This is why we build and test prototypes


I got all the parts in for the prototype SensorBit-C boards, so I built up the first one to see how we did.  There is always this expectation / hope that it will just somehow work the first time you try it out. Rarely if ever does it quite turn out that way.

As you can see above, all the parts fit (well, most of them), and the LED lights up. But there are some significant problems that need to be fixed, which could have been caught at design time but aren’t always:

  • The drill holes for the two jumpers were much too small for the jumper pins I wanted to use.  This is one of the things I should have checked before using the part footprint.  Easy enough to fix, but requires a board re-spin.
  • The two trim pots I selected are very nice parts, but are really too small to be practical for a typical model railroader to work with when upside down underneath a layout.  So I will be replacing them with larger ones.
  • Likewise, the screw terminals are marginally small.  They work well, but require a very small screwdriver.  I *might* consider using a larger part that accepts a larger screwdriver.

The real problem, though, is that the LED is on all of the time, regardless of whether the circuit is detecting  a train.  You’ll note that it is on in the photo, even though there’s not even a detection coil installed.

The reason it’s not working is a design error in the schematic, where I connected the detection circuit to the time-delay circuit the wrong way.  I stared at that schematic a hundred times looking for mistakes, but this one escaped me.

Oh well. No harm, except a few dollars and a few weeks “lost”.  We’ll make the required adjustments and get it right on the next pass.  And the next round of designs will be that much more solid!

Here’s another pic of the PC board on a “standard” Altoids tin, to give you an idea of scale.  This thing is quite small indeed, I think.


This entry was posted in Uncategorized and tagged , , , . Bookmark the permalink.

Leave a Reply