Joe wrote:

I have no experience with the seller you are looking at, but in
general the chance of something you buy on ebay really being
calibrated with NIST traceability are so close to zero that it is not
even worth looking into.  I recommend treating every ebay instrument
as needing calibration regardless of what the seller says.  (Again, I
have no experience with the seller you are looking at.)

If you want to investigate whether there is any chance an instrument
is calibrated with NIST traceability, you need to (i) ask what lab
did the calibration, and when; (ii) ask what accreditation body
accredits the lab; (iii) get a copy of the lab's accreditation
documents; and (iv) get a copy of the calibration certificate for the
particular instrument.  You would then review the accreditation
documents (in particular, the "scope of calibration") to make sure
they are in order and to see what uncertainty the lab is accredited
to for (in the case of a DMM) DC voltage, AC voltage, DC and AC
current, and resistance.  (Somretimes you will find that a lab is
accredited, but not to the uncertainty necessary to calibrate the
instrument in question to the manufacturer's specifications.)

Do all of this before you bid.

If the seller will not tell you what lab did the cal, or you cannot
obtain the accreditation documents and instrument calibration
certificate, treat the instrument as needing calibration and value it
accordingly.

It appears that the seller in this case does its own
calibrations.  If it is an accredited cal lab, it will be able to
supply the documents mentioned above.  If not (most likely because it
is not accredited), treat the instrument as needing calibration and
value it accordingly.

Best regards,

Charles

Well, lets discuss the 3456A I got from goldenrubi, calibrated.  They send
a calibration certificate which claims that their standards are traceable
to NIST.  I have no reason to doubt that.  The list of standards used is
reasonable for a 3456A.  The certificate claims "in tolerance" for
condition received and condition returned.  Not surprising if they cal'd
the unit before listing, then checked it before shipping.  (One of their
later listings now puts a $ value below which they don't check the cal
before shipping.  Fair enough, not worthwhile checking an instrument that
sold for $200.)

Now, how good is their cal?  It's within 10ppm of a Geller SVR-T at 10V.
Said SVR-T is within 3ppm of a new Agilent 34461A (and to be honest, given
the data on the cal certificate for the 34461A, I trust the SVR-T).  Given
the uncertainties involved, I cannot say the 3456A is within 24 hour
specs... but it is well within 90 day specs.

In fact, the two 3455A and 3456As currently in my possession are all mighty
close to 24 hour specs.  These old meters don't seem to drift much and if
cal'ed within 5 years or so, they should be fine for us amateurs.

However, Joe's 3478A does seem to be the exception to this rule.

Orin.

On Sat, Aug 10, 2013 at 8:10 PM, Charles Steinmetz csteinmetz@yandex.comwrote:

On 11 August 2013 05:10, Orin Eman orin.eman@gmail.com wrote:

I can't comment on goldenrubi's ability to calibrate a 3457A, but I'm
pretty sure he had a VNA cal kit on eBay which was calibrated. I was
suspicious, but I can't recall the exact details. I can't find the
auction, but I'm 99% sure it was goldenrubi.

I know Orin knows about VNAs, so he at least will understand this. But
even if you don't, it should be clear that you can't calibrate a
device properly using a device of the same specification.

Anyway, goldenrubi has two items on eBay now, both related to vector
network analyzer, neither of which have cal certificates, but I
doubted the accuracy of some comments on those auctions. So I deceided
to use the "ask seller a question" and query things. As I suspected,
he does not have a clue. It should be noted

1. One kit (85032B) has a part missing, which he is clear about.
2. The same kit shows some adapters (which I think are not the correct
   models), but the adapters are present. The auction says they are not,
   so I query that too.

These two things are not central to the problem.

I also strongly suspect two of these parts in the 85032B are not
original, as they have no HP markings on them at all. I have one of
these kits, and parts clearly have HP markings on them.

1. http://www.ebay.co.uk/itm/111114143954 is a "TESTED" HP 83032B
   typep N calibration kit.
2. http://www.ebay.co.uk/itm/190780743417 is a "TESTED" 85053A 26.5
   GHz 3.5mm verification kit.

Bear in mind, the manuals for both devices states the equipment to
verify the performance of the kits is not commericially available. Now
goldenrubi tells me he can't calibrate them, but his reasons are
totally wong. He seems to think that he could calibrate the kits if he
had other kits of the same model. This is totally wrong.

I know for a fact the open and short standards should be checked by
mechanical dimensions, not electrical measurements. I have got that
from someone at Agilent who is a specialist in this area. If you look
at papers on the calibration of these things, you will find things
like laser micrometers are used.

=== My 1st message to  goldenrubi ====
Dear goldenrubi,

I'm interested in this 85032B. Can you tell me how it was tested? Can
you provide a calibration certificate? The auction states it does not
come with the APC-7 to N adapters, but they are in the picture. Do you
intend removing them? If so,

I'm also interested in the 3.5 mm verification kit you have in another
auction. I'd like to know again how it was tested and whether it is
calibrated.

Dave

• drkirkby
  ==========================

He replies, saying he has the network anayzers these kits belong to,
so that is how he tests them!!!!!  He says to calibate them he would
need another calibrated kit of the same model. This is total rubbish
as you can buy the kits easy enough (excepting the 85032B is
obsolete), but the manuals make it clear the equipment to verify the
performance is not commerically available.

======== goldenrubi  1st reply to me ======
Dear drkirkby,

I do not have the HP - Agilent 85032-60009 Short in this kit. I have
everything else. We have the Network analyzers that these kits belong
too. That is how we test them out. We cannot calibrate them or provide
certificate because we do not have this HP 85032B full kit and we
would need another kit that was calibrated in order to do a one on one
calibration on my kit. You get everything that is in the picture. I
will not remove any item. The 3.5MM verification kit was tested by us.
we have the equipment to test these kits out. WE do not have the
calibration kit to re-cal my kits on a one on one calibration. so we
cannot cal them. NO standards to work with.

• goldenrubi
  ================

I seek clarification, to make sure I have understood him correct

====== My second message to goldenrubi ====
Dear goldenrubi,

OK, I understand the 85032-60009 Short is missing from the 85032B
calibration kit on auction 11114143954. If I understand you correctly,
you can't calibrate the 85032B since you don't have another 85032B -
is that correct? Why does your auction for the 85032B say the adapters
are missing, when there are adapters in the picture? What equipment do
you use to test the 85053A 26.5 GHz 3.5mm verification kit in auction
190780743417? Am I correct in saying you can't calibrate it since you
don't have another 85053A verification kit, but if you had another
85053A you could calibrate it. Let me know if I have understood you
correct. Dave

• drkirkby
  ===========

His reply confirms what I thought.

======= His second reply to me =======
Dear drkirkby,

The short is missing from the picture. There should be 2 shorts, there
is only on. Yes, we need a counter part or other kit that was
calibrated in order to calibrated the two I have. We do not have a
standard (another kit) to use. We use the test equipment we have to
test them out. You understand everything I told you.
GREG

• goldenrubi
  ================================

I don't believe it is possible for anyone to truely verify the
performance of either of these kits other then the manufactuer.  In
the case of the verification kit, if one sends it to Agilent they (or
more likely Maury Microwave who make them for Agilent), would measure
the devices and write the data on them to a floppy disk and USB stick
(formally a tape, but tape is no longer supported). I believe the
format of the data is not public.

I certainly would not trust him to calibrate anything VNA related.

There are things you can do on a VNA, along with connector gages to
check gross errors. You don't need a kit of the same unit. But you
will never truely be able to calibrate one of these expensive high
precison kits. I would suggest one could

1. Check if the mating surfaces of the connectors are in the right
   place using connector gages.
2. Check the attenuation of the two attenuators in the verification
   kits are nominally the right value.
3. Measure the offset delay of the shorts using a VNA calibrated with
   a calibration kit. It would not have to be the same kit.
4. Measure the fringing capacitance of the opens on a VNA and ensure
   it fits the polynomial in the manual reasonably well.
5. If you have a suitable combination of VNA, calibration kit and
   verification kit, then if one calibatrates the VNA, and the
   verification passes, one would have reasonalbe confidence the
   calibration kit is not too bad.

Properly calibrating these sorts of devices is not easy, and is very
expensive to get done. It costs far more to get a VNA calibraiton kit
calibrated than it does to get a VNA calibrated.

I produce a low-cost 3 GHz cal kit myself,

http://www.vnacalibration.co.uk/sales/

so I know things can be done which give reasonable accuracy without
laser micrometers, but I would never claim to be able to get the
accuracy Agilent do and could certanly never calibrate an Agilent kit.

Dave

In message CANX10hA+uhgmQGdX1A+n6=5RsV3oYYzDWHvfLKe+K=JKZV+Sig@mail.gmail.com
, "Dr. David Kirkby" writes:

That is a perception I have met more than once:

"I checked it with my other ${type} instrument, and they showed
(pretty much) the same, so it is calibrated."

People don't understand that "calibrated" doesn't mean that it
shows the right thing, but that you know how wrong it is.

--
Poul-Henning Kamp      | UNIX since Zilog Zeus 3.20
phk@FreeBSD.ORG        | TCP/IP since RFC 956
FreeBSD committer      | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.

Poul-Henning wrote:

A very pedantic expansion of the above (but then, this is volt-nuts):

"Calibrated" means that you know (i) what your best estimate of the
actual measured quantity is, based on a given measurement (i.e., the
predicted "offset" of the calibrated instrument), and (ii) what the
uncertainty of that estimate is.

Best regards,

Charles

Hi Dave,
I find it interesting that goldenrubi uses other dealers websites as sources for datasheets in his auctions!
I looked for a laminated HP panel for you but it seems all the one I have are the earlier painted and screen printed versions.
 
Robert G8RPI.

From: Dr. David Kirkby drkirkby@gmail.com
To: Discussion of precise voltage measurement volt-nuts@febo.com
Sent: Monday, 12 August 2013, 10:44
Subject: Re: [volt-nuts] HP 3457A

On 11 August 2013 05:10, Orin Eman orin.eman@gmail.com wrote:

I can't comment on goldenrubi's ability to calibrate a 3457A, but I'm
pretty sure he had a VNA cal kit on eBay which was calibrated. I was
suspicious, but I can't recall the exact details. I can't find the
auction, but I'm 99% sure it was goldenrubi.

I know Orin knows about VNAs, so he at least will understand this. But
even if you don't, it should be clear that you can't calibrate a
device properly using a device of the same specification.

Anyway, goldenrubi has two items on eBay now, both related to vector
network analyzer, neither of which have cal certificates, but I
doubted the accuracy of some comments on those auctions. So I deceided
to use the "ask seller a question" and query things. As I suspected,
he does not have a clue. It should be noted

1. One kit (85032B) has a part missing, which he is clear about.
2. The same kit shows some adapters (which I think are not the correct
   models), but the adapters are present. The auction says they are not,
   so I query that too.

These two things are not central to the problem.

I also strongly suspect two of these parts in the 85032B are not
original, as they have no HP markings on them at all. I have one of
these kits, and parts clearly have HP markings on them.

1. http://www.ebay.co.uk/itm/111114143954is a "TESTED" HP 83032B
   typep N calibration kit.
2. http://www.ebay.co.uk/itm/190780743417is a "TESTED" 85053A 26.5
   GHz 3.5mm verification kit.

Bear in mind, the manuals for both devices states the equipment to
verify the performance of the kits is not commericially available. Now
goldenrubi tells me he can't calibrate them, but his reasons are
totally wong. He seems to think that he could calibrate the kits if he
had other kits of the same model. This is totally wrong.

I know for a fact the open and short standards should be checked by
mechanical dimensions, not electrical measurements. I have got that
from someone at Agilent who is a specialist in this area. If you look
at papers on the calibration of these things, you will find things
like laser micrometers are used.

=== My 1st message to  goldenrubi ====
Dear goldenrubi,

I'm interested in this 85032B. Can you tell me how it was tested? Can
you provide a calibration certificate? The auction states it does not
come with the APC-7 to N adapters, but they are in the picture. Do you
intend removing them? If so,

I'm also interested in the 3.5 mm verification kit you have in another
auction. I'd like to know again how it was tested and whether it is
calibrated.

Dave

• drkirkby
  ==========================

He replies, saying he has the network anayzers these kits belong to,
so that is how he tests them!!!!!  He says to calibate them he would
need another calibrated kit of the same model. This is total rubbish
as you can buy the kits easy enough (excepting the 85032B is
obsolete), but the manuals make it clear the equipment to verify the
performance is not commerically available.

======== goldenrubi  1st reply to me ======
Dear drkirkby,

I do not have the HP - Agilent 85032-60009 Short in this kit. I have
everything else. We have the Network analyzers that these kits belong
too. That is how we test them out. We cannot calibrate them or provide
certificate because we do not have this HP 85032B full kit and we
would need another kit that was calibrated in order to do a one on one
calibration on my kit. You get everything that is in the picture. I
will not remove any item. The 3.5MM verification kit was tested by us.
we have the equipment to test these kits out. WE do not have the
calibration kit to re-cal my kits on a one on one calibration. so we
cannot cal them. NO standards to work with.

• goldenrubi
  ================

I seek clarification, to make sure I have understood him correct

====== My second message to goldenrubi ====
Dear goldenrubi,

OK, I understand the 85032-60009 Short is missing from the 85032B
calibration kit on auction 11114143954. If I understand you correctly,
you can't calibrate the 85032B since you don't have another 85032B -
is that correct? Why does your auction for the 85032B say the adapters
are missing, when there are adapters in the picture? What equipment do
you use to test the 85053A 26.5 GHz 3.5mm verification kit in auction
190780743417? Am I correct in saying you can't calibrate it since you
don't have another 85053A verification kit, but if you had another
85053A you could calibrate it. Let me know if I have understood you
correct. Dave

• drkirkby
  ===========

His reply confirms what I thought.

======= His second reply to me =======
Dear drkirkby,

The short is missing from the picture. There should be 2 shorts, there
is only on. Yes, we need a counter part or other kit that was
calibrated in order to calibrated the two I have. We do not have a
standard (another kit) to use. We use the test equipment we have to
test them out. You understand everything I told you.
GREG

• goldenrubi
  ================================

I don't believe it is possible for anyone to truely verify the
performance of either of these kits other then the manufactuer.  In
the case of the verification kit, if one sends it to Agilent they (or
more likely Maury Microwave who make them for Agilent), would measure
the devices and write the data on them to a floppy disk and USB stick
(formally a tape, but tape is no longer supported). I believe the
format of the data is not public.

I certainly would not trust him to calibrate anything VNA related.

There are things you can do on a VNA, along with connector gages to
check gross errors. You don't need a kit of the same unit. But you
will never truely be able to calibrate one of these expensive high
precison kits. I would suggest one could

1. Check if the mating surfaces of the connectors are in the right
   place using connector gages.
2. Check the attenuation of the two attenuators in the verification
   kits are nominally the right value.
3. Measure the offset delay of the shorts using a VNA calibrated with
   a calibration kit. It would not have to be the same kit.
4. Measure the fringing capacitance of the opens on a VNA and ensure
   it fits the polynomial in the manual reasonably well.
5. If you have a suitable combination of VNA, calibration kit and
   verification kit, then if one calibatrates the VNA, and the
   verification passes, one would have reasonalbe confidence the
   calibration kit is not too bad.

Properly calibrating these sorts of devices is not easy, and is very
expensive to get done. It costs far more to get a VNA calibraiton kit
calibrated than it does to get a VNA calibrated.

I produce a low-cost 3 GHz cal kit myself,

http://www.vnacalibration.co.uk/sales/

so I know things can be done which give reasonable accuracy without
laser micrometers, but I would never claim to be able to get the
accuracy Agilent do and could certanly never calibrate an Agilent kit.

Dave

volt-nuts mailing list -- volt-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.