Upverter (YC W11): the perfect tool for open-source hardware(venturebeat.com)
venturebeat.com
Upverter (YC W11): the perfect tool for open-source hardware
http://venturebeat.com/2011/09/13/upverter/
12 comments
I took the tour on their site and it looks neat but the first thing I thought was, "Where's my layout tool?"
Open source hardware is no good if I can't manufacture it on a circuit board. Yes, breadboarding is handy for small stuff, but 256 pin BGAs don't fit on breadboards. And repeatability isn't so grand.
With a nice, reasonable to use layout tool (even with just simple routing assistant type features [push and shove]) and tutorials on how to use it, what DRCs are and how they're used, and maybe videos showing people who know how to do this stuff teaching, upverter could be really cool. Heck, for a full-ish featured EDA package in the cloud targeted at open hardware, I'd even pay to use it (provided we're not talking Cadence or Mentor money, more like GitHub paid account money).
Open source hardware is no good if I can't manufacture it on a circuit board. Yes, breadboarding is handy for small stuff, but 256 pin BGAs don't fit on breadboards. And repeatability isn't so grand.
With a nice, reasonable to use layout tool (even with just simple routing assistant type features [push and shove]) and tutorials on how to use it, what DRCs are and how they're used, and maybe videos showing people who know how to do this stuff teaching, upverter could be really cool. Heck, for a full-ish featured EDA package in the cloud targeted at open hardware, I'd even pay to use it (provided we're not talking Cadence or Mentor money, more like GitHub paid account money).
Hey bradfa, Zak from upverter here. We hear you! Layout will be one of the next things we launch (we are working on it right now, hopefully just a few months away).
The hope is to also add a "print button" to connect users designs with rapid manufacturing suppliers.
The hope is to also add a "print button" to connect users designs with rapid manufacturing suppliers.
That's a hell of a business model! Congratulations.
That will be very cool.
Best of luck to you guys.
Best of luck to you guys.
> The 'perfect tool for open-source hardware' would benefit from being open-source itself.
Agreed. Open Hardware needs fewer closed tools, not more.
Agreed. Open Hardware needs fewer closed tools, not more.
As a current EDA user, I hope these guys (or something similar) take off, I really do. It's ridiculously painful working with current EDA tools and not having access to tools I take for granted as a programmer (version control, diff, even text-based data that I can beat up in my own scripts).
But I think they're coming at the wrong end. Their schematic capture was painfully slow when I tried it a couple weeks ago, and I (and I suspect nearly every other electrical designer) rues having to change tools. If someone wants to become the 'github of electronics', I strongly recommend working on getting some sensible versioning and sharing going without also tackling the problem of creating EDA tools.
Having said that, a shared crowdsourced library is a big help, and some others have already noticed the good it can provide: https://github.com/adafruit/Adafruit-Eagle-Library
But I think they're coming at the wrong end. Their schematic capture was painfully slow when I tried it a couple weeks ago, and I (and I suspect nearly every other electrical designer) rues having to change tools. If someone wants to become the 'github of electronics', I strongly recommend working on getting some sensible versioning and sharing going without also tackling the problem of creating EDA tools.
Having said that, a shared crowdsourced library is a big help, and some others have already noticed the good it can provide: https://github.com/adafruit/Adafruit-Eagle-Library
This is the approach we have been taking at CircuitBee (http://www.circuitbee.com). We have Eagle and KiCAD importing, with versioning on the way already.
We're not aiming to replace the tools you already use, just make it easier to share your work on the web. I do love the idea of a web based EDA tool, but I'm yet to believe it can be done with performance characteristics that will work for any reasonably complex circuit.
We're not aiming to replace the tools you already use, just make it easier to share your work on the web. I do love the idea of a web based EDA tool, but I'm yet to believe it can be done with performance characteristics that will work for any reasonably complex circuit.
Hey patrickyeon, its Zak one of the Upverter founders. Thanks for the comment.
We just pushed a major update publicly yesterday that had some incredible performance gains (50 - 100x faster). We have also done a bunch of selective rendering updates which make it feel a ton faster to use.
You should give it a try again and let me know what you think! Hopefully we are a lot closer!
We just pushed a major update publicly yesterday that had some incredible performance gains (50 - 100x faster). We have also done a bunch of selective rendering updates which make it feel a ton faster to use.
You should give it a try again and let me know what you think! Hopefully we are a lot closer!
I'm glad to hear you're not ignoring performance, but please keep in mind that you're putting a big stumbling block in the way for people who would want to use a similar service. github would've gone absolutely nowhere if they forced programmers to re-type everything into HTML forms, or even into extremely well-written web apps.
Github would have also gone nowhere if they only supported code written in one text editor.
Most EDA software is written by companies who have no interest in letting you see how their proprietary formats work.
Most EDA software is written by companies who have no interest in letting you see how their proprietary formats work.
> Most EDA software is written by companies who have no interest in letting you see how their proprietary formats work.
Oh boy, I know it, and it's been a pain in my side. However, EAGLE (lots of hobbiests use it) is migrating to a fully-documented XML-based format [1], and there are two open-source projects for schematic capture, kicad and geda, which would give someone a head start on parsing their design files, at least.
[1] http://www.element14.com/community/message/15751
Oh boy, I know it, and it's been a pain in my side. However, EAGLE (lots of hobbiests use it) is migrating to a fully-documented XML-based format [1], and there are two open-source projects for schematic capture, kicad and geda, which would give someone a head start on parsing their design files, at least.
[1] http://www.element14.com/community/message/15751
Hey rcfox, Zak from Upverter here. We are planning to release import/export and an open file format by the end of the month. We have altium, eagle, kicad and geda converters in the works too!
Do you guys have a plan to make a revision control system here? I really think it would be great to be able to fork projects, look at diffs of projects, and so on. If you already have this, I would suggest highlighting it on the tour page, as I have this suspicion that such a feature would make you guys a killer app.
Other than that, this looks really awesome!
Other than that, this looks really awesome!
We already have forking, both for the current version of your design and previous versions. We cover it on the community tour page.
Diffing has been talked about internally and it would be pretty awesome.
Diffing has been talked about internally and it would be pretty awesome.
Wild fantasy for library & component search - once I find a part - not only can I pull in the symbol and footprint, but also select a canned snippet for one of the common use cases of that part (ie: what everyone copies from the app. note and/or reference designs). Eg: switchers+inductors+caps, max232+caps, ... This could be a value-add for certain parts and suppliers.
Bonus points for parameterising the component selection and adapting to current design rules!
Bonus points for parameterising the component selection and adapting to current design rules!
A few things: when placing parts, allow the user to press escape or other key/mouse combo and jump back directly to the library part selection. It makes it easier to populate a schematic.
Second, separate the parts into different libraries! Other EDA packages have thousands of parts, it would be a mess to have all in one list, even with filtering. It seems you are loading all the parts when creating the library dialog? Probably a different approach to make it manageable...
Preview for the parts, including PCB footprint and variations. What do you plan to do with parts with different blocks, like logical gates, power pins, etc ?
When adding nets, allow the user to move parts around.
I have a few more, but I'll wait when you get layout, drc/erc , auto routing and pcb webgl 3d view working ;) Besides this, keep the good work!
You can try to look at the mailing list archives of the open source packages gEDA, PCB and kicad. There's a trove of nice info and discussions there. I contributed long ago (1998?) to gEDA and use these , eagle and others more or less frequently.
Second, separate the parts into different libraries! Other EDA packages have thousands of parts, it would be a mess to have all in one list, even with filtering. It seems you are loading all the parts when creating the library dialog? Probably a different approach to make it manageable...
Preview for the parts, including PCB footprint and variations. What do you plan to do with parts with different blocks, like logical gates, power pins, etc ?
When adding nets, allow the user to move parts around.
I have a few more, but I'll wait when you get layout, drc/erc , auto routing and pcb webgl 3d view working ;) Besides this, keep the good work!
You can try to look at the mailing list archives of the open source packages gEDA, PCB and kicad. There's a trove of nice info and discussions there. I contributed long ago (1998?) to gEDA and use these , eagle and others more or less frequently.
I really like the focus on getting the open source hardware people using the product first, before taking it to the enterprise. Kind of like GitHub - free for open source, pay if you want to keep your code private.
I just took the really short tour. Looks like a neat tool, as a web-based schematic capture package. I've never run across one of those before. I mostly use Altium in my work, but have considered Eagle and gEDA for this. I don't know how you'd collaborate on schematics though, seems like only one person can work on a schematic at a time.
I don't see any support for hierarchy in a design. For scaling things you need that. You advertise pages that are as large as you want, but that's more a hindrance than a help, I think, because you can't print an infinite-sized page. And how do you print? How do you get the data out at all - is that possible? In order to do layout you'll need to be able to produce a netlist, and the community-provided parts library is not so great if it doesn't carry over into the physical world of PCB layout. It would also be good to be able to do a simulation, but that is going to really be tough on your servers.
It would be really cool if you didn't have to sign up to try it, that stopped me in my tracks even if it did look like an easy form to fill out.
I don't see any support for hierarchy in a design. For scaling things you need that. You advertise pages that are as large as you want, but that's more a hindrance than a help, I think, because you can't print an infinite-sized page. And how do you print? How do you get the data out at all - is that possible? In order to do layout you'll need to be able to produce a netlist, and the community-provided parts library is not so great if it doesn't carry over into the physical world of PCB layout. It would also be good to be able to do a simulation, but that is going to really be tough on your servers.
It would be really cool if you didn't have to sign up to try it, that stopped me in my tracks even if it did look like an easy form to fill out.
So when does interop. with Octopart happen?
The parts library is already integrated with the Octopart API.
Ah - that explains the voluminous library. I missed the references to octopart in some of the library parameters - IMO, that should be called out as feature.
It'd be cool to go from the end of a search via octopart ui to a new symbol in upvert without cutting and mis-pasting part numbers.
It'd be cool to go from the end of a search via octopart ui to a new symbol in upvert without cutting and mis-pasting part numbers.
Really pleased to see a YC startup in this area. One of the secondary ideas I applied to YC this round is a marketplace for new types of computers based on Arduino.
> One of the secondary ideas I applied to YC this round
> is a marketplace for new types of computers based on Arduino.
I'd be interested in learning more about this if you wanted to share.
I'd be interested in learning more about this if you wanted to share.
This is a really neat Idea. I'm working on a 60inch touch screen for doing data analysis. I've tried it on our large system and it works like a charm (Webkit in QT). I think this will be interesting to follow your progress!
I'm a Altium user, can't really use it on a iPad. Upverter can go almost everywhere!
I'm a Altium user, can't really use it on a iPad. Upverter can go almost everywhere!
Login with Facebook would be nice. Also, have you guys thought about having some of the featured content on the front page? It's easier to get an idea of what it's about if you don't start with a clean slate and don't have to create an account..
When I read the name Upverter, I read it like:
U-pervert-er
U-pervert-er
There is a lot more to your average EE CAD package than just wiring up some parts. That's the easy bit. Rules checks, physics, placement, packages, PCB design, (auto)routing, Gerber files and all the other 'goodies' are required before you can take on the big guys.
Schematics is an important element in the whole chain of developing hardware, but it really needs a lot more than that.
That's a large number of man-years, and by going the software-as-a-service route they gain some advantages but are also stuck with a number of disadvantages, the largest of which is that they won't be seeing the benefits of user contributions at a level where it will move the needle. Of course there will be contributions to the parts library that can be shared, which is a step.
The 'perfect tool for open-source hardware' would benefit from being open-source itself.