Sunday, 14 June 2009

Its given birth(Nearly)

After a long absence with no posting I have been busy designing and printing a reprap. I nearly went for a Darwin but having tried to build the original it was fiddly to set up and had a lot of complicated parts which I struggles to separate from the source forge AOI file.

At the same time I discovered that school had a site licence for Solidworks so I thought give it a go build a simple threaded rod 3 axis machine the same as I have at the moment.

I also decided to minimise the different parts. Currently it has

Printed parts
Nema 23 motor holder
Bearing slides
threaded rod bearing block
threaded nut holder

And other parts

MDF20 mm
8mm threaded rod (stainless)
8mm bar
skate bearings
rubber pipe(motor to rod coupling)
Plus various nuts and bolts
In the picture you can see the Z and Y axis. The Y axis is a little big and is not MDF but an old shelf that was lying around.
I have stolen eD's idea about using the skate bearings as bearings for the slides. That works well- (Thanks eD if you read this)
Both the axes move smoothly and I have run the Z axis for 2 hours backwards and forwards to see what fell off (and nothing did)
I tried to make a printed motor to rod coupling but nothing I made had as good a characteristics as the rubber pipe.
I have been improving the quality of the printing in the manufacture of all the components. The Biggest is the Nema 23 motor mount at 8cm X10cm X 4cm. I found that if you make it too thick (1cm) then it tends to delaminate so I redesigned to to 5mm thick and added webs to it and it is now really sturdy. Warping was not a major issue- I am now extruding onto a glass chopping board(Robert Dyas 5 pounds) coated in evostick.

I also found that the accuracy was pretty good. I designed the bearing holes at 22mm diameter and then all I had to do was tap the bearings in with a rubber mallet with no preparation at all. The same for ineserting the long nut (studding connector) that runs on the threaded rod - just tap in .

My major issue was with some ABS I bought which varied in diameter causing many jams. I gave up in the end and bought some from Bits from Bytes and had no trouble at all, but it is very expensive. I would hope to pay 20 pounds for a reel soon.

Anyway next steps make the x axis, the extruder and double my production capability.

Tuesday, 7 April 2009

At the end of 4 days

Well 4 days solidly working on stuff. My family think I am mad and have not seen me apart from meal times. Still I have made some progress. The picture shows the screwable jewelry box bottom made with ABS. Unfortunately I did not lift the nozzle away at the end hence the blob at the back. Still I am pleased with it.

What have I found out?
1. It takes a lot of force to push the filament through the nozzle and my steppers only just have enough torque to drive ABS through.
2. PLA cannot be driven through my design of extruder as it buckles. There may be more to play with here as I think I may have had the temperature too high causing too much friction in the delivery tube.
3. I cannot get enough torque or friction to drive HDPE through the nozzle.
4. When designing something take time to check it as printing takes a long time and is frustrating when you get it wrong.
5. 3D designing is hard it is better to sketch it out in pencil before starting using a programme.
6. Solidworks is an excellent programme easy to use and the dimensioning facility is really easy to use- I am glad that I can use my school's educational site licence as it is much too expensive to buy.
7.Putting plugs on everything saves hours. I have just discovered railway modelling plugs cheap and easy to use.
8 Thermistors are very delicate and difficult to attach to the nozzle. They work well until you remove a nozzle and then they break easily.
9 Warping is an issue
10 The machine is very reliable, no breakdowns, repeatable movements. The threaded rods and long bolts have now probably in excess of 40 hours run time and no failures
11 Drilling a hole in a threaded rod so that it is concentric is very very difficult.

What am I going to do
1. Go to Wales for 3 days and speak to the family!
2. Try putting a thermistor in a drilled out bolt so that it can be removed/replaced easily without damage. Screw it into place on the brass bar.
3 Buy higher torque stepper motors
4. Modify the pinch wheel extruder block design for the 4th time and get it right this time
5 Blag an old oven to see if extruding at a higher ambient temperature stops warping
6. Use the pinch wheel extruder to try and extrude solder, I can melt it and I can move it through the extruder. Can I combine the two?
7. Do some experiments on how strong the bond is in the first layer and the board/plate that it is made from.
8. Try and increase the speed of the machine above 800 mm/s- M8 rods instead of m6, change the stepper driver settings.;
9. Try and find the optimum settings for starting and stopping.
10. Go back to the original screw drive design and see how well that woprks with stepppers, the new nozzle andPLA/HDPE

Saturday, 4 April 2009

Pinch wheel extruder 2

Here is my attempt at a bigger pinch wheel extruder based on Adrian's design. It took about 6 hours to print and had no breakdowns.

The only problems with it (appart from my design flaws- holes in the wrong place was the warping.
I am going to apply heat today using a hairdryer to see if this fixes teh problem and I have also had the evostick curing last night as older evostick seems to grip better.

Better stepper mounting bracket

Here is my next attempt at the stepper mounting bracket. This is much stronger, there is very little warping however as can be seen from the second picture the small holes are not as well defined.

I will experiment with the skeinforge carve preferences to improve the hole quality.

The better warping is due to thinner layers and applying a heat gun to the first two layers to ensure that they stuck. (Ithink!)

First useable part.

My Y axis vibrates like crazy so I designed this mounting block to test out the quality of all my new settings. The whole part is much more sturdy and flat. The only problem was my base plate moved half way through giving an off set. Still I will bolt this on today and reduce the noise/vibration.


Fed up with poor quality I decided to fix all the annoying bits that kept dropping off and then work out the ideal settings (Also to move the whole lot back into the garage to keep the family happy.
The main fixes were to secure the extruder to the machine much more securely. To get the maximum travel on the z axis. To get the feed so that it was consistent using the small screw eyes from B and Q and to tidy up all the wiring. This worked well.
I also changed the acceleration settings on the stepper so that they were on the maximum (increased from 30 mm/s/s to about 600mm/s/s) This has made a massive difference to the quality although I would like more)
I then checked the temperature that the plastic was being extruded at as it comes out of the nozzle and adjusted this to 235 degrees as this seems to be the preferred temperature for ABS.
I then made a series of Nophead's hearts using from left to right 0.7, 0.6 0.55 0.5 and 0.45 settings on the carve layer thickness. The best by far is 0.5 so my settings now are 800 mm/s, 235 degrees0.6 welding tip nozzle .

Adrian's pinch wheel extruder

Here is my attempt at Adrian's pinch wheel extruder. It has worked reasonably well but I

am still not happy with the overall quality but there is very little warping for some reason. Unfortunately I do not have a stepper to fit it to try out! If anyone wants it I will stick it in the post to you!(And it is not to a great standard!)

Weak stepper mount

Here is an attempt of a stepper mount to replace the B and Q brackets I am currently using.
The mounting face plate worked really well and has only a little warping (I am still using Evostick on a glass chopping board as the base)

The bit out to the left in the picture built as a tower however is very weak and would never be strong enough to mount a stepper on. There must have been a blockage in the extruder 2/3 of the way up the tower as there is a fault there.

Big Extruder

Here is my attempt at the biggest object that I have made. Not a great success!
It is about 20 cm long and 8cm tall designed to have a stepper one end and a tube through the middle.
It warped and the darker discolouring at the bottom is where I used a hot air gun to flatten it!
It is made with ABS
I had my whole extruder assembly come loose in the last part hebce the unfinished flattened top.
Also It lacks strength as the layers are not close enough together.

Monday, 30 March 2009

I love ABS

I took delivery of some ABS just before the weekend and I have been running the repstrap for about 6 hrs a day since- what super stuff!

Having worked with HDPE for a long time the major problem was warping of any print. ABS nearly cures this although it smells a lot worse. I am now extruding onto a glass chopping board (heat resistant to 240 degrees ) and evostick.

The first major break through I got was reducing the extrusion rate. This had been puzzling me for a long time as every extrusion seemed to have an excess of plastic. I kept recalculating the and measuring the rate it was laid down at and the size but nothing seemed to work. Then after watching for a long time I noticed that when making small movements the machine never got to full speed- should be 720mm/min but on the on screen display I only got 2-300 mm/min- hence too much plastic! I now need to maximise the acceleration on all my steppers using the EMC set up programme.

I have also programmed the arduino so that when it switches off it goes backwards fast briefly and when it switches on it goes forward 1.5 times the reverse quickly to stop ooze and this seems to wok really well (although it stretched my limited programming capability)

The first thing I printed was Adrian's pinch wheel extruder. (I would show pictures but there are no batteries in my camera I will add them tomorrow). This came out really well and appears to be quiet strong- not quiet at Nopheads quality but good enough for now.

I then decided to print the wine glass as we were going out to a dinner party. Unfortunately the machine does not have enough height so I managed the stem- it looked great and only took 40 minutes. When I showed it to friends they did not understand the rep rap excitement. In fact Alex told her boyfriend Steve "Its a good job he already has a wife or he would stand no chance!"

I am now in the middle of printing a large assembly- it has been printing for over 9 hours now and is looking good despite the various breakdowns.

Next steps
1. Tidy all electronics/wires up.
2. Make sure the feed to the pinch wheel is constant using smaller hooks this appears to cause the pinch wheel jams
3. Fasten the extruder more securely to the machine
4. Fix the acceleration issue
5.Move the whole thing into the garage to reduce noise and smell
6. Increase the z axis movement

Questions to anyone

1. What is the biggest thing that anyone has made?
2. Using a pinch wheel extruder and the resistor/brass bar/welding tip/stainless set up could we just wack in solder and extrude it- what is the melting point?
3.Is acceleration an issue on Darwin?


Sunday, 15 March 2009

Pinch whell stepper extruder

I have battled for a long time to get an extruder drive that works reliably, delivering a steady speed especially when extruding HDPE. The problems were that the DC motor would slow under load and was often variable in speed as the load varied. Also the screw drive did not positivley grip the plastic rod

I have come up with a solution based on all the other pinchwheel extruders but made from easily accesible parts (Thanks Adrian Zak and everyone else).
From the photographs you should be able to see a 90 degree angle bracket (from B and Q) which I have fastened a stepper motor to. I have machined a pulley and given it a rough surface using the technique described in another blog (sorry I can't rember who) where you put a tap into the lathe and the work piece into a bracket so it rotates effectively cutting a bevel gear. This is better than knurling as 1. I don't have a knurler and 2. it gives a groove for the plastic to sit in so keeps it in the correct place.
Initially the stepper did not have enough torque to drive the plasitc consitantly but I rewired the stepper so the coils were bipolar parallel and it worked fine.
I have also made it so that the stainless tube comes as close to the pinch wheels as possible to prevent any buckling.
I just now need to work out how programme the arduino so that when it gets a stop signal the stepper reverses for a few steps then stops to stop oozing.

90 minute nozzle

I have been experimenting with NOPHEAD's nozzle design for a couple of weeks. The second picture shows one made with brass and a Peek tube. I had problems with the thermistor coming out so I held it in with twisted wire and fire cement. This worked OK but I melted the Peek when the thermistor fell out also when the nozzle hit the work piece, it bent.

I then put in place mark 2 using a stainless steel tube. This works very well using virtually the same design as nophead but without the cooling fin. It does have a narrowed wall just above the heater block anmd it is tapered at the end. The cooling comes from the plate that the nozzle is mounted into. The tube is made from 8mm stainless studding and the brass is 3/4" bar. It uses the ceramic resistor as a heater and this is sealed at either end with fire cement.
The welding tip is drilled out but is not drilled in as far as it would go as I suggested earlier on and it seems to work.
The great thing about it all is that from deciding to make it to having it heating on my machine using a pillar drill, lathe and taps it only took 90 minutes. Much faster than the old nozzle. Well done Nophead, genius design.

Friday, 20 February 2009

New nozzle

I have spent a long time messing about with the extruder design trying to get over these problems.

1. The PTFE and nozzle seperate easily under pressure

2. The heater is awkward to make from nichrome wire

3. Nozzles are difficult to make especially the 0.5mm holes which break expensive drills.

In the process I have made a lot of junk.
My new design is based on NOPHEADS. It uses 22mm brass bar with 3 holes. One small hole to hold the thermistor held in with fire cement.
The next hole is 6.5mm and holds the ceramic resistor to heat the nozzle.
The third has a 5mm long 8mm tapped hole for the PEEK insulator and then is tapped all the way at 5mm and has the bottom drilled to 5mmm to insert a 0.5mm drilled out welding tip. The reason for the 5mm drilling out at the bottom is to get the welding tip as far in as possible to maximise the heat transfer.
I have tried this out and have found that I do not need the support that Nophead uses as the Peek/brass threaded section holds well under high temperature (300 degrees) and with it extruding as fast as possible. I have also tried to pull it out and failed. Only a long run at extrusion will proove if it works.
The drive design in the picture above does not yet work reliably as the threaded rod/ drive pulley has too much play and slips. I will stiffen this up with another support pulley when I get time.