Hello PMELIANS. Please tell me if my training is faulty.
Example: 8481 power sensor at 100MHz has cal factor charted at 99%. with an accuracy of 2. 5% from 10Mhz to 500Mhz
Run cal and new cal factor is 96%.
How would you calculate that the new reading is out of tolerance?
If I remember correctly I'd do the following: 99(old chart) x . 025(accuracy for range)= +/-2. 475
limits would be: 101. 475 to 96. 525
This would make the 96% out of tolerance. . . correct?
Thanks in Advance
I will tell you what I do and what I have observed, I make no claim that it is valid for your situation.
When I run a power sensor I generate a report that includes the cal factor & the cal factor uncertainty. If the amount of change from the previous cal factor is less than or equal to the uncertainty of the cal factor I am done. If the change exceeds the uncertainty then I wonder, did I do it last time or am I confident that whoever did used good technique, they are affected by connector torque. Is the connector in good shape, proper pin depth, loose pin? Is the cartridge secure in the bulkhead, I had one that recently was loose enough that I thought the pin was damaged because I noted it moving around when I cleaned it. The previous year it seemed ok enough but I had commented ther was a "hole" in the cal factors from the year before. After I tightened the cartridge in the bulkhead that hole went away, this tells me it was loose the year before but I didn't catch it.
I have noticed Agilent certs & test reports for the E series sensors test SWR, Cal Factors & Linearity. Linearity & SWR are assigned pass/fail, Cal Factors are only assigned "Done" The reports do however include information as to what cal factors were stored in the thing when received and what cal factrs resulted from testing, the difference and wheter or not new cal factors were loaded in the piece. What I don't know is the trigger to load vs. leaving it, I'm guessing the criteria I mentioned above. I have seen them state that new cal factors were loaded to optimize the thing.
So back to your situation, I don't know, I'd test it again to make sure it's stable. I think your thought process is sound, I guess if you are not using a hard & fast written procedure it comes down to you, does that amount of difference matter in the application of the thing? If you've made it this far congratulations and be glad you're not face to face and listening to me rattle on.
Quote from: Moore on 12-11-2009 -- 00:30:26
Hello PMELIANS. Please tell me if my training is faulty.
Example: 8481 power sensor at 100MHz has cal factor charted at 99%. with an accuracy of 2. 5% from 10Mhz to 500Mhz
Run cal and new cal factor is 96%.
How would you calculate that the new reading is out of tolerance?
If I remember correctly I'd do the following: 99(old chart) x . 025(accuracy for range)= +/-2. 475
limits would be: 101. 475 to 96. 525
This would make the 96% out of tolerance. . . correct?
Thanks in Advance
You are correct...take a 'K' action.
On the 8481 you Calibrated.. What was the Ref Cal Factors..
You should have 2.. One from you calibration and one form the previous calibration..
No, it's not the Ref Cal factor. I just choose a frequency. It's more of determining the change between the old cal factors and a fresh run. Comparing the charted readings to see the percentage in change.
But are the Ref Cal Factors the same?