Cold Swappable Nozzles for 3d printing

I'm tired of the involved procedure of nozzle swapping. Nozzles should be:

1. Swapped cold (without heating to 285 like some printers require.

2. Do not require awkward rotation of parts during installation.

3. Install repeatably at the same height.

Here is an idea of how to do it.

The nozzle is made of a rectangular brass block. The cylinder at the top of the block has thread on it, and the nut presses the heatbreak to the nozzle. There should be no need to hot-tighten the nut.

The conical hole in the center of the nozzle serves to clamp it to the heater block using a screw with a matching conical tip (not shown).

If plastic is retracted, the nozzle shoud be able to just detach from the heat break.

I don't have time to pursue developing this idea, so putting it out in the open, with a free license to use to whoever finds it useful.

Vintage tool restoration bandwagon

Soviet and English make in one picture.

When the son has a project at school and father has a 3D printer...

Neck holder for an under-the-table router

A friend has a router under the table to which he wants to add a screw-controlled lift.

The router has a 65-mm neck that can be clamped.
Here's the clamp outline (it's a 1/2 inch slab of aluminium):

Cut it out with a hacksaw, straightening the edges on a mill (yes, those Kant Twists are invaluable).

Note the use of 1-2-3 block to set it up straight. I'm shooting at <1 degree precision, nothing extraordinary needed here so just held it with the hand while clamping.

Time to sketch the clamp with a bit more precision. This is grandpa's tool, first time I do some real work with it!

And now to the rotary table to cut the hole. Need to center it precisely over the axis of the rotary table. Best would be to drill the center hole then to indicate it but no such precision needed here, so I'll eyeball it by pointing a needle at the center and see that the center stays in place while rotating.

Clamping is gonna be tough: there is not enough place to support the washer against skewing. Guess I'll try to cut out a few small pieces of the same slab, maybe it will work.

Updates to follow...

Frame assembly - first attempt

Assembled. Looks mostly fine.

The cross rods turn out to be too short, so changing the plastic ends to fit.

Assembly starting

All the plastics of the gantry (except for the center) printed. Starting assembly of the frame.

First, assembled the legs:

And the XY ends:

Adding some loctite to make sure the bolts stay bolted:

Raspberry PI network stability - resolved

The raspberry PI I have for the central controller of all things home was misbehaving. Once every few reboots, the network just wouldn't start.

Suspecting a power problem, I tried byuing a PS at Aliexpress that was specially desined for Raspi, quoting 2.5A on the box.

Things seemed to be OK, but no, not much stability there either. After a lot of reading logs and tried reboots, I got to this page: http://elinux.org/R-Pi_Troubleshooting#Networking - seems to be a great troubleshooting guide of all things Raspi. And every second problem there is related to power supply.

So I got suspicious, and disassembled the supply. Ha! The transformer inside is quoted at 2A, that before we mention any losses on the way. Little wonder it was browning out.

It actually looks well built, with quality components, good soldering, good mechanical design etc. But it's just too weak for its job.

So far, I connected it to some high-quality USB supply I had lying around that can do 2.1A (but this one likely really can sustain that). In the future, the plan is to connect it to a really powerful 5V supply, one that can even supply 10A if necessary. That's overkill, but then, several devices can feed off it.

Configuring all the computers at home to recognize each other.

I would like all the computers at home to be known by their names, under domain "oasis". Historical reasons, fwiw.

The router I have doesn't support local dns, but there is avahi, which I can install on all the computers at home (Mac mini, two linuxes, Raspi and potentially a slew of ESP8266 and the like).

Unfortunately, avahi normally likes to use the .local suffix for its domain.

Here's what it took to get all machines recognized and searchable under .oasis domain.

In /etc/avahi/avahi-daemon.conf:

[server]
domain-name=oasis

[wide-area]
enable-wide-area=no

[publish]
publish-domain=yes

Then, restart avahi by sudo /etc/init.d/avahi-daemon restart.
After this done on all the computers, I can see them all in avahi-resolve --name <host>.oasis. But ping doesn't work yet. For it to work, need to enable nss integration to work with mdns and to accept the .oasis domain. There are two changes:

In /etc/mdns.allow:

.oasis

In /etc/nsswitch.conf:

hosts:          files mdns4_minimal [NOTFOUND=return] dns mdns4

The important change seems to be the "mdns4" at the end.

Now I can access all the machines on the local network by their names. ping, ssh, browser, all works.

Epoxy pour done

The bases have nice stony feeling about them!

Pouring update

The first pour has hit a trouble. The form got a bit distorted, presumably from the heat of epoxy curing. So, to prevent more distortion, I'm doing the following steps:

1. I'll pour in layers. Each layer produces less pressure, so will distort less. Also, the amount of heat generated by a layer will be easier to dissipate.
2. I'll mix in a lot more sand. It makes the mix harder to pour, to the point where it barely flows if at all, but the amount of heat is again, smaller.
3. I'll add a lot more structure into the forms, so that even if the plastic becomes soft, it will hold shape better.
4. I've added a step at the edges of the leg bases, so even if there is some deformation, it will be hidden away.

To pour in layers, I printed all the parts, so that one batch fills about 3cm of depth. Here are the results of the first batch:

Update: all the XY ends done + small improvement

All the XY ends are printed, ready for installing.
After some thought, I reversed the direction of M8 bolts. Now the bolts sticks from the inside out. This has several advantages:
1. Nothing inside the plastic part has to rotate when I'm tightening  the nuts. So it's much easier to assemble and disassemble.
2. The structure is less sensitive to the bolt size.
3. I think the nylock nuts look way nicer than bolt heads.

Printed the XY ends

Everything fits (well almost):

- The retainer must be some 2mm deeper for the M8 bolts to not to stick out into the slider.
- The stepper mounting plate has a round protrusion that I forgot about, need to make a corresponding hole for the motor to sit tightly.

Otherwise, it is all tight and rigid, everything fits precisely.

Pouring the epoxy and next steps

Mixing and pouring epoxy is simple. But there is one but. Epoxy may seem viscous, but it will go through the smallest of cracks.

So after I poured it, I had a fun session where I ran around with a hot glue gun (thanks whomever necessary for that invention) and plugged the links. Finally, it slowed enough that I could let it dry hoping that something will remain inside. Here it is:



The next life saver was the Fein Multimaster.

1. The spools of epoxy, which I tried to sponge out with wads of paper became a mess of semi-hard papier mache. Didn't succeed to tear them off, so I just cut them out.
2. The edges of the surface weren't flat because of the surface tension. The multitool made an easy task of sawing them off at 45 degrees.

And here is the end product:



Added legs (see the remains of silicon, glued to the frame with epoxy) :( .



The next step is assembling the linear guides. The material is already here.

Straightening the CNC plate - preparation

The metal frame is bent, and very hard to straighten back. So the idea is to pour a layer of epoxy on it to create a flat surface.

3D printer

Assembly of my Prusa (done Sept 2):

My CNC

For a long time, I'm trying to assemble the MPCNC machine http://www.vicious1.com/ Here I'll post my notes.