Preliminary Cable Labor Estimate

1
Preliminary Cable Labor Estimate

• Howard Budd
• University of Rochester
• Rochester Meeting
• Dec 4 2008

2
Attempt to estimate labor

• To estimate the amount of time it takes a person
  to build a part you need to know the number parts
  and amount of labor
• I started by tackling what I thought was the
  hardest, the post polishing labor time
• One the labor sheet, we record seperately
  prepolish and post polish tech time.
• However post polished cables are in various
  stages of production and so the number of parts
  is not defined
• Janina was working on T2K production so
  effectively all post polishing production for
  cables stopped, so there was no way to estimate
  post polishing production time

3
Estimating Labor time

• Continuation of the presentation given in Nov
• Using same technique as I think I can get the
  quickest answer
• Tech lists the number tasks he performs for each
  task and the amount of labor which is charged to
  the project
• With the number of tasks and time I can do a
  least square fit to determine the amount of time
  each task takes per cable
• I presented the after polish results in Nov
• I havent changed the calculation as Janina
  worked on t2k for a significant portion of Nov.
• Next page gives that result

4
Results for Fit for Post Polish Tech Time

• Post Polish time fit
• Preglue 0.93 min
• Boot 13.5 min
• Test Light Tight 6 min
• Sort 0.19 min
• Paint 1.95 min
• Total so far 22.6 minutes/Cable
• Same as presented 3 weeks ago

5
Before polish Labor time

• From the start of when Mike Herron joined us he
  was keeping track of what he did each day. This
  included the tasks and usually the number of
  times he did each task. He created information
  which I would evenually ask him for even though I
  didnt ask at the beginning
• When there was a task which involved a large
  number times, such as reaming connectors,
  sometimes he didnt record the number he did and
  sometime he gave a guess as the number he did. I
  took the guess when it was there as the number of
  times he reamed the connectors
• For assembling connectors he usually gave the
  exact number
• There were some tasks which didnt take much
  time, in which he did not list a number I could
  use, so I just put in some small placeholder

6
Mikes time

• Mike time 10.83 (all times minutes)
• Reaming .61
• Cut tube 0.11
• Assemble Cable 6.9
• Label Cable -0.013
• Inspect before polish 1 ( a placeholder)
• Inspect after polish 1.19
• Repair after polish 1 ( a placeholder)

7
Heng Baos time

• Heng Bao started recording the number of tasks
  when I asked her on the week of Thanksgiving. She
  defined 4 tasks and there was 5 days of task.
• In addition the numbers of these 4 tasks she did
  were about the same each day.
• Hence, the time of each individual tasks from the
  fit is probably not meaningful
• However, the number of these 4 tasks is probably
  meaningful and probably gives a realistic about
  of time she is spending on each cable
• Heng Bao time 9.91 minutes
• Total before polish 20.75 minutes
• Will probably try to look at the labor sheets to
  see if this is in the right ball park, but that
  will take a little while

8
Optical Cable QC Labor Time

• The QC files are written with the date and time
  in the name, Hence I can use this to estimate the
  QC labor time.
• The QC consists of
• Turn on LED and wait atleast 1 hour
• During this time the tech gets to be tested and
  clean the connectors with alcohol and blow them
  dry
• 3 direct measurements 3 goldens
• Measure the cables
• 3 goldens and 1 direct
• Put the cables away

9
Cable QC labor

• We do a linear fit of number of cables measured
  vs the amount of time between the first direct
  measurement and the last direct measurement.
• The constant gives an estimate of the 4 direct
  measurements the golden measurement. There is
  probably other inefficiencies in there
• The linear estimate
• Constant 67.25 minutes
• Time/cable 3.63 minutes
• This time does not include the setup time of
  turning on the LED and cleaning the cables. Lets
  say this takes 1 hour. Hence Constant 127.25
  minutes

10
Labor QC Estimate

• Need to set up a model for QC, could be wrong
• In order to understand the status of production
  better we are trying to group cables in lots of
  50
• Lets see what happens if we measure in groups of
  50 the 12 before booting cables.
• In addition we are tying to QC 25 of the cables
  before booting. This means that the time to QC 50
  booted cables is
• 127.25623.63 minutes 352.31 minutes
• 7 minutes/cable ( with an 25 measured before
  booting)

11
Other Labor Estimates

• We assume that Cables are brought to and taken
  from lab 7 in groups of 100. Each trip takes
  about 1 hour.
• 120 minutes/100 cables 1.2 min/cable
• Putting Cables in boxes
• Dmitre put the Tracking Prototype cables in boxes
  and filled out the QC sheets in about 3-4 day for
  760 cables
• Time 4860/760 .4 min/cable

12
For a Grand Sum of

• The total is
• 20.75 1.2 22.7 7 0.4 52.04 minutes
• 0.868 hours/cable
• In Project we have 1.39 hours/cable
• Note this is preliminary as in doing this I have
  found mistakes as I have done it.
• Note, no RD has been taken place by the techs
  while recording their numbers, so clearly this
  does not reflect what happened during the
  tracking prototype

13
Status

• Cable numbers in this file need to be checked
• Will do it today with Janina
• Using these new time as weights we get we have
  made labor weighted 1853 cables
• 4379 to make (NOT including recent additions or
  our failures, need to put this in)
• Total cables to make 2536 cables
• 2192 hours, 1.25 tech years to finish cables

14
Discussion

• I have done some of this yesterday, so it will
  need some checking
• Have not included additional cables or ODUs
• Cable Labor
• Janina can work probably 40 on mapping
• Mike 50
• Hang Bao about 50
• Not clear on Dmitre or
• Note 50 time of a day is not efficient on clear
  cable QC
• If this were put into effect we would need to
  continue to see if the predictions were born out.