few thoughts about the options

The stepper motor driver 3.1 file looks good, except they left the opto endstop connectors out. This is an easy fix because all files are open source. However, the chip necessary on this board seems to be out of stock everywhere since months.
I can let produce fully assembled stepper motor driver 2.3 board now, but I don't have 3k+ cash spare (maybe anyone interested in buying a bigger part of the batch?)
I could let produce the stepper motor driver 2.3 boards and solder the SMT components in house. But I need someone who can do it my clumsy hands won't do well. [If anyone feels like he can solder 100+ boards at once with lead free solder, contact me.]
The polulu stepper driver could be an option. Problem might be that this chip needs heatsinking. Well I could try using the Shapercube frame for it. Question is: How long will their stock last? The A4983 chip is also nowhere in stock.

What do you think? Did I miss an option?

Water jet parts

2010-03-04T21:12:00.011+01:00

I got the first sets of 5mm water jet parts out of aluminium sheets. Previously we got the parts laser cut, but the cut was not straight and had far too many burrs.
For the first production batch, we get the parts cut by a water jet machine. On the water-jet machine, two sheets are cut at once. The tool head cuts the aluminium with a mix of water and an abrasive which is pushed out of a 0.8mm nozzle with 3000 bar.

The cuts feel a bit rough, but is much more straight compared to the laser cut parts I've got earlier.
However, the surface quality varies. Some of the abrasives will stick to the sheet.

For polishing, I got two polishing tools for my drill. I payed about 25 Euros for both.
The upper one cuts away a bit, the lower one is more for the finishing up. On small parts, it looks nice and almost consistent.
On the larger panels it's quite cumbersome because the drill is heavy. Unless I've something like a modified router to do this automatically, I can't add an option to polish the parts.

A possibility of getting rid of the bad surface is glass bead blasting. I brought a (useless) leftover part to a small sand blast shop nearby. The results are nice:

The surface is very consistent and a bit rough. Looks awesome, but has its price: A whole kit will cost 99 Euro more. Email me if you want to add this option to your Shapercube!

Shapercube coming March, preorders open

2010-02-16T01:09:00.003+01:00

10 pieces of the aluminum RepStrap Shapercube and all of its parts are ordered by mid of March. The complete kit can be pre-ordered here.

The panels are water-jet cut and high quality and need very few post-processing with a file (less than 10 minutes), much better than our previous laser cut panels.

The documentation is stored in the Shapercube Wiki, which is currently filled with images and documentation, day by day.

Mendel Bearing Kit

2010-01-25T23:33:00.003+01:00

While not working on a construction manual for our upcoming Shapercube release, we had some time making a new kit: The Mendel Bearing Kit. And the best part: It's only 29,79 Euro for 50 bearings in total!

Mendel bearing Kit now available on reprapsource.com

That kit contains 48 double-shielded 624 bearings and 2 double-shielded 608 bearings. All you need for one Mendel!

Rapid prototyping with OpenSCAD

2010-01-03T17:49:00.007+01:00

The latest extruder I made had a small bug. When too much force applied, either from the pinch wheel or from below, the filament will bend into a bit of space and eventually break.

Instead of waiting 2 weeks for a fixed part to arrive, I printed my own, which needed less than an hour!

I used OpenSCAD to convert my 2D DXF files. The feature I just needed is linear extrusion. This is a one-liner in OpenSCAD:

dxf_linear_extrude(file = "test.dxf", height = 3, center = true, convexity = 10);

After rendering, one can export the results to STL, which loads into Skeinforge perfectly. 39 minutes later:
Besides it came out 0,3mm too thick, the part fits perfectly and works perfectly. Having a 3D printer really speeds up the development.

Skeinforge and Z-Movement

2009-12-30T17:01:00.007+01:00

I had to try a few skeinforge version before I got working one.

On newer versions, I encountered a bug that the feedrate in the G-Code will decrease when you increase the feedrate and vise versa. The latest version which does not have this problem, is from November 2009.

The G-Code generated from Skeinforge has another problem. The Z movement is not listed seperately, so it will try to move X, Y and Z together with a feedrate for X&Y.

As workaround, I made a small, quick&dirty ruby script to fix the g-code.
Usage:
./z_fix.rb file.gcode

It will change

G1 X38.59 Y-42.09 Z0.9 F55.0
G1 X-38.86 Y-42.31 Z1.61 F1320.0

to

G1 X38.59 Y-42.09 Z0.9 F55.0
G1 Z1.61 F75
G1 X-38.86 Y-42.31 Z1.61 F1320.0

Merry Christmas!

2009-12-24T13:16:00.004+01:00

This is the first usable object our RepStrap printed out:

Available at Thingiverse

While not so great in detail yet (which is mostly software related), it looks great at the distance of 1 meter or higher.

The team of RepRapSource.com wishes you a Merry Christmas and a Happy New Year!

Look what I've got here!

2009-12-02T15:56:00.007+01:00

That's about 300m of isolated NiChrome wire! A short length for making 3-4 extruders is now available on Reprapsource.com

Curing Fire Cement

2009-11-20T14:29:00.012+01:00

I found out that the fire cement I have (branded "KOS Fire Cement") does not cure on room temperature.

So I wrote a small program for my arduino to heat up the wire slowly.

const int pin = 9;
int heat = 14; // start with a little heat. the program should take about 4 hours

void setup() {
pinMode(pin, OUTPUT);
Serial.begin(19200);
}

void loop() {

if (heat < 255){
Serial.print("heating at ");
Serial.print(heat);
Serial.println(" for 1 minute");
analogWrite(pin, heat);
delay(60000);
heat += 1;
if(heat == 255){
Serial.println("maximum heat reached, stay here for 3 minutes");
analogWrite(pin, heat);
delay(180000); // stay at maxium heat 3 minutes
}
}else{
analogWrite(pin, 0);
Serial.println("heat is switched off. Ensure the heater has cooled down before touching!");
}
}

It worked very well. I think one can set the start value higher to start maybe at 70-80°C. I'm not sure if it makes sense to go up to maximum output, because most of the water inside went out on the entire block on 160°C on the inside.

Turning it all the way on, I could find out the optimal voltages for the heater. For example on my heater, analogue value 194 gets about 220°C, 215 about 240°C (my NiChrome wire measures 6,5 Ohm).

The highest temperature I measured inside the heater barrel was 266°C.

After curing, the unit seems to be very strong. I can't get the brass washer on the top off. Also, having just small holes for the wires exiting makes it almost impossible for the fire cement to fall off. And in case this turns out to be a problem, I can seal it with Cerastil.

Although I was sceptical because the inner diameter of my brass tube is 4,0mm, it extrudes nicely on the first test.

Fire Cement Heater

2009-11-19T22:00:00.019+01:00

Good news everyone! We're getting isolated NiChrome wire!

I had a small piece of isolated wire left so I experimented to make a heater with fire cement today.

I begun with a 45mm long brass tube with M6 thread:
Mounted it with two half-size M6 nuts into a 40mm wide spacer (with two 5mm holes drilled):

I used a 15mm inner diameter copper bushing. The copper is so soft that you can peel off with pliers. I did that to have room for the wires exiting.

The NiChrome wire has an isolation. This isolation reacts when getting too hot - in which it turns black. It still isolates still but gets brittle. So we want to have all the Nichrome wire insulated by the fire cement. When it can't move, it's isolation cannot fall off.
For that problem I used a wire and two sleeves:
NiChrome wire cannot be soldered on. Either you can squeeze those to the NiChrome wire and solder it to your normal wire with high temperature solder or you just squeeze both ends in it, like I did:
Now, isolate that joint with some kapton tape:
And isolate the rest of the wire with a high temperature silicone tube:
Now wiring the heater
The bow nut and the spacer is in place for knowing where to stop winding
Now take the bow nut and the spacer off and try to center the copper bushing:

Now, fill the bushing with fire cement. This is not so easy because the viscosity of the cement is pretty high. After wiping off the excessive stuff it looks like this:
Now be sure to clean that thread and put the washer and the bow nut back in place
Now this has to settle. I leave it as it is over the night and start heating it up a bit tomorrow.

Next extruder prototype delayed too weeks :(

2009-11-19T14:03:00.002+01:00

The guy at the company I get my laser cut sheets from is currently on vacation. Apparently nobody else can do that for me there. So I've to wait another 2 weeks to get my parts...

It really would be nice to have a laser-cutter for myself to get prototypes done faster. But the small machines that can at least cut through acrylic are starting at 12000 Euro...

Not enough power with Nema 17 motor

2009-11-15T15:45:00.003+01:00

I've been trying to get my extruder working the last few weeks.

While I've got heater & nozzle working without much problems now, the drive mechanism is still buggy.
With the pinion gear I have been using, the extruder motor can push ABS about approx. 3,75 kg without slipping. That's enough for driving the filament out of a 0,5mm nozzle as long as nothing is blocking it.
Following Murphy's law, something will block the nozzle. That might be some plastic that's heated for too long in the nozzle, that might be an extrusion too near to the last layer.

I tried installing a timing pulley to the Nema 17 motor I use. Since the big teeth of the pulley need a lot more force to bite into the filament, the motor does not have enough power to push it.

An option using the Nema 17 size motor would be using a gearing mechanism. But I fear that this will take months of work getting it right.

To get my prototype ready and perfectly running until end of the month, I've ordered some DC motors (type 1156006) from Kysan. They have 11 times the holding torque I've currently with my Nema 17 motor. That should be bullet-proof.

Heater experiments....

2009-11-06T19:53:00.010+01:00

While experimenting with my heater, I had the idea to isolate my Ni-Chrome wire with Kapton Tape. As it turns out, the wire gets a little bit too hot for the tape...

A more successful expermiment is my Ni-Chrome+Cerastil heater.

On the picture above I tried out winding strands of glass fibre around the heater block for thermal insulation. This worked not so well, a small amount of glass wool gave much better results.

I'm currently making another heater, because my first one has two flaws:
- The thermistor was placed too close to the Ni-Chrome wire. The reading were much higher than the temperature at the nozzle. So to get some more accurate readings, I had to check the temperatures with a thermometer.
- When extruding plastic, the temperature at the nozzle drops quickly by 10-15°C. It could be that I need to reduce the resistance of the wire a bit to produce more heat. But I think that the first layer of Cerastil to the brass barrel may be too thick, since Ceramic is a very good heat insulator.
I'm wondering if switching to Fire Cement on the first layer would be slightly better. But I think it'll be hard to apply as a thin layer.

It would be much easier to build an extruder if I find a factory that produces high-temperature capable insulated Ni-Chrome wire and ships it to Germany. Unfortunately, Pelican Wire in the USA does not and all I can find here is PTFE coated at best.

Heater working, pinch wheel shaft struggling

2009-11-05T21:18:00.005+01:00

I had some nice progress since my last blog entry:

The heating element is up and running now for a few hours. We used un-isolated Nichrome wire embedded into Cerastil (more on that later).
You can see on the picture above, that I currently use some of that ugly glass-whool to isolate the heater barrel. I hope I find something heat insulating as sheet or tape soon.

Currently, I've the splined shaft installed which was described on the RepRap builders wiki. It runs fairly well until it gets some back-pressure from below. As result it'll shred the filament and won't move any more forward. Not that I was not warned that this is not optimal...

Heater Design, Take II ?

2009-10-26T17:26:00.006+01:00

I've got two machined parts today. A PEEK-GF30 thermal barrier on the left, and a brass barrel which is inserted into the PEEK insulator.
Looks all great, the brass barrel is lining up well inside the insulator. There's only one problem: My contractor mixed up the length of the parts. The PEEK part is 40mm long instead of 20mm and the brass barrel is 20mm long instead of 40mm.

In addition to that I've to add a large structural support washer and mount it with a half nut as well as adding a nozzle. I would get something like 5-7mm space for the Nichrome wire. That's way too short ... :/

Stainless steel thermal insulator: Impractical to produce?

2009-10-07T20:18:00.005+02:00

From my last blog post:
- The steel rod with a hole cannot be produced precisely like this. The maximum length possible to drill out the rod is 5x the inner diameter.

As it seems, this is not the only problem. As the end of the stainless steel rod gets warm, the filament will expand and produce large amounts of resistance.
A solution would be a tapered end - but I've been unable to find someone who can do that with stainless steel and such small diameters.

Heater Design, Take I

2009-09-28T14:39:00.004+02:00

The heater design is the last big question mark to be solved. Our first test setup design looks like this:

The grey part is a piece of aluminium (about 19mm diameter, 20-22mm height), the left part has been threaded to M6,
The green part is a power resistor,
The dark grey part is a steel threaded rod (M6, 30mm long, ID 3,3mm)
The orange part is a copper welding tip (M6, 0,6mm nozzle) which is drilled out to 3,3mm

What works ok
- With a bit of glass wool as thermal insulation the nozzle heats up to 240°C in about 5 minutes.

What does not work
- The steel rod with a hole cannot be produced precisely like this. The maximum length possible to drill out the rod is 5x the inner diameter.
- The steel threaded rod heats up too much on the upper part. This results in the filament to expand inside the steel rod and produced a lot of friction. This makes it almost impossible to push the filament further.

Next up I will taper the steel rod on the end which goes into the aluminium housing and try this out.

New parts arrived

2009-09-23T17:36:00.009+02:00

I've received some remade laser cut parts this week and installed them:The Y belt clamp (visible on the right) has been improved and is now very easy to mount with two hands.

The complete X-Carriage (visible on the left) has been replaced by the new version. It has an improved belt clamping and tightening system. The extruder can be mounted in the middle of it. The filament drive part will be mounted on it in the middle, the heater part below it.

The X Opto Endstops are directly mounted to the X drive & idler housings. Unlike all the other screws, one has to tighten the two M3 screws carefully, not to bend or crack the PCB.

Next thing to do is making the heater element right. The idea behind it is to make it similar to the one Hydraptor built.
Unfortunately on my last attempt, I broke my M6 die with no chance of getting it out again.

(Yet another) Pinch Wheel Extruder

2009-09-15T17:16:00.008+02:00

This is the prototype of our pinch wheel extruder. It's fairly big, because we want it to be placed in the center of the X carriage and don't want to interfere with the steel rod & belt drive. The stepper motor is a Nema 17 size one with a knurled screw attached to it.

The knurled screw is a M5 DIN 466 one. Its thread was drilled out to 5mm, on the side a small hole was drilled and threaded.
The pinch wheel is not visible on the pictures (transparent aluminium would help a lot :)). The shaft is a M5 screw which can move freely and is pressed towards the drive gear by two springs.
With enough pressure from the spring loaded pinch wheel, the gear is able to hold more than the motor can. The springs we currently have are too weak, but we'll get some stronger ones in the next days.

Aluminium rods

2009-09-14T23:03:00.000+02:00

Today I tried some precision aluminum rods & precision bearings for the Y axis. As it turns out, 8mm aluminum rods are not suitable for this operation because they bend too much.

Back to steel rods then.

X and Y running, some tweaks will follow

2009-09-13T19:40:00.000+02:00

In the last days we tested the X&Y axis. I found out that I messed up the drawings of the X carriage, so I made it temporary fit until I get my fixed parts.

The belt mounting on the X stage is not very easy with only two hands. I've tweaked the X-Carriage backplate to do a better job in this. Instead of the slot, the fixed version will have two holes for the screws and a hole where you can pull the belt from the outside of the carriage. Also two small extra belt tighteners will be added to the sides to be able to do fine adjustments easily.

The Y belt mounting picture above is not yet perfect. My new design enlarges the nub on the X drive housing big enough to act as one of those belt tightener parts. With having a fixed point, it will be much easier to tighten.

The X drive&idler housings are working fine. The gears currently used are 40 teeth T2,5 plastic gears. I think smaller ones would be better, but that makes them also much more expensive because I've to ask for special made gears. Unfortunately, making them with the laser cutter out of aluminum is not an option that works well.
Will see how well that works when the extruder is running.

Z-Axis working

2009-09-08T22:30:00.005+02:00

After nothing really working out yesterday, I and my partner Rainer finally got the Z-Axis working. With having to synchronize four threaded rods, this is the most time-consuming axis to adjust.

Beginning at the gears, one have to measure the positions starting from the diagonal bracket.The diagonal brackets make a good measuring point.

The plate above the diagonal bracket is used to adjust the shaft position. It's screw holes are a bit larger than the screws so it can adjust a small amount of possible inaccuracy between the diagonal corners on the top and the bottom plate.

In lack of success getting the ends of belts glued together reliable on a ending belt which is standard in the original RepRap "Darwin" design, we go with six large gears and three endless belts. The threaded rods on the back of the machine have two gears each and the ones in the front have one gear each.

During assembly we found out that our Z-bed is not perfectly straight but has a slight bend towards one side by a fraction.To adjust this, we securely mounted a M5x30 screw on every Z bed mount and adjusted the bed position with two nuts to hold the Z-bed in place. This has to be re-adjusted when the extruder is mounted on, to be sure the objects are printed on an even surface.

Next, the opto endstops are directly mounted to the side plate:

A big plastic washer is placed under each screw to prevent shorting with the plain aluminium side plate.
Each opto endstop can be adjusted in the height and the Z-Flag can be adjusted in two directions. The Z-flag can be tightened securely by two M5 screws.

The current distance between MIN and MAX Z-Endstops is approximately 130mm.The Z-distance could be doubled if we would store the Z-drive, the power supply and the electronics in a seperate compartment below the machine. This compartment addon definitely looks like a nice project for early 2010.

While waiting for some stuff to get Y & X axis working, I will install the electronics and hope I get the firmware running tomorrow.

Documentating & Z mounting

2009-09-05T20:42:00.000+02:00

Writing the documentation (which will be public once it reaches beta status)
is currently the most time intensive work on this project. Well unless I'd offer it as a "3D puzzle". But what I'm building here is definitely not a toy.

While not documenting, I've got some more sub-assemblies finished. For example the Z bed mounting bracket:

With the polyamide bushings inserted, the mounting bracket floats freely on the steel rod.

Also I finished the case for the X motor mount:

This one is pretty heavy duty. It weights about 390g and can take up to 4 steel rods for the X-axis. What worries me is that the Y-axis rods may bend too much. That is to be found out next week.

Second prototype arrived

2009-09-04T15:37:00.001+02:00

Our new prototype laser-cut RepStrap device has arrived. So far, the main body has been mostly assembled:

The sheets used for laser-cutting are 5mm (a few 3mm) thick aluminium. A sheet of the smaller parts looks like this:

When prototype phase has finished, the files will be released under GPL.