Hacker News Reader

Comment by weinzierl | Hacker News Reader

This is certainly true in general. I see a niche for the mechanical solution if reliability is a primary goal and there is a (not too complex) gear that can solve your problem.

About ten year ago I worked in project that designed a solar tracker. These trackers need to be cheap and reliable. Every hour an mechanic spends in a remote location is incredibly expensive and cuts into your yield very fast.

Interestingly there is a gear that can follow the path of the sun very well while being driven by motor with constant speed [1].

[1] https://www.youtube.com/watch?v=2NGSL--PmGY

View on HN

Why are mechanical solutions more reliable than a servo ? Aren't mechanical and moving parts less reliable than electronics , in general ?

    Aren't mechanical and moving parts less reliable than 
    electronics, in general?

I don't dare to answer this in general, maybe someone more knowledgeable can offer his or her opinion. In our case the alternatives were:

  1. Simple gear and constant velocity three-phase motor[1]
  2. Two axis mounting and two stepper motors.

First option has less parts and should be more reliable in theory. That being said: I only know of commercially available solutions of the second kind. Maybe this has changed in the last ten years, but our solution never made it to market.

[1] Industrial three-phase motors are real workhorses, produced in large quantities since ages, relatively cheap and incredibly reliable.

You can think of a mechanical solution as a straight line program with no conditionals. The servo solution has a feedback loop that will ultimately face an upset in the feedback loop, a broken sensor, worn brushes on the servo motor, over heated power transistors, etc. Both solutions will suffer from fatigue in the long term, but an active solution with electronics and a servo will suffer from a element failing and possibly causing the active solution to destroy itself.

I was hoping that footnote was a link to the gear, not a demo of its application. Would you happen to have that?

I'm sorry, here is the patent[1].

EDIT: I almost forgot that my old website had a page about a model I built back then. I never migrated it to the new site but it is still available[2].

[1] https://www.google.com/patents/US4585318?dq=inassignee:%22Di...

[2] http://weinzierl.name/2006/en/miniviax

Reply on news.ycombinator.com