“An Engineer is a machine to turn time into computers”

One of the fellows there – unfortunately I did not catch his name – has had a period of unemployment so, rather than head off to the beach or hide in his bedroom suffering from self-pity, he invested the time he had in hand soldering a complete Z-80 based computer. WOWZAH!

The happy maker and his wonderful handsoldered creation.

As I understood it, he has built an IDE interface to use a normal PC hard drive. Because the Z80 is an 8 bit machine and the IDE standard is 16 bit, there was a bit of high-byte / low-byte wrangling to be done. The current screen arrangement is via a serial port, which sounds remarkably low-tech until you remember that serial terminals are rare, so the easiest thing to do is to use an old laptop as a terminal: which means that the terminal is several times more powerful than the machine itself! There was some discussion about building a small (possibly Arduino based) interface for using a TV screen as the terminal, which would be totally retro and quite fantastic.

I am not sure now visible it is in the photo (I see it because I know it is there) but every connection has been hand soldered and the board is Veroboard / stripboard with the traces cut with a knife. This is a level of exactness and detail that, I must admit, I find daunting and amazing to the extreme. Congratulations!

From left to right: CPU, serial interface, IDE interface, plus a lot of logic "glue" to help it all work.

—

Der Drucker druckt wieder. CM hat gestern den Y-Riehmen gewechselt und sich wieder über den Konstrukteur geärgert.

Ich hab auf ReplicatorG 24 upgedatet und ein neues Profil in Skeinforge35 begonnen.

Wieder Würfel gedruckt. Darin sind wir jetzt schon ganz gut. Nebenbei gibts Skeinforge39 und Skeinforge40. Alles nur damit wir nie aufhören müssen den Drucker zu Kalibrieren.

These: Das Objekt das Weltweit am meisten gedruckt wird ist der 20mm Testwürfel. Würde man allen Testwürfel – die Weltweit gedruckt wurden – übereinander Stapeln, würde es einen Turm ergeben der umfällt.

CM hat dann am Stepper Extruder weiter gearbeitet. Wir haben überlegt ob wir ihn überhaupt in runCNC einbauen sollen oder doch lieber in den großen Bruder von WoodyMiller. Hast du eine Meinung dazu?

—

A (not very fluid) translation of my favourite bit:

Thesis: The object that is most commonly printed worldwide is the 20mm Test cube. If one were to pile up all the test cubes in the whole world that have been printed, one would obtain a tower that would fall over.

CM aergert sich ueber den Konstrukteur. (CM is exasperated with the draftsman)

The needle. The gap on the left is where the line is spooled on the needle, stretched between here and the curved part at the rear (right on the image).

There are probably a million netting needle instructions online, this one looks nice. I also like the images from the lovely book from Hervey Garrett Smith.

Lovely graphics.

His book is funny in ways that I am not sure they were intended: from the frontispiece of the author onwards it is a lively read. But that is not the topic here. Tools for making tools is. So how to make the tool? the STL file is on thingiverse, we look forward to hearing about some duplicates, adaptions and mad netting made with it!

Some test netting that was made with too-thick cordage. But it seems to work!

We are using pcb2gcode instead of visolate, which was giving offset problems between the copper traces and the drillholes. pcb2gcode does it in one go, while visolate seems to need an extra step. As well as being one of those complex  “java is a standard that is not completely standard” situations.

Two PCBs milled from plain sheet. The simplicity of the Voronoi layout is wonderful!

What I really like about these are the aesthetics that arise as a complete step away from the standard etching structure that we know from standard PCB production techniques. I am sure some clever engineering types will say something about capacitance and stray signals, but I am sure that, for our electronic needs, such problems are less that significant. And we get sexy boards.

So what is being done here. Let’s start with the style. The Voronoi system has been developed by a few clever people. There is the group at MIT (yes, every clever thing happens at MIT at some point in its career – I have been there but cleverness has not yet descended upon me, obviously the other implication does not apply..) and the strange and wonderful people at Vienna’s Metalab. Seeing as we are only using this system, we will let you read what is going on over there on their pages.

What is it that we are building. New Tools for our Tools. Stepper and other control electronics.

New RunCNC electronics. Left: power area, one stepper controller is mounted. Right is Seeduino Mega, a debug layer and the input/output bit: thermocouples, limit switches and thermistors.

We hope that this new electronics part will help us have more accurate prints, as the main problem with the extruder we are using at the moment (a modified version of the MakerBot Mk4) is the lag when turning on and off.  So we get blobs when the head stops moving and turns off, and gaps when it starts to move and turns on. I am sure there is some parameter to be set up deep in Skeinforge to fix this, but better people that I will have to lead the way down that treacherous path.

Which was precisely the task for today. After replacing one of the toothed belts that got too warm from some building process in the early stages of RunCNC, so our round gears are now round and we can finally assemble the geared extruder that is high on the wish list.

But for this we need more electronics, and more electronics is always a bit hard, so we want to make our own circuit boards, so we have put together the simplest of all mills. CM has been drilling wood and greasing nice threaded rods, CH has been fiddling with some stepper driver boards. My role has been a bit slow. But today I got to write a program with which to create some G-code that would mill out the position in the working surface that the circuit board lies in as the copper is being milled away. Which worked all very nicely except that I forgot that the first round takes off one millimeter outside the tool path as well. Oops. So our boards will need to be wedged in somehow. Double sided tape is pretty impressive stuff.

But try as I might to break stuff, the wonderful “basteln”  has paid off and the first proper milling has taken place.

The first cut! Milling through the wood word surface, a perfectly straight line!

We pumped up the voltage on the Proxxon tool to 18V instead of the normal 12V, raised the feedrate and the tool happily dug through the wooden working surface.The Woody Miller is in great shape.

Now we have it set up, the first circuit board is about to get milled. Time is on our side.

The traces have been etched (so to speak), now we have inserted the drillbit and are setting the zero point.

The milling goes smoothly, the traces separated off from the rest of the copper. We are a bit worried by the occasional copper splinter that crosses the insulating gaps. But they seem not to turn up too often.

The first PCB. Traces fine, holes were drilled a touch too shallow so they do not go through. But close enough!

Now we have the first PCB made, the next ones will not block us too long. Some teething issues with the “tool change” height being zero, so the drill bit traversed the PCB just too low and ended up snapped off. Things are looking good.

This is a very nice position to be in. We can build objects and electronics as we want them, within certain limits of exactness. There is a whole swathe of programs and systems out there to build strange things. People doing strange things and sharing them, encouraging us to do the same.

This looks like fun.

But do we care? No! We have about three dozen Skeinforge parameters to tweak and twiddle, options of rafts, things like “extrusion width over thickness” to decipher and a whole row of experimental sculpture objects to produce that should be recognisable objects like cubes, gears or some other regular polyhedra. All praise Plato.

In all fairness our two 20mm cubes were pretty cubelike. At least the ones that did not die in the middle of extrusion. But the 10mm cubes are somehow Dali meets Kepler and the attempted gears yesterday were more like a melting tower of Babel.

<photo>

Today things are not really any better.

After a series of middling success stories with heated pads running on something between 12V and 30V, we have taking nophead’s advice from the Hydraraptor to build a solid mains 240V AC powered heater. Chris M obtained some thrown out heating wire from a veneer press construction and we did some experiments with passing 240V through it. Heat. Smoke. Movement. A slight amount of fear as we wondered what the movement was about, but no danger and heat was present.This seems to be enough to get us going!

The heating wire, embedded in UHU chimney caulking compound that is drying.

Turning this wire into a heated pad was going to be harder. A piece of aluminium, some material from UHU stuff that is used to fill the gaps in chimneys and suchlike, and we assembled the base. Getting the coils set up halfway right, then holding them in place as we surrounded thrm in the UHU chimney caulking goop, was a bit of an adventure.

But they held!

Wires in place, with the magnets in their positions, the platform  seems solid. Attaching wires generated some heat and (scary!) some smoke. A momentary panic but no real danger.

Today I have been fighting with the electronics and the mounting. Getting all the parts to fit together in safe ways, balancing the stability of the platform, the mobility of the wires, getting earth connections set up properly in case something goes wrong. This is all important stuff but somehow full of contradictions.

The heating system. At the bottom the arduino plus interface board, in the middle the relay board (with an LED because that always helps) and the heating platform itself at the top.

The older temperature control circuit and Ardiuno patch that controls an LCD with three push buttons so we can set up a desired platform temperature, monitor that temperature and have the control work was re-worked for the 220V system: the transistors now drive a relay and the PID system that so cleverly does proportional control of temperature using all sorts of mathematical cleverness was replaced with a simple bang-bang system with hysteresis: if the temperature is more that 5% too low, turn on the heater until it is 5% too high, turn it off then let it cool to 5% too low, repeat. There is a problem that the 5% should probably be 1% because there is a rather large latency in the heating, but I am sure it can be worked out.

So now the problem is to get the platform on the actual build platform. This means that the zero point for the Z axis needs to be re-set and the cables need to be routed so that the whole build surface can be moved safely. The one annoying thing right now is that the surface is not as flat as the build platform surface (or vice versa) so the whole heated bed wobbles around the middle magnet. Solution: remove it! This might end up making the traction of the building surface worse but…less wobble is certainly more important.

Okay, let’s test this thing….

Update: First print. I stuffed up the zero point so the extruder head scratched the surface, digging up the Koptan tape and twisting the whole printing platform (held only with magnets) about 20 degrees. BUT I slid the whole thing up a twist on the Z axis and it printed through and: BINGO! perfect bottom surface, a mirror finish, perfectly sticking. Wowza. So this does help.

Second print: changed the Carve parameter to be a bit higher (0.35 -> 0.45) and notice that the infill seems also to have changed. This might be a side effect of the carve change, as there is less grip. The heated pad makes the Koptan stickier for the object: then once they are removed, they cool down and pop off perfectly. Mirror finish. Full points for the heated pad. Putting the spring steel printing surface back on, it bows up with the effect of the heat: as it finally gets warmed through the bowing disappears.

Third Print: The infill really is different (must have changed automagically when I changed the carve parameter) and with a higher extruder feed rate and temperature the sucking in seems to be a lot better. Th final top surface is not closed: I do not know whether it is meant to be. More like a dense net. Next try: the thin-walled test structure.

Five more prints. Things are getting better. There are problems because at each Zaxis change, a small gap seems to be produced. Trying a large jitter value (50.0 instead of 10.0) I see that the extruder motor actually turns off for each jitter motion. Seeing t there seems to be a certain lag with the extruder, this might tell us why there are gaps in the walls when we go up. So I try Jitter equal to 0.0 and we will see…

Yes. End of layer gives a blob, the start of the next layer has a gap: there is a lag in the extruder so turning it off lets it run for another second or three, giving the blob, and the start is delayed meaning that there is a gap.

Summary of the day (and night: it is coming up to the witching hour!):

• the heated platform ROCKS! in making prints better. I re-did a TUBA nameplate and it came out dead flat, just like a bought one!No need to rough up the print surface or clean with acetone or anything, it seems to simply work, as promosed in the adverts
• the extruder lag is a pain the the ar*e: this seems to give nasty messes. I hope we can fix this with the geared extruder that we have started building. I will re-do the big gear because with any luck the heated bed will keep it flat like it should be.
• The allocation of variables and other parameters in the skeinforge sections is a complete mess. I am sure someone understands it, but they are certainly a bit special, with an amazing ability to remember all sorts of details! Why are the main temperatures stuck in the raft parameters: I do not use a raft, why do I change my main temperature there?? And why does it not work!?!? I presume because I do not use a raft so I have to define it in my start.txt – but I have no idea how I should know this…..
• Going through the calibration steps (like on the thingiverse objects or this) is a good thing to do from time to time, even if it is just to remind you that there are a lot of details to play with and to re-work your preconceptions.

But that is not what we are here to discuss. The latest post on the blog talks about the first proper launching from the TUBA, and as David said, any boat that leaves a shipyard needs a recognition plate. So tonight we got stuck into it and with a mixture of some old school bitmapped text graphics that have been turned into OpenSCAD library for doing text and some high quality faffing around, we got all the parts working.

So for the first boat out of the shipyards we have the first TUBA plate to signify its authenticity.

Huzzah!

The TUBA plate for David Moises Caravan Diving Bell "Karl Stromberg"

We are developing one of the stepper extruders, quite a standard build, so as to have better control over the feedrate of filament. We hope to set this up with a Bowden feed so that the heavy motor does not need to be attached to the moving part of the system. Just the heater will be on the sled, and we are looking at making something like the plastruder Mk5 from MakerBot, which is all steel with a pair of heavy duty resistors creating heat at the business end. Photos will follow as we get closer to completion.

Now to control the stepper motor on the extruder we need another stepper control board, so C.H. has been putting one together. But we are looking at some new stepper controllers that a friend of his is developing and hope to try them out real soon now. Apparently they will be wonderful little SMD things and will be on sale (details to follow as more come to hand). All we need to do is have a circuit board to plug them into.

Which brings us to our other new ongoing project: a tiny milling machine to make circuit boards. Based upon the designs of makeyourbot, this should be a real humdinger, not only because of the chunky woodiness of it, but because we will be able to make our own one off complex circuit boards with it. Such as stepper controllers and arduino shields. But we need to have stepper controllers to do this and they are currently in the RunCNC electronics setup….

Here we reach chicken and egg cyclicity! Oh dear!!

There are a few other things going on, such as a nice little LCD and push button controlled heating pad for the platform to keep our objects warm and avoid too much warpage. Which is slowly coming together as we play with temperature sensors and all sorts of related silliness. Keep your eyes here: now that the summer vacancies are toning down, actions will be appearing. Huzzah!