project planning/design methodology for electronics

Discussion in 'General Electronics Chat' started by bug13, Feb 24, 2013.

  1. bug13

    Thread Starter Well-Known Member

    Feb 13, 2012
    1,208
    38
    Hi guys

    I recently realised that project planning/design methodology is very important, before this point, if I want to built something, I just google a similar/identical circuit, and start building/coding.

    That works well so far for small project, but for a bigger one, stuff starts getting messy.

    So if anyone have tips or something to share for a new guy like me, please do, your help is appreciated.

    or if you have a template that you can share, even better :)

    I am hoping to get some info for early search, goal, concept, to prototype etc...

    Thanks again
     
    Last edited: Feb 24, 2013
  2. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,394
    1,606
    There is a holy trinity of drawings necessary for a successful circuit of any complexity if you are going to reproduce it in numbers greater then one:

    The schematic: shows how the parts interconnect, and establishes the reference designation of every electrical component. The "ref des" are the numbers, R1, R2, U1, D2, etc.

    The layout: Shows the physical location of every component used. Here components is widely expanded to also mean any and everything from screws and nuts, boxes, any and everything.

    The Bill of Material: (or BOM) a list of each and every part or component necessary to build the device. Gives the cross from ref des to manufacturer's name and number, quantity needed, and other optional info such as a description or perhaps a source of supply.

    Typically the schematic leads these drawings, but the BOM is filled in during both schematic synthesis and layout.

    There have been cases where schematic and layout were made essentially at the same time when there was very little room for any parts at all, but it's rare you will ever need to fit a light dimmer in a .1" x .3" outline.
     
    bug13 likes this.
  3. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    In addion to ErnieM's advice above, for the "project planning" part I think you should nut out specs and operation on paper, operation can be done as a flow chart.

    You should clearly define what the thing does, what order it does it in, what type of inputs and outputs it has, look and feel of controls etc. These are all design features that need to be settled before you start work on the actual hardware and software implementation.
     
    thatoneguy and bug13 like this.
  4. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    For every circuit I want to keep and like to look at after some years, there are 20 others which are not worth the storage space.

    Served their purpose to learn about dimensioning and concrete issues, after that, no reason to keep them.

    Schematics are secondary if you have a good knowledge of all the circuit aspects.

    Examining existing designs is essential to know what you don't want, what already exists, how it looks like, or as a starting point.

    Or if you want a single line: Hard work, no way around it. Destroyed components and disposed proto PCBs.
     
    bug13 likes this.
  5. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Hard drive space is cheap. A bit of planning does save headaches later, which is why I came up with my Paint CAD concept. It makes for decent forum based drawings, but if I wasn't planning on doing it as a forum display I would use the schematic software in whatever PCB layout software I like. You may have to break it into pieces to draw it all though.

    Sometimes, if the circuit is really simple, I may not draw the schematic, just a layout. I may draw a schematic for this later, but it is going to be point to point wiring, a rats nest, in other words.

    Here is a current work in progress (by way of example). The sockets on the side are going to be wire ribbon connectors, so there is no electronics except LEDs.

    [​IMG]

    You can find Paint CAD on my blog in my introduction.
     
    takao21203 likes this.
  6. bug13

    Thread Starter Well-Known Member

    Feb 13, 2012
    1,208
    38
    Hi

    Thanks for all the inputs, just let you guys know that I have read all of your friendly advice.
     
  7. bug13

    Thread Starter Well-Known Member

    Feb 13, 2012
    1,208
    38
    Hi Bill,

    I don't quite understand your suggestion, could you explain it in another way?
     
  8. tgil

    New Member

    May 18, 2011
    19
    4
    I do product development consulting. Here is how a project is supposed to flow (but almost never does):

    1) Create Marketing Requirements: these are top level requirements for what features will be marketed to users (e.g. battery lasts all day long)

    2) Create Engineering Requirements: these break down the marketing requirements into more technical detail (e.g. 8 hours of battery life). You should also write a test plan that specifies how to verify each of your requirements are met.

    3) Create Engineering Specifications: the specification lists out what will be needed to meet the engineering requirements (e.g. 400mAh battery Li-Pol battery, average current draw of 50mA)

    3) Design according to the specifications: select components, create schematics, and draw the layout so that each specification is met. Sometimes not all specs can be met due to engineering tradeoffs so the requirements may need to be revisisted at this point.

    4) Prototyping: build some units.

    5) Verification and Test: Execute the test plan from step 2 to ensure all of your requirements are met. You may need to jump back to step 3 and make some changes to ensure everything passes.

    7) Production: start producing your product.
     
  9. bug13

    Thread Starter Well-Known Member

    Feb 13, 2012
    1,208
    38
    Thanks, that's exactally what I am looking for. So why do you think it never work out according to the flow? I am interested to know.
     
  10. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    Today "economic" is a sales term.

    If you design a circuit or PCB or gadget, consider the end of life.

    what if the original purpose no longer exists or a small part breaks?

    Will it render it ewaste or is it possible to use parts for other purposes?

    This is not so much a concern for real mass products. They often all derive from quite a few or just one engineering masterpieces. For instance small boomboxes from different manufacturers are nearly identical except some minor modifications.

    But for smaller scale, think about reusability. For instance a display unit can be used for different designs or it can be detached.

    If you have USB and RS232, and one connector breaks, you still can use the other. Think like that.
     
    bug13 likes this.
  11. DangerousBill

    Active Member

    Jul 21, 2010
    30
    1
    I always begin with a functional block diagram to identify connecting cables, physically separate modules, and separate functions within the modules, with electrical and data flows indicated. (As well as fluid flows, since most of what I do involves gases or liquid flows.)

    From this, I go on to generate the schematic, layout and BOM, but I do them together.

    I'm still a believer in physical prototyping, too. Probably my age.
     
    bug13 likes this.
  12. tgil

    New Member

    May 18, 2011
    19
    4
    The reason it doesn't work out this way is because clients always want to rush things and skip documentation. Depending on the project, sometimes that's OK. Other times it comes back to cost them.

    The easiest thing to skip is the test plan, but it is also one of the most important thing to have. You don't want to place a production order to later find out you have problems with the circuit.
     
  13. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    My current hard drive is 500 Gig, and it is just over ½ full. I expect by the time I need a new one a multi-tera byte drive will be $60, I am already seeing prices like this at Fry's and MicroCenter.

    Some people doodle on paper, I doodle with Paint CAD. The difference is my notes (and brain farts) are stored in such a way I can find them years later.

    Paint CAD is almost a CAD program, but at its most basic level. Absolutely no bells and whistles, just a bunch of drawings you can cut/copy and paste from.

    I have backups of my Paint CAD files (that is everything electronics related) on an 8 Gig thumb drive (price $4).

    Look ma, no papers! Except when I print it out. When I am done I throw it away.

    Another example...

    High Power LED Flasher
     
  14. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    If you do not specify exactly what the inputs and outputs will be, and what functions need to happen between the two, you'll be chasing your tail quite often. Include anticipated operating environment.

    Simulate frequently, then breadboard sub-circuits and inject anticipated signal to ensure concept functions properly.

    At that point, Spec Parts (can be done earlier), do the PCB layout, and order a few prototype boards to be essentially tested to destruction. Fix any issues that allowed interference, or destruction to happen too soon.

    Repeat the prototype test (if this is for large run of critical, skip if for a toy), test to destruction until all requirements are surpassed, not simply met.

    Then do a large run of the board, placed, if you do this too early (frequent mistake for many), you'll find yourself either ditching a ton of boards, or spending a ton of time/money fixing them.
     
    bug13 likes this.
  15. bug13

    Thread Starter Well-Known Member

    Feb 13, 2012
    1,208
    38
    Hi thatoneguy,

    Thanks for your input, I think they are very useful to me. Can you tell me more about the tested to destruction? Do you simply overload the circuit and see which part fail first, or do you have a special precedure to do it?
     
  16. takao21203

    Distinguished Member

    Apr 28, 2012
    3,577
    463
    You'd typically see some components failing after a while for various reasons- not being a good match or not of good making.

    Sometimes as trivial as relays latching, or resistors carbonizing, trimmer resistors burning out, and switches not to work properly anymore (overloading with current).

    It is of course embarassing when this is happen on the customer side and you get PCB returns.

    It is also not good to overdimensionate everything too much.

    Of course you can bury a 5 Amps dc/dc when you only need 200mA.
    You can use a large IRF mosfet when a SOT23 would be enough.
    You can use pushbuttons for 2 dollars each there are even some which cost 12 dollar each.

    Most common put your PCB into an enclosed space and let it run for a few hours, then see where it gets hot or too hot.

    There are so many things you can test.

    From all the experience, you will be able to start off with a better prototype for your next design.

    you'd also know for instance some things on PCBs are not good for handling, or you'd want to have additional things like screw terminals, or proper labels. Even to put the component numbers on the copper is not a bad idea. People will be able to know what the PCB is using without to need a magnifying glass for small ICs.
     
    bug13 likes this.
  17. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Usually tests have these at a minimum:

    Use a noisy supply line, add noise until it glitches (non-destructive)
    Add input noise, see how well rejection works.
    Use near a broadband EMI generator to see how susceptible the circuit is to external noise.
    Make all possible "user mistakes", such as hooking up battery backwards (usually physical protection to prevent that, but difficult if a 9V battery is used).
    swap inputs around
    Heat in an oven during testing, to at least rating of caps (105ºC) do same with freezer. Make sure output doesn't vary more than specified.

    The above shouldn't destroy the circuit in any way, just give you an idea of the environments it can function in reliably. The above should also be monitored on a real full spectrum analyzer to ensure there isn't spurious EMI being radiated from your device.

    If it passes all of the above, you test to ensure it "fails safely".

    Over voltage on power, over voltage on input, high reverse voltage, supply voltage on inputs, etc.

    No capacitors should explode and no resistors should be on fire.

    If circuit still functions, you've done well.

    Increase out of spec inputs until it does catastrophically fail, no fire should be evident, and no damage to supply should occur (such as shorting a lithum battery, making it explode because one component failed).

    If you want something UL/CE Listed, take a look at their requirements, some of them are rather insane, but you'll then know why UL/CE certified products cost twice (or more) as much as a wo-hun-lo knockoff of same item at a store.

    For that listing, several of your items need to pass those tests, at which point you sent several (and a lot of money) to UL, who independently do similar tests to destruction. It's ok if it dies, but it cannot cause interference outside it's zone, start a fire, or damage the enclosure to a point where a human using it could be injured in any way. A UL listing for a hobbyist circuit is ridiculous. There are other standards, even more demanding, if it is to be used in "Life critical" circuits, such as hospital equipment to smart bombs.

    It depends on how far you want to go with it, and if you have a team of lawyers to defend you if somebody gets scratched by your product exploding because they mixed up the power input and signal inputs. Making it "open hardware, as-is, no warranty of any kind, express or implied" adds protection against that.
     
    bug13 likes this.
Loading...