Tuesday, March 2, 2010

Introducing a New RepStrap design: Auger

Introducing Auger
A New and Still Hopeful RepStrap Design

Well, having built a McWire I frankly am not satisfied with the performance. For large parts it print fairly decent even good. The smaller parts have been problematic. Mainly due to the course extrusion control. The smaller Items print better when I use multiply in skeinforge but it lengthens an already too long print time.  

I'm sure that I could probably make due with the McWire but I just can't stand the slow speed. I also think that I can do better starting with a blank slate. That being said I have changed my sights from building a Mendel to a design with the following attributes. 

*The mechanical parts should be available at the local home center (Except bearings, electronics and motors)
* Relatively inexpensive
* No printed parts should be needed anywhere in the build to get up and printing. 
* Keep the part type and parts count to a minimum.
* Speed should be comparable to the MakerBot. (I'm trying to do this without belts and pulleys)* Easy to make (although a small drill press is required)Hopefully this design will be something that would be good enough that it could be your first and last 3d printer if you choose not to make another bot.
So far I think I am on track. I have a sketchup model for the X and Y and am 80% through the build.

Here are the major build steps to get Auger printing from here:
* Mount the X and Y motors
* Join the motors the drive shafts
* Move the Mendel Z sled assembly to the Auger base plate and adjust the height.

Basic Construction
Auger uses a skate and rail system to glide the X and Y axis. #10 hardware is used throughout to hold the bearings in place and mount the skates to the platforms. The X and Y skate and rails are identical. The platforms are 7"x7" made from 3/16 inch luan plywood. The two are identical except for the holes to mount the upper stage to the lower. I think the X and Y sages could be made to work with a rack and pinion system but I am working on creating drive screws to move the stages.
The skate were inspired by the design found here at the GilesBathGate Blog.
They are made of 1"x1"x1/8" Aluminum angle. The skate bearings are inexpensive skate bearings ordered from VXB.  One skate uses opposing 45deg angled bearing to keep alignment while the other uses inline 90 deg to give glide to the other end. #10 hardware is used to hold the bearings in place and allows enough play to tweak in each stage making it level and square. Washers are used to space the bearings from the brakets. The Sketchup file has the ideal distance but I used a number of washers that was closest and that seemed like it was good enough. All the screws bite into tapped holes in the aluminum and medium (blue) thread lock was used because of the high vibe environment.
The rails are made of two parts. The same aluminum U channel that McWire had in its design is used as a base to hold 0.7 inch electrical conduit.  The conduit really inexpensive and is used to provide the rolling surface for the bearings and gives height between the sleds for the drive screws. This conduit is very rigid. We paint it and use it for curtain rods inside our house.  Dry wall or wood screws go through the bottom of the ply wood, through the U channel and grab the conduit to hold the rails in place.

The base is made of 1/2" ply wood. I'm actually using an old cabinet door that was in my scrap wood pile. Anything of that type would work. The main purpose of the base is to hold the rails and give a place to mount the Z sled. The Z sled is not designed yet but eventually will use the same design concepts as the X and Y sleds. In the mean time I will be using the Z sled from my McWire.

Drive System
I have had this wild idea that I could make a drive screw to move the stages. So far it has turned out pretty good but I have hit a few bumps along the way. 

The Jig
To create the drive screw I have come up with a jig. I call it my Augering jig. It basically consists of a platform that fastens the the base of my drill press. It has two rails that form a channel. The channel is adjustable to guide rods 3/4 to 1 inches. Two 1/4inch bolts are used to hold and clamp a bracket over the rods. That bracket holds a razor blade angled by a wedge. The wedge in the picture is 15 degrees. I have made it so the wedge can be swapped out for one of a different angle.  The bracket is tightened down so that the razor blade gets a good bite into the rod. When the rod is rotated the angle of the razor blade translates this into a linear displacement. This displacement drives the rod forward in to a bit that cuts a spiral grove into the rod. The jig works amazingly well. 

The Drive Screw

I first tried the jig out using a 3/4" poplar dowel. After a few attempts I was able to come up with a precise cut. At first I was turning it by hand but that got tiring. Then I drilled a hole in the end of the dowel where I fastened a 1x2 piece of wood. I centered it so I could turn it with two hands like a ships wheel. I was able to turn it easily but the rotation was not smooth because my hands had to trade positions. Then I attached the 1x2 at its end and steadied with my left hand and turning it with my right. I was able to keep constant pressure on it in a smooth rotation. I found out later that pausing had not ill effects on the cut but reversing did as there is no support in the opposite direction  for the razor blade and it flexed easily. 

I used some calipers to measure the variation in thread distance in one rotation and they were less than 0.005" of  one another from thread to thread. (It's really kind of hard to tell. I will have to measure the accuracy under operation.) Then thought it would be better if the rod was made of a lower friction material. I placed an order with McMaster Carr for some 1 and 3/4 inch UHMW HDPE.
 I tried cutting some this weekend but had some problems:

* The 3/4 inch rod I ordered was actually 0.81 inches in diameter or about 13/16".  All I have is a 3/4 paddel bit to make the nut. I cut threads in it anyways. perhaps I will pick up a bit this week.
* The HDPE has a low melting point and as a result likes to be cut slowly. My drill press on its slowest speed and an 1/8" bit still melted it as it cut. I was  able to cut the 3/4"  rod ok but the 1 inch rod melted too much and was not really useable. It seemed it was a slightly different material.
* The 1" inch rod had bulges in the diameter I think caused by the extrusion process.
* All the rods had a bend in them. I was able to correct this by hand.

If the razor blade had bidirectional support I think one could do gradually deeper cuts by going back and forth without too much loss in precision. I think that would be key to cutting plastics without melting.  The bit I was using is a downward cutting spiral bit.  A bit with 2 flutes and less surface area may help the melting issue as well.

Drive Nut
The dive nut is pretty simple in construction. I made it block by glueing two pieces of  plywood together and drilling a hole that matched the drive rod. I marked the diameter of the circle and transferred those marks to the top. I placed a 15 degree wedge under the block and I drilled a hole just inside the markings.  This made a hole that is parallel to the threads of the shaft.  I then put the 1/8" drill bit in smooth side first and cut it off flush to the top using the Dremel tool.  This made the thread of the nut.
On one of the nut and rod pairs there seemed to be some blacklash that is comparable to McWire's rod and coupling nut on the other there was hardly any at all. I will have to  wait for actual operation to see how good it is though. I have some more ideas on how to make these better that I will be exploring.

Sunday, January 17, 2010

Software and First Prints

Lately, work has been crushing me. I have still managed to put some time in almost every day and made some good progress during this last week. Here are the highlights.

Installed Software
  • Arduino Environment - I had this software already installed but I made sure I was using the current version. I had to configure the IDE to be used with a Sanguino Microcontroller. The instructions are here. This all seemed to work fine.
  • Firmware - I had some problems getting the firmware to compile. In fact it never did. While browsing through the MakerBot Google group I found that I did not need to compile these as they come with and can be installed by ReplicatorG. It would have been nice to have known that up front as it would have saved me a bunch of time.
  • ReplicatorG - I down loaded ReplicatorG from here. At first this would not work on my computer. When it ran a Java run time window would pop up and say something like. "Could not find main class, progam terminated." I reinstalled Java and it didn't fix the problem. Then I read on the RepRap forums that a complete uninstall and then reinstall should fix it. So, I tried that and it did. I made a copy of machine.xlm and then added the xml code McWireReplicatorG.xml You need to do "View Page Source to see the XML code in the browser and then copy and paste using a text editor.
  • Python - I down loaded Python from here
  • Skeinforge - I down loaded from here. I decided to change the environment path to include python so Skeinforge could be run from my tool bar. I made a batch file containing "python skeinforge.py" and stuck a shortcut in my tool bar. 
Power up

  • Extruder Test -  I am currently using HDPE for my extrusions. I'm seeing that ABS is popular and PLA is coming on strong but these were not in stock when I ordered my things from MakerBot so I got HDPE. From the MakerBot Store site: "This is HDPE plastic aka High Density Polyethylene aka Milk Jug plastic. This is a nice, smooth, high quality plastic. It comes as a filament in 5lb coils with a diameter of 3mm.
    This plastic is cheaper than ABS, but has a higher shrinkage factor which makes printing large objects more difficult. It does have a much lower coefficient of friction so you can print things that are very smooth.
  • Endstop Test - The endstops work great but are useless. In fact they are worse than useless. When an endstop is triggered, the ReplicatorG and the motherboard freezes up. I think they pickup noise also so I removed all 6 endstops. 

First Prints
My first prints were piles of goo. But, I have learned a bunch about the tool chain during this last week. Basically it's pretty easy. Once you go through it a few times. The instructions are here.

The key is to get the skeinforge settings right. I started with the settings from Binary Construct blog.
"Here are some of the more important skeingforge settings I am using:
PWM Motor Speed: 215 to 225 (depending on the object)
Extruder Temp: 225 C
Speed: 80mm/s (i think the threaded drive has some effect on this no being accurate)
Layer Height: 0.7mm"

These settings allowed me to start printing test cubes that are15mm on a side. These are kind of like the "hello world" of 3D printing.

Tweaking Skeinforge

I have been learning about Skeinforge settings from Bits From Bytes Blog. I think I have tweaked the McWire to the best of it's ability. The prints are very useable but slow and course. The basic problem with it is that the X and Y sleds move to slow because of the gearing with respect to the minimum extruder flow rate. Right now I'm running at 200/255 PWM rate on the extruder motor. Any slower risks a chance of stalling. The 80mm/s in the above settings is being limited by the settings in ReplicatorG. I Adjusted those and found they are best left where they are.  So for right now there is not much, if any, of the desired filament stretch during the extrusion process. This could be fixed by going to a larger thread pitch on the drive screws.

Here are the key skeinforge settings I have going right now.

Tuesday, January 12, 2010

Thunder Cats are GO!

A bunch of progress has been made since my last post. Two major steps have been achieved .

  • Mounting the Gen3 boards
  • Wiring the Electronics

Mounting the Electronics
This started with making a template by exporting the layers out of Eagle to a dxf file and then took that into my CAD program where created the board arrangement. I used some glue stick and pasted it to a piece of plywood and cut it to size.

I made some home made standoffs from drywall mollies. I used some washers to get the desired spacer length and cut them with some good size wire cutters.

Here is a driver board with the standoffs. Also you can see that I added the right angle headers for the opto endstops. I didn't bother taking out the CAT5 connectors.

I had a nice hefty bracket for mounting large curtain rods and  used it to mount the electronics board. I wanted to make sure that the electronics were off the bench so the machine does not take up so much room. All but the power supply is off the bench and I am working on getting that mounted some how.

Wiring the Electronics
Here are some notable points I had for the wiring:

  • I unsoldered the molex connector pulling the pins out one at a time and soldered in screw terminals for hooking up the stepper motors.  Digikey: A98168-ND TERM BLOCK 4POS SIDE ENT 3.81MM
  • I found some sprinkler wire at a home center to hook up the stepper motors. It worked out great. It is multicolor 18AWG solid core wires with 7 wires in the bundle and has black PVC sleeving.
  • I wired up the opto couplers and with Red (+5V), Black (ground) and yellow(signal). This wire was connected to 0.1" female connectors. (Digikey: S7001-ND CONN HEADER FEMALE 3POS .1" TIN) on both ends. 
  • I used right angle 0.1" spaced header pins on the driver boards that are right underneath the cat 5 connectors. I didn’t bother to remove the cat5 connectors since they were not in the way.
At this point the McWire is completely wired and ready for action. The next steps are firmware installs, softwares and then printing!

Monday, January 4, 2010

McWire After Build Issues

Some short comings of the McWire design have shown themselves after the build.

The first one concerns the method of attaching the drive screw to the Z sled. I seems the constant pressure on the tape covered coupling nut made the tape move over time. It telescoped out until the holding pressure in the conduit ran out. Then slipped out of the strap.

My solution was to build a bracket that presses the coupling nut against the Z sled, holding it in place. Overall this is a better way to go as you can adjust the height from the Z sled with a shim to get the drive screw parallel to the sled. This was a pain to do with the tape wound design.The X and Y have not been changed out but I will use the same solution if they become a problem.

The bracket material came from an L bracket that is used to hold drapery rods. A piece of 1/8"thick aluminum would do as well. It's held in place by 2 - #10 screws and tapped holes in the sled.

Another failure that has been occurring is breaking swing arms. They are not standing up to the task. I think it is the constant pressure of the spring and the size of the 5/16 hole plus counter sink relative to width of the swing arm. I have reduced the spring pressure as low as I fell comfortable with but failures are still occurring. I had 4 spares made when I ordered the laser cut parts and those are used up and now I am making parts to replace them.

I think the newly designed swing arms are working great and are easy to build. I used 3/16" press board because that’s what I had on hand. 1/4" thickness might be better.  1/8" aluminum would work nicely.

To make one of these :

  1. Use a template to mark the outline and hole locations. 
  2. Cut it out roughly with the outline markings and the drill 1/16" pilot holes for all 3 holes. (larger for aluminum)
  3. Drill the clearance hole (23/64") for the 5/16" bolt holding it to the sled.
  4. Mount the bearing with a small washer (M5) under it using a dry wall screw. 
  5. Screw in a #2 sheet metal screw to hold the spring.
  6. Dremel off the length the the screws protruding from the other side and grind them down just below the surface so the don't scratch the plastic on the sled.
  7. File the shape of the swing arm down to the outline marks. 

Sunday, January 3, 2010

McWire and MK4 Plastruder

Well it took some tweaking and figuring but I have the MakerBot Mk4 Plastruder successfully mounted to the Z sled. It was not as hard as I thought it would be and it came out better than I envisioned.

I built up the complete assembly according to the instructions at wiki.MakerBot. Then I started to brain storm ways of attaching it. In the end, there was only one practical way of mounting it. Here is the configuration I came up with:

Both the dinos had to be removed. They made the whole assembly as large as the Z sled. I set them free and they are able to rome around the shop and feed on future discarded prints.

I laid the plastruder assembly down on some graph paper to get the relative distance from the extruder nozzle to the mounting holes.

I wanted to create a template for an adapter plate that would hold the extruder board and the remaining of the plastruder assembly. So, I down loaded the Eagle files for the extruder board and exported the Dimension and the Holes layers as a dxf file. This gave me the hole pattern for both the extruder board and the plastruder since they bolt to one another using these.  I then arranged these in my CAD software such that the nozzle was centered on the Z sled. From there it was easy to see that the inside mounting holes for the PTFE bearings would make good mounting holes for the adapter plate. The adapter plate was made out of 3/16" fiber board.  It's what I had on hand.

After mounting the extruder board and the plastruder  to the adapter board and on to the Z sled I saw that it needed to be lowered about a good 2".  A quick run to the hardware store and I got some different lengths of pipe to adjust the Z sled height. I ended up swapping out the 12" piece with two 5" pieces and a coupler.

The Y alignment is not quite centered yet. Good enough for now though.

Over all I think it came out pretty good and will be easy to swap out print heads. I'm thinking of putting longer screws in from the other side and securing the board with some screw on knobs or wing nuts.

Friday, January 1, 2010

Plastruder Build

I was a little nervous when it came to building up the plastruder. It seems like a place that many things could go wrong. I got the MK4 plastruder from MakerBot. The build instructions on their wiki are pretty straight forward. Also,  I am thankful for the makerbot google group for the support they give. I have been trolling and learning some of the pitfalls people have had in building the current revision. I see the wiki has been updated to address those. Awesome!

Now that I have built one and all measurements look good I feel like i can tackle another without too much problem.  The hardest part for me was the fine wire of the thermistor. It was like to hairs  and a spec of glass. I felt that if I sneezed, it would be gone.

I thought I would take pictures of the way I wound the wire on so encase I have any problems I could show the experts and see how it could be improved.

The next challenge is to mount the plastruder to the Z sled.