Calibration equipment design

[VMI]

Hi,
Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.
I guess the equipment manufacturers did the same a long time ago to have
a reliable source to compare their equipment with.
But, how can one be sure that the unit you build is good?

With what to start with?
A voltage source: AC, DC. Just a few volts?
An amplifier with an exact known amplification, like x1, x10, x100?
A nice stable frequency source with exact dividers, x1, x10, x100?

That would be a good starting point I think.

Any idea's, comments and links and hints are welcome.

Zilog

 

[John Tserkezis]

VMI wrote:

I guess the equipment manufacturers did the same a long time ago to have
a reliable source to compare their equipment with.

But, how can one be sure that the unit you build is good?

If you need to ask, you shouldn't be building it.
--
Linux Registered User # 302622
<http://counter.li.org>

 

[Dale H. Cook]

On Fri, 20 Jul 2007 09:46:27 +0200, VMI <VMI@Advalvas.be> wrote:

Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.

They cost that much in part because they are traceable to NIST. It
will cost you an arm and a leg to build traceable equipment -
secondary standards are pricey, and keeping those standards traceable
is pricey.

Dale H. Cook, Chief Engineer, Centennial Broadcasting,
Roanoke/Lynchburg, VA - WZZI / WZZU / WLNI / WLEQ
http://members.cox.net/dalehcook/starcity.shtml

 

[Jim Yanik]

Dale H. Cook <bridgewaterma_remove_@cox.net> wrote in
news:4571a39c72is7k7jjauljr0glrbu1t12di@4ax.com:

On Fri, 20 Jul 2007 09:46:27 +0200, VMI <VMI@Advalvas.be> wrote:

Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.

They cost that much in part because they are traceable to NIST. It
will cost you an arm and a leg to build traceable equipment -
secondary standards are pricey, and keeping those standards traceable
is pricey.

Dale H. Cook, Chief Engineer, Centennial Broadcasting,
Roanoke/Lynchburg, VA - WZZI / WZZU / WLNI / WLEQ
http://members.cox.net/dalehcook/starcity.shtml

Building a frequency standard would be the easiest. an xtal and some
divider ICs. You can cal it using WWVB.
Electronic Goldmine used to sell a xtal timebase kit for scope cal,with
PCB.

You want accuracies at least 4x better than what you are calibrating.

--
Jim Yanik
jyanik
at
kua.net

 

[Steve]

Designing equipment that'll be used for *traceable* calibrations
isn't something most engineers can take on by themselves.
This stuff is expensive for a reason!

You'll find lots of good deals on older calibrations standards
on flea bay. Years ago I bought a very nice Fluke voltage/current
standard for $50.

Steve
"VMI" <VMI@Advalvas.be> wrote in message
news:46a06851$0$14233$ba620e4c@news.skynet.be...

Hi,
Those calibration equipment costs you an arm and a leg, so I was thinking
in building my own.
I guess the equipment manufacturers did the same a long time ago to have a
reliable source to compare their equipment with.
But, how can one be sure that the unit you build is good?

With what to start with?
A voltage source: AC, DC. Just a few volts?
An amplifier with an exact known amplification, like x1, x10, x100?
A nice stable frequency source with exact dividers, x1, x10, x100?

That would be a good starting point I think.

Any idea's, comments and links and hints are welcome.

Zilog

 

[VMI]

VMI wrote:

Hi,
Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.
I guess the equipment manufacturers did the same a long time ago to have
a reliable source to compare their equipment with.
But, how can one be sure that the unit you build is good?

With what to start with?
A voltage source: AC, DC. Just a few volts?
An amplifier with an exact known amplification, like x1, x10, x100?
A nice stable frequency source with exact dividers, x1, x10, x100?

That would be a good starting point I think.

Any idea's, comments and links and hints are welcome.

Zilog
Thanks for the replies,

Reactions:
"If you need to ask, you shouldn't be building it. "

What did you do when you build your first electronic whatever?
You just designed and built it?
I think you asked or looked for info to...


"They cost that much in part because they are traceable to NIST. It
will cost you an arm and a leg to build traceable equipment -
secondary standards are pricey, and keeping those standards traceable
is pricey."

I like to make clear that I do not want to have the equipment to be to
the NIST standards as this is a bit to extreme.
An above standard precision is ok.


"You'll find lots of good deals on older calibrations standards
on flea bay. Years ago I bought a very nice Fluke voltage/current
standard for $50."

That's an option and worth considering.


I ask myself" how did the first test equipment mfg's started to
calibrate and continue to compare the rest of their production with the
first?
It all started somewhere and with an average precision (as precision had
still to be invented).

 

[Fred_Bartoli]

VMI a écrit :

VMI wrote:
Hi,
Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.
I guess the equipment manufacturers did the same a long time ago to
have a reliable source to compare their equipment with.
But, how can one be sure that the unit you build is good?

With what to start with?
A voltage source: AC, DC. Just a few volts?
An amplifier with an exact known amplification, like x1, x10, x100?
A nice stable frequency source with exact dividers, x1, x10, x100?

That would be a good starting point I think.

Any idea's, comments and links and hints are welcome.

Zilog
Thanks for the replies,

Reactions:
"If you need to ask, you shouldn't be building it. "

What did you do when you build your first electronic whatever?
You just designed and built it?
I think you asked or looked for info to...


"They cost that much in part because they are traceable to NIST. It
will cost you an arm and a leg to build traceable equipment -
secondary standards are pricey, and keeping those standards traceable
is pricey."

I like to make clear that I do not want to have the equipment to be to
the NIST standards as this is a bit to extreme.
An above standard precision is ok.


"You'll find lots of good deals on older calibrations standards
on flea bay. Years ago I bought a very nice Fluke voltage/current
standard for $50."

That's an option and worth considering.


I ask myself" how did the first test equipment mfg's started to
calibrate and continue to compare the rest of their production with the
first?
It all started somewhere and with an average precision (as precision had
still to be invented).

Google for "standard cell".

Then when you have this voltage reference you can compare any voltage to
this standard with a bridge composed of a null votage detector and a
Kelvin-Varley divider (KVD).
Now think how about how KVD are built: if you have thermally and age
stable resistors a KVD is easy to build to high precision.


--
Thanks,
Fred.

 

[Michael A. Terrell]

VMI wrote:

VMI wrote:
Hi,
Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.
I guess the equipment manufacturers did the same a long time ago to have
a reliable source to compare their equipment with.
But, how can one be sure that the unit you build is good?

With what to start with?
A voltage source: AC, DC. Just a few volts?
An amplifier with an exact known amplification, like x1, x10, x100?
A nice stable frequency source with exact dividers, x1, x10, x100?

That would be a good starting point I think.

Any idea's, comments and links and hints are welcome.

Zilog
Thanks for the replies,

Reactions:
"If you need to ask, you shouldn't be building it. "

What did you do when you build your first electronic whatever?
You just designed and built it?
I think you asked or looked for info to...

"They cost that much in part because they are traceable to NIST. It
will cost you an arm and a leg to build traceable equipment -
secondary standards are pricey, and keeping those standards traceable
is pricey."

I like to make clear that I do not want to have the equipment to be to
the NIST standards as this is a bit to extreme.
An above standard precision is ok.

"You'll find lots of good deals on older calibrations standards
on flea bay. Years ago I bought a very nice Fluke voltage/current
standard for $50."

That's an option and worth considering.

I ask myself" how did the first test equipment mfg's started to
calibrate and continue to compare the rest of their production with the
first?
It all started somewhere and with an average precision (as precision had
still to be invented).

It may have started out simple over 100 years ago, but now if it
isn't traceable to NIST, it isn't calibrated.


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[VMI]

Michael A. Terrell wrote:

VMI wrote:
VMI wrote:
Hi,
Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.
I guess the equipment manufacturers did the same a long time ago to have
a reliable source to compare their equipment with.
But, how can one be sure that the unit you build is good?

With what to start with?
A voltage source: AC, DC. Just a few volts?
An amplifier with an exact known amplification, like x1, x10, x100?
A nice stable frequency source with exact dividers, x1, x10, x100?

That would be a good starting point I think.

Any idea's, comments and links and hints are welcome.

Zilog
Thanks for the replies,

Reactions:
"If you need to ask, you shouldn't be building it. "

What did you do when you build your first electronic whatever?
You just designed and built it?
I think you asked or looked for info to...

"They cost that much in part because they are traceable to NIST. It
will cost you an arm and a leg to build traceable equipment -
secondary standards are pricey, and keeping those standards traceable
is pricey."

I like to make clear that I do not want to have the equipment to be to
the NIST standards as this is a bit to extreme.
An above standard precision is ok.

"You'll find lots of good deals on older calibrations standards
on flea bay. Years ago I bought a very nice Fluke voltage/current
standard for $50."

That's an option and worth considering.

I ask myself" how did the first test equipment mfg's started to
calibrate and continue to compare the rest of their production with the
first?
It all started somewhere and with an average precision (as precision had
still to be invented).


It may have started out simple over 100 years ago, but now if it
isn't traceable to NIST, it isn't calibrated.


Again you're correct; if only you want to manufacture test equipment and

sell them.
But here it is for a limited use. We like to calibrate our instruments
by our self and not at a price.
It may be exaggerated to some but it is not for us.
Making electronic equipment or calibration standards do not differ much,
except the precision.
And for the later we do not have much expertise.
We prefer to build it by ourself (and for ourself!) so we know what we
are doing.
As written before: getting some on eg Ebay may be an option, but how do
we know they are correct and precise?
Back to square one I would say...
Getting the calibrators calibrated by a NIST recognized institute?

Zilog

 

[Michael A. Terrell]

VMI wrote:

Michael A. Terrell wrote:
VMI wrote:
VMI wrote:
Hi,
Those calibration equipment costs you an arm and a leg, so I was
thinking in building my own.
I guess the equipment manufacturers did the same a long time ago to have
a reliable source to compare their equipment with.
But, how can one be sure that the unit you build is good?

With what to start with?
A voltage source: AC, DC. Just a few volts?
An amplifier with an exact known amplification, like x1, x10, x100?
A nice stable frequency source with exact dividers, x1, x10, x100?

That would be a good starting point I think.

Any idea's, comments and links and hints are welcome.

Zilog
Thanks for the replies,

Reactions:
"If you need to ask, you shouldn't be building it. "

What did you do when you build your first electronic whatever?
You just designed and built it?
I think you asked or looked for info to...

"They cost that much in part because they are traceable to NIST. It
will cost you an arm and a leg to build traceable equipment -
secondary standards are pricey, and keeping those standards traceable
is pricey."

I like to make clear that I do not want to have the equipment to be to
the NIST standards as this is a bit to extreme.
An above standard precision is ok.

"You'll find lots of good deals on older calibrations standards
on flea bay. Years ago I bought a very nice Fluke voltage/current
standard for $50."

That's an option and worth considering.

I ask myself" how did the first test equipment mfg's started to
calibrate and continue to compare the rest of their production with the
first?
It all started somewhere and with an average precision (as precision had
still to be invented).


It may have started out simple over 100 years ago, but now if it
isn't traceable to NIST, it isn't calibrated.


Again you're correct; if only you want to manufacture test equipment and
sell them.
But here it is for a limited use. We like to calibrate our instruments
by our self and not at a price.

That is ok for home brew projects, but not for professional work

It may be exaggerated to some but it is not for us.
Making electronic equipment or calibration standards do not differ much,
except the precision.
And for the later we do not have much expertise.
We prefer to build it by ourself (and for ourself!) so we know what we
are doing.
As written before: getting some on eg Ebay may be an option, but how do
we know they are correct and precise?

You have them calibrated as primary, or secondary standards like
everyone else. You can't just set your own arbitrary standards.

Back to square one I would say...
Getting the calibrators calibrated by a NIST recognized institute?

Its the only way to fly!

Zilog

--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[Dale H. Cook]

On Sat, 21 Jul 2007 11:04:37 +0200, VMI <VMI@Advalvas.be> wrote:

I like to make clear that I do not want to have the equipment to be to
the NIST standards as this is a bit to extreme.
An above standard precision is ok.

If you can't trace it to an internationally recognized standards
organization such as NIST, how will you know that it is accurate? How
will you know what its precision is? In order to know that your
calibrations have a given accuracy those calibrations must be
traceable.

The amateur calibrator of electronic equipment is primarily interested
in two classes of standards:


1) Time / Frequency

You only need one of these, since each can be converted to the other.
In practice the amateur calibrator almost always maintains a local
frequency standard. I have two. The one that I use is the
oven-controlled time base in my frequency counter, which is calibrated
to WWVB using an HP VLH comparator. By running the calibration plot
for 12 hours I can obtain a short-term accuracy for the local standard
of about 1 part in 10^10, and can achieve higher accuracies with
longer calibration plots. This method can also determine the ageing
rate of the local timebase.

My other standard, which I maintain just for fun, is the oscillator in
a very old GR frequency standard.

The appearance of rubidium standards on the used equipment market
offers an alternative for those with deep pockets - they can often be
found in the $500-$1,000 range.


2) Voltage / Current / Resistance

You need two of these in order to determine the third. In practice the
amateur calibrator generally maintains local standards for voltage and
resistance. I own two standard cells, a standard resistor, and the
potentiometers and bridges needed to scale them, as well as a VTVM
calibration set which is calibrated against my local standards. I have
the local standards calibrated periodically by a local cal lab. I can
calibrate instruments such as VTVMs to an accuracy of 1%, which is
what those VTVMs are specified for, and all that I am interested in.


Instruments that I use for work must be traceable to NIST in order to
satisfy the FCC. Those other than my frequency counter are
periodically calibrated by returning them to the manufacturer or by
taking them to a local cal lab.

Dale H. Cook, Chief Engineer, Centennial Broadcasting,
Roanoke/Lynchburg, VA - WZZI / WZZU / WLNI / WLEQ
http://members.cox.net/dalehcook/starcity.shtml

 

[Dale H. Cook]

On Sat, 21 Jul 2007 19:49:01 +0200, VMI <VMI@Advalvas.be> wrote:

As written before: getting some on eg Ebay may be an option, but how do
we know they are correct and precise?
Back to square one I would say...
Getting the calibrators calibrated by a NIST recognized institute?

Precisely - either by the manufacturer or by a local cal lab. Your
local standards will also have to be recalibrated periodically in
order to maintain long-term accuracy.

Dale H. Cook, Chief Engineer, Centennial Broadcasting,
Roanoke/Lynchburg, VA - WZZI / WZZU / WLNI / WLEQ
http://members.cox.net/dalehcook/starcity.shtml

 

[John Tserkezis]

VMI wrote:

As written before: getting some on eg Ebay may be an option, but how do
we know they are correct and precise?
Back to square one I would say...
Getting the calibrators calibrated by a NIST recognized institute?

You've answered your own question. There's no reason why you can't design
it (though it hasn't been recommended here) and send it out to be calibrated
and/or adjusted.

Depending on the calibration lab, some are happy to adjust AND calibrate
providing you have instructions on how to do that. If you've designed your
own, it's obviously mandatory, but if it's a bought box, the cal lab *might*
have procedures on hand to do it - just ask.

Buy and then have adjusted and/or calibrated. This really is the cheapest
and easiest way to go about having _known_ accuracy test kit in your lab.
--
Linux Registered User # 302622
<http://counter.li.org>

 

[John Tserkezis]

Dale H. Cook wrote:

If you can't trace it to an internationally recognized standards
organization such as NIST, how will you know that it is accurate? How
will you know what its precision is? In order to know that your
calibrations have a given accuracy those calibrations must be
traceable.

That's not strictly necessary.

What is also doable here in Australia, having something NATA certified (our
version of your NIST) is not the only way of having a known accuracy.

Some test labs offer tests and/or adjustments that give the same tests as
would the traceable standards, however you don't get a "fancy" certificate at
the end, and you don't have traceable paperwork- at a significant discount.

For those where traceability is not important but cost is, it's a very
viable option.

At the end of the day, you know where you stand, but just don't have the
certificate to prove it. If the traceability doesn't hold much weight for
your workshop, then this is very doable.

I don't know if all test labs offer this, even across any type of equipment,
but it's worth asking.

--
Linux Registered User # 302622
<http://counter.li.org>

 

[Dale H. Cook]

On Sun, 22 Jul 2007 10:20:08 +1000, John Tserkezis
<jt@techniciansyndrome.org.invalid> wrote:

Some test labs offer tests and/or adjustments that give the same tests as
would the traceable standards, however you don't get a "fancy" certificate at
the end, and you don't have traceable paperwork- at a significant discount.

Then you have no way of knowing whether their work is accurate.

For those where traceability is not important but cost is, it's a very
viable option.

But is isn't calibration, because there is no traceability to an
accepted standard.

At the end of the day, you know where you stand, but just don't have the
certificate to prove it

You don't know where you stand, because the work is not traceable.

Dale H. Cook, Chief Engineer, Centennial Broadcasting,
Roanoke/Lynchburg, VA - WZZI / WZZU / WLNI / WLEQ
http://members.cox.net/dalehcook/starcity.shtml

 

[VMI]

Dale H. Cook wrote:

On Sun, 22 Jul 2007 10:20:08 +1000, John Tserkezis
jt@techniciansyndrome.org.invalid> wrote:

Some test labs offer tests and/or adjustments that give the same tests as
would the traceable standards, however you don't get a "fancy" certificate at
the end, and you don't have traceable paperwork- at a significant discount.

Then you have no way of knowing whether their work is accurate.

For those where traceability is not important but cost is, it's a very
viable option.

But is isn't calibration, because there is no traceability to an
accepted standard.

At the end of the day, you know where you stand, but just don't have the
certificate to prove it

You don't know where you stand, because the work is not traceable.

Dale H. Cook, Chief Engineer, Centennial Broadcasting,
Roanoke/Lynchburg, VA - WZZI / WZZU / WLNI / WLEQ
http://members.cox.net/dalehcook/starcity.shtml
Hmm,It seems that the result is as follows:

1. You let your equipment calibrate and do the things you have to do
with that calibrated equipment. You just lost an arm and a leg.
You can allways show everybody that your test equipment is calibrated
but not that your repair or build goods are...

2. You get some calibrated references (or build and let them
calibrated)and go on from there.
Again here your repairs or build goods are within spec (because of the
calibrated calibrators) but can not show that it is due to the missing
paperwork. Youjust lost an arm...

I worked for a multinational company that was very strikt on
calibrations etc. At the end we were working with test equipment that
showed us different results.
Fault by the calibration institute? Due to mal use or aging?

It never happened when the calibration was done in house (with
calibrated calibraters) But no fancy papers.
Iso 9000 etc is no step forward but rather backward at the expence of
the industrie.

So I know enough now to get my instruments calibrated.

Thanks all,

Zilog