Blinky-blinky!

This entry covers the electrical component over the last several session.

Several sessions ago we had little luck getting the microcontroller to talk to the USB device. The FT232R-based adapter we were using was modified from another in-house project (we'll call it the USBBT adapter). The symptom was that 'dmesg' gave some complaint about the USB device when our adapter was attached. (As you can tell, the exact events from several weeks ago are a bit foggy.) I think one of the connectors on the USBBT to MoBo adapter cable was reversed at one point, so the chip may have been damaged - but there was no confirming smoke or sparks.

I should pause here to point out that the bootloader had previously been successfully installed via the USBtiny. Pressing the reset on the MoBo caused the reset LED to give three short and one long blink. This proved the microcontroller was correctly wired and functioning - at least at some level. But we could not download through the USB channel.

Anyhow, a new USB board was assembled with the extensions for RepRap (adding +V, RTS#). This fresh adapter was correctly recognized by the Linux USB driver and so was deemed ready for the next C-Rap build session.

The next session gave us no joy, but we were able to prove that the USB interface adapter was indeed working properly - it was used to work with a commercial motor controller, and performed flawlessly. So why didn't it work with the RepRap board? Someone suggested downloading the bootloader again - just in case.... Well, we had a heck of time getting that to work (lots of timeouts) but it seemed to succeed in the end.

Last night's session we took it step-by-step and verified that the adapter pins were correct, etc., etc. We tried downloading the bootloader via the USBtiny since none of us were confident with the previous attempt, but we kept running into USB timeout errors. An alternative method (command line, not GUI) was used, and that worked like a champ. (Somesh - can you document exactly what you did?)

So, we made some progress. We knew the bootloader was properly installed, and that the microcontroller was still working. Yay. Back to the USB adapter.

Using an oscilloscope to monitor the RTS#, RXD and TXD lines on the mother board, we tried downloading the Menedl software. The only signal activity observed was one (or two) bytes on the RXD line when (I suppose) the host was trying to wake up the Sanguino. What should have happened was that the RTS# line should have strobed low, which would have triggered the reset on the bootloader. This line was always low (that's suspicious!), so there was never a reset pulse. After noodling over why this line was never high we thought to try a different USB adapter. We had on hand a FTDI adapter cable which was pin compatible with the MoBo header, and so connected that. This still did not work but the RTS# was always high!

It was getting toward the end of the session, and many people had already left, when Michael tried what Somesh had come across earlier (but did not work before): manually pressing the RESET button just before the compiler/downloader started the download stage.

Lo and behold, the red status LED on the MoBo began furiously blinking! A program was being downloaded! And then the download program read back the flash area! The program was verified!

Now, when powered up, our RepRap MoBo has a gently pulsing status LED. Michael tried the RepRap interface program even though we had nothing connected to the MoBo. There were some java exceptions (hey! software people - you're on deck!) but gcode commands were recognized.

The electrical team must figure out why the USBBT adapter did not work. My current hypothesis is that the internal EPROM has incompatible settings. Yes - the FT232RL does have programmable internals that can change the behavior on the various output pins. It is possible that the settings in the USBBT device differ from the default settings in the USB adapter module described on the Mendel site.

Cool things other people are doing: Extreme Extruder Edition

Now for a couple of reprap related links that I've been saving up:

Recyclebot

All metal hot end

One guy, making it look easy...

After a long hiatus

Well, well, well. I bet you thought we'd all given up. Not true! It's been a busy couple of months at work, and we've had to cancel several sessions, but we are making progress. The electronics are fully assembled, and are currently undergoing debugging. As of the end of last week's session, the x-axis was assembled enough to get an idea of the shape of the assembly. Progress!

For much of this project, procuring the mechanical parts has been the major bottleneck. We have one milling machine, which means that there is really a fundamental limit to the number of parts we can make in a three hour session. Last week I had an epiphany: I got on ebay looking for part sets. (In my defense, I did this at the beginning of the project, but no one was selling reprap parts then.) Turns out, the street price for a set of printed parts is about $150. Now, we've been hard at work for a while now, and don't need all the parts, so I contacted a couple of sellers and asked if they would be willing to print just the parts we needed. Sure enough, Several of them were. Now, I've ordered some parts, and we'll see what shows up in the mail in the next couple of days!

Progress! Exciting!

Random thought

I think I've figured out why the RepRap parts have all the crazy teardrop shaped holes:

They are designed to be printed on a machine that does not use support material. If you are printing a part with a hole (axis parallel to the build plane), it is difficult to print the "top" of the hole without the material sagging. A teardrop shape has the benefit of keeping the material overhang within reasonable limits.

Cool things other people are doing: The Laser Edition

Selective Laser Sintering (SLS)

Progress you can believe in!

In the interest of preserving the few scraps of sanity that I still posess, I have put together a spreadsheet to track our progress.

At the last build session it became clear that I could not simultaneously keep my darling daughter out of trouble and know what everyone needed to do next at the same time. My hope is that this will give every one a good idea of what remains to be done, and keep me from being the bottleneck.

Build Session

Attended:

• Adam H.
• Adam S.
• Ben
• Branko
• Cedric
• Costa
• Devin
• Jim
• Michael
• Nikola
• Somesh

Food: Castellos Pizza (Supreme, BBQ Chicken, Polynesian) and assorted beer

Casualties: none

Progress:

• Michael, Cedric, and Jim worked on putting together the order for the motherboard components
• Somesh was working on breadboarding something, possibly the motor control circuit?
• Nikola was stripping an old laser printer for parts: we got a motor, some optostops, and a few other interesting bits out.
• Adam S. and Branko worked on modeling parts in Pro/E. We almost have everything, I think.
• Adam H. helped with the modeling then started trying to figure out how we could make some parts out of MDF with the machines we have downstairs
• I (Ben) did my best to keep my daughter (Leah) pacified with the help of Monsters Inc. and Shrek. While I wasn't doing that, I spent some time organizing what needs to be done, and offering Pro/E advice to those who were interested.
• Costa and Devin spent a while in the basement cranking out parts (more on this later. I plan to start posting pictures of parts soon)

A bunch o' pictures from the last couple of sessions:

Marius (we think, it's hard to tell) after spending several hours in the shop doing some precision MDF belt-sanding

Nikola cut all the structural rods.

Somesh worked on breadboarding the new motor control circuits.

Jim and Michael working on soldering up electronics.

Jim, probably destroying the evidence of his latest plot for global domination.

Somesh, with his circuit.

Adam and Branko, plotting something devious, as usual.

A long overdue update

This was intended to be the post for our March 1st build session. I'm a bit late, eh?

So this post is going to cover the work on the session on the 15th also . . .

So here's the short update: We've actually settled into something resembling productivity at these build sessions. A bunch of the guys worked on making more of the misc little pieces out of MDF. We now have enough of them that I'm going to have to sit down at the next meeting on Monday and do some sorting and labeling, just so we know what needs to be done. Unfortunately,we're going to have to start working on the hard pieces now. (The big ones, with lots of angles and holes.)

Michael has been busy working on the various electronics (the AVR programmer and the three optostop boards), and I think they are finished and tested.

Those of us with experience in Pro/Engineer are translating the PAR files into solid models that we can modify ourselves. Adam, Devin, Branko, and I have been working through these at a reasonable rate. We have enough parts that I am starting to build some of the sub-assemblies. (I'll post pictures as they are done.)

Here's the agendas for the last two sessions. I'm going to start posting the session agendas along with the updates, just so that interested parties can figure out what we are currently working on.

-----------------------------

3/15/10 Meeting

Prep:

• Order Pizza & arrange for beer
• Make sure we have appropriate numbers of sketches

Admin:

• C-Rap can
• Set next meeting: 03/29/2010
• Distribute tasks

Mechanical:

• Continue modeling pieces in Pro/E
  
• Modify part designs for work with plywood
• Start building pro/e assembly
• Continue cutting plywood blocks
• Mill vertexes to correct thickness
• Drill out-of-plane holes
• Start making other parts
• Order Bearings

Electrical:

• Assemble and test programmer
• Finalize other board designs in preparation for ordering other boards (Interested in electronics layout? Check this out?)
• Order parts for motherboard

Software:

• Note: Most software stuff tabled until electronics are assembled
• Create sketch outlines

Recycler

• Decide what we want this thing to do.
• Generate potential designs and processes.

-------------------------

3/1/10 Meeting

Prep:

• Order Pizza
• Make sure we have appropriate numbers of sketches

Admin:

• C-Rap can
• Set Next meeting
• Assign people to tasks

Mechanical:

• Start modeling pieces in Pro/E (up to three people. Perhaps: Ben, Devin, Adam S?)
  
• Cut final rod to length
• Continue cutting plywood blocks
• Cut out vertexes
• Mill vertexes to correct thickness
• Mill out-of-plane holes
• Order Bearings

Electrical:

• Assemble programmer
• Assemble optostops
• Finalize other board designs in preparation for ordering other boards
• Order parts for motherboard

Software:

• Create sketch outlines

Recycler

• Have BrainSturm session to generate design ideas
• Get Cracking

Cool things other people are doing: The Bedwarming edition

A heated bed for makerbot

Another nice heated extruder bed (Makerbot)

Next blog

If you press the "next blog" button at the top of the page, then you get a random post.  Here's what I got in order that I clicked
Shards of My Sanity 

International Tina Turner Fanclub
Real Women Have Curves
Prowling Sheep

I had to stop at the sheep.

Construction begins . . .

In the last week, we got the orders from McMaster and Dell Fastener in. We still need to order bearings, motors, and electronics, as well as a few other things.

So, we started in on the grunt work last night.

There was a lot of discussion about how to proceed. One of the big things was what to do about the problem of engineering drawings. Adam suggested that it was probably worthwhile for us to do our own drawings and assembly for two reasons: First, it gives us a better understanding of how the C-Rap goes together, and second, it allows us to be more consistent in our construction and modification of future versions. Those of us with access to a CAD package will start working on that over the next two weeks.

Once we got that out of the way, we broke up into groups: a few of the guys went down to the dungeon and cut the threaded stud and rod to size,

a few started cutting out corner vertices (we decided that we were going to make as many parts as possible out of MDF)

and Devin started looking at making real parts out of these pesky STL files.

Now that we have an idea what needs to happen for each part we make by hand, I think it will go a bit faster.

The next meeting will be on March 1st. Same time, Same place.

Weather

There was no meeting (or pizza or beer) this week, due to a fairly impressive snowstorm that has blanketed the east coast.

We'll try again next week

Nothing is easy

I had not anticipated the difficulty in getting the parts and information that we needed to get going.

On the ordering side of things, we've found that many of the parts we need are on backorder, and in at least one case are discontinued.

Almost as obnoxious is the difficulty in getting CAD data for Mendel. For some reason, the reprap project uses Art of Illusion (a surface based modeling program) files as the native data format. These files are, so far as I can tell, impossible to import into a commercial CAD package, making measuring and modifications difficult at best. While many parts are modeled in SolidEdge, these models are not kept up to date. We have found several cases where the PAR files do not match the STLs, and since we can't effectively measure the STLs, it is difficult to verify that a PAR file is correct.

Orders Aweigh!

I sent out two orders for stuff today.

McMaster-Carr is sending us rods, belts, and a few other misc items.

Dell Fastener is sending all the screws and nuts

I'll add order specifics later.

A bit of background

I thought I'd write down why we are working on the project. I (and, I suspect several others) have been wanting to build a RepRap since I first heard about it. The idea of having a rapid prototyping machine of my own appeals greatly to me, but commercial machines are . . . expensive, despite the fact that the processes used in rapid prototyping are relatively simple. (It's all about the accuracy.)

Since a commercial machine is out of my budget for this decade, my only real option is to build one myself, or live without. Unfortunately there are major hurdles to building one myself:

0) Time

1) Electronics

2) Software

3) The women in my household, particularly my daughter (Shiva, destroyer of worlds)

Any one or two of these I could probably handle on my own, but the combination is a powerful multi-drug cocktail that has side effects of pessimism, malaise, and a powerful tendency to procrastinate.

Oh well. This pattern of thinking: "Maybe when the kiddo is in college" was well entrenched until last Wednesday. That morning Costa came in and was telling me about the most recent issue of Make Magazine (issue 21). The feature project is a "Cupcake" CNC machine. Costa said something to the effect of: "Wouldn't it be great to have one of those?" Someone pointed out that we probably had enough brain power sitting around to do it without any trouble. "We could do it." I though about that for a moment, and I asked Costa something to gauge how serious he was . . .

And we were off. The project had gone critical.

I fired of an email to pretty much everyone in the company, hoping that at least a couple more people would be interested. Well, they were. All of them. The EEs, the Software guys, the interns, the CTO. I admit that I was kind of surprised at the level of interest. I suggested that we meet one evening every two weeks, figuring that that would keep the time commitment to a manageable level for everyone, and floated the idea of beer and pizza as a way to keep people from having other commitments during that time. (Hey, it works, what can I say?)

We had looked at the RepRap website, and it estimated that we could build the newest version of the machine, "Mendel," at a cost of about $520. (I'll talk more about this in future posts.) This sounded perfectly do-able when I was thinking that we'd have two, maybe three guys working on this. Now we had ten. no problem. A few guys suggested that we should talk to Craig, and have him sponsor the project. I didn't think it would happen, but figured that we probably shouldn't count on it. Adam was fairly adamant that it we could give him a good sales pitch, and that it would work, so I told him to work on it. About 5 minutes later, Craig called me over, and asked me: "So, what is this thing you are working on?"

So I got to make the pitch.

Then, lo and behold, Craig agreed to provide us with a budget, which was negotiated up to $750, to be used for parts, pizza, and beer. At this point, all we need to do is get started!

So on Thursday, Jim picked up some beer, I ordered some pizza (from Pizza Parma: a Parma Pie and a Mediterranean Pie), and we all got together after work for the kickoff meeting. There we gave an overview of the project plan, picked out a project name, started this blog, and started laying out what we wanted to get done before the next meeting (to be covered in my next post). Costa managed download and install the current reprap software, and he and Marius took a look. It sounds like they have some work to do to get it up to our usual standards.

Here are a couple of pictures from the meeting.

Here's a picture of about half of everyone. Those are Marius's hands in the lower left. You can't see Somesh or Nikola (those might be his hands on the right edge), or myself.

Here's my daughter. Having a good time. We only gave her empties, I swear. . .

Yay! software built. Runs, but is extreeeeeeemlely primitive from a UI point of view. Time to learn Java.

Everyone left...no picture...sorry.

First bad omen. Synaptic couldn't install Java5. Crashed.

That's what Java 6 is for, I guess....installing now

It's a big one

We voted on names as first point of order, after beer and pizza. Candidates includes
BlueBlet
iRap
others too unpopular to mention
and the winner is:

C-Rap!

Or CRap for short.