Test Suite to begin with a RepRap.

I searched half-heartedly for a way to test my RAMPS firmware step-by-step and came up empty handed.
I was wondering if anyone had a good method for doing this already? A bit of test firmware that would start at the basics. Is the stepper moving, if not do this, is it moving now? If not do this. OK now that it's moving which direction? Trigger the x-opto, I see that it's high, move the x-carriage away from the opto.. OK I see that it's low, it works. OK I'm heating the extruder.... Is it hot to the touch? OK I'm moving the extruder stepper. etc etc.

Maybe print out the finished configuration to copy/paste into a firmware?? Ultimately, doing this confirms that the hardware works and moves you to the next link, the firmware you choose. Then finally, I assume the software you're using to communicate with the firmware.

Let me explain a bit if you're confused. I just started building my Prusa and I'm working with the RAMPS 1.2 hardware. Since there are so many arduino based electronics I've noticed the firmware written for them is kind of generisized to include pinouts, boards, and capabilities of a large number of platforms. This creates a few issues though, in order to troubleshoot basic problems you're left with a really long chain of events to work through. That and every firmware has to write pinouts every time they release a new arduino based thing re-doing that work.

In my case, I'm at the starter level. Before yesterday I wasn't sure if the steppers I bought on e-bay were good so I had a lot of work to do to narrow it down. I haven't even started on the extruder or optos.

I set about starting by stripping pins.h and configuration.h from the sprinter firmware and now I'm on the crest writing a full hardware test suite. A step-by-step walk through to double check pinouts, rotation, and activity coming from the printer before GCode is even involved. I was aware of spiral rectangle and similar firmware but those barely scratch the surface. You have to set pinouts and so on and because they don't mimic a firmware they force you to dig into the firmware code to troubleshoot it separately.

So again, is there already something like this or should I make something like it?

Edited 1 time(s). Last edit at 09/02/2011 11:57AM by attrezzopox.

Seems a good idea. Check the wiki for existing stuff (not so much), thingiverse for calibration & test objects.
What would be the final product: A compilation of test methods, a checklist?

I assume for the first portion the end product would be a program that helps to verify all of your hardware actually works before you try to send gcode.

A secondary goal might be to produce a usable pin-out/config file that a firmware package could use as starter values. The benefit is at least the basic parts of the printer would be known good. That and to break off a portion of the firmware that every firmware version has to duplicate.

As far as calibration might be concerned as an advanced goal. It should be relatively easy to get basic measurements from the reprap to calculate (precisely, regardless of the motors, gearing, and so on) the exact number of steps per mm. More advanced functions might be calibrating to correct x/y skew. (Z is a different beast but that could be possible too. Any ideas?).

If all of that is done at the firmware level, gcode should be more reliable as a whole. Gcode wouldn't have to be tweaked to each individual printer. I know that provided all the measurements during assembly were accurate that they SHOULD have a device that is accurate. But even if the user didn't use jigs to make their reprap or did a slightly sloppy job assembling it software might be able to correct it without a major troubleshooting campaign. Heck, even if you use jigs, if you get squirrelly printed parts you can still impart skew. That's one think that can get WORSE with each generation, if Parent skews the x axis then Child inherits that from the part EVEN if they put it together correctly it could be made worse by the nature of the misprint. A software calibration has a chance at acceptably correcting that, even generations down the road.

attrezzopox Wrote:
-------------------------------------------------------
> I assume for the first portion the end product
> would be a program that helps to verify all of
> your hardware actually works before you try to
> send gcode.
>

do you think the current variability of boards & configs would allow for practical automatization of config testing?

I see more value in suggesting to people a well crafted, step by step method/routine to do that.
(= many uses in future tinkerings. If they do not get to do it by hand at least once, they won't learn)

then a set of small, individual utilities - stepper testing, thermistor testing, etc. (several already exist)


> A secondary goal might be to produce a usable
> pin-out/config file that a firmware package could
> use as starter values. The benefit is at least the
> basic parts of the printer would be known good.
> That and to break off a portion of the firmware
> that every firmware version has to duplicate.

hum...
a hardware abstraction layer... ?

should be unobtrusive enough to entice many (>1) firmware coders to adopt it.

and much of the value of current firmware is that it is still... quite simple.
ppl hesitate to get into hacking mode with more complex stuff.

>
> As far as calibration might be concerned as an
> advanced goal. It should be relatively easy to get
> basic measurements from the reprap to calculate
> (precisely, regardless of the motors, gearing, and
> so on) the exact number of steps per mm.

???

> More
> advanced functions might be calibrating to correct
> x/y skew. (Z is a different beast but that could
> be possible too. Any ideas?).

As long as you have a reference cube to compare it...

>
> If all of that is done at the firmware level,
> gcode should be more reliable as a whole. Gcode
> wouldn't have to be tweaked to each individual
> printer.

that is a general issue in whole CNC sphere.

Gcode hard-codes feedrates. And those can vary significantly if you have, say, an ultimaker and a darwin...

For the rest, you'll need a meta-description of the machines features

and the target keeps moving.

then there's peoples tastes applying upstream of gcode. How they want their printed thing sliced, how many shells etc.

> I know that provided all the measurements
> during assembly were accurate that they SHOULD
> have a device that is accurate.

current method, using a bubble level and a plumb line, provides for some "natural" variability.

and as I have found we do not even have widespread awareness / agreement on key distances between pieces (because currently there is no trusted set of reference 2D drawings) cf. discussion about Z smooth rod to Z motor axis in another thread.

so we have standards issues creeping on several levels.

>But even if the
> user didn't use jigs to make their reprap

Cutting a jig should be more strongly recommended. It is really useful.

> or did a
> slightly sloppy job assembling it software might
> be able to correct it without a major
> troubleshooting campaign.

Doesn't look to me like a sound practice. Adding a layer of increased complexity, one that needs maintenance and coordination between unrelated coders, to a layer of sloppyness and ignorance, leads you straight into micro$oft bob solutions (ever got to know that one?a fraction of it survived in the annoying paperclip-faced e-counsel in msfoffice & openoffice).
Yet this has been done many, many times in the commercial computer software realm.

Problems should remain atomic, be resolved near their origin, and if possible, in simple and meaningful ways.
(sorry to sound so negative, but in real life I have to deal with similar consequences: people who did never get methodic training nor basic notions in due time because "software should take care of that", and/or because "that is a matter for specialists", then are supposed to get creative).


> Heck, even if you use
> jigs, if you get squirrelly printed parts you can
> still impart skew. That's one think that can get
> WORSE with each generation, if Parent skews the x
> axis then Child inherits that from the part EVEN
> if they put it together correctly it could be made
> worse by the nature of the misprint.

there are many, many other concurrent sources for "misprint" and parts degeneration.
But there's a hefty injection of vitamins at each generation.

Then, studs aren't the only -nor the easiest- option for building a 3d printer.

>A software
> calibration has a chance at acceptably correcting
> that, even generations down the road.

I tend to trust more a sound set of triangulation and control points on the frame.
And well written explanations about how to do it.

Get the fundamentals right, and the rest will follow, the basic soft stays simple, and people focused.
(hopefully)

Yes, such a configuration tool whould be a great thing. Doing things step by step, so people can debug along the way.

1. Turn the power supply on.

2. Turn on a heater.

3. Move a Motor at all.

4. Read out a temperature.

5. Calibrate a motor's STEPS_PER_MM.

6. ...

A _very_ basic start is here: [github.com]

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

@Lanthan

> do you think the current variability of boards & configs would allow for practical automatization of config testing?

Yes I do. The utility I'm suggesting would be used on arduino based hardware, I assume other micro-controllers would have to deal with the issue on their own. And I wouldn't consider the process "automated" in the sense that no one is at the controls. I would consider it step-by-step troubleshooting. A software walkthrough (with some intelligence) getting a rep rapper started in first verifying everything works, then printing out a working example of pinouts, and then possibly making calibration measurements to help downstream firmware devs make corrections at the firmware level, I'm specifically thinking of skew.

> I see more value in suggesting to people a well crafted, step by step method/routine to do that.
> (= many uses in future tinkerings. If they do not get to do it by hand at least once, they won't learn)

I'm suggesting that as well. The "automated" portion of this is just that it's a program that steps you through the process and activates motors/heaters etc at the right time.


> and much of the value of current firmware is that it is still... quite simple.
> ppl hesitate to get into hacking mode with more complex stuff.

My point exactly. I don't see this as the end-all be all. Quite the opposite. When you're done using this program, you know all of your hardware works, you can choose your firmware with some intelligence and know where to start if you need to troubleshoot. Additionally, it takes a chunk of work out of firmware dev's hands by offering up a standard set of variables that accurately define pins and may even calculate basic variables to help calculate flow-rate etc. They can choose to use the output, or not. No one is (obviously) going to force anyone.

>And those can vary significantly if you have, say, an ultimaker and a darwin

Sorry I was specifically talking about prusa to prusa or mendel to mendel. As it is each printer is a little different even if it's a few mm difference just because of nature of assembly and human eyes and hands being involved


As far as jigs, measurement, and slopp are concerned. I think it's entirely reasonable to assume that someone might not have the ability to accurately cut a jig, or won't be able to accurately measure to the .01 of a millimeter. If you can correct that without forcing them to buy a good pair of calipers and spend a couple of hours measuring and re-measureing why not? The big theme now is that everyone is building parts that help them AVOID taking apart their printer. That's kind of a travesty in my opinion. What good is a tinkerer when he is inventing things to AVOID tinkering? It seems to come down to the fact that once you take apart a working reprap you give yourself a headache trying to get your print quality back to whatever you've spent months tweaking it to.

Aside from user error, it's even MORE reasonable to assume the printed parts could be misprints. I'm just starting out with my first reprap. I don't know anyone with the ability to print or cnc the reprapable parts so I ordered my set for $100 bucks on ebay. I'm not about to chuck 'em because when I get them I realize they cause my x-axis to skew 3mms. If I can fix it with software, bygod I will, and then I'll print a better set of parts.

Finally, set aside from those to arguments is what I'll call the arduino mentality. Even now, hardcore electronic engineers and tinkerers spit fire about the "bane" of the arduino and how inefficient it is. How horrible it is they cut so many corners so that the users don't even know how to burn a set of fuses. And so many people talk about trying to re-educate these lost souls on "REAL" microcontrollers.
But, the proof is in the pudding. Arduino is successful specifically for all of those reasons, and was designed to get people who aren't interested in burning fuses and learning assembly to use a microcontroller. I think reprap is ideally of the same camp. All reprapers shouldn't be mechanical engineers. We need a few that skip to just making random crap on thingiverse we may never find useful. Like a dremel centrifuge for field biologists, or a planting medium for aeroponics.

attrezzopox Wrote:
-------------------------------------------------------
> @Lanthan
>
> > do you think the current variability of boards &
> configs would allow for practical automatization
> of config testing?
>
> Yes I do. The utility I'm suggesting would be used
> on arduino based hardware, I assume other
> micro-controllers would have to deal with the
> issue on their own. And I wouldn't consider the
> process "automated" in the sense that no one is at
> the controls. I would consider it step-by-step
> troubleshooting. A software walkthrough (with some
> intelligence) getting a rep rapper started in
> first verifying everything works, then printing
> out a working example of pinouts, and then
> possibly making calibration measurements to help
> downstream firmware devs make corrections at the
> firmware level, I'm specifically thinking of skew.

As we all know, technical / practical culture seems to be waning in many western countries. The causes are multifactorial, but schools and teachers, and a certain "technophobic" attitude have much to do with it, automation may have also contributed its share.
To the point that major engineering schools have been alarmed, and launched initiatives trying to reverse the trend.

I see in the diffusion of repraps & hacklabs & similar (including Arduino, and the good tutorials of Adafruit, etc), the whole make movement, a superb occasion to distill back some systematics & good practices into (motivated) people.

Consequences depend on the tool. If you're not systematic enough with a mill or a lathe, even a mini lathe, it may easily chew a finger or two off you. Some tools call for respect: "DO learn good practices before using me".
The mendel isn't impressive enough in that way. There's little major damage you can do to yourself messing with a reprap. But it is a good opportunity to get some method onboard.

Yes, software may help. but many people will inevitably get to the shortcut: if they can just use the software and avoid reading the d... tutorial and avoid having to think about it, they will do. It is mostly self-defeating.

On the other hands, if they have to fetch different utilities form a bootstrapping "get started with your mendel" tutorial, this requires more active (and persistent) learning.
Yes, you could even automate the tutorial itself, for example with a processing front-end and firmata on the arduino side.

But I'd go the wiki, generic way, first and above all because mendel is not a fixed design, but a moving target. more on this later.

>
> > I see more value in suggesting to people a well
> crafted, step by step method/routine to do that.
> > (= many uses in future tinkerings. If they do
> not get to do it by hand at least once, they won't
> learn)
>
> I'm suggesting that as well. The "automated"
> portion of this is just that it's a program that
> steps you through the process and activates
> motors/heaters etc at the right time.
>
>
> > and much of the value of current firmware is
> that it is still... quite simple.
> > ppl hesitate to get into hacking mode with more
> complex stuff.
>
> My point exactly. I don't see this as the end-all
> be all. Quite the opposite. When you're done using
> this program, you know all of your hardware works,
> you can choose your firmware with some
> intelligence

another moving target.

> and know where to start if you need
> to troubleshoot. Additionally, it takes a chunk of
> work out of firmware dev's hands by offering up a
> standard set of variables that accurately define
> pins and may even calculate basic variables to
> help calculate flow-rate etc.

hum... some stuff is better layered. (have we defined our OSI layers equivalent?)

> They can choose to
> use the output, or not. No one is (obviously)
> going to force anyone.

then you have to create something so enticing that most developers will adopt it.

>
> >And those can vary significantly if you have,
> say, an ultimaker and a darwin
>
> Sorry I was specifically talking about prusa to
> prusa or mendel to mendel. As it is each printer
> is a little different even if it's a few mm
> difference just because of nature of assembly and
> human eyes and hands being involved
>
there are signs that a correctly adjusted mendel (nonobstant stiffness issues) works OK.

the emphasis is on "correctly adjusted". I found the mendel much fiddlier to adjust than a routerized/asercut design.
Yet you absolutely need to do this for the further stages. Error multiplies faster than rabbits.

> As far as jigs, measurement, and slopp are
> concerned. I think it's entirely reasonable to
> assume that someone might not have the ability to
> accurately cut a jig,

For the many people buying kits out there (many of them with lasercut parts), the minimum they should be asking for the money they pay is to be getting a couple of convenient lasercut jigs. It seems some kits at least include such.

>or won't be able to
> accurately measure to the .01 of a millimeter.

cheap, easy to use capacitive calipers with that accuracy are a reality. Anyone with a taste for tinkering should have one, or get one. With a reprap, this is an item for everyday's use (let alone for checking prints, calibration and the much variable filament diameter)

>If
> you can correct that without forcing them to buy a
> good pair of calipers

they should get calipers (USD 10 to 30 depending on the sources). They'll find many, many good uses for them. Although it won't make mechanic engineers out of 'em, it is part of the most basic tooling.

> and spend a couple of hours
> measuring and re-measureing why not? The big theme
> now is that everyone is building parts that help
> them AVOID taking apart their printer. That's kind
> of a travesty in my opinion.

I too got mildly pissed off with the fiddly frame adjustment. I cut myself a jig, but that possibility stems from the initial assessment of the whole project. And I have also designed parts for the mendel in such ways that I do not need to get again in frame adjustment mode. I believe those design changes do not result in any loss of key structural qualities, so I see no problem there.

The mendel frame is a compromise: you regenerate accuracy lost in plastic replication by fiddling with studs and jigs. that makes it sort of "self replicating" as in other designs you must use a meta- source of accuracy (laser cutter, cnc mill or router etc) as a master.

Anyway, there may be a problem with that frame design (actually there are several, but none of them overwhelming).

Should this design get too much in the way, I might switch to another frame. I like nophead's design in another thread. The point is: I do not see the mendel's current frame (however kewl its looks) as _the_ definitive answer for a diy 3D printer frame. Eventually, the community may move on to other frame designs at some future point.


> What good is a
> tinkerer when he is inventing things to AVOID
> tinkering?

This is a basic motor of innovation: you invent ways to get around the unsavory tasks.

> It seems to come down to the fact that
> once you take apart a working reprap you give
> yourself a headache trying to get your print
> quality back to whatever you've spent months
> tweaking it to.

I have yet to see many cases of that illness reported in the forums.

> Aside from user error, it's even MORE reasonable
> to assume the printed parts could be misprints.
> I'm just starting out with my first reprap. I
> don't know anyone with the ability to print or cnc
> the reprapable parts so I ordered my set for $100
> bucks on ebay. I'm not about to chuck 'em because
> when I get them I realize they cause my x-axis to
> skew 3mms.

I hope you complained loudly and reported the sloppy irresponsible seller (after discussing the matter and unsuccsessfully trying to get replacement parts of course). Good pictures, schemes, and a technically accurate description of what is wrong may help. Please go ahead!

> If I can fix it with software, bygod I

First get the sloppy seller fixed. Others may thank you.

> will, and then I'll print a better set of parts.
>
> Finally, set aside from those to arguments is what
> I'll call the arduino mentality. Even now,
> hardcore electronic engineers and tinkerers spit
> fire about the "bane" of the arduino and how
> inefficient it is. How horrible it is they cut so
> many corners so that the users don't even know how
> to burn a set of fuses. And so many people talk
> about trying to re-educate these lost souls on
> "REAL" microcontrollers.

Yeah I have heard parts of this very same discourse in friends teaching electronics.
Arduino is insanely great because as an ecosystem it has helped masses of people get bootstraped into the microcontrollers realm. Whish I had had this in the eighties!
It's not just the open source hardware & software, it's the tutorials, the forums, the connexions with processing & openframeworks & matlab...
At some point - especially if you are supposed to get into electronics as a professional- you'll need more than Arduino. But that is a wholly different purpose. The point is, there is an upgrade and learning path from arduino.

Also, inside the Arduino, you may start with the standard board, then get into more involved things like, for example, gen7 electronics.

But let's not mix mechanics and electronics. You can fix, swap, replace program electronics (thanks to the arduino thing) in easier ways than you can mess with structures.

> But, the proof is in the pudding. Arduino is
> successful specifically for all of those reasons,
> and was designed to get people who aren't
> interested in burning fuses and learning assembly
> to use a microcontroller. I think reprap is
> ideally of the same camp.

notice that Arduino does not have that much ready-for-purpose code as it has tutorials, howtos and specific-point examples. Solutions do not come pre-cooked, and even if you are to adapt some existing code, you have to dig deep enough in the libraries documentation. There's a gap between the Arduino and the NXT...

>All reprapers shouldn't
> be mechanical engineers.

no, but, as said in the beginning, as happens with the arduino, with convenient help & tutorials, you do not have to be a mech engineer to understand a very basic structure, or the workings of a stepper, or (hopefully) to adapt parts design to mechanic & accuracy constraints.

>We need a few that skip
> to just making random crap on thingiverse we may
> never find useful. Like a dremel centrifuge for
> field biologists, or a planting medium for
> aeroponics.

personally I have more issues with the gangsta crap or the self-promoting company logos (some would say it is the same vein). But I wouldn't want to discourage you. Maybe you're on the way to lay the foundations of the micro$oft of 3D printing (no offense intended) as anybody who intends to replace accuracy and materials and learning with software will get quite some attention from potential investors.

Let's be honest that wall of text wasn't meant to do anything but discourage me.

Either way you missed the arduino reference. The comparison lies in the fact that the arduino was created as a short cut way to teach people about microcontrollers. Just like an EE could say (and many DO say) that the arduino skips important steps and "prevents" newbies from learning, you could say that writing a simple program that turns a few steppers prevents others from learning about 3d printing.

I think that's extreme. I think it's unrealistic to hold back amateurs who might enjoy their own 3d printer because of a righteous belief that they follow all of the steps You had to take because it will make them better people in your eyes. I have this crazy notion that some of those people might not care. They may see those steps as I do, an unnecessary roadblock to just printing.

Following that chain of thought to the extreme has us mining for ore to smelt the rods that make up the printer, or synthesizing the plastic that reprappers extrude. As noble as it is to teach someone a new trick, it's only useful as a lesson when they care enough to remember it. And as easy as it seems to discard them from a community for their apathy. That's an easy, shortsighted way of discarding any enrichment they might have provided later. The truth is that sometimes catching a man a fish can give him just the boost he needed to figure out how to hunt a deer. More often than not, knowledge begets knowledge without requiring someone to follow all of the steps of their ancestors. It's a concept generally referred to as progress. A general theme of invention.

How any of this relates to "micro$oft" (I do take offence btw, you pretty clearly implied that you're comparing me to a company you believe to be greedy, blundering fools with an intent to keep people stupid. Just the way you misspelled the name practically screams that.) is far beyond me. I must be missing something. I really can't stretch my mind enough to believe there's any money value in this. Open source is a pretty good educational start. Seems in the spirit of reprap anyway. Or that I'd be exploiting people or encouraging them to remain ignorant!? really? That's aside from the fact that I don't want any money and if anything, I see this as educational to a beginner.

But you can't please everyone. I think I understand what your trying to say, and I still think your wrong, if not being a bit of an elitist. Either way, you can keep then next diatribe to yourself unless you have something constructive to add.

attrezzopox Wrote:
-------------------------------------------------------
> Let's be honest that wall of text wasn't meant to
> do anything but discourage me.
>

ok let's make it short. ;-)


> Just like an EE could say (and
> many DO say) that the arduino skips important
> steps and "prevents" newbies from learning, you
> could say that writing a simple program that turns
> a few steppers prevents others from learning about
> 3d printing.

you're talking of several things. while I fully agree that a suite of "a simple program that turns steppers" (there are already some of those outside there) would be great, especially if it is presented in a systematical way, a software that supposedly allows people to get away with mechanical slop is (probably) dead wrong. At least as long as we are running an open loop operation (the microcontroller does not have any information on actual positions of the carriages excepted at the endstops, motor steps are taken as a best guess). On the other hand, if you're thinking closed loop (linear encoders, image analysis et al.) you may get somewhere - probably with a more powerful microcontroller to simultaneously take care of the interrupts _and_ the stepping. There's a very nice Austria Microsystems magnetic linear encoder chip just crying for code out there.

> I think that's extreme. I think it's unrealistic
> to hold back amateurs who might enjoy their own 3d
> printer because of a righteous belief that they
> follow all of the steps

there's a lot of choices already. from pre-assembled printers with proprietary hardware and software and consumables, to the reprap wiki's many creations. An amateur whishing to 3D-print can already gravitate towards his preferences. Don't want to measure filament diameter and calibrate? pay for ready-to-insert cartridges. etc.

> You had to take because it
> will make them better people in your eyes.

"better people" is a judgement about value.
They do not get "better", they get "empowered", and VERY interesting stuff starts to happen when you reach a critical mass of empowered people. I'm just curious. That's at the very opposite of elitism.

> I have
> this crazy notion that some of those people might
> not care. They may see those steps as I do, an
> unnecessary roadblock to just printing.

the market already provides appropriate responses to this particular urge, and prices them accordingly.

> How any of this relates to "micro$oft" (I do take
> offence btw, you pretty clearly implied that
> you're comparing me to a company you believe to be
> greedy, blundering fools with an intent to keep
> people stupid. Just the way you misspelled the
> name practically screams that.)

Dont misintepret. In the past I've had to run an operation depending on their stuff, which failed at least once every three weeks. Unreliable enough to cause major trouble.
What lessons?
go for simple, atomic solutions. Avoid monoliths. and remember, commercial or not, nobody's in charge, anyway.

> is far beyond me.
> I must be missing something. I really can't
> stretch my mind enough to believe there's any
> money value in this.

well, from a commercial point of view, if you get a way to reduce the inputs in expertise, time and/or materials while keeping grossly the same functionality, you can increase profits (and/or underprice your concurrents). This has been a major driver of hardware-based to software-based solutions.
In personal fabrication, this might not apply the same way.

Speaking of commercial ways, did you finally get the faulty parts replaced or refunded?

Young minds consider all ppls around them should think just like they do.

Obviously, none of the ppls that do disagree in general, just around any principle(s) ... is actually married

I believe longer on the way, it just becomes easier and easier to agree to disagree.