Archive for August, 2011
Even though there’s not much to tell, some folks asked about the differences in the new v3.2 uIEC/SD design. A picture is worth its weight in gold here, but I’ll also point out some less apparent details.
- Due to the new SD socket footprint, I was able to push the edge of the socket further from the edge of the board. This should help with implementations sitting behind thick plastic cases.
- Two small half moons (on the top left and bottom right) should allow the unit to be mounted in a Hammond 1551RBK enclosure.
- Although not populated on the PCB, there are pads for a Dallas DS1307 (or compatible) RTC with battery backup. The battery pins are shown on the right of the new PCB, while you can make out the watch crystal footprint below them on the right side.
- The LEDs have been pushed further outside the PCB. Truly, the assembly house went overboard on the first batch, but they should stick out to the edge of the SD socket.
Nothing else, I am afraid. I tried to add device jumpers to the design, but ran out of space and time to route the pads. The rest remains the sames, including:
- Pinout. v3.2 shares the same pinout as v3.1 and v3.0
- Mounting location. The mounting holes are in exactly the same place. Though the SD socket has mover 1/8″ further out, the PCB will fit in exactly the same place as previous designs.
- Same uC. The Atmel ATMEGA1281 is still in use, as is the 74LVC06 serial bus driver
Designing “piggyback” PCBs can prove challenging at times. For example, extending a piggyback PCB outside of the original ICs footprint must contend with physical obstructions (connectors beside the IC, passives like capacitors or resistors sticking out of the original PCB, etc.). Luckily, it’s often easy to print out a 1:1 version of the design on paper, cut to dimensions, and test fit onto the original board. Often, obstructions can be mitigated by switching to SMT components and using a smaller footprint or changing the PCB outline to for around obstacles.
However, less easy to design around are height issues. The best approach involves replacing the original ICs function entirely. Replacement solutions like ROM-el rise little further than the original IC they replace. If the original IC is still needed, but a pass-through of all signals to the original IC is acceptable, the design can utilize a socket both to contain the original IC, and let the pins from the socket function as the header for the piggyback PCB connector to the original socket. This solution typically rises a bit further, but rarely far enough to cause issues.
Designing a piggyback PCB that requires the original IC and must reroute some of the original IC signals forms the worst case. If the original IC height is defined as ‘X’, and we ignore the .8mm to 1.6mm of the PCB, this design still requires a header (‘X’ height), a socket for the original IC (‘X’ height), and the original IC (‘X’ height).
Mind you, most 70’s and 80’s systems designs left lots of room between board or above the original PCB. Thus, 3X height, while potentially looking out of place, will often continue to work. However, as designs became more compact, such luxuries are often unavailable.
The Commodore 1541II disk drive PCB is one such example of the last case. Commodore, in order to reduce overall drive height as much as possible, left little room between the control PCB and the drive mechanism mounted immediately above it. Designing a piggyback board for use in this device is particularly challenging, as there is simply no way to add any height; the drive mechanism won’t rotate the disk if the piggyback board touches it, and you simply cannot relocate the mechanism.
Months ago, I pulled a CMD JiffyDOS ROM out of a 1541II, and was surprised to see an unusual header being used to minimize height. Instead of a normal header, which is simply a batch of double ended pins stuck in a plastic socket, this header used a thin flexible “film” to hold the pins in place, adding no height at all. I had no idea of the source, but a quick request on the CBM Hackers mailing list yielded a source. Vanessa Ezekowitz tracked down the manufacturer (Advanced Interconnections) and the product (Peel-a-Way Removable Carriers).
Of course, this doesn’t solve every problem, but it does offer new hope for designs in height constrained locations. The solution is not cheap (10X the price of a normal IC header), but it can turn an impossible design into a possible product.
As of tonight, the last of the uIEC/SD pre-orders have finally shipped. In fact, for the first time since early May, we are caught up on order fulfillment. Now, I can relate some features of the new uIEC/SD daughtercard option:
- Two (2) IEC connectors. No need to ensure the uIEC is the last item on the bus
- 3 uIEC/SD connectors (one populated by default). One is designed to point backwards from the daughtercard (for a horizontal setup), while the other two are vertical. (This means users can reposition the unit for ease of use, or can utilize more than 1 uIEC on the same daughtercard)
- Integrated power plug. No more pigtail wire to break.
- RESET button on board.
- Selectable uIEC/SD RESET operation. Removing the on-board jumper will prevent computer resets from affecting uIEC/SD unit.
Of course, the original Daughtercard remains available for those who prefer a minimal approach. The original daughtercard works best for C128D/DCR users, while the new unit works best for other machines.
The new unit will be available as an option in the store shortly.