The FPmini spent a good couple weeks in the auto router. It probably could run longer but wouldn't have any drastic improvements. The via count got down to less than 200 which is good for a board this size.
Quotes are back from the PCB manufacturer and were ready to pull the trigger. Our only decision left is to determine how many boards we want. An email went out to the group looking for a few additional builders/testers.
If you'd like to see images of the completed board use the links below:
Have you ever wonderd how all the traces on a PCB board are done? What's a trace you ask? A trace is a electrical connection that connects the electronic components on the board. Traces are very thin copper layers (think of them as flat wires) on the top and bottom of the board. More complex boards can have these traces sandwidtched inbetween the board to create three, four, five and more layers.
There is a fantastic tool out there called Freerouter (http://freeerouting.net).
When looking at the board's rats nest (connections from component to component) we see this:
Here is a screen shot from a running autorouter session as it is running and figuring out how to place the traces. The autorouter can take several days to complete a board. It performs numerous passes to optimize the traces. One function of the auto-router is to eliminate as many of the vias as possible. A via is a tiny hole that connect traces from one side of the board to the other. We like to have as few of these via's as possible. Another function of the auto router is to have the shortest path possible between components.
Pretty cool huh?
The auto-router generates a session file. This can then be imported into the KiCad program and you end up with a routed board. It's not quite the final version which is sent off to be manufactured but it's close. Here's a screen shot of the routed board from the design program (KiCad):
Good news! I just wrapped up the final schematic and board layout
changes for version two prototype board. I will still need to verify
all of the changes got included and nothing got left out so there's a
bit of verification to be done. Once that's complete I will start
auto-routing the board. Version one board took about a week to
auto-route however the actual time was more like three weeks. When
auto-routing is complete the board design is run through a DFM (Design
for manufacturability ) tool to identify any possible manufacturing
problems. This includes things like traces being to close together,
problem with ground and power plans, etc. Once any DFM errors are
corrected the auto-routing has to be restarted and the entire cycle
starts again. Since this is the second version I'm not anticipating
too many restarts.
We are feeling real good about prototype number two and are hoping it
will be the final prototype (oh, did I just jinx us by saying that).
Prototype PCB manufacturing takes about three weeks. Were hoping to
start building and testing by mid April. After the board has been
finalized I'll post the schematic and board layout for your viewing
One significant change we made was to remove the AC power switch.
Having AC current flow on the same board with low voltage TTL made a
some people nervous. All that current flowing through those large
traces can cause some serious damage should there be a short. It also
presented a safety issue since the power pads are at the top of the
board (I can testify to that. I made the mistake of grabbing the board
and learned some new respect for it). Most S-100 systems already have
an external power switch so I doubt it would be missed. We feel this
is a positive change plus it enhances reliability. It also freed up
precious board real estate which is needed to add additional driver
logic for the LED displays (future enhancement). This will help the
board run at faster clock speeds. The design is pretty much limited to
Hey guys sorry I haven't posted for a while. I got very busy at work and have had very little time to work on the FP Mini board. I'll try and post more often since it seems that there is some interest in this project. If you would like to be added to the mailing list and recieve updates and progress reports please subscribe to the FPMini mailing list. You can subscribe yourself by going to this Link.
The board testing continues to shake out errors. Several errors were identified in the schematic. Schematic errors can manifest themselfs in all sorts of ways on the PCB board. Some of these errors can be corrected by applying 'cuts' and 'jumpers' to the PCB board. Cuts are accomplished by taking a small exacto knofe blade and carefully cutting the copper trace thus breaking the connection. Traces are the small copper path ways on the top and bottom of the board. Jumpers are easier since all is needed is to sold small wires to the component pads and/or traces. Some of the cut and jumper modificarion's can be a challenge. Traces run on both sides of the board and are connected with what are called via's. Via's are small holes filled with solder to connect traces from one side of the board to the other. Your typical two layer PCB board has traces running left to right on one side while the other side run from top to bottom.
In addition to the schematic errors there were numerous board layout issues. Most noteably the toggle switches where oriented opposite to their function on the silt screen. Other minor issues such as some capacitor footprints were laid out as radial when they should of been disc. This was my first production PCB board. I'm sure that experience would of helped identify these obvious errors. To compound the problem I had to learn the KiCad software which is used to develop the schematic, board layout and routing.
If you recall from my last post, the switches where difficult to align squarely and solder in. I had made a jig from wood however it was a bit crud and did not yield good results. I needed a more precision jig that would hold numerous switches if not all. My thought was to build one out of metal or perhaps have one fabricated. I sketched up this cad design. I sent it out to a couple fabricators for a quote. Well as usual this was probably a pipe dream. The jig was going to be way more money than I wanted to spend. Besides I wanted it now.
I headed on down to Ace hardware to see what I could cook up. I found some nice little 1/2 inch 20 gauge brass stock. I was also able to locate some real tiny screws (2-56) machine screws. Its near impossible to get good precision drilling the holes by hand. I built a jig for my drill press to help align the holes.
The holes had to be on .400 inch centers. I have no measuring device that has a scale of 1/10 inches. I layed the PCB board on the brass stock and used a safety pin to mark the centers.
Here's a look at the jig after drilling the holes:
Well I got the jig made but as hard as I tried a couple of the holes where a bit off by 1/64" or so. Doesn't seem like much but it is noticable.
Here is the switch jig in use. I used it to install banks of eight switches at a time:
The board after all the paddle switches installed: