It was a good week. A red-letter week.
- My son returned to college. I am so excited for him, and everything he is learning. I am even more excited for the man he is becoming: responsible, diligent, imaginative, exploring, compassionate. I miss him already.
- I shipped Seder plates this week to a new customer — a nerve-wracking experience. Making something is such right-brain process — measuring, improving, analyzing. But to ship something — to fire your internal engineer and unleash the salesman — is wrenching. And exciting.
- I cut some two 11″ plates using newly measured, un-calibrated joint settings. The surface finish of the plates didn’t improve as much as I had hoped, but it seemed as if there was less sanding required. Surprisingly though, the size was wrong. The plate I made was roughly 1/4″ short of the 11″ radius, for an error of about 2%. Since the newly measured joint positions are close, even though not yet calibrated to a finer accuracy, that much error must be coming from something else. What if the actuators lead-screw pitch that is wrong? The lead-screw specification is 13/16″ per turn, with an accuracy of 0.004″/foot, so I didn’t consider that as an error source. But, if anywhere in the software I have used five turns per inch as the screw rate, then I would be off by about that amount in the right direction.
- It is now after Shabbat, and I checked… the screw rate looks correct.
- I checked the specifications, and they clearly show 13/64″ screw advance per turn.
- I measure the ball screws which I purchased per that spec, and the screw advance is clearly 1/5″ per turn. When I measured over 6″, the number of threads was 30, not 29.5. I could be off by 0.010 over the six inches, but I am not off by 0.100.
- Can I believe that the spec is wrong? Did it change between when I purchased the screws and now?
- News groups on the net point to several references that the McMaster Carr listings have converted to the nearest 1/64
th of an inch, making 13/64 be an approximate measure. Why they combine that with an accuracy spec of +/-0.004″/ft I don’t understand.
- This explains why my earlier calibration numbers had moved the joint positions to be so distant. The calibration process may have tried to correct the scale error by moving the joints.