Před nějakou dobou jsem řešil delikátní problém, studenti Střední školy aplikované kybernetiky Hradec Králové si postavili Prusa Mendel s updaty různě postahovanými online a pustili se s ní do soutěžení. Nejdříve vyhráli SOČ a pak České hlavičky. Problém je v tom, že mají všude podepsaná čestná prohlášení, že všechna práce je jejich nebo uvedli původní autory, nicméně, žádné citace zde nejsou a vydávají to za své.

Řešil jsem to pár týdnů a byl z toho akorát naštvaný a nevěřím, že se něco změní k lepšímu. Dávám proto alespoň online dokument, se kterým jsem bojoval a můžete sami posoudit, o co jde.

Ve zkratce ještě napíšu věci, které nejsou v oficiálním dokumentu.

Nejdříve jsem navštívil pana ředitele pana Ing. Jana Langa, abych mu řekl o problému, zde taky jsem dostal nabídku na “finanční odškodnění” za podmínky dalšího neřešení věci. Samozřejmě jsme odmítl. Oficiální vyjádření je, že škola v tom nemá žádný zájem a je to věc studentů. (Pozdějí jsem zjistil, že za vyhrané soutěže má škola body a peníze další rok.)

Kontaktoval jsem pana Václava Marka ml., pořadatele soutěže České Hlavičky, od kterého jsem nakonec i dostal informaci, že porota celý případ přezkoumala, uznala pochybení, ale zároveň nebylo tak velké, aby se nějaké ceny odebíraly. Dále bych podotknul oficiální tiskovou zprávu z vyhlašení, kde chlapci dostali ocenění za, cituji:

Za vlastní konstrukci, výrobu i softwarové vybavení 3D tiskárny získali, 
Václav Holler, Miloš Jedlička a spol., studenti kybernetiky ze 
Střední školy aplikované kybernetiky cenu Futura.

Nicméně v televizním přenosu pár týdnů po tom co jsem začal věc řešit, cenu dostali za konstrukci páky, která údajně sundavá předměty a systém na zprávu tiskové fronty  Jen pro poznámku, objekty pákou prezentovanou v práci nejdou sundat, pokud je předem uživatel neodloupne.

V neposlední řadě jsem také kontaktoval porotu SOČ (Středoškolská odborná činnost), a to jmenovitě pana Miloslava Hlaváčka, který přislíbil, že porota věc bude řešit, ale on sám mi nemůže podat oficiální vyjádření. Nicméně odcituji jedno z jeho osobních vyjádření:

Mou snahou je, aby se problém vyřešil a přitom nebyly poškozeny již mnou citované
soutěže pro talentované středoškoláky a mám za to, že máte i se mnou v 
této věci shodné stanovisko.

Možná by stálo za úvahu, co mně nyní napadá, že byste autory práce "vzali 
pod svá křídla" a tyto, pokud se zajímají o tiskárny 3D a jsou schopni 
dále na sobě a problému pracovat, zapojit do Vámi řešených úkolů a zadáním 
drobných dílčích úkolů by Vám tito mohli být v ledasčem nápomocni a tím 
využít jejich potenciálu ke vzájemné spolupráci. Co Vy na to? 
Nebylo by to nakonec nejvhodnější řešení?

Oficiální vyjádření SOČ se mi bohužel zatím nedostalo.

Kdybych to měl shrnout, nejvíc mě mrzí, že jsem na Střední škole aplikované kybernetiky HK chtěl kdysi sám studovat. Děkuji všem zúčastněním za krásnou šarádu co sehráli a že se vlastně všechno vyřešilo i když se nic nevyřešilo

Díky Miloš Jedlička, Václav Holler, Jan Kellner, Ing. Jana Lang, Václav Marek ml., Miloslava Hlaváčka a další.

Update (26.3.2012)

Našel jsem v archivu velice zajímavé emaily ze samého počátku akce. Konkrétně s panem Ondřejem Šintákem (konzultant práce). Tehdá to vypadalo ještě nadějně.

A nyní už samotný dokument.

Continue reading →

Few weeks ago I stumbled upon interesting problem, we were trying out Solidworks on friends computer (MacBook with integrated Intel HD4000) and everything was very slow. Turned out, that OpenGL acceleration isn’t working. Well, obviously, you have to buy a special certified graphics card for that. Which ain’t easy with a MacBook. I looked around the internet and found few registry workarounds. None of them worked though.

Solidworks on Intel HD4000

After some digging, it turned out that you need to set Workarounds in registry, but there was no key for our current setup. When you set Workarounds in the “Current” key, it resets to default value after, which disables the OpenGL.

HKEY_CURRENT_USER\Software\SolidWorks\SolidWorks 2012\Performance\Graphics\Hardware\Current

Solution turned out to be extremely easy! We need to create a new key for the Renderer  you find under Current key. In our case it is Parallels using Intel HD Graphics 4000 OpenGL Engine.

For Parallels and Intel HD4000, you can create “solidworks.reg” with test from bellow and double click it, or you can follow the screenshots.

[HKEY_CURRENT_USER\Software\SolidWorks\SolidWorks 2012\Performance\Graphics\Hardware\Parallels using Intel HD Graphics 4000 OpenGL Engine]
"Workarounds"=dword:04000480

Set the value to 4000480

Start Solidworks and enjoy!

So today is the day. I’ve wanted to write about this for a long time now. Some of you may know, some may not, but I started my own RepRap company. It’s called Prusa Research, maybe it is the first open hardware based company in the Czech Republic, but certainly not the first worldwide. There are many companies/projects doing that very well, for example, Arduino (I’m so grateful that thanks to RepRap I can now call Massimo my friend. I started doing hardware because of Arduino), Adafruit, Sparkfun, etc.

RepRap and especially 3D printing is now full of bullshit, this is true. There are so many new companies that are being created these days. Most of the guys haven’t contributed a single thing and yet they call themselves inventors , but enough trolling. I’’ve been in the RepRap community for a very long time by todays standards and helped to spread it a lot. Prusa Mendel is probably the most wide spread 3D printer on the Earth . I do have different position when I want to dive in then others, I have a lot to lose if I do it wrong. Since I applied for my business license (dunno how to call it in english) I thought of a way how to keep myself from doing any bullshit marketing and do it right. Trust me, sometimes it’s very tempting to claim stuff you haven’t done or isn’t true? Like “Hey, my printer can print at 50 micron resolution!” LOL. I designed a bad ass tattoo to help me with that. It’s Open Hardware logo with standard RepRap infill inside. It’s on my right forearm, everyone can see it when I shake their hand, when we meet or sign a deal, it doesn’t matter. It has a great deal of symbolism for me. Open Hardware and RepRap got me where I am. If you are starting with Open Hardware, don’t be afraid to do it right, you can still succeed.

What finally motivated me to write this article is a something very sad that happened today. Makerbot turned closed source, or at least all the signals lead to it. Makerbot was started by one of the RepRap core developers, Zach Smith, Bre Pettis and Adam Mayer. Adrian Bowyer, founder of the RepRap Project, gave them some cash along with others to start. Makerbot is based on RepRap, and they weren’t ashamed about it. That started to change and Makerbot started to distance itself. When I gave talks about RepRap people started to accuse me of ripping off Makerbot. I was like WTF! I even exchanged a few emails with Bre. I still saw Makerbot as a powerful friend on Open Hardware field, as a shield which will get the potential law suits by big 3D print companies sorted out.

Bre was giving keynote last year on Open Hardware Summit, etc etc. Company that shows to others that the Open Source way is possible. Even though I grind my teeth about the PR which make it seem like they invented the thing, see this crap Makerbot invented stepper extruder They even did a nice vid about Open Hardware

Later on they got some $10M of funding. And things started to change in the background. Just recently they did one Open Hardware achievement, with Tangibot, clone of Replicator on Kickstarter (google it if you want to know more) where community totally crashed the clone and prove that they do not need any closed source bullshit to succeed.
But surprise, surprise, we now have a Replicator 2 and it is closed source. Hey look, we took all your improvements you shared on thingiverse, compiled it into one package and closed it for you . Same with MakerWare. (They finally, after several years, stopped using Skeinforge, software done by Brazilian who doesn’t even have a printer )

And you know what is the biggest, sneaky move? Not mentioning it while they announced it. My guess is, that they will mention it when first pre-orders ship out. Which is after OPEN HARDWARE Summit where Bre gives a talk (I wonder about what, lol) and Makerfaire and all magazines writes about them as Open Hardware heroes . That is the sickest move. I had to call their support and ask them directly. I got the answer, that Replicator 2 is closed source. Call them yourself, I wonder if the guys already got instructions to bluff ? +1-347-334-6800
I asked on their facebook page, asked Bre directly and tried to ask some other employees but didn’t get any answer. If it’s Open, why don’t they say it, right?

Open Letter to Bre Pettis.

Hi Bre,
we know each other for some time. I want to ask you about the Replicator 2, and if it’s closed source? If so, then why? I would also love to shoot an interview with you for my RepRap interviews show on youtube, I promise I will be neutral. But you have to explain weird behavior I wrote about in the linked article.

Jo Prusa, RepRap core dev.

EDIT: OMG and look at http://www.thingiverse.com/legal Looks like I’m going to pull down all my stuff from Thingiverse … I feel very sorry that I have to rant this way, but this makes me really sad!

UPDATE 20.9.2012 9:22 CET

Bre released statement  here http://www.makerbot.com/blog/2012/09/20/fixing-misinformation-with-information/. Basically saying “We might be open or might be not.” I don’t know how you can already sell stuff to ppl you are not sure about license. I get a big NO from that statement :-(

$streetcred – -;

UPDATE 21.9.2012 10:20 CET

Zachary Smith, one of the Makerbot original founders, published an article about this incident. He is concerned about Makerbot future. Even to the point that he added sad kitty picture to the post! http://www.hoektronics.com/2012/09/21/makerbot-and-open-source-a-founder-perspective/

UPDATE 22.9.2012 7:26 CET

As I’m going deeper, I find more and more interesting stuff. Did you know that Makerbot has patent? Surprise for me too  Automated build platform. Idea pitched year before patenting by Adrian on RepRap blog? Continuous belt production.

Well in the meantime I looked at MakerWare if it can be made to work with older Makerbots. It turned out it is possible and ridiculously simple. Adding support officially would  take minutes. But It seems noone want’s you to see that. Makerbot had to buy commercial Qt license (if they didn’t they are violating Qt license) to be able to make MakerWare closed source. Qt is library for making user interface. I’m starting to think, that closed source is not misinformation but was planned for some time. But this is a speculation and my personal feeling. (UPDATE: Official support should come “soon” according to http://www.makerbot.com/docs/makerware/use/)

Check out my video how to use MakerWare with Thing-o-matic and Cupcake.

Update 24.9.2012 6:18 CET

New statement released by Makerbot. Saying “What was open stays open, new stuff will be close.” Which is what we expected as the worst case. And we were completely right with our predictions.

What enrages me is this statement “Specifically the one that states that “cloning ain’t cool”. The electronics are nearly identical to our original Mighty Board electronics, the extruder is nearly identical to our original Replicator extruder with only minor tweaks to optimize manufacturing of injection molded parts.”

Why? Mighty Board is one board version of Arduino + Ramps + Pololu RepRap electronics. So don’t get angry on someone who did same thing as you. And their super cool extruder is a copycat of UP! Chinese 3D printer I’ve got that even confirmed from one of the employees last year at Makerfaire. Whooops

We still have no statement on the patent they have on Heated Conveyor Belt  looking very very similar to stuff posted on RepRap blog. 

We still have no statement on the reason why they didn’t announce this information during launch.

We just got bold statements how their printers are the best …

For latest updates, follow me on twitter

Follow @josefprusa

Previous RepRap interview with Sound, Slic3r developer!

When I first gave a talk about RepRap (TEDxPrague2010 and TEDxVienna2010) I thought it would be wonderful idea to give every visitor some small printed thing. I love simple whistles, they print quickly and show that RepRap prints are functional. As simple the idea sounds the actual execution was hard. Back then I was printing one whistle roughly 30 minutes and the conference had 500 visitors! Do the math. I was printing them for almost two weeks on two printers

But when you think about it, it is not really using the potential of 3D printing. Making 500 instances of same objects is a job for injection moulding. Back then I started to play with idea that every guest should get whistle with his own initials. Printer does not care what initials it puts on whistle, it will take same time. Since that time, two years passed. I designed Prusa Mendel, probably the most build RepRap ever, I sped up the printers to print one whistle under 8 minutes and learned some more coding.

Only thing I was waiting for, was right conference with not too wild number of visitors and actual list of visitors. Fortunately guys making INFOTRENDY conference in Bratislava wanted me to talk about RepRap. I had great time there, even met Adrian Bowyer there.

Conference had only 150 visitors and full list of them one week before. I freed some of my time from writing my Getting Started with RepRap book and coded small python script generating plates of whistles to print.

Well and it worked! I printed all 150 on my Prusa Mendel the day before I left for conference. Everyone was happy to get custom made gift and it was much easier to show them the versatility of 3D printing! No-one has to give out boring prints anymore!!

You can download WhistleGen in my Github repository.

You should check out sample chapter of my Getting Started with RepRap book! The book will help you choose right breed of RepRap to build, build it, configure it and design for it. Sample: How RepRap prints.

So finally I’ve found great and easy way how to mount my PCB heatbed.

Problems is that PCBs aren’t flat. If you’ve got your PCB and it’s not flat, its not a big deal.
You always have to help it to be flat. Ideal stuff is glass sheet, since it’s cheap and pretty flat.

What we also want is quick swap of the print surface. Before I tried mounting whole heatbed on magnets, unfortunately, it was more PITA then good solution. The glass in this solution wins again.

Many people used the four holes in the corners to mount the bed on standoffs, which again wasn’t the most brilliant idea under the sun, since it wasn’t enough to keep the bed flat enough.

What you need:
PCB Heatbed
20x20cm glass sheets, 3mm or 2mm
Two part epoxy glue
4 bulldog clips

How to do it:
1) Place the heatbed on the wooden top plate on your Y axis and mark the corner holes and its outline on wood, remove heatbed and drill the holes with 3mm drill.
2) Place the bulldog clips on the marked contour of the heatbed and mark how deep it goes into wood. Then cut this pieces of wood away, it will allow you to freely hold the glass on bed with these bulldog clips.
3) Mount the pcb. Add three washers between the wood and PCB. It will allow the pcb to be more flexible = easier to make it flat with glass and also, air is better heat insulator then wood.

4) Use bulldog clips to hold the glass on the PCB.

5) Mount thermistor with some kapton tape on the glass. I place some wool and alu foil on that and then I tape it down again. There is also possibility to glue the thermistor on the bottom of PCB Heatbed with the epoxy, which will be more handy if the glass is changed frequently.
6) Put some kapton on the top of the glass or just clean it to print pla.

Electrical wiring:
I suggest to use my PCB heatbed with RAMPS 1.3 which has great support for heatbeds.Use some thick wires, the PCB will eat almost 11Amps.

There are two 0805 surface mounts LEDs. Idea is, to have each of the in different direction, so the PCB will shine if you connect it both ways.
You also need one 0805 SMD resistor, 1kohm should be fine for every led to dimly signal you the activity of heatbed.

So, if you are wondering whats new in Prusa printer development, I’ll answer some of your questions today. And maybe even bust some myths First I want to show you results!

(Darker blue is print from Dimension uPrint, Lighter blue is print from Prusa with felt bushings)

Most of the development is done lately in felt branchhttps://github.com/prusajr/PrusaMendel/tree/felt on my githubhttps://github.com/prusajr, which started as idea to remove need of the PLA bushings and replace them with easier to get stuff. After all it ended up much better then I thought.

New bushing look.

It’s printable directly on the carriage! No glueing needed, you just need to print all x-ends and x-carriage in same direction.

Another amazing thing is, that it can be used with more materials then just felt! Which I found pretty amazing!

In the same printed object you can use:
- felt (100bushings for $5 if you use best felt)
- brass tube (80bushings for $9)
- pla bushings (one prints in about 30seconds )
- igus bushings (as Shapercube use, its pretty expensive because of matching rods)
- sku bushings (steel with teflon inside. expensive because of rods $50full set incl rods)

All of this stuff remains snap in! So you can remove your carriage without disassembling half of printer.

During the big demand I will incorporate the possibility to use linear bearings too, but it will be few days after this release. And it’s not going to be the main stuff. Simply because it’s not better and sourcing is harder and pricier.
ETA to merging it to master branch is about two weeks. Catching last bugs out. And also, from this point, every commit will have its own ThingDOC documentation, which is under huge development during moving documentation to the files!

ThingDOC is now Ready for its prime time!

n this article I want to help you with SF40 calibration.

Main difference over older version is volumetric flowrate.

It basically means that SF counts how much of the plastic its should extrude for given layer height and width of trace (height*width*lenght = volume).

Most confusing on all this is to realize SF uses the length of filament which goes in instead of extruded thread which goes out. For that you need to set your e_steps_per_mm in your Firmware to count with filament instead of extruded thread. And I’ll show you how.

Tools you need:

1. Some permanent marker which can make visible marks on your filament
2. Calipers
3. Brain
4. My RepRap calculator http://calculator.josefprusa.cz/#ExtrusionSF40

So how to do it?!

1. Set your e_steps_per_mm to some nice value. Lets say 50. You can do that either by re-flashing FW or simply by sending “M92 E50″ G-code (Works only on tonokip based FWs = Tonokip, Klimentkip, Sprinter, …) (See note 1)
2. Heat up your nozzle to temp you normally use for the print. (Or you need to remove nozzle. But we are lazy, right? Or atleast I am )
3. Find the speed at which it can safely extruder without skipping steps. E.g. hit the extrude button in Repsnapper, if the filament extrudes, you are fine, if not, reduce speed.
4. Mark some length on the filament with the permanent marker and calipers. I use 3cm, but longer it’ll be more it will reduce our measurements error. Then mark another segment right after the 3cm one. Let it be 1cm, this will later be used for validation of our new settings.
5. Extrude just enough material to get the first mark on filament aligned with intake of extruder.
6. What we will do now is extruding of the marked 3cm. So hit extrude button as many times as it needs and count how many times you pressed the button. DON’T CHANGE THE EXTRUDED DISTANCE DURING THIS. (You actually can, see note 2)
7. Now we have all the info we needed to count our new e_teps_per_mm Which you can count ( (extrude button clicked times)*(extruded lenght per click in host)*(e_steps_per_mm we set [50]) ) / (marked length on filament in mm we have [30]) Or you can use interactive calculator athttp://calculator.josefprusa.cz/#ExtrusionSF40 Explanation: ( (extrude button clicked times)*(extruded lenght per click in host)*(e_steps_per_mm we set [50]) ) Counts how many steps we used to extrude our marked distance on filament. When you divide it by the distance in mm, you’ll get number of needed steps per mm.
8. We should test the new e_steps_per_mm. So send “M92 E???” to your printer, where ??? is the number you’ve got from calculator.
9. Now we want to move the filament for exactly 1cm (we marked it before) so do it in your host SW GUI. For example in Repsnapper set distance to 10mm, and speed somewhere around 60. If it’s exactly 1cm, you’ve won! “G1 E10 F60″
10. This still doesn’t have to be exactly right. You’ll need to tweak it sometimes a bit. If the lines aren’t fused in solid layers, just add 5% to the e_steps_per_mm It’s not rocket science  

    Setting up Skeinforge:

    1. In dimension tab, set precise filament diameter. Filament packing density leave at 1.
    2. In speed tab. Set same Flow rate as feedrate.

     

    There are few alternative ways how to do it. Best alternative is calibration script written by Nathan shipped with Kliment’s amazing Pronsole. Or araspitfire’s posthere.

    Note 1:

    You don’t have to change the e_steps_per_mm during the experiment at all, you just need to be sure what number you use now, and at the end use your number instead of 50.

    Note2: You can use multiple distances during the 3cm lenght. What you really need to know after this is the overall length you extruded. But interactive calculator wont allow you to do this, you’ll need to use your brain and  this formula ( (extruded distance in mm)*(e_steps_per_mm we set [50]) ) / (marked length on filament in mm we have [30])

Less is sometimes more. I want to publish really simple hack how to make part of object stronger when you use Skeinforge for slicing. (It’ll probably work even in other slicers). This is part of Skeinforge how-to I’m preparing to publish here.

Why less is sometimes more? When you have void or hole in your object, Skeinforge has to do shells around it. And even will bind some of infill to it. If the hole has perimeter smaller then lets say 0.5mm radius, the overall density in this part of object will increase, the hole wont appear on object anyway and you can save some plastic in compare to increasing overall infill density!!

Picture is more then 1000 words

Will produce this

More holes you add, more stronger the part of object will be. Of course till you hit state where the holes are too close to each other. Experimentation is needed. In my case, it reinforced the weak parts of X-ENDs in my felt branch of Prusa Mendel.

What I will try next, is small slots in the part, where by rotation of them you could control in which direction object will be stronger. To get an idea, I did a small mockup pic. Again the slots will be small enough to be unprintable, but big enough to force Skeinforge to making stronger infill

I managed to get some thermal images of printing RepRap, thanks to David Kuboš.

Here you can see my Prusa Mendel printing Buddha. Notice the last line is really hot. But notice the heat coming from inside object? Its from heatbed.

Another nice picture. Line seems almost same hot as the insulated heater. Also look that object is same hot as heatbed.

Pretty cool overlay. Temperature trashold was set to 50°C I would say.

1/4 of heatbed is visible here, you can see that in temperature in corner is about 10°C lower then in center Temps are not precise, since the heatbed doesn’t have right coating for precise measurements, but the relative difference should remain same.

Printing with real image overlay.

Heat coming from bottom. Printer is doing perimeter in the red pointer it has 133°, while the rest of object is 75°C.