Ed,
I have dealt with this before. I will make the case first for not going to "no
calibration required." Being ISO certified your company no doubt has written
procedures for all production work. Any procedures that reference equipment
that would go to a "no cal" status will have to be examined and possibly
modified. No production measurement can be done with the "no cal" equipment,
even if the accuracy needed is only 50%. With no cal, it is not traceable to
NIST at any accuracy. If the procedures already use other equipment to confirm
the settings of the "no cal" equipment then you are OK. Otherwise, procedures
must be rewritten. You will have to compare the cost of modifying the
procedures- the time for the sustaining engineers to look over all of them,
change some of them, run the changes through the ECO system, and the
opportunity-cost of all of these people being taken away from their normal
work flow. So, perhaps the "no cal" approach is too costly. But, there might
be another way.
Based on the records of the calibration department the calibration cycles
could be extended on some equipment. If the past history (they do keep the
data from all calibrations, I assume) shows that the equipment is likely to
remain in cal for a longer period of time then this can work. A good cal lab
will measure a piece of equipment before making any adjustments. This data is
used to see if the equipment is falling out of cal during the cal interval or
is broken. There is a difference between calibration and a calibration
adjustment. Some standards never receive an adjustment (such as a Thomas 1 ohm
standard). The calibration is plotted over time. The cost here is the time
taken by the calibration department to check the cal records of the 125 pieces
of gear. I use this method for things like pulse generators and pattern
generators. The main reason being to not incur the downtime of having the gear
sent to be calibrated.
Another technique to lower costs is to perform a "limited calibration". Only
those parameters that matter to you are checked. For example, if you never use
any handheld digital meters for AC current measurement this can be skipped. A
"limited calibration" sticker should be affixed and records kept in the cal
department. The cost here is that procedures and record keeping in the cal
department need modification and all production procedures need reviewing to
ensure that no written procedure is compromised.
So, in my opinion, the cleanest path is to leave things as they are. No extra
work involved, the cal lab is happy, and having everything calibrated makes
things simpler for future production and procedure development.
Dave Cuthbert
Micron Technology
From: [email protected]
[mailto:[email protected]]On Behalf Of Price, Ed
Sent: Friday, November 07, 2003 2:59 PM
To: '[email protected]'
Subject: Equipment Calibration
In some ways, I have the luxury of having a Metrology Department that
maintains the periodic calibration on all of my test equipment. OTOH, as a
"customer" of this Metrology Department's "product", I would like to have some
control over my overhead costs. And my latest bright idea has me getting
stomped by the gurus of the status quo. I need to get smarter about how a
calibration system works, and how flexible it can be.
My lab has about 500 pieces of capital equipment, and the way I see it, all my
equipment falls into one of two categories. The first category consists of
those instruments which are used to measure the parameters of our company's
products, and determine if the performance of those products falls within a
range of acceptable tolerance. Data from these measurements is often
contractually reported to our customers. Every equipment within this category
needs to be maintained on a program of periodic, traceable calibration.
But then there's the second category; which consists of support and stimulus
equipment. Items here are old analog signal generators, function generators,
amplifiers, pulse generators, sweepers and power supplies. To me, none of this
equipment needs ANY periodic calibration. I base this on practical usage. Who
can accurately read a power supply mechanical 80-amp ammeter that has a 1.5"
long scale? Who can set a function generator frequency control that covers 2
decades, logarithmically, in 270 degrees of rotation? If I need to apply a 100
kHz signal in bursts of 2 milliseconds at a 1 Hz rate, I'll use a calibrated,
traceable oscilloscope to set the uncalibrated generator to exactly what I
need. The same for that power supply; if I need to know the current to 2% or
better, I'll use a calibrated resistor and a calibrated DMM. And I couldn't
care less about the gain of an RF power amplifier, as long as it pumps out
enough power to create the field I need.
Now, I'm not trying to justify the use of distorted, unstable or junky
equipment. I'm just trying to spend my calibration dollars the most efficient
way. And the way I see it, about 1/4 of my equipment fits my definition of not
needing periodic calibration because I can monitor the results with calibrated
equipment.
So I proposed that these items be tagged with some kind of "uncalibrated" or
"user verified" or "no calibration required" label. The gurus of Metrology say
this can't be done, our ISO9000 Quality System will not allow this. I can't
understand how a customer-oriented quality system can't be crafted to meet the
needs of all of the customers of that system. And I suppose I'm felling a bit
squeezed, what with my customers expecting me to use COTS equipment to
function in military environments. I have to get more out of what I have, and
the old military concept of everything in sight is on periodic calibration has
to yield to current reality.
So, am I getting shoveled upon regarding the impossibility of having a
category of officially non-calibrated equipment alongside my calibrated
equipment? How have you dealt with calibration program costs?
Regards,
Ed
Ed Price
[email protected] WB6WSN
NARTE Certified EMC Engineer & Technician
Electromagnetic Compatibility Lab
Cubic Defense Applications
San Diego, CA USA
858-505-2780 (Voice)
858-505-1583 (Fax)
Military & Avionics EMC Is Our Specialty
