test equipment calibration stability, older vs newer

[[email protected]]

X-No-Archive: Yes

I have an older digital power meter and there are 21 potentiometers
inside and I'm guessing the calibration is made to the analog side of
the circuit. Some newer instruments use EEPROM to store calibration
offsets of other part and these have very few or no potentiometers.
The stored values within the EEPROM is digital and it does not change
over time and it is only changed to make an adjustment to drift of
other components.

Do newer instruments using ASICs and electronic calibration offer
better calibration stability in general compared to older ones made
with lots of general purpose ICs, discrete components and dozens of
adjustment potentiometers on the analog side?

Having a lot of discrete components and potentiometers seems like a lot
more room for calibration drift.

 

[**THE-RFI-EMI-GUY**]

Are there multiple power ranges? Perhaps there are fewer pot's involved
in setting each power range?

X-No-Archive: Yes

I have an older digital power meter and there are 21 potentiometers
inside and I'm guessing the calibration is made to the analog side of
the circuit. Some newer instruments use EEPROM to store calibration
offsets of other part and these have very few or no potentiometers.
The stored values within the EEPROM is digital and it does not change
over time and it is only changed to make an adjustment to drift of
other components.

Do newer instruments using ASICs and electronic calibration offer
better calibration stability in general compared to older ones made
with lots of general purpose ICs, discrete components and dozens of
adjustment potentiometers on the analog side?

Having a lot of discrete components and potentiometers seems like a lot
more room for calibration drift.

--
Joe Leikhim K4SAT
"The RFI-EMI-GUY"

"Treason doth never prosper: what's the reason?
For if it prosper, none dare call it treason."

"Follow The Money" ;-P

 

[Dan Hollands]

X-No-Archive: Yes

I have an older digital power meter and there are 21 potentiometers
inside and I'm guessing the calibration is made to the analog side of
the circuit. Some newer instruments use EEPROM to store calibration
offsets of other part and these have very few or no potentiometers.
The stored values within the EEPROM is digital and it does not change
over time and it is only changed to make an adjustment to drift of
other components.

Do newer instruments using ASICs and electronic calibration offer
better calibration stability in general compared to older ones made
with lots of general purpose ICs, discrete components and dozens of
adjustment potentiometers on the analog side?

Having a lot of discrete components and potentiometers seems like a lot
more room for calibration drift.

New designs have several advantages:

Fewer components + many of the newer components are more temperature stable

Calibration is usually done automatically during test by applying known
signals, monitoring the response and storing an adjustment value. When
calibration is done with potentiometers one is relying on the test tech to
do the adjustment properly and in the correct sequence. My experience has
been that the test people don't always follow the proper procedure but look
for short cuts to reduce the test time.

Some modern equipment does periodic self tests to verify that calibration is
still good

Dan



--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
[email protected]
www.QuickScoreRace.com

 

[Dan Hollands]

X-No-Archive: Yes

I have an older digital power meter and there are 21 potentiometers
inside and I'm guessing the calibration is made to the analog side of
the circuit. Some newer instruments use EEPROM to store calibration
offsets of other part and these have very few or no potentiometers.
The stored values within the EEPROM is digital and it does not change
over time and it is only changed to make an adjustment to drift of
other components.

Do newer instruments using ASICs and electronic calibration offer
better calibration stability in general compared to older ones made
with lots of general purpose ICs, discrete components and dozens of
adjustment potentiometers on the analog side?

Having a lot of discrete components and potentiometers seems like a lot
more room for calibration drift.

Another advantage of the not pot design is that it eliminates the temptation
for the customer to mess around with pots and then complain about the unit
not operating properly

Dan

--
Dan Hollands
1120 S Creek Dr
Webster NY 14580
585-872-2606
[email protected]
www.QuickScoreRace.com

 

[Richard Henry]

X-No-Archive: Yes

I have an older digital power meter and there are 21 potentiometers
inside and I'm guessing the calibration is made to the analog side of
the circuit. Some newer instruments use EEPROM to store calibration
offsets of other part and these have very few or no potentiometers.
The stored values within the EEPROM is digital and it does not change
over time and it is only changed to make an adjustment to drift of
other components.

Do newer instruments using ASICs and electronic calibration offer
better calibration stability in general compared to older ones made
with lots of general purpose ICs, discrete components and dozens of
adjustment potentiometers on the analog side?

Having a lot of discrete components and potentiometers seems like a lot
more room for calibration drift.

I got gigged by QA the other day for using an uncalibrated decade box. (I
was simulating a thermistor to test circuit and firmware response.) Out of
curiosity, I popped it open - there are no adjustments whatsoever, although
I guess you could trim some of the elements with a soldering iron.

 

[Robert Baer]

New designs have several advantages:

Fewer components + many of the newer components are more temperature stable

Calibration is usually done automatically during test by applying known
signals, monitoring the response and storing an adjustment value. When
calibration is done with potentiometers one is relying on the test tech to
do the adjustment properly and in the correct sequence. My experience has
been that the test people don't always follow the proper procedure but look
for short cuts to reduce the test time.

Some modern equipment does periodic self tests to verify that calibration is
still good

Dan

Perhaps *test* people do a sloppy job; a cal lab is a bit
different,to say the least (worked in one got a good part of a year
during a RIF period).

 

[Robert Baer]

Another advantage of the not pot design is that it eliminates the temptation
for the customer to mess around with pots and then complain about the unit
not operating properly

Dan

That is what a calibration seal is for; proof of idiot tampering.

 

[Robert Baer]

I got gigged by QA the other day for using an uncalibrated decade box. (I
was simulating a thermistor to test circuit and firmware response.) Out of
curiosity, I popped it open - there are no adjustments whatsoever, although
I guess you could trim some of the elements with a soldering iron.

But if you do not kow enough about the characteristics of each of
ther brand and types of resistors used, you would not know which ones to
leave alone, and which ones that *might* tolerate such abuse.

 

[John Gilmer]

I got gigged by QA the other day for using an uncalibrated decade box.

Well, the "gig" was for using a box that didn't have a current inspection
sticker.

Something like a decade box would not require frequent testing but it should
be subject to some independent evaluation before being set out among the
engineers and the techs as a de facto secondary standard.

And things like that need to at least be checked every now and then. It
just ain't hard to "cook" a resistor without knowing it. It should only
take a few minutes to ensure that every digit of every decade is "gud enuf",
the contact resistance is "in the noise" and such.

 

[BFoelsch]

Well, the "gig" was for using a box that didn't have a current inspection
sticker.

Something like a decade box would not require frequent testing but it
should
be subject to some independent evaluation before being set out among the
engineers and the techs as a de facto secondary standard.

And things like that need to at least be checked every now and then. It
just ain't hard to "cook" a resistor without knowing it. It should only
take a few minutes to ensure that every digit of every decade is "gud
enuf",
the contact resistance is "in the noise" and such.

Yes, good points. Equipment can be out of spec even if it has no
adjustments.

I am going crazy right now with a few Tektronix plug-ins. The 1% resistors
are dead on, but the 5% ABs are drifting like crazy, out of spec by 200% to
300%. Admittedly, these problems took 25 years to develop, but, how old is
the decade box in question?

 

[Keith]

I got gigged by QA the other day for using an uncalibrated decade box. (I
was simulating a thermistor to test circuit and firmware response.) Out of
curiosity, I popped it open - there are no adjustments whatsoever, although
I guess you could trim some of the elements with a soldering iron.

At another site we had lab inspections as part of ISO 9000 audits.
For equipment that wasn't used as standards there were red "ISO-NO"
stickers put in place of the calibration stickers. The one thing they
really got the audit-police twisted were the calibration stickers on
torque screwdrivers. Get caught with an expired sticker anywhere around
customer hardware and you got your head handed to you. I haven't seen
such things since I moved though.