FTIStrap - a RepStrap built from FischerTechnik parts

Hello!

I always had the idea of constructing a 3D printer, but never found the time to build it (or, as some people might say: I did not want it hard enough). Last week, I had some time...and here is the result:

FTIStrap - a RepStrap machine primarily built from FischerTechnik parts

(see [www.andreas-rozek.de])

As pointed out in the title, it consists almost completely of FischerTechnik parts (for non-Germans: FischerTechnik is similar to Lego or Meccano, but with an emphasis on industrial models rather than private ones. This was different several years ago, but with the success of Lego (technic and MindStorms), FischerTechnik vanished from the stores) - including the steppers and even the electronic interface (no welding required!). All parts can be easily bought (online), the whole machine can be easily assembled within a single working day.

The extruder is made from a small hot-melt gun, which has been disassembled and then placed within two plates of balsa wood - a sharp knife and a bit of heat-resistant glue is required to complete this - nothing more.

The base, on which the parts are built, is simply a glass plate taken from a "frame-less picture frame" (don't blame me for that name) which can also be easily bought in many stores. Before every print, I wipe the glass with a cleaning tissue which contains a bit of iso-propanole - the printout can then easily be taken apart later.

Currently, the driver software runs on a Macintosh, and a USB-to-Serial converter is used to send any control signals to the FTIStrap - but the program may also be compiled for Windows and Linux platforms.

As you can see on the pictures, the FTIStrap is already able to "print" 3D objects - and the results already resemble the foreseen shapes. Nevertheless, there is still a lot of fine-tuning required (especially the extruder still causes some problems). Hopefully, I will be able to qualify as a real "RepStrap machine" (by printing the RepRap components) in a few weeks.

Here are some pictures:

- [www.andreas-rozek.de]
an overview of the machine
- [www.andreas-rozek.de]
it's currently printing a cubic "glass" (at least, it tries to) Unfortunately, the camera did not focus on the object but on the chains in the background
- [www.andreas-rozek.de]
after each layer, the extruder moves away to let the plastic cool down a bit
- [www.andreas-rozek.de]
this is the electronics needed to drive the FTIStrap - everything can be bought, there is no need to build the electronics oneself!
- [www.andreas-rozek.de]
FTIStrap in action - from the top
- [www.andreas-rozek.de]
a close-up of the printed object (still far too much material, the extruder has to be slowed down a bit). Please note the balsa wood which covers the heater of the (former) hot-melt gun!

If time permits, I will provide some documentation and assembly instructions within two or three weeks. Before, however, I will first try to enhance the printing quality.

Kind regards,

Andreas Rozek

L

Hello Sid!

Now that the machine is basically working, I am open to all suggestions for improvement.

Indeed, I am currently using a small Proxxon hot-melt gun with the smallest nozzle it came with and standard hot-glue sticks.

But I will definitely go out and look for alternatives as well - thanks for the suggestions. I'm afraid, I'm still at the beginning of the road...but one which is worth walking along!

Kind regards,

Andreas Rozek

Hi Andreas,

... good start

I see, you have the new Robo-Pro and -expander - so enough ports for future needs ...

Some ten years ago Fischertechnik sold 2D-plotter-kits with aluminium-bars and stepper-motors, so if you'll eventually find some of this sets (ebay or elsewere), you could build more rigis mechanics - some big companies (Volkswagen, BMW, Siemens, ...) developed most of their electromechanics and automatics (e.g. complete car-assembling plants) first in FiTe

I have two of the older controllers, one expander-module, two IR-bricks and some ten kilograms of the plastic-parts (i think more than 50 sets) and made most of my 'first-tries' in Fischertechnic too - much more 'technical' than Lego

When you need some special parts or assemblies, don't hesitate to ask - either i'll find something in my scrapery or i can mill some fittings.

I think it's a good way testing your thoughts and ideas in a 'sandbox' before investing time and money in serious hardware ...

Viktor

Hello Viktor!

Thanks for the drinks ;-)

Indeed, I am currently using the RoboInterface - but it's running in "online" mode and, thus, my software works with the old "Intelligent Interface" as well.

I needed the extension module as every stepper occupies four output ports - and I have four steppers (x,y,z plus extruder) There is no special electronics needed: eveything is done with the interfaces and some software - therefore the need for so many ports...

Currently, I only own two aluminium bars - I decided to start with the plastic parts (as I can easily modify everything) and then replace some of them with aluminium parts as soon as the basic design is "stable" enough.

The extruder seems to need some redesign as the material feed is not constant enough when the machine has to switch from one stick to the next. I already have some ideas - but in the moment, family has priority (my little son has a cold and is coughing a lot)

Later, I will run some systematic tests in order to find the parameter combinations (xy speed, extruder speed, time for the material to cool down etc.) which produce the best results

---

Kind regards,

Andreas Rozek

Hi Andreas,

fast recovery for your son! - My two consume most of my time even when they're well

For your design - when you want to reduce the count of used IO-pins, add discrete stepper-drivers with only two pins per motor (DIR and CLCK). You can find cheap IMT901-drivers or better the SMC11-drivers from Nanotec ( [de.nanotec.com] for nearly 30 Euros each) - they manage 1/8 (1/16) microstepping and drive motors until 2 Amps ...

I'm thinkin to build a paste-extruder for room-temp-fabbing with a syringe squeezed by a stepper-driven plunge - this should work good with your design too.

If i only can find some more time

Viktor

PS: are you from germany? - maybe we can exchange in the german forum in plain german too ...

Edited 1 time(s). Last edit at 09/20/2008 02:32PM by Viktor.

Hello!

I think, it's time to inform you about my progress: after building a slightly different extruder housing, redesigning the extruder feed and reducing the vibrations of the cartesian bot, I am now getting closer to "production quality".

Here are some photographs of my most recent "print":

[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]

The current characteristics of my little FTIStrap are:

- position accuracy: approx. 0.1 mm in all directions
- material thickness in xy direction: approx 1.0 mm
- material layer thickness (z direction): approx 0.5 mm

In the shown print, the corners are too massive, as I had the impression that I should wait before changing the direction of the tool head - which was definitely wrong. Additionally, there seems to be a slight positioning problem in x direction - but this might also be a software error.

Currently, my biggest problem is that the FischerTechnik interfaces sometimes refuse to drive the stepper motors - causing the whole print to fail miserably! Presumably, they simply get too hot and modifying the driving technique may help.

More in a few days, hold on!

Edited 1 time(s). Last edit at 09/25/2008 09:04AM by Andreas Rozek.

---

Kind regards,

Andreas Rozek

Sorry,

but I could not resist: now, that I am switching off the current through the steppers as soon as possible, the FischerTechnik interfaces are able to drive the FTIStrap for a complete print - here are my latest results:

[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]

The corners are getting better - but still suffer from too much material. The repeat accuracy in y direction is ok, but in x direction there are sporadic offsets whose source is completely unknown to me. Because of their non-regular nature, it should not be a software error. But as the offsets are also below the distance of a full stepper step, the loss of steps should also not be the reason. Does anybody have an idea?

---

Kind regards,

Andreas Rozek

Ok, I think I got it:

in x direction, the stepper sometimes(!) does not reach its final step. By simply keeping the current on a bit longer for the x stepper, I was able to get rid of the x offset - yeah!

Now it's time to have a look at the software!

---

Kind regards,

Andreas Rozek

Very nice. Do the objects stay flat when you remove them from the glass, or do they warp?

---

[www.hydraraptor.blogspot.com]

Nophead,

they can be easily removed and stay flat - I wipe the glass with a cleaning tissue (the same which are also used to clean computer monitors) right before printing. I do not know what is in the cleaning liquid, but I guess it's isopropanole.

Wait, I have a new print ready - I'll take some pictures.

Interesting, I don't think any of the thermoplastics we have tried so far will stick to glass and they all warp to some extent.

---

[www.hydraraptor.blogspot.com]

Nophead,

here are some pictures - again my excuses as the camera always focuses something else rather than the actual object :-(

[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]

As you can see, the base is completely flat!

The x offset has gone but the walls are stil la bit "bumpy" due to the limited glue flow control (I might have to reduce the flow a bit). The corners are getting better and better although there is still room for optimization.

During the print, there was no need for some extra time to let the printed layer cool down - nor did I use a fan. It's just the thin layer and the small amount of material pressed through the nozzle whaich lets the glue cool down fast enough.

(P.S.: in fact, the base is flat like a mirror (or like the glass plate), flatter than any other surface of the printed object ;-) )

Edited 1 time(s). Last edit at 09/25/2008 12:26PM by Andreas Rozek.

---

Kind regards,

Andreas Rozek

Hi Andreas,

... a hint for sharp photography with IR-translucent objects: - position something substantial (e.g. a flat box) as dummy-target beside your part in the same distance. Then focus the target with half-pressing the camera-button. Then hold the button, turn the camera to focus your part and finally press the button to get the image ...

Beside - read through this thread resp. glue-gun-fabbing: [forums.reprap.org]

Viktor

The glue does seem to work particularly well. The translucency is also better than I have seen in examples made of HDPE. What is the strength like?

In trying to find out what glue sticks are made from I've discovered that there is no single recipe; Ethylene-vinyl acetate EVA seems to be used a lot, but on it's own it is like rubber.

Do you have any idea what the sticks you are using are made from? Have you tried any different brands?

... i have different coloured glue-sticks (see the image in the linked post) with different hardnes and rigidity. Maybe they are made from different materials too?

I think its mainly EVA, but i have some special glue-sticks and granules meant for wood, they are really stinky ...

Most of the translucent ones are elastic, the yellowish and some of the brown are stiffer, so would be a good material for rigid objects.

You can expect the rigidity of the fabbed object to be like the original sticks, only thin walled objects are more like stiff silicone-rubber.

Viktor

Good morning!

Viktor: thanks for your hints - it seems that I'll have to learn taking photographs now ;-)

Peter: currently, I have no idea of the material my hot glue sticks are made of. I am using a small Proxxon hot-melt gun (just because it is so small) and also use their hot glue sticks - I'll try and contact them, perhaps they inform me about the material.

The translucency is pretty good, as long as you don't print layers thicker than approx 1 mm. Indeed, if you are a model maker, you may use the material to print windows ;-)

The material itself is tough and elastic - but as soon as you print thicker layers, the result get's stiffer (one of my first prints used far too much material - see the first picture set - but the resulting object is really stable). For that reason, I have no doubts that I can produce the parts for a reprap with my little setup.

What might be a problem, though, is the inherent adhesiveness of the material - after all, it's a "glue". As a consequence, the friction between a hot glue object and another one (even a metallic one) might be too high for technical applications - you might have to place a spacer or something else in between.

Kind regards,

Andreas Rozek

... you can embedd other parts and materials by placing them on the bed and fabbing the body 'upside down'.

As hint for changing/finishing the surface you can heat the fabbed object with a hotair-gun or gas-torch (be very carefully then) and stick some thin sheets of your 'finishing' material on the melted surface (for example PTFE-stripes, aluminium or even cutted fitting steel-film parts) or simply scatter some dust over the molten areas, so it's 'sanded' with a different material ...

Viktor

Good evening!

I am slowly getting better...

I reduced the size of my test object in order to speed up printing - and had to install a fan because the (now shorter) lines did no longer have the time to cool down enough.

As a side effect, the objects started to "sparkle" as the individual layers became more distinct and produced a "ripple effect" on the outside which reflects the light in many different directions - quite interesting.

Many prints with different parameter sets did not yield any substantial improvement. However, the whole set-up (mechanics, hot-melt gun, electronics) were able to "survive" a full day of almost continuous printing (it's really nice to start a print and then have it running while doing something else - with only one important duty left: to refill the hot glue sticks in time(!))

Some silicon oil reduced the friction in the mechanics, some modifications of the driver software additionally reduced the spurious step losses - now, I'm just still loosing steps in the z direction (which is really astonishing as the stepper motor has to *lower* the print bed rather to raise it - thus, gravity should help)

Corner handling has also improved a bit: due to the cohesion of the hot glue, the melt tends to cut corners - here I may have to apply some software tricks in order to further enhance the corner quality.

And, well, I still have to learn how to take pictures (sorry) - the current ones could be better:

[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]

Whenever z steps are lost, the resulting layer becomes too thin and glue wells to the sides of the line, producing the "bumpy" effect you may notice in the pictures. When the layers have their proper size, the walls look fine.

Hopefully, I'll find some time next week to solve the remaining problems and start the "production phase"!

Edited 1 time(s). Last edit at 09/26/2008 12:27PM by Andreas Rozek.

---

Kind regards,

Andreas Rozek

Oops,

it was faster than expected: the spurious z step losses have been eliminated - but, well, corners could be better and lines perpendicular to the x axis are sometimes slightly off their foreseen position.

Here are some pictures:

[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]
[www.andreas-rozek.de]

Lines parallel to the x axis look perfect (i.e. y position is accurate), lines perpendicular to the x axis could be better (this might be a side effect of my corner enhancement technique)

Let's see what I can improve next week!

---

Kind regards,

Andreas Rozek

Hello together!

It took four weeks to receive the "fischertechnik designer", but it finally arrived. I now began to create a model of the FTIStrap in order to allow people to build their own copy of this kind of RepStrap.

Please, have a look at [objects.reprap.org]

The documentation is not yet complete - but the remaining descriptions will follow during the next two weeks.

---

Kind regards,

Andreas Rozek

Hi Andreas,

... great job!

It's interesting, how distributed Fischertechnik is - i know only three guys in my radius (maybe 1000 people), which have some and only one of them (beside me ) has enough parts and interest building complex machinery ...

AFAIK FiTe is a mainly German brand and market.

Maybe an intersting question --- can anyone out in the world knowing or even possessing FiTe send a ping?

Viktor

Good morning!

Thanks for the honor, Viktor!

I don't want to start a discussion about FischerTechnik, but: several years ago, FischerTechnik vanished from the toy stores (at least, many of them). Since then, FT is even more a tool for professional modellers. Nvertheless, anybody can order individual FT parts online (through Knobloch GmbH, the "official" FT distributor)

For my little son (6 years old), I prefer Lego technic, which is easier to handle and has a better concept to integrate computers and electronics.

As far as I have seen, there are already beginnings of a Lego RepStrap in the wiki. It would be interesting to see such a model as well (not necessarily as a comparison to the FTIStrap, but as another alternative to build a RepStrap)

And what about Meccano and all the other construction kits?

---

Kind regards,

Andreas Rozek

Hi Andreas,

... it's not essential, which brand or type of construction kits you use, but sometimes ist's a more than philosophical question

My sons (8 and 10 years old) have some lego-technics and lego-educational kits and can build/construct on their own ...

But it's a dispute every three or four months, if i'll give free my FiTe - it's more parts, more stable and somehow easier to build rigid frames, gear-boxes and transmissions.

See here: [forums.reprap.org] - the image "Fischertechnic-Session1.jpg" ...

Viktor

Edited 1 time(s). Last edit at 11/26/2008 03:15AM by Viktor.

Hello Viktor et al.

If you are interested in FischerTechnik, you may also join (or have a look at, at least) the "FischerTechnik Fanclub" at [www.FischerTechnik.de]

The FTIStrap is mentioned in the "Forum" under the topic "Modellideen"

---

Kind regards,

Andreas Rozek

Hi Andreas,

... thanks for the info, i made a glimpse.

I think i'll join the FiTe-forum in the next days - so long ...

Viktor

Hello Andreas,

... here is the perfect material for your FTIStrap - found in the "ALDI-S

Hello Viktor!

Great! Unfortunately, these sticks have a diameter of 11mm, whereas I am using a smaller gun which needs 7mm sticks.

If I manage to visit an Aldi store, I'll still buy a bunch of these packs - in principle, I am looking for a solution to remelt unused hot glue anyway. With these sticks, I would have coloured hot glue, which is a nifty idea.

Thanks for the hint!

With three extruders feeding into a single in-line mixer and nozzle, you could print the full spectrum !

I think you would need yellow, cyan and megenta to do that.

---

[www.hydraraptor.blogspot.com]