A Helping Hand

A few days ago ‘Schulz’ made a post on the Australian forums. He was looking to buy a modified Wade’s Extruder that would fit on the Arcol hot-end. Those who’ve read through this blog think that this story sounds quite familiar.

So, taking pity on a fellow sufferer, I offered to print him up a Greg’s accessible extruder for him for free.

First thing up was to find the STL files. Looking at the parts posted on Thingiverse, it looked like the STL files posted there weren’t pre-compiled for the Arcol hotend mount.

I get around that limitation, I downloaded the OpenScad file also published along with it. Opening up the file it looked good, but wouldn’t compile without a bunch of warnings. It was missing some configuration files that were meant to go along with it.

At first I tried having a look at Thingiverse, to see if the missing configuration files could be found there. Unfortunately not. I then had a revelation. I went to Github, and did a search for Greg Frost. There he had his fork of the Prusa files there. Looking in the list, I saw a newer version of the extruder file, and all the configuration files that it was looking for.

Once I had all the right configuration files, the extruder compiled just fine. I then added in the options for the groovemount, and the Arcol hot-end. For some reason, the groovemount option didn’t compile in correctly, but the Arcol mount did, which is all that I was looking for.

I then exported the file to an STL, and loaded it up in Pronterface. The estimation came up as three hours, fifty minutes. Three hours later, the print finished.

What struck me during the print was all the solid layers. I only used my standard 35% fill, but it printed with a lot more solid layers. It must be a feature of Greg’s design to include lots of solid reinforcing layers.

In all, I was pretty happy with the print. It doesn’t look quite as good as Greg’s extruder that’s on my machine, but it’s relatively close. Interestingly, mine required less cleanup than Greg’s did, but his surface finish is superior.

Printing out the object gave me a great sense of admiration for Greg’s design. It’s by far the most complex print I’ve ever done. Whilst the snowflake’s were complex, they were only complex in a two-dimensional manner, and were just a extrusion of that two-dimensional shape. The extruder is a true 3-D shape of high complexity.

After I did the extruder, I went to do the gears. I didn’t want to print out the standard gears, so I found some STL files for some herringbone gears. The herringbone gears mesh together more tightly, so there’s less backlash from the extruder. Also, even though they are a more complex shape, I think that they’re more forgiving for poor printing.

The gears printed out fairly well. My main concern for the print was that the top of the small gear would get too mushy from too small a print area. To get around that, I put in some ‘orbit’ around the print. Unfortunately, I forgot how oozy my hot-end is. It meant that there was a lot of fine strings all around the print that needed to be cleaned up at the end.

The teeth on the large gear printed fairly well. The teeth on the small gears not so well. They were a bit blobby, and needed some cleanup. They tidied up pretty well, and I think that they’ll work fine. They seemed to mesh together nicely.

The completed extruder pieces:

All up, it was a very good test for my printer, one that I think it passed fairly well. Yesterday, I threw all the pieces in the mail. Hopefully, they’ll work great for Shulz and get him on the path to printing!

Hall-Effect Endstops

I made a little purchase in the middle of December, and there was a nice package waiting for me when I got home from holidays – Hall Effect endstops!

For those who aren’t familiar with the Hall Effect, it’s the voltage difference generated by a magnetic field. They can be used as a non-contact distance-measuring device. Hall effect sensors are used in various devices, most notably DC motors with fixed magnets, as a timing measure. The wiki page can be found here.

Its advantage is that it’s a non-contact, adjustable endstop. The problem that I have with the current endstops is that adjusting them is a pain. You have to have them attached fairly tightly to prevent them from moving during normal operation, but then trying to loosen them and adjust them by a small fraction of a mm is fiddly. This isn’t a problem for the X or Y-axes, which don’t require so much precision, but I find adjusting the Z-axis ‘just so’ to be very time-consuming and inaccurate. Lots of back-and-forth until the distance is ‘right enough’.

So, I purchased these Hall-effect endstops from mauk.cc. He makes these, and sells them fully assembled for 9 Euros each. At that price, I couldn’t resist giving them a go. I’m going to use this endstop to replace the Z-axis microswitch.

Due to limitations in the firmware, you can’t just set the printer to stop at a particular distance from the endstop. (Though that would be extremely nice.) It operates as a standard endstop, with an ‘on’ or ‘off’ signal. However, you can adjust (via a trimpot), the distance at which the ‘on’ signal activates. Thus, you can easily and quickly adjust the distance of the printbed to the nozzle.

I’ll set this up over the weekend, see how well it works.

A Welcome Return

Back!

I had a good break over Christmas, going down to Brisbane for a couple of weeks. Didn’t receive any reprap-related Christmas presents. This was the closest one I got.

Mad science! The essence of 3D-printing!

Haven’t done anything Reprap-related for the last while. The grandmothers really enjoyed the printed snowflakes that the girls had decorated, and everyone was very impressed that I’d made them on a 3D printer. Except my brother-in-law, of course. ‘Yeah, we’ve got about four 3D-printers at work.’

I had a guest around just a couple of days ago. He was interested in the Prusa, particularly since the last time he was around, the printer was down with a broken extruder. I printed out one of my ear-bud holders for him. It was good to just sit down and do a print, without having to mess around with settings, or software updates, or any other drama. The print turned out beautifully, too. A great demonstration for my guest.

I’ve got a couple of interesting projects in line for this year, I can’t wait to get started on them!

Praxis

This weekend, I’ve just been having fun printing out things.

First up, I printed out a few of these snowflakes. This piece gave me a lot of grief during the week, trying to figure out the right settings to print it. It’s a real testing piece. Lots of sharp corners will test the limits of your printbed for adhesion on the first layer, and it’s an absolute nightmare for retraction. If your retraction isn’t perfect, it turns ugly, extremely quickly.

At one time during the week, I had my retraction spot-on. However, I hit the perimeter-speed bug in SFact, which prints out the perimeter of the first layer at 60mm/s. That was way too quick for this complex piece, so I had to upgrade SFact, which then means all the retraction settings were obsolete, and I had to start back from scratch. I got them sorted out on Saturday morning, and I’ve been having fun just printing ever since.

I’ll be printing out four of these snowflakes. My daughters are going to decorate them, and them give them to their grandmothers. Should be a nice Christmas present from them. Here’s a video of the snowflake in print. Have a look at all the retractions / relocations during this one layer!

Next up I printed up a screwable jewellery box. I thought that this would be an excellent piece to demonstrate just how well the printer is working now.

The box turned out beautifully, and screws together perfectly.

Just for comparison, here’s how far I’ve come in two months:

Next up, I printed out a pair of pliers. Just because!

I also lost a couple of prints during the week. For some reason the extruder was skipping steps. I tightened everything up on the extruder, re-mounted the motor and motor gear, but still had issues. Turned out one of the extruder motor wires was loose. It was being pulled too tight by the x-carriage and losing connection. So I rotated the motor to give the cables more slack, and re-connected the wires. All good!

Slic3r, redux

I was thinking that I didn’t give Slic3r a fair shake last time. Doing 60mm/s and 0.2mm layers was probably a bit too much. I don’t think any skeining program could have held up to those requirements on my printer.

To be fairer, I thought I’d do another earbud holder, but this time at 30mm/s and 0.33mm layers, which are my current SFact settings, do a direct side-by-side. Here’s the results:

The earbud holder turned out fairly well (especially once cleaned up).The primary difference between the SFact and the Slic3r output is the flyaways – a function of the retraction settings. SFact’s retraction is clearly superior at this point in time. However, if your hot-end isn’t particularly oozy like mine is, then Slic3r may be a better option.

Watching the print, Slic3r definitely had some areas in which it was superior to SFact. SFact always starts each new layer at the same point. Slic3r instead just goes to the nearest point on the perimeter to where the fill finished. This helps reduce the ‘blobby corner’ effect that SFact produces with me. Its fill logic is a lot better as well, reducing unproductive travelling.

I’d really like to have a try at some of the other fill types that Slic3r has – it might produce better fill performance, and maybe end up with some better surface finishes than you get with SFact’s standard fill.

I’m thinking that the Slic3r might actually be a good choice for beginners. It’s definitely not as imposing or scary for people starting off, and produces quite good results.

Here’s all the earbud holders I’ve printed so far:

This is one thing that I’m really enjoying about doing this lot of earbud holders for Christmas presents – it’s giving me a good chance to refine my printer, and do a lot more side-by-sides, experimenting with just one variable at a time.

Backing up Godaddy with Rsync

Like many, I’ve got my webhosting with Godaddy. I quite like them, as they’re cheap, you can host multiple websites very easily and they provide you with a really good level of control. My only issue with them is that you don’t have rsync functionality to backup. I much prefer rsync for all of my backups, as the incremental process is so much faster than a full ftp backup.

I was having a look at this problem this morning when I found this page here. It outlines a way to get a copy of rsync onto godaddy for use. Unfortunately, it was a bit thin on details, so I thought I’d expound them a bit more here, so that people from the future may be able to learn from what I’ve found.

Godaddy’s hosting is currently using CentOS 5.5, so you’ll need to find a copy of rsync which is compatible with that OS. Version 2.6.8 currently works. You can find it here.

This is an RPM package, so it’s not easy to open. Grab 7-zip (freeware) program, which can read it just fine. Go to the ‘bin’ directory, and copy the ‘rsync’ executable. Copy it to Godaddy.

I had troubles when I copied it. For some reason, my FTP program copied it in ASCII mode, which corrupted it. I had to manually set my FTP program to transfer in Binary mode. Make sure the filesize comes out at 313688 bytes.

Once it’s on Godaddy, log in with SSH. SSH isn’t automatic with Godaddy. You need to go to your hosting control panel and turn it on.

home$ ssh genericusername@example.com

Create a new directory called ‘bin’.

example$ mkdir bin

Move the rsync executable to the bin directory

example$ mv ~/html/rsync bin/

Make the program executable.

example$ chmod 755 rsync

That’s all you need to do! This then is the command I use to backup my websites.

home$ rsync -aviPh --progress --delete-after -e ssh --rsync-path=bin/rsync genericuser@example.com:~/html/ /mnt/Backups/Websites/

Argh!

I was printing out the 3D-knot as a bit of a demonstration / vanity piece. However, when the print was at 77%, I accidentally disconnected the USB cable!

Unfortunately, the cable I have at the moment isn’t long enough to reach my printer, so I use a USB-extension cable which is a bit prone to coming apart.

It was looking really good, with only a little bit of lifting up at one corner.

 I noticed that a few of my recent pieces have had this lift-up issue. To combat it, I’ve made the following changes:

  • Add in a few extra shells on first layer only.
  • Drop down the nozzle so its a bit closer to the bed (0.13 to 0.18mm)
  • Reduce the ‘Additional height (ratio)’ in ‘bottom’ tab to 0.3
  • Give the bed a really good clean. There were a few grease marks causing the filament to stick badly, which would then start off the lift-off.

That’s really helped on the last couple of earbud holders I’ve printed out. I’ve also figured out how to take some really nice videos:

Okay, I’m really going to have a good look at SD card printing now.

Removing Extra Shells

I’ve had the suspicion for a while that the ‘extra shells’ setting in SFact was causing a few of my prints to print out sub-optimally. Now, I’ve got some proof. Some 608 bearings arrived during the week, so I thought that I’d print out some Z-axis stabilisers. I’ve seen that the walls on my prints have a bit of ‘wobble’ in them. This wobble isn’t on the Y-axis, it’s only confined to the X. I’m thinking that the Z-axis is moving a bit, creating the X-axis wobble.

However, when my first print was done, it looked like this:

There was a clear gap between the inside wall and the outside wall of the bearing holder. This gap went all the way down. I think what’s happened is that SFact creates extra shells. However, once these shells are created, there isn’t enough space left for SFact to then do some infill.

To try and combat the problem, I went into SFact, and removed the extra shells on all layers except the base layer. I then tried out the same print again, and it came out perfect.

At the moment, I’m wondering why SFact even has the ‘extra shells’ feature at all. Just how necessary is it?

With another print of the earbud holders, I tried removing the extra shell on the base layer as well, but I ended up having trouble with my first layer. Once the infill started, the head would sometimes pick up the outside line, tearing it off the bed. Adding back in one extra shell fixed this problem.

Once two Z-axis stabilisers were finished, I fitted bearings in them. (They just pushed in with only a moderate amount of force. Perfect sizing.) I then mounted one on the X-motor side. (The one on the X-idler side will have to wait for some endstop fiddling). I then printed out another earbud holder for comparison. It came out beautifully. The combination of the Daily Branch of SFact, Marlin 1.0 Beta, and the Z-stabilisers is making my prints look absolutely fantastic.

This part needed absolutely no clean-up at all. This is the way it came out of the printer.

Before Z-stabiliser on left. After Z-stabiliser on right. Notice particularly the fill goes right to the edges much better.

Stabilisers on top holder.

In related news, the ooze problem seems to have disappeared. Have a look at the photo below. The vertical hole has printed perfectly, no strings at all. I’ve got my suspicions as to what’s fixed the problem, but I’ll leave that for another post.

 

Slic3r

Okay, I’m a bit sick of fighting ooze at the moment (see last post for more details), so I thought I’d get in and print some useful objects.

When browsing Thingiverse the other day, I saw this design for a holder for Apple earphones.

I thought that these might make ‘stocking stuffers’, and a good answer to the question ‘Well, what can you make with a 3D printer?’ that you often get asked.

Not willing to just print something out (heaven forbid), I used the opportunity to also give Slic3r a try. Slic3r is an alternative skeining program, different from Skeinforge and SFact. Unlike SFact, which basically works as a ‘plug-in’ to Pronterface, Slic3r is a stand-alone program. I’m not sure what advantages this gives, but it’s extremely fast to skein, which is probably due to running natively vs running interpreted Python code.

One of the supposed strength’s of Slic3r is in thin layers, so I thought I’d try something a little different.

I printed up two earbud holders. One, using Slic3r, would be at 0.2mm layers, but printed at 60mm/s. The other, using SFact, would use 0.4mm layers, and print at 30mm/s. So they should have the same printing time, but each have a different area of focus.

Here’s the results.

Slic3r’s output is on the right, SFact on the left.

I’m not totally impressed with Slic3r’s results. It’s clearly missing a lot of the fine tuning which has gone into SFact. I think that the 0.2mm layer was just a bit too thin, and the walls came out a bit unevenly.

I think that it’ll definitely be worth keeping an eye on Slic3r. It’s only very new, but doing pretty well so far.

More Ooze Fighting

Unfortunately, I didn’t have much luck getting help for my ooze problem in the Reprap forums. The only respondent suggested that the best way to fight it is to replace the hot-end, with a J-head nozzle.

While I do have a J-head spare, I’m not quite willing to give up on the Budasnozzle yet. I thought I’d sit down with the ‘daily branch’ version of SFact (which has more retraction-related settings), and battle my way through the various settings, see what the effects are.

Here’s my output from my tests:

Retraction settings:

v1 – Use Fixed Retraction
Retraction speed – 13.3mm / s
Retraction distance – 1mm
Restart Extra Distance – 0.1mm
Result – Corner bulge worse than SF default

v2
Retraction speed – 13.3mm / s
Retraction distance – 2mm
Restart Extra Distance – 0.2mm
Result – Even worse.

v3
Retraction speed – 13.3mm / s
Retraction distance – 0.1mm
Restart Extra Distance – 0.2mm
Result – Really terrible – filament everywhere. Did I save the second one correctly before skeining?

v4 – retry v2 again, slow down filament speed to 30mm/s just to get a better idea of what’s happening.
Retraction speed – 13.3mm / s
Retraction distance – 2mm
Restart Extra Distance – 0.2mm
Result – The same as V2. Notice that the Z-axis bars are moving around a bit. Probably need to tighten up X-axis bars again. Maybe print some Z-stabilisers.

v5 – Change settings again
Retraction distance – 3mm
Restart Extra Distance – 1mm
Result – Okay, it looks like the ‘restart extra distance’ is causing a lot of the blobbing. It’s the extra filament that gets fed into the extruder before starting the next section. Take this back to 0, I think.

v6 – Restart Extra distance at zero.
Retraction distance – 3mm
Restart Extra Distance – 0mm
Result – Much better. Internal blobs lowest of all prints. External still higher than default settings. Can you set a negative Restart extra distance?

v7 – Restart Extra distance at -1.5.
Retraction distance – 3mm
Restart Extra Distance – -1.5mm
Result –  Very good. Improved internal blobs even further. Corner blob improved but still larger than defaults

Here’s some of the results from the tests;

While I wasn’t successful in removing the ooze, I’d still call tonight a success. I’ve learnt a lot about SFact’s retraction settings, and feel like I have more control over the machine, in general.