Page 3 of 5

Re: stm32f4 MPU module LQFP 100

Posted: Sun Dec 25, 2016 3:22 am
by kb1gtt
So how would you connect the USB? If this is to plug directly into the board, it would be hard to get the laptop connected to the USB. So it will either need an USB cable, or a ribbon cable. If it has a ribbon cable, would that ribbon cable mirror it? Would we expect they use a USB extension cable? Perhaps something like the below.

http://www.ebay.com/itm/2ft-60cm-Black- ... Sw6n5XrQ9C

Re: stm32f4 MPU module LQFP 100

Posted: Sun Dec 25, 2016 12:48 pm
by kb1gtt
@russian, you got it through my thick head. The SWD is now mirrored and looks like this. It's kind of annoying because KICAD's pin 1 and the programming device's pin 1 are mirrored. But as far as I can tell this is now routed properly.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Dec 25, 2016 2:09 pm
by russian
Thank you Santa! :)

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 01, 2017 2:36 pm
by russian
8) Check if there is a chance to match discovery board reset button location

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 12:35 am
by russian
Got the boards and we have a problem with vertical USB connectors - instead of slots they've produced little holes :(

Damn. Emailing oshpark support.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 1:01 am
by kb1gtt
From here http://docs.oshpark.com/submitting-orde ... and-slots/
As mentioned above, there is no guarantee that this call out will produce the expected results, and the use of plated slots is discouraged.
It appears that OSHPark has some special and less standard expectations for plated slots. Even if you follow some funky rules, it's kind of a crap shoot if they happen to notice it's a plated slot.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 1:21 am
by kb1gtt
As well, even if you follow the odd OSHPark rule(s), and you happen to have a slot that's less than 40mils wide, you still get drilled holes instead of a slot. In our case the slot is 20mil wide, and making the slot 20mils wider is a problem. See the original below. I'm not sure that OSHPark can MFG this board. Either that or we have to commit to either horizontal USB or vertical USB, but it appears we can't get both from OSHPark.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 1:23 am
by russian
discovery: button center is 27.5mm above lowest 25x2 pad hole center
discovery: 24mm between button centers

brain 0.2: button center is 24mm above lowerst 25x2 hole center
brain 0.2: 17.5mm between button centers

on the brain board, can we lift buttons 3.5mm higher?
on the brain board, can we place the buttons 6.5mm futher apart?

we only need oshpart for prototyping. I am planning to snip these stands off the connector and see how good it would hold with just the five vertical terminals

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 1:30 am
by kb1gtt
Per the below thread, OSHPark is aware that slots use a G85 command, and as of 2015 they were working on flagging this issue. However apparently that has not happened yet. See the drill file, which includes the G85 commands. Yet you got no warnings about this problem. This issue could be easily programically checked for and could easily prevent bad productions. I do not know why this is not part of the Dynamic Rule Checking (DRC). It's a simple items to check for when compared to the other stuff they do check for.

https://forum.kicad.info/t/oval-pads-no ... park/835/5

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 2:36 pm
by russian
Another drilling issue - see how red Frankenso five holes are the right size while brainboard board five holes are smaller?

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 3:25 pm
by kb1gtt
Can you measure your wire lead?

I see Frankenso R0.4 has the drill at 0.015, same for the OSHPark. So I expect them to be the same. Is the fat-est part of your lead 0.015in? It's possible that the Chinese board used the 0.15 as the finish diameter, and that OSHPark used it as the drill, such that plating makes it smaller.

As far as I can tell, these hole diameters should have been the same.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 15, 2017 9:06 pm
by russian
Why so small pads for D5 and else Ds?

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 8:34 pm
by russian
Getting closer to reality! With the SD grand-daughter board we will have logs and tuning access via the top plastic cover.

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 8:52 pm
by sleepingAwake
This is relevant to my interests.

Can we put the the MCU right on Frankenso 0.4? I was wondering if i could just take off the discovery board and installmit right on the main board.

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 8:55 pm
by russian
sleepingAwake wrote:Can we put the the MCU right on Frankenso 0.4? I was wondering if i could just take off the discovery board and installmit right on the main board.
This was done. Once. Two years ago :)

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 9:49 pm
by tomiata
russian wrote:Getting closer to reality! With the SD grand-daughter board we will have logs and tuning access via the top plastic cover.
Where can I get one those SD card thingies, and how does it connect to the board?

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 9:59 pm
by russian
tomiata wrote:Where can I get one those SD card thingies, and how does it connect to the board?
Pictures coming very soon. Granddaughter board only works with rusEfi own brain board (not discovery), but actually now that you ask we can make a little adapter board for the SD card granddaughter board.

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 10:30 pm
by sleepingAwake
russian wrote:
sleepingAwake wrote:Can we put the the MCU right on Frankenso 0.4? I was wondering if i could just take off the discovery board and installmit right on the main board.
This was done. Once. Two years ago :)

Anyone reason it wasnt pursued further?

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 11:16 pm
by russian
Lack of time & capital. Two simpler and cheaper boards is easier for now.

Re: stm32f4 MPU module LQFP 100

Posted: Sat Jan 21, 2017 11:37 pm
by kb1gtt
It's kind of funny, the discovery board can be purchased for less $ than we can populate it's components in low qty. So even though it can be put on the PCB directly, we don't typically go that route because it's lower cost. This would change if we ever end up with higher quantity builds.

On the rusEFI version of the STM32F407 brain board, we added some headers on the edge of the board. Then we made an adapter board for the SD module which allows putting that SD card up through the case like that. This is currently a first spin kind of thing. We are currently waiting for some boards from China. However the initial mock ups are looking good.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 4:04 am
by russian
0.2 brain board is a total success

@ reports
1) vertical SD with vertical daughter board works
2) st-link 5x2 pinout is right
3) vertical USB pinout is right
4) RTC Xtal works!

Since we are waiting for the end of Chinese New Year anyway, we will have a 0.3 revision with button locations matching discovery.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 4:10 am
by russian
tomiata wrote:Where can I get one those SD card thingies, and how does it connect to the board?
https://oshpark.com/shared_projects/KCdXUUDD link added into top post, but as is this would only work with our own brain board, not discovery.

If we make another little board matching the red outline, such a board could be soldered flat onto discovery and allow mounting of the existing vertical SD card board in a similar way. Need to double-check where are all stm USB pins are located - maybe same adapter board could be used for vertical USB placement or routing USB to the vertical board.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 12:04 pm
by kb1gtt
A key feature of the discovery is that it's a replaceable board. Which is handy for diagnostics and such, as you can isolate the board from the Frankenso. I would think if we are looking for a vertical on the discovery instead of the side mounted one on the Frankenso, we would probably want a connector similar to this.

http://www.digikey.com/product-detail/e ... ND/1111298

That could be put on top of the discovery. Would this work?

If we use a soldered discovery, as noted in @russian's graphic it would jack the SD card up about 0.063in (1.6mm) Would that be a concern?

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 1:21 pm
by russian
kb1gtt wrote:connector...That could be put on top of the discovery. Would this work?
...jack the SD card up about 0.063in (1.6mm) Would that be a concern?
top side pins are pretty short and inserting/ejecting an SD card takes a surprising amount of force. I suspect a connector just would not withstand.

1.6mm lift is probably fine since we can always face the angled mounting pins down and lower the vertical board by that.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 4:52 pm
by kb1gtt
Do you mean something like this?

Image

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 4:58 pm
by russian
Yes!

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 5:20 pm
by russian
Can we make it less sexy but help reduce the chance of twisting the board while inserting/ejecting?

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 6:25 pm
by kb1gtt
Something like this?

Image

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 6:35 pm
by russian
Yep, in this direction. I can either order as is or we can add some at the bottom pin area.

Re: stm32f4 MPU module LQFP 100

Posted: Sun Jan 22, 2017 7:07 pm
by kb1gtt
Something like this?

Image