Deployment strategy

Having recently made the journey from breadboard to proto board to PCB, perhaps some of these suggestions will help.

First, it is cheap and easy to add a "bare arduino" from parts - 1 chip, 2-3 caps, a crystal, and a resistor if you want a reset button.
(see art364 / Standalone Arduino)

For my project, I combined this on a proto board with the rest of my project. You can program the chip in your Arduino and move it to the board, or use serial.
Since my project was battery powered, I added a 3v-5v switcher from Sparkfun. (Obviously, there are other options.)

Later, I wanted to explore making a PCB, so I bit the bullet and learned Eagle. (There is a Mac version, I believe.) For me, it was frustrating at first, then it clicked, and it was a lot of fun.) I combined the switcher and "bare arduino" on one PCB and made a daughter board for the other stuff. I used Sparkfun's PCB service - slow and cheep.

The point is, that I'm glad I did the proto board first. Like doing a breadboard before a proto, it helped me make a better PCB. One is as good as their options, so you can stop with the proto board if you decide not to go forward.

Perhaps you asked for the time, and I told you how to build a clock. :slight_smile: So in summary, I'm suggesting to start with #3 and play your option on #4.
John

Thanks for your explanation. That's exactly what I am looking for. just wish the cost of custom boards are not that high. You especially confirmed that doing a "bare" arduino setup first would be a good thing instead
of using another kit.

Thanks,
P.K.

If you're doing less than 10 boards, it's probably not worth the effort of making a PCB unless you're doing something special (like surface mount components).

Once you're doing more than 10 boards, the price per board becomes quite reasonable.

Thanks. Good rule-of-thumb on the 10 board minimum.

About perf board, I think using "hole only" perf board resulting in the need for point-to-point wiring. While fun my gut is that it is going to be pretty painful and messy. I looked at some of the cheaper protoboards, at Jamesco, for example, which has "rows of three" pads a little like a breadboard. Those may make wiring a little easier, plus they sell a matching case for it. The board is very small though, 2x3 inch or something like that. I guess I can stack them together.

Just typing up this message makes me wonder -- won't it be great if there is a arduino board that has a large prototype area? like merging the BBB with 2 protoshields?

P.K.

. I looked at some of the cheaper protoboards, at Jamesco, for example, which has "rows of three" pads a little like a breadboard. Those may make wiring a little easier, plus they sell a matching case for it. The board is very small though, 2x3 inch or something like that. I guess I can stack them together.

Those rows of three holes boards are great. They're also available in larger sizes, I usually buy them at Sayal, a local electronics store. They have power lines between the rows too that will fit between the rows on an IC which is also very helpful. They're available in larger sizes and are quite cheap. I'm sure there's plenty of online stores that have them, or even ebay.

I've recently finished a project using the Arduino to control LED lights, and I would highly recommend the RBBB boards. It depends on what you're connecting the board to, but for my particular application, it worked out great. I had tried to assemble the components for a standalone atmega168 on Radioshack perf boards (two different kinds, the ones with copper pads around each hole, and the kind set up like a breadboard), and the traces kept separating from the board and breaking connections. In the end, buying an $11 RBBB turned out to be way better.

I didn't realize from the website's description and picture that the board would be so tiny, it's great. It's not very much bigger than the size of the chip itself. It took about 15 minutes to assemble and worked right away.

Like I said, I don't know what you need for your particular application, but my experience has been great with the RBBB. Not much more than you'd pay for the parts themselves for all the parts + a PCB.

I thought about using the RBBB as well. from the picture of the PCB, it looks like it just have connections in a row down one side? Do you connect the RBBB via a row of headers to another perf board? or do you ribbon wire them together? (I assume the rest of the electronics is on another board.

I had tried to assemble the components for a standalone atmega168 on Radioshack perf boards (two different kinds, the ones with copper pads around each hole, and the kind set up like a breadboard), and the traces kept separating from the board and breaking connections.

Sounds about right. Don't but anything from Radio Shack, their stuff is all garbage. That will not happen with a decent board (which those are not) unless you really overheat it with your iron. I hope for your sake you're not using a cheap radio shack style soldering iron. About all those are good for is turning people off electronics for life.

Remember that the bond between that pad and the board is all that mechanically holds your component in place on the board. It has to be strong.

I don't see how the RBBB will help in this case, you'll end up making almost as many connections between this board and the other as you would to put the ATMEGA right on the other board.

The 10 board minimum doesnt really apply with some places like BatchPCB.com, if youre willing to wait. Theres a 10$ minimum and $2.50 per sq in so for small simple boards its not that bad to pay something like $15 to avoid the hassle of wiring perf boards. I cant tell you how much I despise wiring perf boards or hand-etching my own pcbs! Yeah itll take a week to get through the learning curve of Eagle but once you know it, its like riding a bicycle. So I guess I would suggest if this is a one off thing that just needs to be done quickly, it's not too bad to wire up a single perf board. But if this is something you might be doing again in the future and you have the time, I would suggest looking into making your own pcbs.

-B

The 10 board minimum doesnt really apply with some places like BatchPCB.com, if youre willing to wait. Theres a 10$ minimum and $2.50 per sq in so for small simple boards its not that bad to pay something like $15 to avoid the hassle of wiring perf boards. I cant tell you how much I despise wiring perf boards or hand-etching my own pcbs! Yeah itll take a week to get through the learning curve of Eagle but once you know it, its like riding a bicycle. So I guess I would suggest if this is a one off thing that just needs to be done quickly, it's not too bad to wire up a single perf board. But if this is something you might be doing again in the future and you have the time, I would suggest looking into making your own pcbs.

-B

IMO, BatchPCB is a perfect example of why the 10 board minimum does count.
At BatchPCB, it's $10 + $2.50/square inch, so a board as small as an Arduino shield (about 2x3 inches) is $25. Add to that the time messing around with Eagle to create the file, and it's also a lot more work to go the PCB route. Why in the world wouldn't you just buy one of the fine protoboards on the market if you only need a few? It will take less time, be easier, and cost a whole lot less.

The only exception is, as I noted above, if you have special parts like surface mount.

Even then, at places like Olimex, you can get 4 shield-sized PCBs for not much more than that $25. And Olimex lets you choose your soldermask colour.

I really don't understand why people use batchPCB.

I really don't understand why people use batchPCB.

Because people love and trust Sparkfun, other than that, I have no idea. Unless you have something small (say under 4 sq in) your likely cheaper getting someone else to make it. I use them every now and then when I need a breakout board for -that-fscking-small-weird-footprint-device-I'm-tryingout- usually this means I can get 4 breakout boards in less than 3sq in, there's no point paying someone else (or wasting a panel slot) to make 3sqin of something if I only need one set.

This said, if you make mistakes (knock wood), it's nice to mess up one board then mess up 20.

Just to express an opinion, I think it's difficult to generalize about when to get a PCB made. Seems like too many variables - the complexity of the circuit, where the board will be used, and any desire to simply experience the process are a few.

The same might be said concerning where you should get boards made. I used BatchPCB to make two different 2" x 2" boards for a little less than $33 US - shipped. I could be wrong, but I doubt that I could have gotten these for substantially less elsewhere.

Perhaps there is a difference between those of us who get boards made occasionally and those who do it on a regular basis.
(Personally, it was worth it all just to see my name on the silkscreen!)

Because people love and trust Sparkfun, other than that, I have no idea. Unless you have something small (say under 4 sq in) your likely cheaper getting someone else to make it. I use them every now and then when I need a breakout board for -that-fscking-small-weird-footprint-device-I'm-tryingout- usually this means I can get 4 breakout boards in less than 3sq in, there's no point paying someone else (or wasting a panel slot) to make 3sqin of something if I only need one set.

fwiw, I do like sparkfun itself very much. It's just their board service that seems to make no sense.

That small weird footprint device is exactly the exception I've mentioned though where you would want a pcb even for a one off.

I do want to comment that if you need that breakout, there's probably 20 other people on here who would be interested but don't want to take the time with Eagle or the expense of $17.50 to make 3. So, for something as general-purpose as a breakout, why not have a 24 square inch panel of them made at Olimex or custompcb.com or some such and sell the extras on your site? Even if you do end up with coasters, it's really not that much more than the batchPCB price for a much smaller amount.

Really?! Im not going to blow a hundred bucks on a whole batch of a single board that may or may not work and then have to go through the hassle of liquidating all the ones I dont use. Look, its your preference.

If Ive got 2 or 3 weeks to play with I would rather spend the 30 minutes some evening to whip up a little board for something in Eagle, zip it over to sparkfun, and when it comes back, happily put together a nice professional board in just a few minutes. The alternative is hassling with perf boards, making all the jumper wires (Ive always hated stripping wire... ::)), and then spending the rest of the afternoon troubleshooting some errant solder bridge. Plus, Ive always felt anything Ive ever put on a perf board just seems so kludged together and altogether uncool.

When I consider what my time is worth, and the fact that I have very little of it, it makes sense to effectively have someone else do it. $25 + an hour or two work for some standalone board is a bargain compared to what, maybe a day or two's worth of work? What do you charge by the hour?

Anyway, the added benefit is that whatever I do in Eagle is repeatable, if I, or someone else, find themselves needing a similar solution to one that Ive already drawn up. This way I use sparkfun as a cheaper prototype and then I take the design to Gold Phoenix when I need a larger batch done. Id rather spend little and wait awhile just to know the design works before having a 100sq in coaster.

But honestly, YMMV. Just saying that for anything smaller than about 5"x7" batchpcb makes a lot of sense and is in fact a good alternative for final, professional designs.

Cheers,
Brian

Really?! Im not going to blow a hundred bucks on a whole batch of a single board that may or may not work and then have to go through the hassle of liquidating all the ones I dont use. Look, its your preference.

There's not much to say here since the tone of your thread makes it sound like you'd rather rant then have productive discussion.

I will merely point out that this "hundred bucks figure" seems to be a figment of your imagination as I mentioned a $30 Olimex product several times -- that's $30 for 4 shield-sized boards, not an unreasonable amount, if you can "whip up" a little Eagle board in 30 minutes, the design is so trivial you're better off with perfboard (unless it's one of those adapters for a weird part), I never told you to liquidate anything, that was aimed at spiffed, who has a very nice webstore all ready set up.

Are you really, really saying you can whip up a 30 minute board that would take you 1-2 days to build on perfboard? That fails the sanity check.

Not that that's even the point since I didn't say don't get a board made, all I said was don't get a board made at batchPCB, so your reply fails the rationality test.

Anyway, I'm done replying here. You'd rather rant than read, and if this thread continues it will only degenerate.

I don't think it sounded like a rant; don't forget, we can't really hear tone of voice and so forth online...

I personally hate perfboard, and stripboard isn't a lot better. It's difficult for me to get it right, whereas if I draw the schematic correctly in eagle, and the PCB matches the schematic, my prototype's gonna be what I intended (it may be wrong, but at least it will match the schematic).

On the Olimex cost, please correct me if I'm wrong, but that's 30 euro plus at least 8.50 euro to ship to the US, which works out to US$60 for that single 4"x6" panel.

-j

On the Olimex cost, please correct me if I'm wrong, but that's 30 euro plus at least 8.50 euro to ship to the US, which works out to US$60 for that single 4"x6" panel.

I really hadn't wanted to post in this thread again, but I don't like to run away from my mistakes either.

You're absolutely right there. I haven't orderd from Olimex in ages (euros were practially american dollars back then).

There are a lot of places that do small runs now though. PCB-Pool is a direct competitor to batchPCB, but around $1.45/square inch + shipping. I've never used them and can't recommend them.

I don't think it sounded like a rant; don't forget, we can't really hear tone of voice and so forth online...

"Really?! Im not going to blow a hundred bucks on a whole batch of a single board that may or may not work and then have to go through the hassle of liquidating all the ones I dont use. Look, its your preference. "

Sounds a lot like a rant to me, but I suppose you can be right.

I'm not a fan of perfboard either. But if I've only got 10-20 components and just need a 1-off, it's a lot faster and easier than waiting for a PCB.

I'm not a fan of perfboard either. But if I've only got 10-20 components and just need a 1-off, it's a lot faster and easier than waiting for a PCB.

Yep, faster than waiting, and I haven't found sending out to a board house cost effective enough for a 1-off - my budget is too tight.

I sort of got interrupted and left off part of my post. I was actually going to argue for home etching. Even for simple stuff, it seems easier to deal with toner transfer and acid/peroxide etching than to deal with perfboard. I was doing a simple power supply (LM317, 2 caps, 2 resistors) and wished I had time to etch a board instead of dealing with the stripboard (I managed to goof up the first one). I route the boards so that they're single sided, or quite nearly single sided (a few traces on side 2, all of them straight lines if I can manage). I also use the press-n-peel paper, rather than screw around with photo paper. I'm beginning to like SMT parts, too (SOIC and 1206 size please) just so I don't have to drill holes in the PCB. Drilling is the worst part of the home etching process, IMO.

-j

Yeah, I actually imagine SMT parts will make home etching more reasonable in the future. Drilling all those holes was such a drag conventionally. With pre-sensitized pcbs, all you have to do is print out your art on a transparency, slap it on the board, shine a bright light on it, develop and etch.

No, Oracle it was not a rant. Maybe a little flippant but whatever. Youre right I didnt acknowledge your comment about Olimex. Thing about them, is that its 24EUR for a ~6"x4" single sided panel with a maximum of 4 pieces per panel. That works out to almost 40USD or $1.6 per inch. Work that out in BatchPCB prices thats about $60, but you also get double side pcbs, double soldermask, and double sided silkscreen. Olimex is also pretty picky/limited about drill sizes where BatchPCB works with anything over 20mil diameter.

The $100 I "imagined" is for Gold Phoenix (ultimately the board house BatchPCB uses) which is more like $90 for 100sq in panel. This is double sided, double solder mask, and they route as many boards as you have for free. An extra $30 means you can have up to 6 different designs that they will panelize for you.

Anyway, the whole point of this was just to let the original poster know that fabricating nifty production quality boards wasnt entirely out of reach. That was all. And depending on the deployment, ie if there were multiples needed or this was some sort of commercial application, a pro level pcb might be worth the extra work (to learn Eagle) and expense. Okay, thats all I wanted to say.

Cheers,
Brian

I sort of got interrupted and left off part of my post. I was actually going to argue for home etching. Even for simple stuff, it seems easier to deal with toner transfer and acid/peroxide etching than to deal with perfboard.

I've tried just about every way of making boards with home etching. I used the photo method for years and it was tolerable but not great. The press n' peel does work extremely well, but it does eat up a lot of time. I do especially hate drilling the holes, but the best way I've found to do that is with a rotary tool.