Tri Hotend

[Generic Default]

I used dead time at the beginning of this year when repetier host updated to it. I switched back to regular PID because I got a smoothieboard. Dead time is SO much better than regular PID control, but both will work. Dead time centers on the target temperature within one or two oscillations; with PID, you have to overshoot then slowly come back down.

Smoothieware needs to integrate dead time for temperature control.

[bot]

I think you don't have anything patentable here. The application of tungsten disulphide as a coating? I just don't buy it. Nor do I see the need for you to pursue a patent. I'm personally very off-put from supporting your good ideas when right off the bat, you're seeking to patent them...

So much for an open-source community.

[Generic Default]

The patent isn't claiming the concept of applying a layer of anything to a any surface, it's specific to certain parts of a certain type of hotend and limited to several materials that I experimented with over the summer. I excluded a bunch intentionally. I spent a ton of time checking for prior art to make sure that nobody else had tried the same thing. In the end, I found that nobody had done it within the ~30 year time period that 3d printing has been around, and I took that as an indication that it is novel in the field.

I also thought a lot about what people, especially reprappers and the open source community, would think about it. The general attitude toward patents in open source is negative, but I changed my mind a little bit from everyone I know telling me to get a patent as well as the thread I started a few months ago on these forums. Before June I was opposed to getting a patent entirely; recent lawsuits and product cloning within the additive manufacturing field is ultimately what convinced me to go for a patent. I gotta eat!

I was given IP advice specifically not to post the entire patent word for word on the internet, but if I did, you would be less offended by it. My intent is to make improvements in 3d printing technology that everyone befits from, and I didn't want to take the risk of investing so much of my time and money into development only to get screwed over by clone products, trolls, or huge corporations who would gladly take my work and monopolize it for their own exclusive use. Business heavily favors establishment and I'm poor and new to this stuff. Patents are unenforceable when it comes to DIYers. It is only manufacturers (big money) who have to check for infringement before mass producing something.

One last point about open source. When I made that first post about WS2 a few days ago, I made it clear -exactly what I was doing. If this were truly closed source, I would have come at you guys with a trademarked name for the coating and no explanation of what it actually is.

[redoverred]

I would say a patent is prudent, as long as you release the specs freely for private, non-commercial use AND license the technology at fair rates to others who would like to commercialize it. Otherwise you're just as bad as the large corporations which slowed down the 3D printing technology we have now by 30 years. Yes, I am saying that the patent on FDM printing fucked over the community we have today since we could have had that DECADES ago. Don't be like that. Also, DO NOT become a patent troll. If you're patenting this, you should produce and sell a product.

[bot]

I guess we are in a catch 22. As long as you are pursuing a patent, I won't even give you an ear to listen.

The catch 22 being that I don't believe you have something patentable/worth patenting and you won't tell me what it is. I'm not trying to imply that me listening to you matters at all. I wish you the best of luck going forward.

[Mac The Knife]

I think you don't have anything patentable here. The application of tungsten disulphide as a coating? I just don't buy it. Nor do I see the need for you to pursue a patent. I'm personally very off-put from supporting your good ideas when right off the bat, you're seeking to patent them...

So much for an open-source community.

So you think an individual shouldn't be able to protect his intellectual property? Just because it's patented, doesn't mean he can't share it, nor does it stop anyone from duplicating it for themselves. A patent does offer some protection from others PROFITING from his work. Wasn't it Stratasys that created an uproar when they started patenting a bunch of opensource stuff developed by others?

[barry99705]

I think you don't have anything patentable here. The application of tungsten disulphide as a coating? I just don't buy it. Nor do I see the need for you to pursue a patent. I'm personally very off-put from supporting your good ideas when right off the bat, you're seeking to patent them...

So much for an open-source community.

So you think an individual shouldn't be able to protect his intellectual property? Just because it's patented, doesn't mean he can't share it, nor does it stop anyone from duplicating it for themselves. A patent does offer some protection from others PROFITING from his work. Wasn't it Stratasys that created an uproar when they started patenting a bunch of opensource stuff developed by others?

Bingo!

A very wise man told me once,
"All knowledge is free. Sometimes you have to pay for the delivery."

[geneb]

Any chance of you producing a purpose-built single nozzle version of this hotend?

g.

[bdjohns1]

I think you don't have anything patentable here. The application of tungsten disulphide as a coating? I just don't buy it. Nor do I see the need for you to pursue a patent. I'm personally very off-put from supporting your good ideas when right off the bat, you're seeking to patent them...

So much for an open-source community.

Stop being a knee-jerk zealot. Open source and patents can coexist. Look at http://www.google.com/patents/opnpledge/pledge/ for an example. I think Tesla is doing the same thing with a bunch of their patents.

It's actually a good idea for open hardware people to do this kind of stuff with a commercial company like Stratasys out there, because Stratasys can't use an open-patent product and assert their own patent claims. For example, if GD's patent is granted and Stratasys wanted to use it in their printers, it would preclude them from bringing any kind of patent action against GD if he were to market a printer with a heated enclosure (which at least in the Cartesian printer market is something they have a patent on).

If enough good ideas in the 3D printing world are protected under open non-assertion agreements, it puts the open folks at a competitive advantage, because patent-asserting companies like Stratasys can't incorporate those offerings unless they agree to not assert their patent claims.

In the pedantic Stallman-esque sense of "Free Software/Hardware", it's not perfect because getting a patent isn't cheap. I'm a realist, so I have no problem telling Stallman acolytes to get bent when they gripe about that.

[bdjohns1]

Smoothieware needs to integrate dead time for temperature control.

They also need to include integral wind-up prevention, which would prevent overshoot at startup. Right now the integral error term can drive you to max_pwm.

[McSlappy]

Not sure where the hate on patents came from. Maybe from companies like Stratasys doing rotten things with them.. Or patent trolls... or Apple v Samsung... In all these cases there are questionable tactics surrounding the ownership of patents. However none of these companies HAD to be dicks about the patents, they just chose to for the pursuit of money...

Let's just say that Generic decides NOT to patent anything (assuming he has anything that can be patented...) and along comes Stratasys (surprise surprise) and says "Hey there's a nice idea, lets implement this into our next product! Oh and lets patent it while we're at it!" Where would that leave the open source community and Generic? Both will lose.
If Generic can patent what he's working on, he can then choose to treat it as open source by not prosecuting anyone who uses it, while at the same time preventing anyone else from taking and locking it down.

It all depends on whether you trust Generic to stick to his word on this.

I don't know Generic, but if I were to wager, I'd bet on it remaining more 'open source' having a patent on it than left to a larger company to do the same. Just look at the recent Stratasys patent trolling (stealing) to see what happens to great open source ideas when a company with a massive budget and team of lawyers gets a hold of it.

Sometimes you need to play the game that the big boys are playing in order to keep the big boys honest.

I give tons of props to Generic for going this route. It shows he's serious about his development and serious about the future. It would be foolish to put a tempting treat on the table, walk away and be mad when the dog steals it. Corporations are psychopaths and they will not respect the open community, we should all be very aware of this.

[astroboy907]

Makes you wonder - we have open source, maybe it's time to start the ball rolling on open patents? Like a group dedicated to the preservation of patents for the "little guy" to make sure that Stratysys doesn't happen.

[bot]

The cost of actually receiving a patent is probably more than it's worth to attempt. I can't foresee how there could really be anything that could be enforced as a patent here. The application might help establish prior art, for when/if startsys decides to "rip off" this idea. But, to be honest, exactly who might be ripping off whom here? This design doesn't actually look that original. Filament paths, nozzles, heat sinks, pretty standard stuff for hot ends.

I don't know the specifics of the cases of patents based off of open-source work, and these cases are obviously unfortunate, but that shouldn't deter people from going the fully open-source route. Patents are expensive and they ultimately represent an obsolete idea of intellectual property. We all build off of each-other's ideas. This is very obvious in the design of this hot end. Our contributions to the open-source community, and our ability to capitalize on them before we release the open source files should be our only reward. Closing off elements of your design to others wishing to build from it, in any way, is absolutely antithetical to the idea of open-source.

A patent, if one were to be given in this case, would likely do little to protect the concept anyway. Unless the one holding the patent has better lawyers than Stratasys, that is. They will just describe it a different way and patent it anyway. Having the technology/idea as pervasive concept in the open source community would be a much greater defense to a Stratsys theft attempt than one designer with one measly patent and no lawyer on retainer, vs. Stratasys' team of lawyers on retainer, eager to pursue an intellectual property battle.

And to be clear, this is not meant as an attack on anything but patents. Patents are evil, and they bankrupt people while convincing them they are protected. They end up being ripped off anyway and are already bankrupt... so what left? The lawyers only win, when patents come into play. The key is to simply execute the idea better than anyone could before they could... and allow others to develop the idea. You then benefit from their developments too, and can incorporate the developments into your commercial product.

[Generic Default]

They also need to include integral wind-up prevention, which would prevent overshoot at startup. Right now the integral error term can drive you to max_pwm.

AHHH that's the term I've always been trying to figure out! The integral accumulates the temperature difference even when it is passing the target, and even when it is inactive. I never knew there was a special word for it though. The code I used to write for helicopter gimbals had a bunch of logic to deal with similar problems.


This thread has taken a path towards patent discussion. I wish more companies would discuss the technical details of their products with transparency, but everyone should understand that there is a lot of risk involved with it. My goal is not, has never been, and will never be to stockpile patents for legal extortion purposes (ie. no trollin'). I want to stay open source and prosper. The patent is just here to decrease my chance of being a startup business failure; if I ever get a demand for licensing, I'll keep any royalties dirt cheap and avoid exclusive contracts.

The benefits from tungsten disulphide (and several of the other improvements) are obvious for 3d printers, and I want everyone to have access to them. It has already taken me longer than I'd like to bring this to market; if I had access to the tools and money earlier this year, progress would have been much faster. So once I start shipping orders (should start late next week) I expect progress to accelerate. I apologize for all of the delays up until now, thanks for being patient everyone!

EDIT; 3 people have posted since I started writing this. Hot topic I guess?


On the performance of the hotend, I've concluded that PLA can still be jammed if the heatsink fan is turned off and the PLA is left to cook in the thermal isolator at excessive temperatures. I let it sit in the hotend at 280 degrees for around 6 minutes, and I couldn't extrude it with the cold end after that. When I let it cool down to room temperature, I was able to pull the filament out of the hotend. It releases once it gets below its rubbery temperature (~55 degrees C). So I guess that the moral of the story is to not burn plastic in a tight space. It didn't leave any residue on the interior of the thermal isolator.

The external nozzle coating had an unusual effect. The ABS and PLA are very hesitant to stick to the nozzle, but bringing them above their typical printing temperature can allow them to stick on the outside. Translucent ABS discolored to a light brown after a few hours at 250-280 degrees, but it never turned into crusty black chunks like it did on an uncoated nozzle. The coated nozzle seems to stay much cleaner than the uncoated one, but not spotless. Keep in mind these tests are not under normal printing conditions. I'm pushing some limits here.

[bdjohns1]

They also need to include integral wind-up prevention, which would prevent overshoot at startup. Right now the integral error term can drive you to max_pwm.

AHHH that's the term I've always been trying to figure out! The integral accumulates the temperature difference even when it is passing the target, and even when it is inactive. I never knew there was a special word for it though. The code I used to write for helicopter gimbals had a bunch of logic to deal with similar problems.

The external nozzle coating had an unusual effect. The ABS and PLA are very hesitant to stick to the nozzle, but bringing them above their typical printing temperature can allow them to stick on the outside. Translucent ABS discolored to a light brown after a few hours at 250-280 degrees, but it never turned into crusty black chunks like it did on an uncoated nozzle. The coated nozzle seems to stay much cleaner than the uncoated one, but not spotless. Keep in mind these tests are not under normal printing conditions. I'm pushing some limits here.

The joys of a degree in chemical engineering - I had to take a course on controls theory. I forgot all of the ugly math to derive transfer functions and crap, but I still do enough PID work in my day job (process engineer/manager at a large food plant) that the practical parts of PID are still front of mind. I'm usually playing in ladder logic on PLCs.

Looking at the code in the Temperaturecontrol module in Smoothie, it looks like it would be easy enough to add the code in - just need an if statement to force the integral term to zero outside the desired control band. I just need to download the toolchain so I can build my own firmware, I guess. I did put in a request in the Smoothie google group, but it might be faster to do it myself.

Back on topic for your hotend, how wear-resistant is the coating? If I do get black crud on it and hit it with a mild brass brush, will that wear away the coating? What about if we ever muck with something exotic like glass-filled filament?

[bdjohns1]

Looking at the code in the Temperaturecontrol module in Smoothie, it looks like it would be easy enough to add the code in - just need an if statement to force the integral term to zero outside the desired control band. I just need to download the toolchain so I can build my own firmware, I guess. I did put in a request in the Smoothie google group, but it might be faster to do it myself.

Turns out it was dirt simple, even for a hack like me who hasn't touched C++ since my undergrad CS requirement. I just tried it out - on my first go (setting the windup limit to 15C from the setpoint), I reduced my overshoot from >13C to <6C.

https://github.com/Smoothieware/Smoothieware/pull/540

One new parameter for your config file (temperature_control.hotend.i_windup). I might try narrowing the windup band further, then see if I can get my temperature rock solid with PID parameter tweaks.

edit: nevermind - they just properly documented i_max instead, which is the other way to prevent integral windup.

[redoverred]

All I want to know is when can I buy one! at only $50-odd bucks over an E3D for a triple extrusion that is lightweight and easy to use, I am sold. All I'd need to do is figure out where to mount a couple more steppers, spools, and how to run more wires now that the printer is all put together already!

On that note, how would I run 3 extruders on the stock board? Is it even possible? I could probably run 3 steppers for the cold end (put one in the 2nd z axis driver) but there are only two extruder power and thermistor connections, IIRC.

[bdjohns1]

You would need an extension board to do this. Fortunately, since the Kraken predates the Tri, one of our fellow users already figured out how to get 4 extruders, even:

http://forum.seemecnc.com/viewtopic.php?f=55&t=4522

Over on the Smoothieboard, you just need to pick three available GPIO pins per stepper. If you're using a graphic LCD, then the following are already used: 1.30, 1.31, 2.11, 3.25, 3.26. I'd probably use 1.23 and 1.22 (next to one of the LCD hookups) and then one of the unused endstop pins. You could probably also commandeer unused i2c pins as another option, since i2c on the Smoothie isn't terribly usable.

[redoverred]

You would need an extension board to do this. Fortunately, since the Kraken predates the Tri, one of our fellow users already figured out how to get 4 extruders, even:

http://forum.seemecnc.com/viewtopic.php?f=55&t=4522

Over on the Smoothieboard, you just need to pick three available GPIO pins per stepper. If you're using a graphic LCD, then the following are already used: 1.30, 1.31, 2.11, 3.25, 3.26. I'd probably use 1.23 and 1.22 (next to one of the LCD hookups) and then one of the unused endstop pins. You could probably also commandeer unused i2c pins as another option, since i2c on the Smoothie isn't terribly usable.

Cool, Thanks!

[Generic Default]

Back on topic for your hotend, how wear-resistant is the coating? If I do get black crud on it and hit it with a mild brass brush, will that wear away the coating? What about if we ever muck with something exotic like glass-filled filament?

WS2 has been used on injection molds for quite a while now, so while my own Tri hotend experiments don't have the benefits of long study times, I know that WS2 lasts for years under the right conditions. It doesn't come off easily, you usually have to scrape off the material under it to remove the stuff. So a wire brush would probably take a lot of it off the outside.
Glass fiber resins will definitely wear it out of the brass nozzles, and they will expand the orifice hole too (over time). Unfortunately I can't buy glass filled filaments right now so I'm unable to test this for at least a few more weeks.

I have plans for producing nozzles that can withstand this for extended periods of time. My experiments showed WS2 to be the best solution for the current filaments being used in 3d printers; I messed around with multiple other coatings and some of them have excellent abrasion resistance. The strongest, stiffest plastics are usually glass filled, so we may have to start viewing nozzles as consumables when it comes to high end plastics. The price of these "super nozzles" that last a long time will be at most a few dollars more than the current brass nozzles. Changing nozzles on the Tri hotend takes ~1-5 minutes.


But for PLA, ABS, nylon, PET, ect. the current nozzles should last a long time. I really have no idea how they will interact with metal powder filaments, but I'm guessing they won't last indefinitely for those either.

[0110-m-p]

So what are you using to control 3 extruders? Can the RAMBo's extra Z-motor be used for a 3rd extruder?

[artexmg]

So what are you using to control 3 extruders? Can the RAMBo's extra Z-motor be used for a 3rd extruder?

The answer just three posts back

http://forum.seemecnc.com/viewtopic.php ... 110#p52458

[0110-m-p]

So what are you using to control 3 extruders? Can the RAMBo's extra Z-motor be used for a 3rd extruder?

The answer just three posts back

http://forum.seemecnc.com/viewtopic.php ... 110#p52458

Thanks, I have been keeping track of that thread as well because I was originally considering a Kraken. Just wasn't sure if that was necessary for 3 extruders or if there was some work around other than adding an auxiliary board. I remember reading something about using various other pins on the RAMBo to make it work.

I've got my printer calibrated and running great with RAMBo and Repetier 0.91 firmware, so I'm just hoping to change as little as possible.

EDIT: Although...I have been thinking about building a second machine and the Azteeg X3 Pro's schematic makes it look pretty badass.....G4 S(time for brain to process)....

[artexmg]

So what are you using to control 3 extruders? Can the RAMBo's extra Z-motor be used for a 3rd extruder?

The answer just three posts back

http://forum.seemecnc.com/viewtopic.php ... 110#p52458

Thanks, I have been keeping track of that thread as well because I was originally considering a Kraken. Just wasn't sure if that was necessary for 3 extruders or if there was some work around other than adding an auxiliary board. I remember reading something about using various other pins on the RAMBo to make it work.

I've got my printer calibrated and running great with RAMBo and Repetier 0.91 firmware, so I'm just hoping to change as little as possible.

EDIT: Although...I have been thinking about building a second machine and the Azteeg X3 Pro's schematic makes it look pretty badass.....G4 S(time for brain to process)....

Well, not sure if I understood correctly, but the solution given in the thread is for a RAMBO, just adding an extra controller for the extra extruder

I am in the same very dilemma of using a Smoothie board (or Azteeg, same thing) or get stick to RAMBO.

Probably will have to have one Rostock with Rambo and other with Smoothie board

Cheers!

[teoman]

Would some one minde sharing gcode of 3riple or double extrusion?