Puma board for FreeEMS

[GartnerProspect]

I just created a footprint for the combo u-mini-header, and pushed to github.

About all this stress relief on a data connector, shouldn't we jump the wireless gap? I'd still like to see a chip we can put on puma, that emulates 232 to the MCU, and a software driver on a PC that emulates a 232 connection, but have the actual medium be 802.11. Such that we simply tell the existing soft to use say com9, which a driver then translates into 802.11, and sends out on your laptops wireless. Then the chip receives the 802.11 signals, and relays 232 to the MCU. I'm sure such a device has to exist, but I haven't found it yet.

I could be wrong but it seems like the issue with that is the software stack overhead required for all the handshaking and authorization. Plus Wifi isn't exactly great for AdHoc peer-to-peer. But I do agree that wireless is worth investigating. Seems that Bluetooth is better suited and I know there are products out there that utilize it for what amounts to a wireless serial port. Unfortunately, nothing is quite as ubiquitous as 802.11.

Also, Full Disclosure: I haven't read the entire 50 pages of this thread yet...

[Fred]

Also, Full Disclosure: I haven't read the entire 50 pages of this thread yet...

Nor have I! I ordered my boards and told the world about them simply on the basis that Marcos seemed like he had a few clues. Now he has a few more, spin 2 will be better :-)

Fred.

[Fred]

Also:

Mate, where the hell have you been for the last 6 months??? :-p

Too many projects, too many women! :mrgreen:[/quote]
LOLOLOL :-)

[DaWaN]

Wireless isn't really important at this point in my opinion.

Doing Wireless properly requires using a decent standard in my opinion.
Using a decent standard means having some sort of software stack, something that should not be added to the main processor at this point in my opinion.
At this point I think only Bluetooth offers a industry standard serial connection, so I vote for that one.
Maybe 802.15.4 (zigbee) can be used if it's fast enough, will bring down complexity and BOM of the ECU itself, but it will also require a 802.15.4 dongle and some non-standard network and application layer.

[jharvey]

Here's a couple quick pictures of the connector combo.







Understood, you haven't read the 50 pages of this thread, or the pile of posts in DFH/ect. If it's really important for communication purposes, it should be put on the wiki, or web page. I'm not one of those people that will tell you to go read hundreds of pages to figure out stuff I can relay in a couple words. I don't think that's beneficial to progress. However, I would like to encourage skimming the pages, and I typically recommend starting with the wiki. The best thing do to is to check out the repo's, and critique away. I'd be happy if someone notes that resistor blah, would be better as value blah, because of x,y,z.

About wireless, I tossed it out there to scratch around. I do not plan to add this to the design, I doubt Marcos will plan for it either. However, if someone comes up with an option, it might be handy for the add on card. You know the one that includes the connector.

About the header, if you don't put in the USB connector, you can always install a header. It just won't be a nice straight line like you might be used to.

[GartnerProspect]

Oh, I'll get to it eventually. I had a bit of a hiatus and need to get back into the swings of things.

I like the look of the connector combo - ALOT. It gives flexibility but also if somebody was dead set on a panel mount port, it looks like it would be very easy to use the through-holes for header pins. What is the spacing on that feature? I also really like the spec of USB connectors with through-hole mounting. Should greatly ad to the robustness in my opinion.

Win Win. What's the next challenge

Also to spare me digging too long for it, what PCB-CAD and where is the repository? It looks like Kicad to me, but I've been wrong most my life. I'll have to see how Kicad works on OSX, I've only used it on Windows but now spend most my time on a Mac... at least for now.

[Spudmn]

Great job jharvey. That should keep every one happy.

[jharvey]

I like the look of the connector combo - ALOT. It gives flexibility but also if somebody was dead set on a panel mount port,

The only potentail down side I see with the combo is that the mini connector would probably want an extra insulation step during assembly. The pads for the micro could potentially short if you don't add some form of insulation, but I don't think that's a hug problem. A dab of paint or tape should work well. The solder mask thickness would create a natural air gap, but I would prefer a solid insulator in case things get bent a bit.

it looks like it would be very easy to use the through-holes for header pins. What is the spacing on that feature?

.1 in just like the other headers. Should allow for standard connector options.

Win Win. What's the next challenge

I think Fred would say noise analysis / robustness tests of the circuits, and detailing information about spin 1, such that folks can build their spin1 boards. I would say addressing misc items detailed in the spin2 document would be a good focus. Such that we can get the next spin in the works. Marcos is probably the best one to do those changes, but is limited in time as he's currently in school and has exams and such work loads to deal with. So I think I'll hear your comment as an inquiry of how can I help? How about drafting up a connector card? I'd say that's a good place where you might be able to help, do you feel OK with drawing it up in KICAD?

Also to spare me digging too long for it, what PCB-CAD and where is the repository? It looks like Kicad to me, but I've been wrong most my life. I'll have to see how Kicad works on OSX, I've only used it on Windows but now spend most my time on a Mac... at least for now.

You are correct KICAD. Marco's repo is the official copy, but I have a repo that I use to pass work to Marcos. So until Marcos gets to merge my work, you would really need to look at both repos to get all the parts. I believe the connection points between the processing board and the connector card are solid, so you could probably do a connector card with out worry of things changing on the processing board. So if you check out Marco's repo, you can modify it into a connector card.

https://github.com/nitrousnrg/puma
https://github.com/jharvey/puma_JAH

[jharvey]

Hmmm, here are some options for those that are concerned about the USB cord.

http://www.amazon.com/Cables-Go-29571-T ... 04&sr=8-13

http://www.amazon.com/Cables-Go-29570-T ... 904&sr=8-8

[Fred]

Win Win. What's the next challenge :-D

I think Fred would say noise analysis / robustness tests of the circuits, and detailing information about spin 1, such that folks can build their spin1 boards. I would say addressing misc items detailed in the spin2 document would be a good focus. Such that we can get the next spin in the works.

You just earned some points with me! Best thing you've ever written? Possibly not, but I like it! :-) The spin 1 docs are crucial because if we piss off the people that put their hands up, spin 2 will be harder to push. Not to mention software can't get tested without people trying it. 3 years has proven that NO ONE tests shit on the bench, and 2 days of trying to get people to pickup boards with no known quality level produced a massive amount of interest. It would be fair to say that I fucked up there, but my intentions were good... SO, let's get people testing software and spin 1 boards (BIG priority) and then worry about making spin 2 happen WHILE THEY build and test spin1. I think that's the only smart way to roll, personally. If too many people get boards and I have to change routing, it looks bad. So we need to get some specific units of work done before moving onto that.

Marcos has been awol for a week or so, but he did warn us, and he is working on documenting stuff while he's disconnected.

Fred.