Transition from Prototype to finished boards
EmptyBit
Posts: 72
I am interested to read some process stages that take one of my projects to the next level. More specifically to reduce labor on populating and wiring protoboards one at a time. There is still fabrication and package assembly I will need to do, but a drop in board would sure help.
One proposed plan is to start with the GG propeller platform and have the stack on boards made for future project applications. Not knowing the market demand for these finished products will limit board orders to lots of 10 to get them out there and see where it goes.
This basic module requirments
4 watt mono audio with pot connection
Composite video and/or VGA in a future rev
3 I/O for various user PB inputs and 1 for power loss sensing
Besides a schematic and GG basic stack on module layout, what is next?
Can our organization afford to have this design stage outsourced?
Who offers quotes to do this service and still make it economically feasible for non-profits?
Volunteers for the board design concepts ready to submit RFQ? I have no circuit software experience other than CAD/CAM.
Much appreciated if this thread discussion can point us in the right direction.
$0
One proposed plan is to start with the GG propeller platform and have the stack on boards made for future project applications. Not knowing the market demand for these finished products will limit board orders to lots of 10 to get them out there and see where it goes.
This basic module requirments
4 watt mono audio with pot connection
Composite video and/or VGA in a future rev
3 I/O for various user PB inputs and 1 for power loss sensing
Besides a schematic and GG basic stack on module layout, what is next?
Can our organization afford to have this design stage outsourced?
Who offers quotes to do this service and still make it economically feasible for non-profits?
Volunteers for the board design concepts ready to submit RFQ? I have no circuit software experience other than CAD/CAM.
Much appreciated if this thread discussion can point us in the right direction.
$0
Comments
Do you need a person to assemble a set of boards?
Do you need a person to do the board layout?
Do you need a person to do the whole package?
You ask for volunteers with a RFQ? Are you asking for volunteers or a quote?
I would be willing to help out a non-profit if you could tell me what you want to achieve.
Jim
Sorry for such a vague description. I am working on adaptive devices for the Blind, Deaf and Deaf/Blind in industrial manufacturing environments. Simple but effective large displays with voicing and possibly add on Braille displays off in the distant future.
Currently I am using the Parallax protoboard and I was thinking of switching over to the Gadget Gangster Propeller USB platform with its stacking module feature.
Here is a link to the GG ProtoPlus module as the basic footprint to emulate. They have the layout for download in PDF or DIP.
http://gadgetgangster.com/find-a-project/56?projectnum=254
The ultimate goal is creating my own custom stack-on modules with component layout and traces to fit various applications.
The volunteer portion:
Guide me through the process.
I would offer the schematic of my as-built prototype for some critique of my I/O circuits used on the Propeller ProtoBoard are what they should be. My current prototype does function as expected, yet best to make sure all is correct before diving in with board design.
Then I need assistance in taking the schematic and GG module footprint, generate a component and trace layout; put in a digital format for RFQ submital. Maybe Pspice or some other digital data they require? I have no software or experience to get that part done.
The RFQ portion:
I can submit the RFQ. I just have no contacts that make and/or populate boards or experience getting it from prototype stage to a finished board.
I can send off the data to a board maker to get quotes in small batch quantities. Reasonably priced board Vendors anyone?
I can send off the boards and BOM to someone that can return them soldered and ready to stack. I could solder by hand and be more effiicient than using a prototype boards point to point wiring drudgery. It would be nice to have a contact and know the cost to achieve it ready to use.
A one-stop-shop that can do both economically would be dreaming?
$0
Do you have a schematic of what you would like to build?
I downloaded the Freeware version of Diptrace offered through Parallax only two days ago and was able to design a simple board very quickly (~4 hours) after following the very decent Tutorial.
The Free Lite version allows you to place up to 500 pins on two separate layers (top and bottom) and may cover a great deal of what us amateurs like to do.
In addition I found that you can download a free Gerber viewer from Pentalogix and you can get a free .dxf viewer directly from Autodesk which both greatly help verifying dimensions and correctness of the design.
If you can sketch the schematic then you can very rapidly enter it into Diptrace and then you are on your way. Also, the Pentalogix software has a feature where you can directly submit your entire project for quote from them for prototype manufacture.
I would be happy to help your project and exchange files. It sounds like a very worthwhile arena to be playing in even though I have a hard time picturing a Braille HMI.
Good luck to you my friend,
SM
You are only about 13 miles north of me on SR169. How's that for a small world! I'll PM you contact info.
I do use AutoCAD 2000 and a few of the older 2004-07LT versions. I've never used any of the add-ons. I'd need to check compatibility on those.
500 points would be plenty for a stack on module. Sounds like I'd need to learn a bit in software, but the price is right. A Diptrace file and schematic is all I need?
Pentalogix is a one stop shop or just more data preparation software?
$0
If you have a schematic you can reproduce it in Diptrace along with the correct component footprint for each item. Then, it is pretty straight forward to generate an optimized PCB layout and copper trace layers from the schematic. Diptrace is a full featured package and can also generate every other feature that goes into a PCB... e.g. via locations along with the NC drill files. It can provide files for every single board layer necessary for a vendor to completely define the manufacturing process and generate a quote for you. This can get complex with through hole plating, solder masks, silk screen print layers etc.
I would suggest starting really simple at first and build a board with just holes and copper traces to replace your protoboard set-up.
I am no expert by any means.
However at the price (free) you should probably investigate for yourself just how powerful the program actually is.
I think there are some local shops that do one-off PCBs in the area too.
There are experts on this Forum who can help with evaluating your circuit design too.
SM
There are many tutorials on the internet (eg in Instructables). You learn many things along the way as well, eg things like making the autorouting grid smaller (10mil), making power traces fatter, setting the via-to-pad distance higher so that you don't accidentally bridge from a pad to a via when soldering, and 'smashing' the labels for components so you can move them where they are more legible. I prefer the autorouter as I am constantly changing things on the design. A board like the one below takes about 30 seconds to route.
Then there are all the custom library components you can design, eg you put a RCA socket on a board and it doesn't fit with the RCA socket you have bought, so you design a new component, but the holes are too small when it comes back from the manufacturer, so you have to redesign the library again with bigger holes.
So when getting this sort of thing right, it helps to use one of the 'batch' fab houses that can make one board cheaply by sharing space on a panel with other customers.
And even when you have it all perfect, some clever person on the forum does something like design a better VGA driver with different resistor values so more colors are visible, or point out the low pass filter on the audio circuit is incorrect, so you have to redesign it all again!
I've been designing Gadget Gangster boards for 7 months now, and still finding faults. Some boards are up to revision 6. But some I think are getting close to something I'd be prepared to put on the market. YMMV and you may be able to get it right with less iterations than I have *grin*.
And when you think it is correct, it can sometimes be humbling (but save a lot of money) to post a design on the forum and ask for constructive criticism.
I have 5 boards I'm working on in parallel - a 'motherboard' which is the same as the GG USB board except it uses switching regulators, a Keyboard/Mouse/SD card board, a touchscreen board, an external memory board, and a TV/VGA board. I see in your original post you mentioned something about a TV and VGA board so if you are interested, the schematic and board are below and the design is open source, so if you want the eagle source files I can post these too.
How many layers and pins does Eagle PCB allow?
Isn't there a list of recommended components that could fit most of our needs (like connectors, resistors, capacitors, etc.) that are contained in a 'Parallax Library' for use by people in this Forum. Similar to OBEX but for components that have been used, been converted to a standard component library and have been known to work as hoped for (and we could buy through Parallax)?
I would really like to see your touchscreen design for the GG board.
SM
I am still on my rather basic projects from a year ago and nothing on the level you are up to. Then again, I still have a day occupation so I only get part time to work on these. I'll be happy just to finally get the first prototypes mounted and in use. I'm so close, but alas, I need to make another parts order.
I know all to well the redesign issues you mention. I've had several occurances of mold modifications made without my knowledge and suddenly parts do not feed through my robotic automation feeders, pick and place projects I used to design and build.
My hope with custom boards is a better use of resources in both time and funding. A slightly higher board cost would offset the labor hours required to mod a protoboard. Although my labor hours have been pretty much free, I'm trying to avoid burnout too. I still need to account for that time in real world value to the projects once numbers warrant a production run. I anticipate any volunteer portion I can recruit is minimal as far as taking a schematic to a board layout. I hope that is not asking too much now that you have given it proportion with the unknowns/changes and outside influences.
I guess learning another piece of software may be unavoidable. That is going to have to wait for the next project or between lulls.
The free version of Eagle is limited to 2 layers and board area 100x80mm, and for "non commercial use". Otherwise no limits.
SM, I was hoping whatever boards I have made would at least meet the same quality as Parallax or GG. Sounds like that is expecting too much?
$0
I guess the catch is the price of parts at sparkfun - sometimes 10x what you can get if you wander down the Silicon Alley of Hong Kong/China/Taiwan/Thailand on Ebay, or go to places like Futurlec. I've tended to do prototyping with expensive parts, but then try to find similar parts at much cheaper prices to get the cost down. Stocking parts does cost money in terms of handling and in terms of stock that might not be sold, but having said that, there are a few parts that would be very handy to be able to buy in one place, like SD sockets, VGA sockets, through hole sockets with long pins, memory chips.
Bundling these up into a single library is a good idea. I'll see if this can be done.
Re
Yes, absolutely. The boards you get from fab houses are of very high quality. No problems there.
If you learn one piece of software, the skills will be transferrable to other software packages. The general concepts of things like fat traces, routing optimisation, neat layout, decoupling capacitor placement etc will be the same for all packages.
Time vs money. The equation changes depending on how well your day job pays!
Ballpark, I pay about $50 for one board from BatchPCB and it takes about 3 weeks to go from wherever they make them to the US and then to Australia. (BatchPCB send you two boards if you order one, so you get a free one).
For OurPCB, that board in post #7 cost $1.13. But then there is a fixed tooling cost of $80, so if you want one board then BatchPCB is cheaper, but if you want 100 boards, then OurPCB is a lot cheaper. The numbers would be similar for other fab houses. Sometimes it is still a lot quicker and cheaper to build a prototype using a protoboard and point to point wiring.
Maybe the best way is to just leap in and have a go? First step is to design a schematic. Then post it here and people could suggest any changes. Then design the PCB from the schematic, which involves placing all the components and then running the autorouter (or doing the routing manually if you are keen!). Then take a PCB design and turn it into a Gerber and a Drill file ready to send to the fab house.
Lots of help is available here on the forum.
If you're coming from AutoCAD, Diptrace is very useful in that you can import a DXF. So it is possible to design up your board mechanical layout in AutoCAD, with the benefit of commands you're familiar with. In the mid 80's I used Autocad for all sorts of things including pcbs. You can use traces, polylines, donuts etc to produce pcbs. You can also do fancy board outlines - this is how I did my ninja board which has the OSHW logo as its outline recently.
Regarding getting the PCBs made, you can send "gerber files" (the graphics files that describe your pcb) to all sorts of board houses to get fabricated.
* I like silver circuits in Malaysia - they would make 8 boards up to 3.15" x 4" for you for about us$89 including 2 day fedex shipping. Turnaround about 1 week total,
* ExpressPCB have a great little service with 3 boards for $51 in a very short time ( 2-3 days) - but you need to use their free software to design the board. The free software is quite usable. Jonnymac's original Propeller Platform used this.
* Dorkbotpdx would do 3, 2.8"x2.5" boards for $35 including postage, and these are US made, but would take slightly longer.
If you need help getting the boards assembled, and its nonprofit work I would suggest getting a college/university on board. They are always on the lookout for final year and summer projects that benefit the community.
And certainly ask the forum for help whenever you need it
I have been a liason with a local University on a couple Senior year projects.
The first one was some voiced GUI/HMI software work on a dedicated CNC mill. A couple years later, I designed a simple braille display out of an adding machine 18 wheel-printer. New wheels were made in SLA octagon shapes where each face had 0-3 braille dots. 2 wheels made up one braille cell. I drove it with the Parallax SX48 chip as a means of display for Mitutoyo digimatic output precision measuring equipment. The students did the mechanical design work while I was doing electronics and software. It was a fully functional 9 character display, maybe impractical for shop environments. The point was to experiment with new methods for cheap braille displays. You can buy a complete printer calculator for $40. A typical braille cell can be $40-70ea.
That program usually starts at the beginning of the school year, so that would be an option for next year.
Welcome to the Forums T-Bone!
It is customary to start a new thread with a question that is not directly related to the original post. You will get more replies if you start a new thread with a title as such: Basic Soldering Questions.