Too Good to be True?
Posted by Anonymous User
|
Re: Too Good to be True? June 29, 2007 11:55PM |
Andrew,
"Darwin" right now is a theoretical machine. No one has made one yet because the design specifications for it aren't *quite* done yet. Adrian is currently completing what are (hopefully) the finishing touches on the hardware side, and there are still several tasks on the software to-do list before it can be considered truly releasable (Johnathan M has become the go-to man for the host software, at least). That being said, several people (see the thread "how many repraps are there?") have created RepStraps, machines that are "making things." Adrian has an almost-fully-functional Darwin, and I suspect we'll know almost as soon as he does when that's finished. Forrest of course has his Tommelise, you can refer to his site for more information about that. Zach has a mostly-complete set of darwin parts (I assume it doesn't include pieces for the support extruder, because I don't think said extruder existed when he received his parts) and just in the past few days a couple more people have come out of the woodwork with part sets they printed on university machines.
So in summary, if you're looking for a practically guaranteed machine with almost no bugs ("undocumented features"
), it would probably be best to wait a few months at least before attempting to build one of these. However, if you're a dedicated early-adopter, feel like contributing to a world-changing technology, or just plain can't wait to start printing your own objects, I encourage you to dive in and join the fun! (Says the guy sitting on his bottom with no parts in hand...)
"Darwin" right now is a theoretical machine. No one has made one yet because the design specifications for it aren't *quite* done yet. Adrian is currently completing what are (hopefully) the finishing touches on the hardware side, and there are still several tasks on the software to-do list before it can be considered truly releasable (Johnathan M has become the go-to man for the host software, at least). That being said, several people (see the thread "how many repraps are there?") have created RepStraps, machines that are "making things." Adrian has an almost-fully-functional Darwin, and I suspect we'll know almost as soon as he does when that's finished. Forrest of course has his Tommelise, you can refer to his site for more information about that. Zach has a mostly-complete set of darwin parts (I assume it doesn't include pieces for the support extruder, because I don't think said extruder existed when he received his parts) and just in the past few days a couple more people have come out of the woodwork with part sets they printed on university machines.
So in summary, if you're looking for a practically guaranteed machine with almost no bugs ("undocumented features"
), it would probably be best to wait a few months at least before attempting to build one of these. However, if you're a dedicated early-adopter, feel like contributing to a world-changing technology, or just plain can't wait to start printing your own objects, I encourage you to dive in and join the fun! (Says the guy sitting on his bottom with no parts in hand...)
|
Re: Too Good to be True? June 30, 2007 07:24PM |
I thought I would add my story in here so that the project developers can gain an understanding of the experiences of an essentially unskilled end user (I am no expert in programming or electronics, but I know enough to be able to follow directions), and end users can get an idea of what state the project is currently in.
I stumbled across the RepRap project several months ago, and I was instantly fascinated by the idea. Being able to simply draw an object on the PC and have a machine print it out for me without huge expense, and without having to tell the machine exactly what I want it to do at each step really appeals to me. After reading every single page on the RepRap.org site, the forums, the main blog, and the builder's blog, I decided that I wanted to build one, and from what I saw I had a good chance of success.
Since no one is currently making the parts available, and I do not have access to an FDM machine, I looked at costs for having a commercial rapid prototyping company make all of the parts. That cost was far too high, but the cost for having them create just the parts for the extruder was within my budget. AFter discussing it on the reprap IRC channel, I dug out my Sherline mill from the garage, and looked into what it would cost to convert it to CNC and purchase the RepRap control boards so that I could control it with the RepRap software. I estimated the costs to be within my budget, and I estimated my skillset to be at least close to adequate (I know some basic electronics, a very little bit of programming, I know which end of a tool goes where, and I can follow directions), so I went ahead and ordered all the parts.
When I ordered the control boards, the new 1.2 version of the PCBs was available, so I purchased those, but the BOMs for those boards were not complete yet. Luckily, the parts were essentially the same as the previous version of boards (except the connectors), so I was able to get the parts together and order them. Once all the parts came in, I started assembly. At this point, the construction pages for the 1.2 boards were still being created, so I had to figure out the assembly order and parts placement partially from the 1.1 pics, and partially from the silkscreen on the PCBs. As I was soldering things together, I kept switching pages back and forth to look at various things, and the pages were being edited, created, and changed as I was assembling! Sometimes I would see the picture I needed, sometimes not, and I would have to cross-reference the 1.1 and 1.2 construction pages to look for the differences, and then the page I was looking at a minute ago would disappear, then a picture would no longer be there, a test procedure would be incomplete, etc. It was rather frustrating, but I was able to figure things out for the most part (now that the instructions ae complete and stable, it's a lot easier to follow them).
Once I got the boards complete, I tried attaching a stepper motor and seeing if I could get it to turn. I unfortunately had some issues with the software install (mostly due to my own mistakes), but with help from Jonathan, I was able to talk to my boards correctly, and tried the stepper exerciser test. I would get a flashing light on the stepper controller, and a hum from the motor, but no turning. After some frustration and discussion with Joost on IRC (who was having the same problem), we figured out that it was due to the torque setting in the RepRap software being set to 50% (I think someone should change that default to %75 or so to guarantee that a motor moves on first test). After confirming that the motor moves, and getting advice from the forums (thanks nophead!) on how to connect my non-standard Sherline motors to the control boards, I was able to assemble everything and get things to move in a meaningful manner.
After dealing with some nasty communications problems (that appear to have been primarily due to my own mistakes), I was able to spend the next week or so tweaking the settings and fiddling with it to get my RepStrap to put plastic on the table in an organized manner. Unfortunately, a number of the preferences and buttons in the software are not adequately (or in some cases at all) documented, and some of the connections between the boards have to be guessed at (last time I checked, there is no info on how to connect the stepper sync lines in the construction pages), so it takes a lot of trial and error to get things going correctly. The forums and IRC channel are great resources for troubleshooting and tweaking, but unfortunately there are only a few people who REALLY know what is going on in the software and hardware, so unless you can slog through the code yourself and understand it, all you can do when you have a question is ask around, and sometimes you get an answer immediately, sometimes a few days later or not at all.
As it stands right now, my RepStrap (named Lamarck) is able to print lines parallel to the X or Y axis for an object quite nicely, but angled lines are causing problems. From what I can tell, this can easily be caused by me not connecting things correctly due to lack of documentation (sync lines!), hardware failure (I kinda doubt this, but it's possible), the settings for my machine in the software putting it into untested territory and displaying a new bug, or something odd I need to do to compensate for my non-standard hardware. From what I can tell, if it was a Darwin machine, I could tweak the settings in a way to get it to work.
From all of the experience I have, I am confident that if you were able to get an FDM machine to make all of the Darwin parts, ordered all of the correct non-RP parts, and assembled them exactly as the current instructions describe, you will end up with a computer-controlled machine that will deposit melted plastic in an organized manner. To get it to print what you expect correctly, you will likely need at least a little bit of help from the forums to get the settings tweaked right, or problems resolved, but you should be able to at least print out small parts of a few layers that represent your object file fairly accurately. I am not sure if Darwin in its current state will be able to print larger objects, or as well as a commercial FDM, but it should at least be able to make pretty darn good parts.
In summary, this is how I feel about the current state of the project: The parts appear to be able to extrude useful small objects, though they may or may not have some minor innaccuracies. The software is now much easier to install and does what it is supposed to do under perfect conditions, but has a few bugs and missing features and/or protections against user error, and has little to no documentation (what the heck does the 'calibrate' button do in the stepper exerciser?). Troubleshooting procedures are still being developed, but with help from the construction page troubleshooting techniques, forum postings, and IRC channel, most problems can generally be solved within a few days. Basically, I see parts of the project as being complete but in need of testing and tweaking, and other parts in need of completion but mostly working. In a few months, it should be to a definite Beta stage - everything pretty much done, but not tweaked to perfection.
If you have decent tinkering ability and are willing to go through some frustrating issues at times, you can create a machine that will extrude *something* at the very least with what is available now. If you have programming and/or electronics skills, you'll probably have an easier time. If you prefer to just be able to connect things up, flip a switch, and have it all work, I would suggest waiting at least six months or so.
OK, enough rambling... Time to get back to beating my head against a wall with Lamarck.
I stumbled across the RepRap project several months ago, and I was instantly fascinated by the idea. Being able to simply draw an object on the PC and have a machine print it out for me without huge expense, and without having to tell the machine exactly what I want it to do at each step really appeals to me. After reading every single page on the RepRap.org site, the forums, the main blog, and the builder's blog, I decided that I wanted to build one, and from what I saw I had a good chance of success.
Since no one is currently making the parts available, and I do not have access to an FDM machine, I looked at costs for having a commercial rapid prototyping company make all of the parts. That cost was far too high, but the cost for having them create just the parts for the extruder was within my budget. AFter discussing it on the reprap IRC channel, I dug out my Sherline mill from the garage, and looked into what it would cost to convert it to CNC and purchase the RepRap control boards so that I could control it with the RepRap software. I estimated the costs to be within my budget, and I estimated my skillset to be at least close to adequate (I know some basic electronics, a very little bit of programming, I know which end of a tool goes where, and I can follow directions), so I went ahead and ordered all the parts.
When I ordered the control boards, the new 1.2 version of the PCBs was available, so I purchased those, but the BOMs for those boards were not complete yet. Luckily, the parts were essentially the same as the previous version of boards (except the connectors), so I was able to get the parts together and order them. Once all the parts came in, I started assembly. At this point, the construction pages for the 1.2 boards were still being created, so I had to figure out the assembly order and parts placement partially from the 1.1 pics, and partially from the silkscreen on the PCBs. As I was soldering things together, I kept switching pages back and forth to look at various things, and the pages were being edited, created, and changed as I was assembling! Sometimes I would see the picture I needed, sometimes not, and I would have to cross-reference the 1.1 and 1.2 construction pages to look for the differences, and then the page I was looking at a minute ago would disappear, then a picture would no longer be there, a test procedure would be incomplete, etc. It was rather frustrating, but I was able to figure things out for the most part (now that the instructions ae complete and stable, it's a lot easier to follow them).
Once I got the boards complete, I tried attaching a stepper motor and seeing if I could get it to turn. I unfortunately had some issues with the software install (mostly due to my own mistakes), but with help from Jonathan, I was able to talk to my boards correctly, and tried the stepper exerciser test. I would get a flashing light on the stepper controller, and a hum from the motor, but no turning. After some frustration and discussion with Joost on IRC (who was having the same problem), we figured out that it was due to the torque setting in the RepRap software being set to 50% (I think someone should change that default to %75 or so to guarantee that a motor moves on first test). After confirming that the motor moves, and getting advice from the forums (thanks nophead!) on how to connect my non-standard Sherline motors to the control boards, I was able to assemble everything and get things to move in a meaningful manner.
After dealing with some nasty communications problems (that appear to have been primarily due to my own mistakes), I was able to spend the next week or so tweaking the settings and fiddling with it to get my RepStrap to put plastic on the table in an organized manner. Unfortunately, a number of the preferences and buttons in the software are not adequately (or in some cases at all) documented, and some of the connections between the boards have to be guessed at (last time I checked, there is no info on how to connect the stepper sync lines in the construction pages), so it takes a lot of trial and error to get things going correctly. The forums and IRC channel are great resources for troubleshooting and tweaking, but unfortunately there are only a few people who REALLY know what is going on in the software and hardware, so unless you can slog through the code yourself and understand it, all you can do when you have a question is ask around, and sometimes you get an answer immediately, sometimes a few days later or not at all.
As it stands right now, my RepStrap (named Lamarck) is able to print lines parallel to the X or Y axis for an object quite nicely, but angled lines are causing problems. From what I can tell, this can easily be caused by me not connecting things correctly due to lack of documentation (sync lines!), hardware failure (I kinda doubt this, but it's possible), the settings for my machine in the software putting it into untested territory and displaying a new bug, or something odd I need to do to compensate for my non-standard hardware. From what I can tell, if it was a Darwin machine, I could tweak the settings in a way to get it to work.
From all of the experience I have, I am confident that if you were able to get an FDM machine to make all of the Darwin parts, ordered all of the correct non-RP parts, and assembled them exactly as the current instructions describe, you will end up with a computer-controlled machine that will deposit melted plastic in an organized manner. To get it to print what you expect correctly, you will likely need at least a little bit of help from the forums to get the settings tweaked right, or problems resolved, but you should be able to at least print out small parts of a few layers that represent your object file fairly accurately. I am not sure if Darwin in its current state will be able to print larger objects, or as well as a commercial FDM, but it should at least be able to make pretty darn good parts.
In summary, this is how I feel about the current state of the project: The parts appear to be able to extrude useful small objects, though they may or may not have some minor innaccuracies. The software is now much easier to install and does what it is supposed to do under perfect conditions, but has a few bugs and missing features and/or protections against user error, and has little to no documentation (what the heck does the 'calibrate' button do in the stepper exerciser?). Troubleshooting procedures are still being developed, but with help from the construction page troubleshooting techniques, forum postings, and IRC channel, most problems can generally be solved within a few days. Basically, I see parts of the project as being complete but in need of testing and tweaking, and other parts in need of completion but mostly working. In a few months, it should be to a definite Beta stage - everything pretty much done, but not tweaked to perfection.
If you have decent tinkering ability and are willing to go through some frustrating issues at times, you can create a machine that will extrude *something* at the very least with what is available now. If you have programming and/or electronics skills, you'll probably have an easier time. If you prefer to just be able to connect things up, flip a switch, and have it all work, I would suggest waiting at least six months or so.
OK, enough rambling... Time to get back to beating my head against a wall with Lamarck.
|
Re: Too Good to be True? July 01, 2007 06:26PM |
I think the issue with apparently missing page links is that a large number of the title bars for some of the sections are links to the page, and some are not. If you're not used to this, and don't happen to mouse over the title and notice it's a link, you may get confused. I'm not sure if this is something that can be readily changed, or should be changed, but a good example of this is at [reprap.org] - if you just load the page and look at it, the top several sections appear to have no linked pages, but if you mouse over the page, you can see the cursor change over the links. If someone has a different browser setup, or different web experience, it may be non-obvious. I don't know if it would be worth it, but perhaps going through all of the link titles and adding '(Link)' to them may make it more obvious.
Just my 0.00000002 cents
Just my 0.00000002 cents
Sorry, only registered users may post in this forum.