How to fix everything

[thoughtfix]

RetireeJay and I are going over my stepper motor issue and he reminded me of a method I use to train new Linux systems administrators. Since we're all working on Printrbots here, we are all enginners: electronics, mechanical, and material science. While we all have different levels of skill on each of those (I'm pretty good with electronics, but weak on material science) we can use the same troubleshooting steps involved in any kind of engineering:

Reproduce the problem. Try to make it happen again under the same circumstances and observe it closely. Try to collect as much information as possible just by watching.

Identify the problem. With the information in the last step, try to find a specific, short way of describing the problem. That will help you with forum searches, Google searches, or just as a way to continue through this list. Is filament not extruding? Are your models warping? Is the software crashing? Is your filament not sticking? Is your print head not moving?

Isolate the problem. This is where you get to be an engineer. Step through the process in your mind and address each part one at a time. My first major problem was that my plastic was not extruding so I had to test the individual components that make the plastic move:

• The filament is pushed by a stepper motor. Is the stepper motor spinning? Test it all by itself with no filament.
• The stepper turns a hobbed bolt which causes friction on the filament. Is the hobbed bolt clean of plastic debris and does it spin freely?
• The filament is held in place with tension on the idler. Does everything line up there? Does filament move when you turn it back and forth with your hand?
• The filament must melt to the right temperature before being extruded. Is your software set to the right temperature? Can you measure it with an IR thermometer?
• The melted filament must go through the nozzle. Is any filament at all going through? If you remove the nozzle tip and tell it to extrude, does melted plastic come out the bottom?

Fix the problem. That last step was the solution for me: My nozzle had some burnt debris in it because I didn't protect it well enough when building the printer. Because melted filament didn't extrude, it caused the hobbed bolt to "strip" filament and prevent it from pushing in more. I removed the nozzle, cleaned it out, reassembled it, and had a working printer. By testing every component individually, I was able to find the problem and fix it.

How to be a BETTER engineer: Learn about all the things your 3D printer does. There are no mysterious elements to Printrbots: every part is well understood by the community. Learn all about the combination of events that make a 3D print: movement, heat, bed adhesion, filament, and design. Learn about the combination all those elements: Bed adhesion is a combination of heat, bed texture/smoothness, model footprint, and sometimes add-ons like blue painter's tape or hairspray.

Final thought: You have US! With all the Printrbots out there, chances are your problem is not totally new and unique. The community has probably seen it before, figured it out, and fixed it. If you can't get as far as fixing the problem yourself, just collect all the information you have and search the forum. If you have no luck there, write a post being as specific and detailed as possible and post all the steps you took to troubleshoot the issue. Once you fix it, come back to the forum and tell us what worked: you'll be helping the next person who has the same problem.

[thoughtfix]

Thanks for the rep, guys.

[drawcut]

Excellent post. A great troubleshooting write up that could be applied to most any area.

[thoughtfix]

Here's the example with my X axis:

REPRODUCE: That's easy. Tell the motor to move and it doesn't move.
IDENTIFY: Also easy: It stutters a bit and hums. Tried running a print job and tried manual movement with two different software suites. No luck.
ISOLATE The X axis movement has several parts: The X axis assembly, Signals from the PC to the Printrboard over USB, signals from the Printrboard to the Motor over the four-wire connection, the motor's coils, and the mechanical components of the axis.

• Move the X axis by hand to check for obstructions. It moved smoothly.
• Check that other motors are moving (to test software/USB) - They were.
• Physically inspect all wiring and check the four wire connection and if it's plugged in properly. - It was fine.
• Swap the motor's position with another motor (in this case, the extruder motor) and see if commands to move X move the other motor. They did.
• See if commands to move the other motor moved the troublesome motor It didn't.
• Remove the pulley and belt from the motor to test for friction. There was none.
• Optional if you have the tools: Check for continuity and resistance over the coils with a multimeter and check for shorts. Continuity was fine at 3.1 ohms and there were no shorts.

FIX: That leaves one possibility: Something INSIDE the motor is broken. Replace the motor.

[RetireeJay]

Another way to think of this process is taught as the Scientific Method:

Observe the phenomenon, and find conditions under which it can be reliably reproduced

Develop one or several Hypotheses - theories that might explain your observations

Design a Test (or experiment) which will test the theory or theories
- NOTE 1: If possible, make ONLY ONE CHANGE at a time. When you have multiple interacting influences, this is the only way to isolate them and narrow down the possibilities
- NOTE 2: Sometimes you can't design a test which will unambiguously verify one theory, but the test should at least allow you to narrow down the number of theories.

Run the Test and observe the results carefully
- NOTE 1: It's very important to keep an open mind. If test results don't support your theory, or the results are ambiguous, then keep on developing theories.
- NOTE 2: Even if your test is flawed, such as you failed to set up some condition the way you intended, never ignore the observation step. I can personally vouch for the fact that significant inventions and even patents have resulted from "flawed" tests. That's because the test revealed some aspect of the problem that was totally unexpected and not anticipated in the design of the test.

If the observations are not explained, develop more theories

NOTE: Most of the time when you have equipment which has been operating correctly for a while and suddenly starts to misbehave, there is only a single root cause for the misbehavior. Resist the urge to make wholesale changes to the equipment all at once because that may mask or obscure the root cause of the problem, or introduce new and confounding problems.

[allanonmage]

There are no mysterious elements to Printrbots:

Oh, there is MUCH mysterious-ness to 3D printing.

[thoughtfix]

Heh - that's true. There ARE mysteries about 3D printing but most of them are not in "how they work" but rather "what can we do with it?" If we're just using a PrintrBot to make PLA prints of little castles, we have all the information we need to get that done. If we want to do something new like modify a PrintrBot to sculpt icing on a cake, we have some mysteries to solve.

Damn ... now I want cake.

[Bats1939]

But not all of us are engineers. I'm an artist. I make video games and toys. I think assuming that we're all engineers assumes that we all learn the same way or have a similar skill set... and that creates misunderstandings and confusion.

The problem with trouble shooting is that often times I'm not sure what to call parts or how to explain what's going on. Sometimes I don't even know the jargon to look up under the search. I'd love to learn how all the parts of my printer work- but that would involve me learning an entire new skill set and I'm not really sure I have time for that. I'm very focused on getting better at art and that takes up most of my time. I'm not expecting someone to just come over and fix my machine, although that would be nice, but I would hope that people realize I'm not being lazy, I'm just coming at the problem from an entirely different discipline.

Generally I have to see someone using the same machine and program fix the problem in order to really understand how to fix something myself. I've never been very good at the inner workings of machines themselves- it's art and modeling that I pick up and understand quickly. So things have to be dumbed down pretty far for me to understand it and I don't always understand written instruction. But I wouldn't expect you to understand how to bake normal maps or construct a mold with severe undercuts, because that's not your job that's mine! Everyone has a different skill set, so while you might understand how the printer works it, it's still a mystery to me. I think the hardest thing for me is the programming. RetireeJay has been really patient with me and my frustration with my machine and I really, really appreciate it. I wish I could contribute more, but it seems like most people are more interested in tinkering with the machine rather than the object to print. Which I can totally understand! For me, I understand the creation process better than the machine.

I think your breakdown will help to bridge the gap between, at least it gives me a better idea of how to explain the problem I have.

[allanonmage]

But not all of us are engineers. I'm an artist. I make video games and toys. I think assuming that we're all engineers assumes that we all learn the same way or have a similar skill set... and that creates misunderstandings and confusion.

And for those of us that are engineers, it's maddening because we have to deal with software. I've developed a peculiar aversion to software on account that programmers never know how it works, even if they just wrote it. This meme sums it up perfectly:


And in the field of mechatronics, robotics, and composite fields, hardware will only get you so far. You have to rely on software, and you can't complain too loudly because it's free software, and no one knows how to work on the software. :-/

Try asking some programmers what "Python" is sometime. They go into some form of feedback loop for a while talking about scripting and languages. I'm still trying to figure out what they're talking about.

[Mooselake]

Think of software as a machine with thousands to millions of moving parts. It's to complicated to remember exactly what each one does, aand some are always broken or put in upside down. Good testing (and documentation) isn't much fun, and is often rushed or neglected.

Python (or whatever) is a fad language of the moment. It's what the cool kids use, and because of that it must be the best there ever was. When it loses it's coolness you have to move on. That's why you can't get a good explanation, along with if it's made too hard for normal people to understand then only the cool people can do it <ducks, runs for cover>

Pretty soon Watson and brain simulating chips will take over programming. Hopefully it'll take longer to replace artists.

Kirk

[Mooselake]

Coincidentally, this was just on slashdot. Microsoft may have found the solution to buggy code, although I couldn't find the reference to the large electrical shock correction for behavior modification. Connect electrodes to the programmer's head and measure their brainwaves.

Glad I'm retired.

Kirk

[JavierVanegas]

Well imagine my case, where I am almost at the position of the Artist (instead I am a future Politologist/Political Scientist). I do have a spirit for tinkering and electronics, but don't have the time to study it deeper either. Also my mother language is Spanish, and I live in South America. Most of the information on the web is 99% English (for western users I suppose), and at least 30% / 50% technical depending on the page you refer to. I do know how to disassemble my pc and put it back together, but from that to coming in touch with RepRap and Arduino for the first time, knowing what a G-code is (which I still don't think I fully understand). My Printrbot plus v2.1 is an investment which I need to make the most of, and as soon as possible (as things in my country aren't favorable to any kind of entrepreneurship), and if I take too long, "next semester is coming", and well, that explains itself. I love the whole concept of 3D printing, rapid prototyping or as you want to call it, but it's a whole new world, it needs quite an investment of time. One of the things I read from a book by Christopher D. Winnan is that "you need someone to share the frustration with", as in someone who either helps you or at least listens to you. I would add that you find a companion/partner that is in he same wave as you and helps you out in any way he can, or a group of guys. If all can have their own printers, even better. But for me it's hard, I am trying to do my best to move out of my country, but just like all countries that are being driven into communism, that is harder everyday.