Tuesday, July 17, 2018

Screwed Holes

Mornin' Guv'nor!

It's 3 posts in one day!!! (Half of the first one was already written - but hey).

I'm still waiting for a response from the RFI I sent out yesterday (I don't even think they sent it along until today), but there was another aspect to this system layout that I forgot to mention - and that is how it is actually going to be installed.

It mostly consists of pipe routed below the roof deck, but there places where it runs perpendicular to the beams, so it either has to run across the bottom of the beams (which requires compensation), or it has to go across the roof deck and have four elbows at each beam to go down and back up.

As it turns out, it's going to be a combination of these two - but fortunately in the large server room (and I do mean large - over 71,000 s.f.) [Edit: There is another 71,000+ square feet  of server space on the second floor - the building itself is over 280,000 s.f.]  the pipes will run parallel to the large beams (and despite the .pdf of the 3D Revit Structural model that I received showing solid beams and roof joists - the joists are supposedly open - so our pipe can pass through them).

There will still be some places where the pipes have to route across the bottom of the beams to get to the places where they turn and run parallel (where I can simply turn it up and get up to the roof deck - but comically, the client doesn't want anything attached to the roof deck, so the installer is going to get to have a field day figuring out how to support the pipe (every 5' - despite it being filled with nothing but air).

Fortunately that's not my problem - but someone pointed out that some other equipment we need to put up at the roof was going to require a solution.  Fortunately the electrician is going to be installing 4" square boxes at each location a device will be installed, so we decided to sketch up a modified cover plate that would allow the device to bolted to the cover with another hole for a wiring gland.

Another guy actually sketched the detail, but a brief glance at the data sheet set my OCD on edge.  I'm not so bad as to need to flip lights on/off a certain number of times to leave a room - but I have spent a lot of time honing my pattern matching and spatial dimension skills, so I often spot stuff that other people completely miss.

The sheet showed three 9/32" holes - the first two were 2" apart on the left side with a third on the right 1-3/4" from the midpoint of the first two - but the diagram was fucking WRONG.  The dimensions were very close - but whoever drew the diagram had made the hole on the left 2" away as well, resulting in a skewed picture.  Ironically, directly to the right of this diagram was a note indicating 3 holes 120 degrees apart on a 2.312 bolt circle.

Purely for my own amusement, I decided to sketch the holes in CAD.  First with the 2" and 1-3/4" dimensions - then with a 2.312 circle and a polar array of circles 120 degrees apart (it was about this time that the guy who drew the detail told me I was getting obsessed with it).  At first glance they looked identical, but when I overlayed them, there was a discrepancy of 1/32".

It's not enough to worry about (in fact, the default dimension resolution in CAD doesn't even register the difference - even though you can clearly see two holes sitting next to each other) and despite the actual dimension being 1.71875" rather than 1.75", it will probably bolt right up.  Comically, we received a CAD file showing side/top/front and a '3d' view - none of which showed the actual mounting holes (you could only make out two on the '3d' view).

I'm not a mathematician, but I do hold a fascination with the ability of mathematics to define things down to a ridiculous level of accuracy.  In this particular case, I know 1/32" doesn't matter - but there are situations where it absolutely fucking does.  I've seen machine tools capable of accuracy up to 1/1000" (and now with the ability to work at the nano level - even more accurate than that).

It's not necessary for every day use - but as long as we are here, let's at least get it as close as possible.

Obsessively Yours,

-.000000000001 SkullFucks

Next Time: Can You Get Any Stupider?

P.S. - Because my brain literally never stops thinking about this stuff, I had a slight epiphany this morning while putting my shoes on that while lining up either the two holes or the one hole does result in a 1/32" discrepancy on the other side, splitting the difference might only result in 1/64" offset at each hole.  I tested this theory by redrawing it, only to find that in reality, none of the holes actually line up - but the discrepancy is so small as to be negligible.  My point still stands though - I'm fundamentally insane.

No comments:

Post a Comment