Friday, November 21, 2008

A wiring solution

Here is a birds-eye view of the class layout with the front of the room being at the bottom of the page and lab space w/fixed lab 'islands' (drawn as squares) starting at the top of the drawing and going off the page w/2 more rows of tables--OK so this drawing is pretty unclear. Ignore the drawings crossed out at bottom. The idea is to tie the left column of 3 tables together so as to make a single structure. Then, tie in that triple-table to the square-shaped, built-in lab island behind it. Each table would have a power strip w/a sufficiently long cord to reach back, threaded under each table and along the 2x6's connecting them, to an outlet in the lab island. This solution would be repeated on the right side of the room as well. The solution for the middle tables is explain a ways below.Directly above is a picture of the system to connect 2 adjoining tables. The 2, 2x6's sandwiched in the middle serve as spacers for the 2x6's coming from the tables. There would be a 12 inch gap, then, between the table legs from adjoining tables.

Getting connectivity between tables 4 and 5 in the middle would require a different solution. The side view drawing below shows that the tables would be about 3 feet apart, and a cord protector would span the distance on the floor where both power and data cables would travel. Again, the back table would be tied in to the middle lab station. This is not a heavily trafficked area.
Below is an attempted perspective drawing, standing at the front of the room looking towards the back not showing any of the classroom computer embedded tables, just showing the lab islands in the back. In the upper left-hand corner you can see the classroom switch and the network cables coming from the right hand triple tables. The idea is to use Plenum data cables so that they could be above the hanging ceiling panels.

You can also see the network cables coming from the middle island and the far left island. Cord protectors would be on the floor, spanning the short distance between the lab islands and moving on to the wall. Upon reaching the wall the cables would travel up to the switch (this is not shown in the drawing).
Cody came up with an interesting solution where we wouldn't have to run so many network cables to the switch. As seen in the drawing below there would be an 8 port switch hidden in each one of lab islands. The gigabit port would then be connected to the built-in wiring in the lab island which would connect to the main switch in the building tech closet. The building wiring is Category 5e, unofficially rated up to gigabit speed. I'm not sure how well this would work in a thin client environment however, but I'll find out.
A few points remain to be investigated:
  • Can we have the built-in "power strips" inspected instead of buying and installing after market power strips?
  • If this were so, then we would simply cut down the length of the power cord and still be able to use the nicer built-in solution.
  • Is it true that the length of cord on a power strip can not exceed 25 feet?
In conclusion, the fire marshal, when asked by the district IT people for information on safety code only offered a link to a page with, "suggestions". It was in these suggestions that I found out that you can use power strips as permanent wiring as long as they're not daisy chained. It ambiguously stated that, "power strips are often found in lengths less than 25 feet." What does that mean? I trust that this combination of solutions will relieve the fire marshal's concerns.

I can't help but wonder if instead of using orange power cords for our power strips we had instead used white ones, would the fire marshal have been so upset and instead worked with us to find a solution?

Finally, big thanks to Fred, Cody and of course Christen for their help in taking down all of the previous wiring that Fred had so beautifully installed. Thanks too, to Judy the school secretary and the building engineer Ramon who helped in getting us the 8 foot ladder without which, of course, nothing would've gotten done!


  1. This blog entry refers to the GCOS installation at Nellie Stone Johnson school (but something similar might work in David's room to solve the vexing wiring problem - see below)

    In conjunction with the idea of using a switch for
    each row of 3 tables (6 computers - thin clients) ,
    I wonder how the Ethernet plugs on the lab tables at NSJ are wired and whether that wiring could be appropriated for GCOS use to eliminate the need to run any ethernet cables up from the tables to the ceiling.

    In David's room there are outlets flush in the floor
    but they are prone to being disconnected when bumped.
    If a "cover plate" for these outlets was crafted the covered and protected the plugged in cords, maybe this problem could be eliminated and thus eliminating the need for power cords going up to the ceiling. I imagine a sligly modified plastic junction box attached via the screw that normally holds the regular coverplate (and maybe addition attachment where the
    duplex outlet attaches to the regular box. This assume that these cords do no ned to be disconnected.

    A similar "cover plate could be used at NJS where the cords plug into the lab tables.

    The power coming from the lab tables.


  2. Good point, I will talk with the Dir of Fac about installing some kind of protection around the outlets.

    Fred, do you know of any source of info for 'advanced topics on thin clients'? Specifically, can you try and find out about the practicality of using an 8-port switch for each row of 3 tables, then connecting that switch via a giga port, through the building wiring (Cat 5e) to a final switch which would have (at least) 4 giga ports: 3 connecting to the 8-port switches and 1 connecting to the thin client server.

    In other words, how significant would the performance hit be with having 2 levels of switches between the servers and the thin clients?

    See you at Roosevelt soon!

  3. I used Open Office draw to create this image ) to represent the two possible ways of wiring a GCOS classroom. I thought it would help ask the question of whether using additional switches to simplify wiring would impose a significant performance penalty.