I'd like to know from more experienced users that is it wise to use protoboard in final product that I can sell.
It would be for automotive use, so expected to get some hits, bumbs and vibration. Is protoboard a hopeless idea? The board will be enhoused in waterproof container and tightened.
I will produce probably a quantity of 2-5 units. Is it worth it to etch PCBs? I want to get fairly reliable product, that would work atleast 5 years.
My component count is very minimal, only MCU, few transistors and resistors and wiring to main PCB module(produced by real manufacturer).
Thanks for reading!
Bonus question: what kind of overlay glue/protective spray I could use for the board, to protect it from vibration/environment?
If you mean a perforated pcb with proper solderable connections, whereby the tracks are replaced by wires, that is a reasonable solution. It is just as reliable as a pcb, only more tedious to assemble.
If the waterproof container doesn't give you enough confidence, you could fill it with pitch. It was common practice for harsh environments.
These days a PCB is almost cheaper than perfboard, they will be more reliable in that there is no user error involved with cutting and drilling tracks and leaving a little bit of metal in the wrong place etc.
These days a PCB is almost cheaper than perfboard, they will be more reliable in that there is no user error involved with cutting and drilling tracks and leaving a little bit of metal in the wrong place etc.
Rob
I'd assume that in such small quantities PCB cannot compete with 2$/pcs(8cm x 8cm) pricepoint of perfboard. I might be wrong though...
Personally I would consider it a major challenge to design a custom PCB and would not bother for a small number of units - but then I'm not familiar with PCB design software.
Robin brings up an excellent point. PCB CAD software is not difficult to use, but it isn't trivial to learn either. Also, as far as I know, there is nothing both usuable and free (And that you can then sell). Advanced Circuits (www.4pcb.com) has their own software but I'm not sure you can export a gerber (I haven't used it in 3 years). DipTrace used to have a free version that wasn't too restrictive, but I haven't checked on that recently. Eagle will let you make really tiny boards for free. That being said, I support you using PCBs if possible, its a useful skill and will save you time and money in the long run if you are making more than 5 or 10. However, that does include learning how to use it, and how to make your own parts. THe libraries are always lacking.
I would point out that on a manufactured board, you can put your name, the company, the model number.
it should have a solder mask to protect traces. and silk screen to show the top and parts layout. makes assembly much easier.
also, if you are selling for any sizable profit, anyone looking at a home-brew style assembly will not be impressed.
@mirith, out of curiosity (and so as not to remain a dinosaur through ignorance) given that you are familiar with the design software do you expect to design a PCB (for manufacture by someone else) without any significant errors first time, or is it usual to incur the cost of two or three versions before you get a usable product?
Robin2: @mirith, out of curiosity (and so as not to remain a dinosaur through ignorance) given that you are familiar with the design software do you expect to design a PCB (for manufacture by someone else) without any significant errors first time, or is it usual to incur the cost of two or three versions before you get a usable product?
...R
You do bring up an excellent point. I know I screwed it up my first time. Professionally, the company I'm at has actually produced shippable boards with, at most, minor rework the first time for the last 5 or so major boards (And the same with all of the minor stuff as well.) I personally didn't do the specific layout for most of the major work (We contracted out to someone with decades of experience), but I was involved in the design and review processes. I also had to do the rework, and that usually happened because I didn't participate in the double-checking.
As long as you are careful and double check EVERYTHING (Net by net, component by component, pattern by pattern), and assuming your circuit design isn't flawed, It will probably work out fairly well. Make absolutely no assumptions until you've checked it. My own personal projects, I've been less accurate (I made assumptions), but have still been relatively successful the last few times I've built PCBs. Again, its a matter of double checking everything. The other big assumption is that your design is functional. Most of the designs we've been working on have been significant reworks of the same basic design to add features or simplify the design. And sometimes you cannot test that without building the board, especially for high frequency stuff. That all being said, those designs weren't perfect, and always had room for improvement, but they were still usable.
Graynomad:
2 layer, 10x10cm, 10 pieces, $19.90, that's < $2 each even if you throw half of them out (plus freight, $8.98)
So it seems! I definitely have to consider the PCB if I get more demand. Thanks for the eye opener :.
OSH Park's price is $5 per square inch, boards are ordered in multiples of three. A little more than from China, but if you are in the USA they are probably faster and their quality is top-notch.
PC boards are so inexpensive and I am so bad at perfboard that I have been known to order ten from China when I really only needed a couple.
I am totally doing that for the next board I need made. I only need 2, but its cheaper and easier to order 10 from china. (Its big enough that OSHPark is less reasonable)
mirith:
I am totally doing that for the next board I need made. I only need 2, but its cheaper and easier to order 10 from china. (Its big enough that OSHPark is less reasonable)
Yep, I always weigh the size vs. price as well. And the extras can be quite useful:
• Use them as coasters (of course this is SOP if the design has errors)
• Use them as Christmas tree ornaments
• Teach a kid to solder
• Sometimes I even find uses for more and end up building them out!
I to don't think I've ever seen a PCB be 100% correct first time, that said if it's a simple design, no hairy stuff like RF, and proven as a prototype there's no real reason it shouldn't be right first time.
Mine never are because I don't prototype, the PCB is my prototype
As for learning the PCB software, yes that is a major hurdle that might be difficult to justify for a one-off project and 3 units. OTOH if you have experience and a good library of parts you can do a simple board in no time at all. A bit longer if you have to design some new components.
I used to make a lot of my own PCBs when I was younger and they nearly always needed something fixing due to my failure to convert the circuit diagram to a layout. With the advent of PCB CAD where the layout is built FROM the circuit, it's much easier to get something functional first time - provided your circuit is right to start with.
The problems I have with the finished PCB these days are far more likely to be mechanical - stuff doesn't physically fit because I've used the wrong library part, mouting holes in the wrong place that sort of thing - but which most of is easy to check if you just print the board on paper first and try it out! The need for later revisions often come from changes that simplify or speed up assembly when doing a production 'run', but are not essential.
After getting over the hurdle of learing the PCB software (Eagle for me) I now really enjoy the layout process - I find it really theraputic, possibly to the point of obession when trying to get everything looking neat...
As a personal aside, if I was buying something and it turned up on perfboard, I'd be really disappointed! Especially if there was no indication up front!
stuff doesn't physically fit because I've used the wrong library part, mounting holes in the wrong place that sort of thing
If you spend time getting a good library of 3D models that part is pretty much solved as well. You can even get models for the enclosures and "put" your PCB inside to check the fit. Great stuff.
I am currently making a board that I hope will be successful on kickstarter.
I am in the process of doing a prototype on a perfboard with all the chips and parts ordered. not using anything I already have. hoping to eliminate some out of date part from creating problems.
then once that is shown to work, I will have a run of 3 or so board made. the site listed offered an on-line quote of about $10 per board (3.5"x6", single sided) if you are selling something, this process of staging should work.
the full boards should be 100%.
if you are using some odd layout, you could use a laser printer and print it on a sheet of paper and use foam under that to hold chips.
