Interest check: An updated MiniITX-MiniMig ( Pics!!! )

[Illuwatar]

The MiniITX-MiniMig project did generate some cashback for me as I was able to sell off two of the prototypes. I also got requests to make more of them. But as the the prototypes does contain some errors I have started to correct these and do some major changes at the same point, making a version 2 of the board, suitable for production. For the original design, please have a look at my web page (www.illuwatar.se).

This thread is about asking for interest to make a larger run of boards. To create an order, at least 10 boards must be ordered to keep the price at a decent level. A initial run of 10 boards will cost at least €250, going for 25 boards will be €345.


Above corrections, following things will be modified or changed:

* The PIC18LF452 will be changed to a DIP-version, making it easier to reprogram in case of failed bootloader (or own development). The ISP-connector will be removed as it is useless when the PIC runs at 3.3V (it can't be erased and reprogrammed). The PIC18LF452 is a larger but 100% compatible version of the 18LF252 used in Dennis design, adding more I/O-pins for future enhancement.

Maybe even change to a PIC18LF4520 as the 452 is marked as "Mature Product" by Microchip.

* All connectors will be moved to the rear of the PCB as it should on a MiniITX. I managed to source what I need to pull this off at DigiKey.

* The LCD-connector will be removed - it was just that stupid idea I got sometimes. A generic header will be available for access to the added ports from the larger PIC. They are free for whatever fun You can think of.

* SD-slot moved to the board front. For off-board connection of a SD-holder, a pin-header will be available to connect to a break-out board with a SD-holder. Nothing should be at the board sides that needs external access.

* Changed several 1206 to 0805 - they aren't much harder to solder but it is easier to optimize the design with smaller parts.

* Improved ESD-protection at joysticks and PS/2-connectors.


What will NOT be changed:

* The form factor - it is still MiniITX.

* The color (of course it will be flat black).

* The core functions - it will still be 100% compatible with the Original MiniMig and it uses the same core parts (memories, FPGA, CPU & clock generator).

* 2-layer board, basically if you are skilled and have the tools, you can make one yourself.

* Integrated regulators - no need for a hard-to-find 5V PSU. Take whatever you have that gives out 9 - 15V DC.

* The S-Video/C-Video outputs and the buffered line-out audio. The buffered VGA contains a flaw in the current design that will be corrected. When done, three video outputs are available in the 15kHz mode (in VGA-mode/31kHz, the converter does not work - it is a limit in the AD724), showing the same picture simultaneous (good for demos at gathering).


Please note that this is not a build for a future MiniMig (there are others doing that) so I will not implement features like USB or larger FPGA. I will just take the nice design made by Dennis and move it into the easy accessible world of PC by using the MiniITX form factor.

[Illuwatar]

Thank you for that link - I will look through that one.

Regarding the LCD-interface - are You meaning that you will have direct access to the FPGA pins used for the video DAC? It could be done by placing a pin header just in front of the resistors. Just skip the resistors at the board and you will have an additional 14 FPGA I/O for your own joy.

Edit: do you have a link to where to buy these screens? It could solve the issue making that portable MiniMig I'm thinking of.

[Illuwatar]

The price examples are totals - for a board price, divide it with the board number. More boards are of course cheaper per unit.

My idea is to primary sell bare PCB as these are rather tedious to hand-build. These two I sold (complete, tested boards) went for €300 each.

A TFT is basically a LCD too, just a different way to control the pixels...

The header does not have to be a 2.54 mm spacing, hole mount part. I could use a SMD connector with tighter spacing if it would improve the EMI-part. Or just a line of pads with 1.27 mm spacing...

[Illuwatar]

Lot of good ideas, but many of them breaks the compatibility with the original design. And there is already an other designer working with a "Super MiniMig", doing a lot of the things you want:

http://www.amiga.org/forums/showthread.php?t=35270

My idea is more to make a solution that takes Dennis original design and make it more user friendly by changing physical aspects of the PCB and adding a few features.

That ARM-plugin board sounds interesting. I maybe should change back to the PIC18LF252 to make my board compatible with that extension.

Going smaller is possible - look at the on-going project Mini-MiniMig at my homepage. But to do that, a 4-layer design was required. The MiniITX is 2-layer, making it harder to shrink it to a decent size. Also, going smaller means getting rid of the connectors and replace them with pin headers. The MiniITX will fit fine in a standard PC-case - that's the whole idea of this project.

[Illuwatar]

All buggfixes are now implemented in the schematics. Still some connector issues to solve to be able to put all connectors at the rear. This correction release will spawn an entirely new PCB layout. By moving all the connectors, I have to basically re-route the entire board. That is actually good as I can optimize the design even more, specially the power supply will go through a rather serious change. I will also add this header/connector that allows access to the RGB-ports at the FPGA (for direct LCD/TFT connection).

For the PIC, I will make it possible to use both the 252 and the 452 (not at the same time of course). This will make the board really 100% Dennis-compatible (with the 252) or you can go for the larger PIC and have a few more port for fun. With this solution, this ARM-expansion will fit my board too (not just the Dennis original one).

[Illuwatar]

yaqube - that would be nice. You could send both file formats, but the AutoCAD version would be preferred in case you have to select.

[Illuwatar]

Update:

Got the drawing of the ARM-board from yaqube - many thanks!

The PCB-design have been started, currently laying out the parts in a more optimized way than before. It will be an entirely new board as none of the old layout will be re-used. And I managed to put all connectors where they belong: at the rear face of the PCB.

[Illuwatar]

The project is ongoing, it just takes some time as I have to split the time with other things. The board is routed to about 90% (I do it entirely by hand).

[Illuwatar]

Board is 100% routed and it passes the DRC-testing... :-D

What left to do are some minor tweaking of the ground planes, then they are done for manufacturing.

Before actually making any boards, I need to get some sense of interest. A few have indicated that they will buy boards, but should I make 10, 25 or maybe more? Total cost will of course increase for each board added but the price per unit will decrease. Before shelling out for a 25-board run, I must be sure to sell them all (except for a few I keep for my own joy). Calculated prices based on board size and manufacturing parameters are: €25/each @ 10 boards, €14/each @ 25 boards. This price includes shipping from China to me but not from me to You. Also, there are zero-profit prices - the job is done for free if I sell for these prices. I'm unsure what is fair to take as a "reward" for the work making these boards - the Amiga-scene is little bit like a family...

Maybe the first run should be like five boards or so to prevent a large economical damage in case of board errors found during testing. If the board runs fine, I can do a second order. One thing i sure - I'm not going to sell any untested designs.

Edit: What color should i choose? Keep the flat black or select something else... :-D

[Illuwatar]

I don't think I will delay the board even more by re-designing the memory section. Swapping the existing SRAM's to faster ones with the same pin layout is fine, but not re-designing the board. My other board (the Mini MiniMig) will be using 10 ns SRAM (1M x 16), but that project is currently on-hold until some other projects are done (that includes this Mini-ITX board).

Regarding the colors, I can choose from: standard PCB green (flat or shiny), black (flat och shiny), blue, red, white or yellow. The text could be white, yellow, black or red. My personal choices are black (flat), blue or red (in that order). The colors aren't changing the price of the boards (green may be a little bit cheaper).

[Illuwatar]

It seems like the first initial board run can be sold without problems (judging from the PM's I have received). I will do a "debug-run" of five boards first. And I will do them flat black as I did with the previous version (the blue PCB should stay as a trade mark for the GBA1000 project). I will put a note in this thread when the order is done (it will take about two weeks form order to boards in my hand).

[Illuwatar]

The first run of five boards are ordered now. In about two weeks I will know if this is a successful project...

[Illuwatar]

The first set of five boards has arrived. So, now I will start the hard work of assembling and testing. And I have already found a fault in the design - a mechanical one that I could bypass with some brute force... :madashell:

So, if these boards are working fine, I will assemble all five and sell them as complete units as these boards are not suitable for public release. A second run will be done with all corrections included and that one will be (hopefully) sellable.

A part order have been done too - to three different places to get all parts needed for a complete board. If it works fine, I will order parts for the remaining four boards.

Pictures will come, both here and at my homepage...

[Illuwatar]

The pictures:





The currently known problem is related to the VGA/PS2-combo connector. The holes aren't properly aligned with the connector so I have to change the footprint a little. With some cutting, bending and forcing, I was able to mount it anyway so this board batch could be used for further testing.

[Illuwatar]

The problem with the back cover plate is manufacturing - everything I do that you see here is made in a home environment with tools you could find in the nearby hardware store (except for the PCB manufacturing that is done by www.pcbcart.com). Mass producing sheet metal parts like these plates are out of my workshop's capabilities.

Fully populated boards are currently build by hand, so it is impossible for me to compete with companies like ACube in both production speed and cost. The first five boards will be built and tested (if I don't find more problems that affects the functionality), then I may put them out as auctions to see what people may pay. The corrected version will be available as bare PCB's for those who like to put it together by themselves.