liam.ll - Thanks for the link to that vertical adapter board before. Brilliant!
I assume this is the PIC 32-bit Whacker: http://www.schmalzhaus.com/UBW32/
Food for thought
liam.ll - Thanks for the link to that vertical adapter board before. Brilliant!
I assume this is the PIC 32-bit Whacker: http://www.schmalzhaus.com/UBW32/
Food for thought
gbulmer - You found the correct PIC Whacker. Narrow and full use of input/output pins are very attractive. On the other end of the spectrum, I've seen companies emulate the 24-pin Basic Stamp with a 32-bit mcu but then end up eliminating input/output pins. Very wasteful. It seems designing a longer module actually simplifies fitting components. Here's a good example with 64+ pins:
http://www.alvidi.de/avr32_module_bmb.html
I want to point out that a major turnoff with the Whacker and AVR32 module is that they don't offer the Arduino-compatible format. I'm holding out for the Mini.
Eh, sorry for not responding to this more quickly. There's a little confusion about the "Maple Mini HD". It's not a LeafLabs project, and we're not actually sure who added it to the wiki -- I'm leaving it up for now with a request for additional information. But right now there's unfortunately no plan to produce anything like the "Maple Mini HD", or bread-boardable versions of the other boards. Production of Mini should be wrapping up, and we're hoping to have them available in the next couple of weeks. Native's a little more up in the air -- we're waiting for one more prototype to come back, and if everything works as expected it will be going into production after that.
I see the Maple products are assembled in China. As far as parts on the Maple boards, the microcontroller has links to Japan; STM expanded their operations there in the early 90s. Will Japan's tragic events affect Maple products, particularly the Maple Mini?
(gbulmer) I think it's gonna be very hard to get a 64-pin part into the same 0.6" width as the Maple mini.
The mini's 48-pin LQFP part is 7mm x 7mm, or 9.9mm across the diagonal corners (almost 0.4")
Looking at the pcb (http://leaflabs.com/wp-content/uploads/2010/02/maplemini-r21.png) it is very tight. I am impressed!A 64-pin LQFP part is 10mm x 10mm, or 14.1mm across the corners, almost 0.2" wider than 48-pin, over 0.5".
It might be doable with a Ball-Grid-Array (BGA) part, but I think that using BGA packages make the board harder to manufacture. Even then, it would be a lot of signals to route in a very narrow area.If you'd accept something much wider, but still breadboard friendly, then it'd be very doable.
What would be an acceptable width between rows of pins? 1.5 inches?
If we try to keep to the "nano" form factor (as small as possible). Then on a single regular breadboard, 0.9" would leave exactly 1 row of pins free on either side of the stamp. Which is the upper width limit for a single breadboard. Which is big enough to fit a 64-pin chip. Its enough room either side to place a jumper cable to a 2nd adjacent breadboard and still only 1.5x wider than the Mini. You could keep the 40-pin here, but be sure to give priority to ensure all the extra handy UART, DAC, ADC and INT are fully broken out. The Maple Mini is good form factor but has already maxed out its future RAM limit in its first revision. We will desperately need more RAM for TCP/IP stack, video, audio etc. So 64-pin 0.8" wide shall probably be the only available option (for that form factor) in future. Note: as far as we are aware, thats exactly what this mythical "Maple Mini HD" clone claims to be. Personally would not be so concerned about getting 100% all of the pins broken out for these nano form factor. So long as all the UART, DAC, interrupt etc are fully represented. For more pins we should diverge to add a 2nd larger stamp form factor. Some breadboard friendly equivalent of the Maple Native.
For these wider stamp form factors you're always going to straddle 2 similar breadboards side-by-side. Any 2 similar breadboard. For perhaps the widest common BB - its dual-power line on either side. Those require a MIN stamp width of 1.2". So gbulmer's suggestion of 1.5" isnt too bad idea. In fact any stamp wider than 1.2" will work, since 2 adjacent breadboards aren't required to be butted up right against each other. It always leaves the maximum of 4 available pins in the rows to which the stamp sits. Thats assuming there is no over-hang of the pcb. Anyway the remaining 2 (outer) rows are definitely completely free.
In fact - why cant we just change over the female header pins from the maple native to be male pins? How wide is the Maple Native compared to our ideal stamp width of 1.2"? Its approx 1.9". Therefore you (a 3rd party vendor probably) should just take the finalized r1 native and tweak it to be as narrow as possible, including rotating the MCU by 45 degree similar like the Maple Mini design. And in the photo of the Native, whats those 3 rows at the end used for? (not-broken-out). They make the native wider than it needs to be. And are on wrong axis to break out to breadboard. So they must be eliminated, or re-positioned if any remaining space available after duplicating the Mini's killer silkscreen. That would be a worthwhile future progression in terms of the form factors, I rekon.
dreamcat4 - I could go for 1.2".
I said 1.5" because I have a 1.5" wide STM32F board from ETT in Thailand:
http://www.ett.co.th/product/ARM/ET-STAMP-STM32.html
which is available through Futurlec, so I know it is practical
It is a bit weird because it doesn't have a USB plug, but does have a proper RS232.
Which type pf pins would you be willing to forgo on the 'stamp-format' 0.1" rows?
I assume you are happy with USB only appearing on the USB socket!-)
(gbulmer) Which type pf pins would you be willing to forgo on the 'stamp-format' 0.1" rows?
Clearly the least-useful pins are those which are not internally connected to any special register (eg UART, SPI, INT, DAC and so on). Personally I've come to accept that the 40-pin form factor of the Mini is usually good enough. Otherwise for more pins the discovery board is now working and is 1.9" wide.
The drawback of both boards is 128k/20k of flash/ram so thats probably going to be the largest incentive to make a new board. As previously cited - the best way is re-working MapleMini pcb design in Eagle with a view to accomodate larger LQFP64 sized chips. I was initially thinking remain at 40pins to fit a 50mm x 50mm size restriction for cheap(est) pcb manufacture ($2 per board, 10pcs).
However the Mini is on 4-layers so that particular PCB scheme (2-layers) would not work. Indeed considering further, when you open in Eagle files of the Mini it becomes clearer just how cramped the on the pcb space is. Everything on it is at such a high density. Its worth considering the possibility that other PCB manufacturers may run into difficulties to be making this kind of board. If you were to use the same manufacturer as leaflabs however then you'd probably be allright though.
It makes me wonder if we cant just take STM32 Discovery board, use hot air gun to remove the existing chip, and replace with STM32F103RET6. ST Microelectronics page says the 128k chip on the Discovery is a STM32F100RB (rather than the more similar STM32F103RxT6). So am hesitating to state they're compatible. I just wonder what the difference is exactly.
(gbulmer) I assume you are happy with USB only appearing on the USB socket!-)
Personally yes, however USB D+ and D- are indeed broken out on the Mini. Some people may see benefits in keeping MapleMini pin compatibility. Its probably too early to assume that everybody else is necessarily going to be happy with that kind of a choice. Given that the easiest way to build this board is to start with the Mini rev2 layout, then its going to be a sensible tack to defer such decision (changing / replacing pins) until its clear in Eagle what the alternative pin options are for replacing them.
dreamcat4 - I am working on a 64-pin STM32F103 board which is not based on the mini. It is a two layer board.
I've built prototypes, and 'learned lots of lessons' (aka made lots of mistakes :-). So I am not planning to change that to base off the mini.
I see your point about maple mini pin compatibility, but they aren't footprint compatible, so it seems like less of an issue at the moment.
Currently my board has 60 pin headers, but I was wondering about 48-pin, 2.5" x 1.5" PCB.
With USB, JTAG, etc. removed from the header pins onto other sockets, it would fit (Maple rev 5 is only 44 header pins). I wouldn't like to lose 8 more to reach 40-pin unless there is a significant benefit.
gbulmer - Sounds great. Really fabulous. Then doesnt really matter (about the pins). I still suggest aiming for 1.2" width is well worth it. The board will feel significantly smaller than the current alternatives (which are 1.9"). By comparison, 1.5" sits squarely in the middle therefore not as compelling as the mini.
Along the other dimension:
40pin = 1.9" tall
48pin = 2.3" tall
60pin = 2.9" tall
Any of those choices will not impact the board area as much as reducing the board width. Therefore if possible, I would suggest to try to explore that possibility first of all. It would make best sense to do so before positioning those top-sockets (for dropping down to 48pin). As (im just guessing here) you really would want to fix the width for doing that.
I wouldn't like to lose 8 more to reach 40-pin unless there is a significant benefit.
Since different pins to the mini then theres no pressing reason to go as low as 40pins. The only other cited benefit was related to cost of pcb manufacture. Fitting into the seeedstudio 5cm x 5cm constraint. (http://www.seeedstudio.com/depot/services-c-185.html). Perhaps a non-issue if seeedstudio allow multiple boards per piece. Perhaps that ambiguity warrants contacting seeedstudio.
In the case that we do end up with a 1.2" board. Then it would be good news indeed because their 10cm x 10cm process work well for 60pin, 1.2" wide. A 48pin (thats 2.3" not 2.5") x 1.2" design would fit 4 boards per piece with very little wasted area and allows the unit cost (for 4x10pcs) to drop down to only $1. So in response to your question - yes, it seems there are reasons to reduce down from 60 to 48-pin. And more breadboard line to play for your circuits since youre moving the less frequently used pins "out of the way" onto top sockets.
Generally in balance, the benefits of 48pin (2.3" x 1.2") design would be highest out of all those options. One question... Do you share on Github? that would be awesome. Or perhaps you choose to keep such design closed. Then I also wish you success and hope you can take away from these suggestion positively. Many thanks.
dreamcat4 - a 1.5" wide PCB would be 1.4" centres, which is the maximum we'd accept (at the moment :-).
I am aiming for 1.1" centres on the basis that we did a "breadboard calculation", and for us, 1.1"/1.2" is "the right width". But (on a 60 pin board) it is a tad tight under our other constraints. My 60 pin board brings out all signals onto the 0.1"headers, which is not essential, and hence is why I was pondering 48/50 pins, and asking questions. Ditching some header pins may be enough to get it to route.
40pin = 1.9" tall
48pin = 2.3" tall
60pin = 2.9" tall
I disagree. A 48 pin PCB would be 2.4inches + border clearance.
The 24-pin headers would be 2.4", not 2.3", to allow for the plastic structure of header pins or sockets, and a bit of PCB to drill for the pins. Similarly, 1.4" pitch would be a 1.5" PCB + border.
I'm trying to avoid targetting the design too much to a single PCB maker, so the design rules are relatively tolerant while I am making prototypes.
My tolerant design rules have a 1mm border clearance around the PCB to allow for separation (scoring or routing). This reduces the chance of accidentally shaving into a copper track, and causing a short (My 'guru' says he had that happen once, and it was a nightmare to track down, and worse, can happen in use). Once the prototype is right, and I am looking for a production supplier, I'll likely revisit the width of border clearance.
This means that 1.1" pitch board would be 1.2inches + 2mm in prototype.
It could end up 1.2" pitch. That would translate to 1.3"+1mm x 2.4"+1mm PCB, with a 0.5mm border, depending on the manufacturer, for production runs.
I haven't yet sorted out the license for the PCB. Part of the goal is to make them for an educational charity, who have some constraints. So maintaining closed source makes it easier for me to make progress until we can figure out the "whole game plan". A problem is, once 'Open Sourced', it is hard to change the license. Sorry for the vague answer, but it may change.
gbulmer - do you have pictures of your 60-pin board design? I get the impression it will take some time before you release anything. Will it be Arduino compatible? Will it be under the Maple brand?
I posted this "Mini vs. Netburner" topic in late December anticipating a Maple Mini release in late January. Delays are natural and expected with new products. But we're well into April now and my project can no longer wait - I need to go with the Netburner 5213 module.
I'm still rooting for the Mini and will purchase it when ready (and maybe wait for a rev2).
lliam.ll - I have pictures of the existing board, but it far enough away from where we now feel we need to get to, I haven't posted. I keep meaning to update my blog, so maybe I could fix that?
It was not Arduino compatible. It is designed to be a self-contained robot controller. I use robots as a focus to introduce people to electronics and programming. Some folks become very engaged when they discover this stuff; one of my 'students' has won UK championships.
You are right, it is still some way away. But the idea of a 48pin board may make it easier and small enough that some of the other constraints may be soluble.
Unfortunately, Mini isn't quite what we need.
(full disclosure: I am not a member of the LeafLabs staff)
liam.ll - Here is a link to my blog, showing a version of the original OROne board.
(full disclosure: I am not a member of the LeafLabs staff)
gbulmer - regarding the OROne, is it for sale? Do you have a breadboard-able stamp version? Any plans for one?
leafblowers - I see that you plan on testing the Mini this week. Have you decided on price?
First release price of Maple Mini is going to be $34.99, but we are going to do a really good launch promotion so stay tuned ;)
You must log in to post.