Below are pictures of my board (just to get better understanding how it looks in RL):
http://we.easyelectronics.ru/uploads/images/00/22/11/2012/01/09/cf9d56.jpg
http://we.easyelectronics.ru/uploads/images/00/22/11/2012/01/09/9d5e93.jpg
http://we.easyelectronics.ru/uploads/images/00/22/11/2012/01/09/acc84a.jpg
http://we.easyelectronics.ru/uploads/images/00/22/11/2012/01/09/8ca813.jpg
This is not most recent version, but main pin headers are at fixed locations for a rather long time now. So, in general, answer to your question is yes, this is possible to use my board like you're using Mini. Nevertheless there are two additional rows of headers and if you need these pins, you'll need also additional DIP16 socket cut through long side in 2 parts. And, of course, in this case both pairs of pin header rows should be soldered identically. I usually soldering them in opposite directions (as seen at pictures) and using inner headers for tiny shields (like 0.96" OLED).
Is Leaf and the Maple dead or dying?
(158 posts) (25 voices)-
siy
MemberPosted 3 years ago # -
siy
MemberRows in each pair are aligned to 0.1" pitch (distance between outher rows - 1.2", between inner - 0.7"), but rows from different pairs are not aligned to same 0.1" pitch grid, so both rows will not fit in same breadboard even if pins will be soldered in same direction. Using different breadboards for each pair must be possible (not sure if this will be convenient though).
Posted 3 years ago # -
Rod
MemberWider than 0.6 inches is fine. You just need to use 0.1 inch strip sockets that will take 0.025 square pins. I've actually been using the long 311 series mill max sockets as both lower pins and upper socket. It does need to fit in a breadboard, so the 0.1 inch pins need to be aligned.
I think the right design is just to take the existing Maple Mini, spread it out to 0.8, 0.9 or 1.0 and keep all the pins the same. And then install an F4 or an L1.
I'd be mightily tempted to add a winbond 1 uA serial flash for data logging, but I'd probably take a pill and calm down first...
http://www.mill-max.com/images/products/pdf/089N.PDF
http://search.digikey.com/ca/en/products/W25Q80BVSSIG/W25Q80BVSSIG-ND/2202664
At Digi-Key "Rectangular - Headers, Receptacles, Female Sockets"Posted 3 years ago # -
siy
MemberWell, my board mostly follows this approach, outer rows have 20 pins each. But I've tried to make as much MCU pins available as possible and to achieve this goal I did need few more header pins. Additional 16 pins serve this purpose and at present my board has routed to headers all MCU pins except VBAT (since board has battery holder this should not be an issue) and main crystal pins (should not be an issue too). All other pins, including DISC, NRST, BOOT0 and BOOT1 are accessible. Also, for both LEDs and RTC crystal provided solder jumpers which allow user to disconnect on-board circuitry and use pins in application. I've decided to add two LEDs (one connected as in Maple and one as in Mini) instead of one just because LED pin used in Mini supports PWM output (unlike Maple LED pin).
Posted 3 years ago # -
gbulmer
ModeratorFolks I have started to set up a repository at github for everything
I thought I would avoid any confusion, and not use LeafLabs or Maple in the name.
We have said software and hardware, so no 'lib' either.
I picked openstm32 because it seems to encompass everything, and be specific and clear.git@github.com:gbulmer/openstm32.git
I thought I'd come back and see if this is okay before going any further. It only contains a draft README.
Is this okay? Should I continue and add ala42's stuff? Any views on the name of that?
Posted 3 years ago # -
gbulmer
ModeratorRod -
I think the right design is just to take the existing Maple Mini, spread it out to 0.8, 0.9 or 1.0 and keep all the pins the same. And then install an F4 or an L1.
Please let me check I understand. As you know, the STM32F4 isn't in a 48pin maple-mini compatible package, so I assume we must be talking about 64pin.
So I think you are saying, don't bring the extra pins of the 64pin part out to the 'DIL' two rows of headers.
Instead maintain *EXACT* 1 to 1 pin mapping to maple-mini, the pins are further apart, but there will be no other changes.
Yes?Do you need the other pins, or are you okay with leaving them unconnected?
If they are left unconnected, I think 0.9" pitch is very doable. I started an 0.9" pitch, 64pin STM32F103 double-sided PCB, but gave up when we couldn't DIY the PCB, the track/space was too fine for us. The rules are widely available commercially, and a PCB should be under $6 (+P&P) in small quantity, but that wasn't my goal at that time.
0.9" pitch which would fit on the small breadboards I use.
Posted 3 years ago # -
Rod
MemberYes, exactly the same pin out, using the 64 F4 pin part, just wider spacing so the traces can get down to 7/7 and the one off board cost can be $20 or so. Then make a goal to get the software ported to that.
(Can't really see the point of DIY, when one offs are so cheap and fast (like 2 days). I am thoroughly addicted.)
Can't tell you how much I can do with the Maple Mini F1 just as it is.
And yes, leave the extra pins un-connected.
Its really tempting to add extra stuff, for example pads for a 32K crystal (wish they hadn't used those as I/O), but saying no is the real discipline that makes things work.
On the free version of Eagle? Is that what you meant by double sided. That would be awesome.
I can contribute an untested, 2 layer Maple Mini design as a starting point if that's helpful.
Posted 3 years ago # -
gbulmer
ModeratorRod - I have been talking to one of the senior tech support people at ST UK.
He says that the pin-out of the STM32's are deliberately designed to 'nest'.
So a 48pin footprint can fit inside a 64pin footprint on a single sided PCB.
I've had a look, and it seems to be true for parts in the same family. I have not thoroughly checked between 48pin STM32F1xx to 64pin STM32F4xx, but at first glance it looks correct (allowing for the power supply modification).
So, if we made a 48pin/64pin nested footprint, we could make one PCB which takes any of those parts. Like a SMD to DIL adapter on steroids :-)
Edit: I understand what you are saying about DIY vs commercial PCB's but the goals are to enable people to make things themselves starting at relatively primitive levels. We did make DIY double sided PCB's though.
Posted 3 years ago # -
gbulmer
Moderatorsiy - I am having second thoughts. Maybe have a second repository for hardware?
I am thinking we might end up with weird looking hardware test software which might confuse people.How about a second repo, called something like openstm32hw for hardware designs, and its hardware test software? The keep openstm32 relatively clean?
Posted 3 years ago # -
siy
Membergbulber - Yes, packages are nesting just fine. I got same conclusion while working on 100-pin version of the board. This also explains why there is no 80-pin version (although LQFP-80 exists and used for MCU, some LPC models using it, if I remember correctly) - 80-pin package will interfere with the 64-pin, while 100-pin can be placed without intersection.
Rod, your approach makes sense for purposes of creation of direct Maple Mini replacement. It might and might not be suitable for some particular purpose, so I think that boards created with both approaches will find their users. Also, I think that making all pins fit into 0.1" grid makes perfect sense, so I'm going to tweak my design a bit in order to follow this approach.
As for DIY: it makes perfect sense for me not only because it is much cheaper, but because I can get board in 3-4 hours (including solder mask) instead of week or so (fastest PCB service available to me). In two days I can prototype (not just get boards, but assemble and test them) at least two versions of design :)
Posted 3 years ago #
Reply »
You must log in to post.