Good news all around regarding the enclosures and previously mentioned reliability problem, though there will be a slight delay in delivery as a result.
I recently received the enclosure samples and I am very happy with them. The finish is good, the PCB’s fit well, the halves fit together tightly and securely. I have a minor complaint about the silkscreen. The "S" in the logo isn’t as legible as my source graphic. I’ve asked ICOMold to improve the quality. The bad news is that the screw does not thread into the top half of the enclosure because the hole is too large. This was a problem in the model that went unnoticed in part because in the resin prints walls came out thicker than the model’s. Luckily this problem can be fixed by simply removing some material from the mold to shrink the hole the right amount and ICOMold is already in the process of making this fix.
Below are some photos of the samples that I took today. I’ll probably make the button shorter for production. You might notice that the top surface is not as polished as the bottom. For production, it will have the same level of polish. For some reason, when they silkscreened the samples, they mistakenly used a solvent to clean them which damaged the finish.
Signet enclosure sample (top)
Signet enclosure sample (side)
Signet enclosure sample (bottom)
I have a better understanding of the reliability problem I mentioned in the last update. When the microcontroller resets, it is reporting the cause as a power brownout. I’ve done a number of experiments and found that increasing the capacitor values on the input and output of the voltage regulator and changing one or both to polarized capacitors reduces the unexpected reset frequency to near zero. I’m still not sure what the ideal values are. Since Macrofab will be assembling the boards soon, I have instructed them not to populate the capacitors for the power supply so that I can hand-solder the values that I will determine to produce the most stable power output under a variety of conditions. I have ordered an assortment of capacitor values in bulk so I will be ready to finish the boards when they come in.
Looking at the delays with the enclosures and the PCBs, I think a delivery date of December 8, 2017 is realistic. The lesson I’m taking from this is for is that spending extra time testing and double checking would have saved more time than it would have cost.