Webinar on open hardware in biology

[Carlo Quinonez]

I'm giving a webinar Tueday morning at 11am PST on open hardware in
science. I'll be presenting some of the hardware I've been developing,
including a stage-top incubator, a toolset for making and operating
microfluidic devices WITHOUT a cleanroom (but you do need a 3d-
printer!). Here's a link to the webinar info -
http://blog.neuinfo.org/index.php/general-information/webinar-02-07-201

Anyways, I just stumbled on DIYbio a few weeks ago. I think it's
exciting to have so many citizen scientists out there! I'd appreciate
some feedback on the webinar, the feedback from scientists usually
runs along the "looks good" or some other comment that's equally
useless...

Regards,
Carlo

[Bryan Bishop]

On Sun, Feb 5, 2012 at 10:11 PM, Carlo Quinonez <cqui...@ucsd.edu> wrote:

science. I'll be presenting some of the hardware I've been developing,
including a stage-top incubator, a toolset for making and operating
microfluidic devices WITHOUT a cleanroom (but you do need a 3d-
printer!). Here's a link to the webinar info -

can you drop us a link to this work on microfluidic devices? sounds vague, but maybe you have a specific 3d printing strategy in mind.

- Bryan
http://heybryan.org/
1 512 203 0507

[Mac Cowell]

I can't make the webinar, but I'd love to read a blog post or other writeup about the techniques.

231.313.9062 // @100ideas // sent from my rotary phone

--
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diy...@googlegroups.com.
To unsubscribe from this group, send email to diybio+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/diybio?hl=en.

[Chris]

Sounds like an interesting talk. Look forward to listening in on the
webinar.

Chris

On Feb 5, 11:11 pm, Carlo Quinonez <cquino...@ucsd.edu> wrote:
> I'm giving a webinar Tueday morning at 11am PST on open hardware in
> science. I'll be presenting some of the hardware I've been developing,
> including a stage-top incubator, a toolset for making and operating
> microfluidic devices WITHOUT a cleanroom (but you do need a 3d-

> printer!). Here's a link to the webinar info -http://blog.neuinfo.org/index.php/general-information/webinar-02-07-201

[Carlo Quinonez]

I'm planning on submitting the manuscript to PLoSOne with the next
month.

I don't have anything posted online about the system, I just haven't
had the time. But the webinar will be recorded if you cant make it.

A word of warning about the content... Although its an
interdisciplinary group, I'm going to be emphasizing the informatics
aspects of open hardware because I think we need some hardcore
information scientists involved to drive open hardware forward. So
some parts may get a little slow unless your into that sort of thing.

[Patrik]

Hi Carlo - sounds interesting! You may want to consider submitting you
3D printed designs to Tekla Labs' DIYbio 3D printing competition:

http://www.teklalabs.org/3dprinting/

On Feb 5, 8:11 pm, Carlo Quinonez <cquino...@ucsd.edu> wrote:
> I'm giving a webinar Tueday morning at 11am PST on open hardware in
> science. I'll be presenting some of the hardware I've been developing,
> including a stage-top incubator, a toolset for making and operating
> microfluidic devices WITHOUT a cleanroom (but you do need a 3d-

> printer!). Here's a link to the webinar info -http://blog.neuinfo.org/index.php/general-information/webinar-02-07-201

[Carlo Quinonez]

Here's a link to the webinar:

http://nif.adobeconnect.com/p4o23zwdulf/

The tekkalabs contest is interesting, and publicity is great. It's the
"Legal/Intellectual Property Rights" section that's the problem. It
says that by submitting a design, you "acknowledge that you are solely
responsible for the protection of any intellectual property rights you
may have in the work." But I don't own the IP, the people of
California own it...

It's probably not a big deal, but respecting other people's IP and the
rules are important...

Carlo

[Derek]

Great webinar, Carlo. Really inspired by what you're working on. The
UCSD licensing issue is really unfortunate from a number of
perspectives, not just the competition, but I hope that they are just
patenting the specifics of the designs rather than pursuing a very
broad "we claim the patent on 3D printed microfluidics injection
molding platforms."

I'm mostly working with .5mm resolution FDM printers at this point, so
I suspect my ability to use your designs will be limited until I get
some better printing technology into my space, but I'm certainly
excited by what you're doing!

Thanks again for the talk.

--Derek

[Chris]

Carlo, Thanks for the talk it was very interesting and informative.

Following up on what Derek mentioned about printer resolution sort
of...

Have you thought about a simple micro-fludic design, perhaps just a
proof of concept, that was lower resolution and had a minimum number
of parts that
we in the diybio community could test out?

Chris

[kingjacob]

In the vein of simple microfluidic systems, you should check out the work by the awesome people at hackteria,

link: http://hackteria.org/wiki/index.php/DIY_Microfluidics#Advanced_DIY_Microfluidics

--
You received this message because you are subscribed to the Google Groups "DIYbio" group.
To post to this group, send email to diy...@googlegroups.com.
To unsubscribe from this group, send email to diybio+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/diybio?hl=en.

--

Cheers,
Jacob Shiach

editor-in-chief: Citizen Science Quarterly

twitter: @jacobshiach

[Carlo Quinonez]

It would be AWESOME if a set of features could be developed for
different printers. One set of parts that will print on printer X,
another for printer Y, etc. Just like code that only works on a
certain architecture... Most of the parts I showed could be printed on
with printer architecture, it's just the molds that aren't portable.

I'll let you guys as soon as I get the manuscript submitted, and I'll
make sure to include STEP files for all the templates and features
I've made so far.

But you should know that my wife and I are both out of jobs in
November (we're in the same lab, that's shutting down). Add in our 1
yr old toddler, and now you're talking pressure! So don't take it the
wrong way when I say I've got to focus on getting a funding rather
than continuing to tinker with more building/refining more...

C

[Carlo Quinonez]

There's a lot of clever and innovative approaches to building
microfluidic devices. They all deserve recognition. I wish there was a
source of information for this. The discoverability of information is
a real problem.

On Feb 7, 2:46 pm, kingjacob <kingja...@gmail.com> wrote:
> In the vein of simple microfluidic systems, you should check out the work
> by the awesome people at hackteria,
>

> link:http://hackteria.org/wiki/index.php/DIY_Microfluidics#Advanced_DIY_Mi...

> editor-in-chief: Citizen Science Quarterly<http://citizensciencequarterly.com/>
> twitter: @jacobshiach <http://twitter.com/#!/jacobshiach>

[Bryan Bishop]

On Tue, Feb 7, 2012 at 5:38 PM, Carlo Quinonez <cqui...@ucsd.edu> wrote:

There's a lot of clever and innovative approaches to building
microfluidic devices. They all deserve recognition. I wish there was a
source of information for this. The discoverability of information is
a real problem.

A useful resource is the mems-talk list, but unfortunately it's become more of a place where grad students go to ask questions that die and never get answered. Once in a while there's a good discussion, but overall I'd trust diybio's collective knowledge of microfluidics over that any day.

http://mail.mems-exchange.org/mailman/listinfo/mems-talk

http://mems-talk.memsnet.org/mems-talk/

Also this resource:

http://diyhpl.us/~bryan/papers2/microfluidics/

[Carlo Quinonez]

On a related topic... I forgot to mention that the student version of
CATIA only costs $100/yr and all you need a .edu email address. The
bigger problem is the electrical (EDA) software, I'm using Altium and
there's no student version. Hopefully that will change, but I can get
a "lab" copy from UCSD's license for $125/yr. And I use Arduino for
all the firmware.

Did you recognize the arduino nano on the valve controller? :-)

C

On Feb 7, 3:45 pm, Bryan Bishop <kanz...@gmail.com> wrote:

> On Tue, Feb 7, 2012 at 5:38 PM, Carlo Quinonez <cquino...@ucsd.edu> wrote:
> > There's a lot of clever and innovative approaches to building
> > microfluidic devices. They all deserve recognition. I wish there was a
> > source of information for this. The discoverability of information is
> > a real problem.
>
> A useful resource is the mems-talk list, but unfortunately it's become more
> of a place where grad students go to ask questions that die and never get
> answered. Once in a while there's a good discussion, but overall I'd trust
> diybio's collective knowledge of microfluidics over that any day.
>

> http://mail.mems-exchange.org/mailman/listinfo/mems-talkhttp://mems-talk.memsnet.org/mems-talk/

>
> Also this resource:http://diyhpl.us/~bryan/papers2/microfluidics/
>

> - Bryanhttp://heybryan.org/
> 1 512 203 0507

[John Griessen]

On 02/07/2012 05:52 PM, Carlo Quinonez wrote:
> On a related topic... I forgot to mention that the student version of
> CATIA only costs $100/yr and all you need a .edu email address. The
> bigger problem is the electrical (EDA) software, I'm using Altium and
> there's no student version. Hopefully that will change, but I can get
> a "lab" copy from UCSD's license for $125/yr. And I use Arduino for
> all the firmware.

I was stopped from listening to your webinar for lack of an ubuntu installation
for compatibility. Student versions still don't let you make product for sale,
do they? The EDA part is not really a bad hurdle --- the info for circuits
and circuit boards can be translated to FOSS tools kicad and gEDA.

I'd say lack of open easy solid modeling CAD is the biggest hurdle.

If you got around licensing blocks, and closed tools, a kickstarter campaign
might get you into the manufacturing business in November instead of
job hunting if you wanted. The closed tools arenot even a block to kickstarter --
most funders don't care if it's open hardware or not. At least the
open hardware definition that the data and tools to use it are all open.

John Griessen

[Bryan Bishop]

On Tue, Feb 7, 2012 at 5:52 PM, Carlo Quinonez <cqui...@ucsd.edu> wrote:

On a related topic... I forgot to mention that the student version of
CATIA only costs $100/yr and all you need a .edu email address. The
bigger problem is the electrical (EDA) software, I'm using Altium and
there's no student version. Hopefully that will change, but I can get
a "lab" copy from UCSD's license for $125/yr. And I use Arduino for
all the firmware.

Maybe you should consider using open source software instead, like gEDA. I know nothing* in the open source world compares to CATIA, but that just means we need to write more code ;-).

* almost nothing.. the opencascade engine might be able to match it, if opencascade had better code quality, and didn't have the weird licensing issues.

- Bryan

[John Griessen]

On 02/08/2012 07:04 AM, Alex Hoekstra wrote:
> Was the webinar recorded at all?
He said replaying it would be possible, so try the process of using adobe connect.
It needs Windows or Ubuntu 10.

[Carlo Quinonez]

I've worked in manufacturing, and it's not something I want to get
back into! I'm not a bad builder, but I'm a much better designer...
There's legions of un- and under-employed fabricators in the US and
worldwide - why would I want to get into that business myself?

I'm working towards a set of plans that any machine shop can build
themselves. Then I just worry about maintaining and electronically
distributing the plans. It's a distributed manufacturing model, where
the demand for products is met locally rather than by some central
facility. Saves on distribution costs. I'm already into
"manufacturing" because my local machine shop can build them for half
to quarter the cost of commercially available competing designs. With
the distributed manufacturing model, I can scale up production to meet
demand easily by recruiting more machine shops. Distributed
manufacturing scales up from a single lab to a world-wide scale. Of
course there's going to be some hurdles, like ensuring quality and
consistency, but I prefer to see those as opportunities...

As far as the open hardware definition is that the tools have to be
open as well. I don't believe that's a part of the definition. I know
projects licensed under the CERN OHL are developed in Altium. Open
hardware should emphasize the process of building, rather then an
axiomatic requirement for openness. If you look at the openpcr
project, you'll notice they give you access to all the data, but if
you want to build one yourself, you had better be prepared to spend a
few hours sorting through all the files to extract necessary
information in the right information in the right format. That's just
an unnecessary barrier to openness that was most likely a deliberate
decision to protect their business model. And that's OK. We all need
to make money to live. My preference is to make as easy as possible
for someone completely unfamiliar with the modeling tools and my
device, to build a copy themselves.

Starting a business is something I'm interested in, but not on the
kick-starter level. It's cool and at another point in my life I would
have jumped at the chance. I know that some open-hardware startups
have gone on to make millions (AdaFruit et al) but I'd much rather get
a tenure-track position. Plan B is to raise serious VC money for a
startup, and plan C is an engineering job in industry.

C

[Bryan Bishop]

On Wed, Feb 8, 2012 at 1:02 PM, Carlo Quinonez <cqui...@ucsd.edu> wrote:

project, you'll notice they give you access to all the data, but if
you want to build one yourself, you had better be prepared to spend a
few hours sorting through all the files to extract necessary
information in the right information in the right format. That's just
an unnecessary barrier to openness that was most likely a deliberate
decision to protect their business model. And that's OK.

using solidworks files isn't how tito and josh sell their thermocycler, they do that by selling thermocyclers..

- Bryan

[Carlo Quinonez]

This was in response to Derek's reply from yesterday. I accidentally
hit "reply to author" instead of "reply".

-----------
The big problem with FDM printers are the voids left between adjacent
strands... You can work around it by vapor polishing the finished
parts, but it's genuinely hazardous because the vapor can be
explosive. You could also dip the finished parts in a strong solvent
to polish them as well, that's a little safer, but requires stronger
solvents (like chloroform) which are bad for you. And the geometry is
going to be distorted in minor but unpredictable ways by any polishing
process. That said, FDM it might not matter if you just scale up the
size of the features. In fact, it would be cool to have different
feature sets for different processes. Besides FDM is a great process
for some of the other parts, like the stands, syringe holder, and
casting jig, where airtightness and surface quality isn't important.
Is there an open-source project for a jetted 3d-printer?

I prefer to think of patents as a very good thing, by the way. We have
to prepare for a war with the companies that prefer the status quo
versus the innovators. Consider what happened in open source
software... Microsoft declared all out war on Linux about 10 years
ago. They spent hundreds of millions of dollar buying up patents and
filling lawsuits against the corporate allies of Linux, but it was
already too late. By that time, Linux had enough momentum and
supporters to win the war. If Microsoft had perceived a threat to it's
business back in the early 90's, when linux was still in it's infancy,
I believe the world would be a very different place.

I personally like that the IP is patented and owned by the public,
we're just formalizing the public ownership. And just because it's
patented, it doesn't mean it can't be open. But it does mean that non-
compliance with licenses can be severely punished. I'm hoping for the
best, but if all the time I spent playing FPS games taught me
anything, it's that it's always nice to have a rocket launcher in
reserve... just in case... :-)

C

[Carlo Quinonez]

Sorry I was wrong, I thought they only sold kits, I didn't realize
they sold assembled thermocyclers.

On Feb 8, 11:09 am, Bryan Bishop <kanz...@gmail.com> wrote:

> On Wed, Feb 8, 2012 at 1:02 PM, Carlo Quinonez <cquino...@ucsd.edu> wrote:
> > project, you'll notice they give you access to all the data, but if
> > you want to build one yourself, you had better be prepared to spend a
> > few hours sorting through all the files to extract necessary
> > information in the right information in the right format. That's just
> > an unnecessary barrier to openness that was most likely a deliberate
> > decision to protect their business model. And that's OK.
>
> using solidworks files isn't how tito and josh sell their thermocycler,
> they do that by selling thermocyclers..
>

> - Bryanhttp://heybryan.org/
> 1 512 203 0507

[Bryan Bishop]

On Wed, Feb 8, 2012 at 1:17 PM, Carlo Quinonez <cqui...@ucsd.edu> wrote:

Sorry I was wrong, I thought they only sold kits, I didn't realize
they sold assembled thermocyclers.

My point isn't any different: they aren't selling those files, so I don't see how it's "protecting their business model". They used solidworks because they couldn't find anything better- and it's true, there's really a lot of problems with STEP and IGES and the available open source CAD stuff out there. I agree that they should have posted STEP/IGES versions, but don't assign to malice/bad business models what could easily be attributed to incompetence and/or not caring about open source tools. Being competitive in business is fine, but not when you're using open source as the marketing; this is why e.g. Upverter has had some public backlash for claiming to be "Open Source".

- Bryan

[Bryan Bishop]

On Wed, Feb 8, 2012 at 1:21 PM, Bryan Bishop <kan...@gmail.com> wrote:

Being competitive in business is fine, but not when you're using open source as the marketing; this is why e.g. Upverter has had some public backlash for claiming to be "Open Source".

I should clarify that, I meant "being competitive by holding back your sourcefiles is fine, but not when you're marketing them as open source." It's like someone posting eagle files instead of gEDA files. John would be happy to rant for many hours about this I'm sure.

--

[Carlo Quinonez]

Here's a link to a RECORDING of the webinar:

http://nif.adobeconnect.com/p4o23zwdulf/

Sorry I didn't make it more clear in my earlier post.

[Carlo Quinonez]

I apologize for not being clear. What I intended to imply was that
open hardware should ALSO focus on documenting and RELEASING the
process of fabrication, rather than focusing exclusively on the
geometrical data. Manufacturing a product is a PROCESS that is much
more than just the geometric description of the final shape, and that
process is NOT captured in a CAD file, regardless of the format. It
takes work for a skilled person to figure out manufacturing process/
strategy that is most appropriate for that part and material.

I can give you all of the "source files" in the form of the CAD
documents, and you could still be lost on how to build the real
product. As the complexity of the product increases, the degree of
additional information increases dramatically. There are some parts
I've designed that require a specific orientation on the build tray
and the manual addition/subtraction of support material. NONE of that
information is in the CAD files...

Regarding OpenPCR. There are several hours of work required to get
from downloading openpcr data -> ordering parts. The business model
seems to rely on that fact, and that most people would rather just pay
them a reasonable sum of money in exchange for saving themselves those
few hours of time. That's all. I NEVER said OpenPCR's business model
is bad or attributed it to malice... I actually said it's OK.

To be completely honest, I admire their product and have to confess to
some minor jealousy and envy of their success!

C

>
> My point isn't any different: they aren't selling those files, so I don't
> see how it's "protecting their business model". They used solidworks
> because they couldn't find anything better- and it's true, there's really a
> lot of problems with STEP and IGES and the available open source CAD stuff
> out there. I agree that they should have posted STEP/IGES versions, but
> don't assign to malice/bad business models what could easily be attributed
> to incompetence and/or not caring about open source tools. Being
> competitive in business is fine, but not when you're using open source as
> the marketing; this is why e.g. Upverter has had some public backlash for
> claiming to be "Open Source".
>

> - Bryanhttp://heybryan.org/
> 1 512 203 0507

[Bryan Bishop]

The other stuff is standard agreeable stuff about open hardware,

On Wed, Feb 8, 2012 at 3:19 PM, Carlo Quinonez <cqui...@ucsd.edu> wrote:

Regarding OpenPCR. There are several hours of work required to get
from downloading openpcr data -> ordering parts. The business model
seems to rely on that fact, and that most people would rather just pay

Their is no particular reason they should rely on that, really. Their business is selling things, not selling plans. And in fact, the branding of their product is that it's "Open". So as much as possible should be open.

 

them a reasonable sum of money in exchange for saving themselves those

Sure, and that's fine when it happens, and it happens every day for non-open-source products as well.

 

few hours of time. That's all. I NEVER said OpenPCR's business model
is bad or attributed it to malice... I actually said it's OK.

From your original message, you said:

information in the right information in the right format. That's just
an unnecessary barrier to openness that was most likely a deliberate
decision to protect their business model. And that's OK.

Sooo it seems to me like there's a conflict between them claiming to be open, and this artificial barrier which is counterproductive to the open source ethos and licensing. The artificial barrier doesn't make sense, because the 'business model' is selling kits... not selling information. Just for clarification, the discrepancy this is elucidating isn't between you and me, Carlos. Like you said, "most likely a deliberate decision".. well so was calling it "open source". So... it's entirely possible that they are being deliberately not completely open, while still claiming open source on the entire product.

My original message was pointing out how odd this was, since they supposedly are in the business of selling.. not information restriction. *shrug*

- Bryan

[Bryan Bishop]

On Wed, Feb 8, 2012 at 4:15 PM, Carlo Quinonez <cqui...@ucsd.edu> wrote:

I've been curious about this because I know so little about the
limitations of these open formats. What have you heard? I know that
IGES replaced STEP, but I couldn't tell you too much more!

I'd be happy to. But first! Always use IGES or STEP, at the very least, for transmitting solid models on the internet. Proprietary file formats are okay too, as long as you include at least one of those two. 

As it happens, though, there's no good/dissected format for parametric solid modeling. STEP does not keep parametric information like the solidworks or alibre or catia file formats do. So if you get a STEP object, you can get your perfectly smooth features but not your edit history or the parameterized basis shapes. I was working on reverse engineering solidworks' file format, but the bytestream is a pain in the butt. The parasolid XT file format has better documentation that is somewhat available online, but BRLCAD, OpenCASCADE, etc. haven't actually implemented it. It's a pretty old format, back before Solidworks was a thing IIRC..

There's some recent open source work on SCL (NIST's STEP Class Library) from the oce-dev group, OCE being a fork of OpenCASCADE since OpenCASCADE upstream doesn't really accept community contributions. BRLCAD has been working on integrating these new additions into their code as well, so that's very exciting. There's some additional specs for STEP like ISO 10303 AP2xxsomething that specifies parametric information in STEP files, but nobody has implemented that. I'm pretty sure Solidworks, Alibre, CATIA, Pro/Engineer only do AP203 and AP214.

Honestly, I think a different file format should be used, but it's unfortunate that we still can't translate parametric information out of the proprietary CAD packages..

[Quinonez, Carlo]

Do you know what the limit of the STEP geometry is? There's always a limit to any computer representation… CATIA is upfront about this, they claim 1 micron accuracy, unless you enable Small Scale, then the limit is 1 nm. It makes a big difference in the micro- and nano-fields. I've wondered how other programs and formats stack up. And yeah, CATIA v5 only supports AP203 and AP214. The only difference I can see is the AP214 supports color information

I'm just planning on releasing STEP files. It's a start at least… And it also creates a business model, paying for access to the parametric data. It's not ideal, but I think its a fair balance as long as the manufacturing data that I talked about is also bundled with the STEP files. Things like READMEs and pregenerated DXFs for cutting and gerbers+XY pickNplace files for the board manufacturing. BOMs and suggested vendor lists too.

What do you mean by OpenCASCADE doesn't support user contributions? Are they just advertising that they do, but that no user generated features manage to make into the trunk?

Do you know much about the OpenPLM project? Are any of the FOSS tools for CAD/EDA ready for "production"?

Honestly, the use of FOSS tools is a "nice to have" for me. I'm more concerned with capturing and curating the DNA of the machines that biologists need to cure diseases and continue making scientific break-thrus. I've been staring at this fork… what PLM toolchain should I commit to? It's a decision that's going to have long-term consequences. I chose CATIA because I know it can do ANYTHING and I also have a small, but deep, network of contacts that can help with CATIA. But as I try to expand to embrace a broad development community, using FOSS tools has been a concern.

And why do the French seem to have a monopoly on CAD software? :-)

C

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

Carlo Quinonez, Ph.D. 

IRACDA Fellow

tel: 760-576-4589

fax: 858-246-0761

email: cqui...@ucsd.edu

Department of Pharmacology

University of California, San Diego

BSB 3020

9500 Gilman Drive, MC 0636

La Jolla, California 92093-0636

[Bryan Bishop]

On Wed, Feb 8, 2012 at 5:26 PM, Quinonez, Carlo <cqui...@ucsd.edu> wrote:

Do you know what the limit of the STEP geometry is?

I remember seeing at least millimeters in the spec, but it shouldn't matter as long as you multiply everything by 1eN, right?

 

There's always a limit to any computer representation… CATIA is upfront about this, they claim 1 micron accuracy, unless you enable Small Scale, then the limit is 1 nm. It makes a big difference in the micro- and nano-fields. I've wondered how other programs and formats stack up. And yeah, CATIA v5 only supports AP203 and AP214. The only difference I can see is the AP214 supports color information

iirc AP214 added car configuration manufacturing features. Weird, I know..

 

I'm just planning on releasing STEP files. It's a start at least… And it also creates a business model, paying for access to the parametric data. It's not ideal

um... screw that. I'd just copy your data, and release my own parametric models. So.. I don't see why you would do that. It just makes people like me sad and grumpy.

 

but I think its a fair balance as long as the manufacturing data that I talked about is also bundled with the STEP files. Things like READMEs and pregenerated DXFs for cutting and gerbers+XY pickNplace files for the board manufacturing. BOMs and suggested vendor lists too.

I put some work into a standard git repo structure for open hardware if you want to look:

http://gnusha.org/skdb/

but there are lots of improvements that could be made.

 

What do you mean by OpenCASCADE doesn't support user contributions? Are they just advertising that they do, but that no user generated features manage to make into the trunk?

Right.. and plus, their trunk is closed source. They used to have a public svn repository, but they hated all of the contributions. However, their own code isn't that fantastic either.

 

Do you know much about the OpenPLM project? Are any of the FOSS tools for CAD/EDA ready for "production"?

Some of the FOSS CAD tools are used in production, sure. 

 

Honestly, the use of FOSS tools is a "nice to have" for me.

But why not just use them?

 

And why do the French seem to have a monopoly on CAD software? :-)

I think it's because someone- I'm not sure who- started to perpetuate the lie that CAD is impossibly hard to write. "You'll need 100s of programmers!" but in reality, you just need one or two that know what they are doing. I've tried implementing some NURBS boolean operations based on the papers, but it's a lot of work to do without testing in between stages. Ultimately I think ESOLID is a good base to start from, but it needs to be rewritten because of the potential licensing issues (even though the source code is public), and there needs to be unit tests to confirm that each part of the monstrosity is working correctly.

[Patrik]

On Feb 7, 12:34 pm, Carlo Quinonez <cquino...@ucsd.edu> wrote:
> The teklalabs contest is interesting, and publicity is great. It's the

> "Legal/Intellectual Property Rights" section that's the problem. It
> says that by submitting a design, you "acknowledge that you are solely
> responsible for the protection of any intellectual property rights you
> may have in the work." But I don't own the IP, the people of
> California own it...

I assume that section is there primarily to absolve Tekla Labs of any
blame if anyone breaks any IP laws.

Also, Tekla is run by a bunch of University of California people, and
I'm sure they didn't intend to exclude other UC participants. Although
if these are designs you developed as part of your job at UCSD, then
you may not be eligible for a DIY competition.

[John Griessen]

On 02/08/2012 03:19 PM, Carlo Quinonez wrote:
> There are some parts
> I've designed that require a specific orientation on the build tray
> and the manual addition/subtraction of support material. NONE of that
> information is in the CAD files...
>
> Regarding OpenPCR. There are several hours of work required to get
> from downloading openpcr data -> ordering parts. The business model
> seems to rely on that fact, and that most people would rather just pay
> them a reasonable sum of money in exchange for saving themselves those
> few hours of time. That's all.

I've found that none of my kit product buyers wanted to improve it
beyond making comments to me, so all the to do about open hardware
and definitions sometimes seems like "much ado about nothing".

So far my one kit product is stupid simple though. Maybe as you go
up in complexity there are collaborators to be found -- the adafruit
people say so, and have updated products with newer designs, though
most of theirs just run a short product life cycle and disappear.
It's easy to start thinking of protecting your developments from
copying instead in order to extract a living from them, and I don't
mind that by anyone. I just wish there were easy ways to publish
design info with Bryan's SKDB right now in usable tool formats.
I'd like to do open hardware manufacturing with a model like many FOSS
developers, where you free publish what already works and sell new updates
and add ons for cash for a while, then fold them into what is free published.

I like open circuit board tools of the gEDA suite because they are
set up like the chip design tools -- so you can script them and define
your own methods and sequences to repeat during design cycles.

The free open 3D CAD tools I tried fall down at usability. Freecad
is supposed to have improved enough in the last year to be usable, so
I'm going to test it again.

John Griessen

[Quinonez, Carlo]

On Feb 8, 2012, at 3:38 PM, Bryan Bishop wrote:

On Wed, Feb 8, 2012 at 5:26 PM, Quinonez, Carlo <cqui...@ucsd.edu> wrote:

Do you know what the limit of the STEP geometry is?

I remember seeing at least millimeters in the spec, but it shouldn't matter as long as you multiply everything by 1eN, right?

 

Actually it does matter, there are a lot of non-linear properties that micro- and nano- designers rely on. So you just can't scale models linearly and expect the same results.  So if I'm doing FEA analysis the part, it's going to behave differently at 1000x smaller scale, even if all the forces are scaled appropriately - the physics of how matter behaves is just different at a small scale.

um... screw that. I'd just copy your data, and release my own parametric models. So.. I don't see why you would do that. It just makes people like me sad and grumpy.

But only if you have access to the parametric models…I realized a long time ago that it's impossible to make everyone happy simultaneously. I bet the ELF folks protest Sierra Club meetings because they're not green enough. The doesn't mean ELF should be running the green movement. I don't know if this is a good business model, which is why I'm asking. I'm trying to figure this out. And feedback is an important part!

I thought that if a STEP model, like the ones I demoed in the webinar, were available, they can serve as a robust starting point for forks. 

 

but I think its a fair balance as long as the manufacturing data that I talked about is also bundled with the STEP files. Things like READMEs and pregenerated DXFs for cutting and gerbers+XY pickNplace files for the board manufacturing. BOMs and suggested vendor lists too.

I put some work into a standard git repo structure for open hardware if you want to look:

http://gnusha.org/skdb/

but there are lots of improvements that could be made.

 

I thought git didn't support file locking on large binaries? How do you handle merges on CAD files? I briefly looked at Veracity because it seems like one could write a plugin for your favorite CAD program to handle interfacing to the database.

And thank you for pointing out the http://gnusha.org/skdb/ site. Do you mind if I use your repo structure? How would you like to be cited?

What do you mean by OpenCASCADE doesn't support user contributions? Are they just advertising that they do, but that no user generated features manage to make into the trunk?

Right.. and plus, their trunk is closed source. They used to have a public svn repository, but they hated all of the contributions. However, their own code isn't that fantastic either.

 

Do you know much about the OpenPLM project? Are any of the FOSS tools for CAD/EDA ready for "production"?

Some of the FOSS CAD tools are used in production, sure. 

 

By who? References and specifics. Basically, I have no experience in FOSS CAD tools and don't know which vendor to trust. To put this another way, if I were shopping for an OS to build my latest and greatest web portal, I'd probably go with RHEL and gladly pay for it. The security and peace of mind of having a single entity responsible for the OS is worth it. 

Honestly, the use of FOSS tools is a "nice to have" for me.

But why not just use them?

 

Because >I< don't have the resources to address the limitations of the CAD software. If I need feature X to do something, I can't do it. Like the length limitation I mentioned, if I get started on a project and then 2 years later, I realize that the FOSS cad designers don't understand that it's not as simple as they thought, then I'm boned.

There's also the learning curve and longevity of the data. I'm loathe to start learning a new CAD package only to have the project languish with no updates or bug fixes because all the developers have starting working on the next great thing or left because they were pissed over some dogmatic disagreement. Once again, I'm boned. With an open source project, the developers are fickle. At least commercial software packages, I trust that as long as there's customers with money, the software will continue to exist. In many open source software projects, the interests of the developers are closely aligned with the interests of the users, so this isn't a risk. But there is a massive gap between the fabricators/engineers and software developers.

If I had a big company, then a FOSS tool makes a lot more sense. Then I'd be able to afford to put couple of kick-ass programmers to add features/bugfixes etc. But with any company comes business models… 

And why do the French seem to have a monopoly on CAD software? :-)

I think it's because someone- I'm not sure who- started to perpetuate the lie that CAD is impossibly hard to write. "You'll need 100s of programmers!" but in reality, you just need one or two that know what they are doing. I've tried implementing some NURBS boolean operations based on the papers, but it's a lot of work to do without testing in between stages. Ultimately I think ESOLID is a good base to start from, but it needs to be rewritten because of the potential licensing issues (even though the source code is public), and there needs to be unit tests to confirm that each part of the monstrosity is working correctly.

- Bryan
http://heybryan.org/
1 512 203 0507

[Carlo Quinonez]

> So far my one kit product is stupid simple though.  Maybe as you go
> up in complexity there are collaborators to be found

It takes a lot of people to design high-end instruments like an
Illumina sequencer or an automated microscope. Maybe the sequencer is
too big to even think about, but an automated microscope seems like a
good project for a collaborative development effort. It's pretty easy
to modularize the problem into distinct subsystems that could be
developed somewhat independently, as long some thought went into how
to break up the problem and the interfaces between the subsystems.

Hopefully, more people will be interested in collaborating on hardware
in the future. It took legions of programmers to make open-source
software the juggernaut of innovation and wealth creation that it is
today. We just need to hope that practical engineering education
becomes better in the future. And that high schools start teaching
shop class again... And that society stops telling kids that blue
collar work is for chumps. Otherwise, there's going to be no one left
to build our hardware.

> The free open 3D CAD tools I tried fall down at usability.  Freecad
> is supposed to have improved enough in the last year to be usable, so
> I'm going to test it again.
>

That's good to know, thanks for the information...

[Carlo Quinonez]

> I put some work into a standard git repo structure for open hardware if you
> want to look:http://gnusha.org/skdb/

I just realized that you presented at the 2009 H+ summit, I gave a
talk there too! It's funny how small the world can be sometimes...

-C