Assembled Minimig v1.1, larger FPGA (BGA-package)

[wesssper]

Hi,

I would be interested in an assembled board. I am ok with the original design since it's part of the fun to make a custom case for it.

--
Per-Erik

[freqmax]

@ThomasML:
I think you have to measure some existing Nano-ITX board :-)
(when via.com documentation is so nonexistant)

There are infact two "standard" layouts for the serial-ports (DE9).

Maybe we could arrange the RAM_xx signals in order for the v1.1 board aswell?

[yorgle]

i'm interested in one.

I use FPGAs at work, and it'd be neat to have those extra IO pins available on a header, so that other peripherals can be made for it (framegrabbers and such) or perhaps parallel port, floppy port, etc.

What kind of price point are you looking at for the board/kit/package?

I think i would prefer all of the io connectors on headers, but i'd be fine with on-board connectors

[EvilGuy]

Put me down for one, depending upon price. PM sent.

[Mikkel]

Why not go with an XC3S500E instead of the 3S400? It's cheaper, has more logic cells and is newer, so probably available longer.

[ThomasML]

freqmax wrote:
@ThomasML:
I think you have to measure some existing Nano-ITX board :-)
(when via.com documentation is so nonexistant)

There are infact two "standard" layouts for the serial-ports (DE9).

I'm leaning more and more towards using pinouts on the board, actually.  This will be the most flexible option (make two diff. layouts is too expensive :-(  )

Maybe we could arrange the RAM_xx signals in order for the v1.1 board aswell?

Anything special you have in mind?

--
Thomas

[freqmax]

@ThomasML:
I don't suggest making two versions. Just that you pay attention to which PinHead<->DE9 wiring scheme you will use.

As for the RAM_xx signals, just look at the current wiring:
http://www.opencircuits.com/Minimig_Board_v1.0_FPGA_connections
I think that if the signals are A0, A1, A2, A3 etc.. routing will be much easier and electrically sound.

If you go for the BGA route. I suggest you have the pcb board DRC checked at fab. And the FPGA-BGA assembled aswell at fab. However the rest people could do by themselfes.

[ThomasML]

Mikkel wrote:
Why not go with an XC3S500E instead of the 3S400? It's cheaper, has more logic cells and is newer, so probably available longer.

Unless the 3E is missing some features that the 3 has, and that's used by the Minimig, I don't see any reason not to choose the XC3S500E for the design.   Good catch :-)

--
Thomas

[ThomasML]

freqmax wrote:
@ThomasML:
I don't suggest making two version. Just that you pay attention to which PinHead<->DE9 wiring scheme you will use.

Ok

As for the RAM_xx signals, just look at the current wiring:
http://www.opencircuits.com/Minimig_Board_v1.0_FPGA_connections
I think that if the signals are A0, A1, A2, A3 etc.. routing will be much easier and electrically sound.

You're thinking about where the RAM-signals physically connects to the FPGA-package?

If you go for the BGA route. I suggest you have the pcb board DRC checked at fab. And the FPGA-BGA assembled aswell at fab. However the rest people could do by themselfes.

My plan was to get pricing on the fully assembled board (and that includes checking, but I'll get a quote for only assembling the BGA-package as well, but I think most people would want a fully assembled board).

--
Thomas

[McVenco]

ThomasML wrote:

but I think most people would want a fully assembled board.

How very true. Although there are quite some people who are not really afraid to do some basic soldering, I think the majority is pretty much "solderingly challenged" (just like me).

I very much would like to have a fully-assembled Minimig, no matter the size, as long as it works :-)

[freqmax]

ThomasML wrote:
You're thinking about where the RAM-signals physically connects to the FPGA-package?

Yes.
Less vias => cleaner design => less emi etc..
And easier to track wires if the need arise.

[Hans_]

McVenco wrote:

ThomasML wrote:

but I think most people would want a fully assembled board.

How very true. Although there are quite some people who are not really afraid to do some basic soldering, I think the majority is pretty much "solderingly challenged" (just like me).

Soldering a BGA isn't some basic soldering. AFAIK, it's impossible to do with a soldering iron; you need the right equipment to do it.

Hans

[alexh]

You cannot solder a BGA with a soldering iron, you need reflow oven. Prices start at about £150 and go up.

[koaftder]

alexh wrote:
You cannot solder a BGA with a soldering iron, you need reflow oven. Prices start at about £150 and go up.

Or you could reflow in a nice toaster oven, but thats for us hackjobs (:

[Hans_]

koaftder wrote:

alexh wrote:
You cannot solder a BGA with a soldering iron, you need reflow oven. Prices start at about £150 and go up.

Or you could reflow in a nice toaster oven, but thats for us hackjobs (:

Which is the technique that I'd use, if I had an oven set up for it. I'd be interested to hear of any "hackjob" method of verifying that all the solder balls are correctly soldered; you can't exactly inspect them visually.

@ThomasML
Are you considering putting a proper 24-bit RGB DAC on the board for video? I'd definitely be interested in the board if it included that. Such a feature would provide make adding AGA support a possibility. That would mean a certain amount of extra work though.

EDIT: Adding 24-bit RGB would be pointless without a 68020+ processor, so it's probably too much work for now, particularly as you're using the v1.1 schematics.

Hans

[Belial6]

OK, maybe this is a stupid idea, but what about having the BGA FPGA put onto a daughter card, so that other revisions of the MiniMig could use the same FPGA without the BGA worries?  Are the electrical tolerances to tight for that? :crazy:

Of course I certainly don't want to discourage a run of fully factory assembled MiniMigs! :-D