Archive for December, 2010
Ingo Korb has released a minor update to sd2iec (used in uIEC, among other solutions). Users are encouraged to update to this release. Fixes/Features include:
- U1/2, B-R/W commands now work without a trailing CR character (reported by Draco)
- UI now always reports the dos version no matter what characters are following unless it’s a + or -. Previously anything but ±/nothing would result in a SYNTAX ERROR which doesn’t match what a 1541 would do. (reported by skoe)
- JiffyDOS timing optimized, now loads a massive 1.8% faster!
Although parts were delayed due to a Blizzard in the Midwest this past weekend, I was able to construct ZoomFloppy #1 Monday night. Though one hopes for first attempt success, that is not often the case. It was indeed not the case for this construction.
When I designed the board, I knew the 0402-sized surface mount components were small, 1/4 the size of the 0804-sized components I normally utilize for SMT designs. Still, it’s a bit academic until one actually tries to solder the parts. At a size that is seems near my eye’s minimum ability to resolve details, the components truly tested my soldering abilities. For comparison, my finest iron tip is 1/32″, and the parts were about the same size as the tip. Still, I was able to place each component on the board.
When first powering up the unit, it did register as a USB device, and I was quickly able to load the required firmware. However, upon re-insertion, the unit registered as “xum1541 (ZOOMFLOPPY)” and demanded USB drivers I did not have. Since it was late, I left the project at that point and solicited help from Nate (the project designer) and others.
Tuesday night, I had learned what version of OpenCBM to load on the PC, the correct USB drivers had been sent to me in a ZIP file, and progress was made. The correct drivers were loaded, and OpenCBM commands were issues to the device.
Sadly, initial tests failed. Before assuming the worst, I checked all solder joints, and measured impedences, on the assumption I had soldered a component incorrectly. During the inspection, I noticed two resistors attached to the IEC lines were shorted to each other, thus effectively shorting the IEC lines together. After resolving that issue, the unit successfully passed the tests by transferring data from the drive ROM to the PC.
With no need to spin a new PCB, I released the 98 first batch units to production and ordered the required parts for assembly. The plan is to quickly assemble 25 units and potentially ship before end of year, with the rest coming quickly in January. Since the 2×8 header X3 was nominal in cost, I ordered it for inclusion in assembly.
Sale price target is $35.00, and I will add a pre-order option in the store shortly. Given the ease of installation and configuration, I predict significant sales. This device eliminates the need to fiddle with parallel port settings, trying to remember a myriad of differently lettered adapters, and a need to maintain older systems with legacy ports for disk access purposes.
PCBs for ZoomFloppy are finished and the first 2 are enroute to the US for manual assembly and testing. Those following the blog for a while or having chatted with me at shows know the sequence for a new product:
- Initial prototyping
- Unit and system testing
- Schematic capture
- PCB CAD design
- PCB manufacturing
- Limited run hand assembly
- Final testing
- Automated Assembly
Key commitment points include step step 5 and step 8. Before step 5, little expense is involved, and truly, many projects cease before that step. Step #5 has become much more economical, so much so that initial production board runs approximate the cost for limited quantity prototype manufacturing. Thus, I don’t often run a true prototype phase. At worst, the design has an issue and I need to run another batch of boards, which costs no more than a prototype run + production PCB run. At best, the first revision works fine, and the money saved can be used elsewhere.
Step #8 is a larger commitment. Whereas a board run of 100 boards might cost a couple hundred dollars, parts acquisition alone for step #8 can easily cost 1 to 2 thousand, more if the design is a higher end product. Thus, it makes sense to, in the absence of a prototype PCB phase, hand assemble a few production boards to check final placement, fit, and function.
One the design passes testing, it becomes a waiting game for final assembled boards to arrive for sales. As in all things, there are 3 criteria: speed, quality, and price. You can select 2 of the 3 for your project. I select quality and price, so I typically wait a bit longer for assembly. In perspective, some assembly houses can do 100 units in 5-10 days, while I typically figure 15-20 days.
Still, things should progress smoothly and quickly once the final testing is done.
QuadPortIEC boards arrived today from the manufacturer. They look very clean, though the design didn’t exactly push the envelope. At $100.00 for 100 sq inches of PCB (shipping included), they were not cheap, but I saw much less competitive offerings in the marketplace. Test fitment in the Hammond case looks good, though I need to order a few DIN sockets to test for sure.