1. Forum
  2. Usenet
  3. SCI.PHYSICS.ELECTROMAG

  • NanoVNA : Three papers on ResearchGate regarding the calibration techni

    From gin-pez arg@21:1/5 to All on Fri Feb 4 01:26:55 2022
    NanoVNA : Three papers on ResearchGate regarding the calibration technique of NanoVNA

    Hello.

    It may be of some interest to you that NanoVNA uses the calibration
    technique which is the subject of the following three papers by us now available on ResearchGate.

    Sincerely,

    73

    Nikolitsa, OE3ZGN|SV7DMC
    Petros, OE3ZZP|SV7BAX

    - - - - -

    1/3
    Building Complex Differential Error Regions

    October 2008

    Conference: ANAMET Meeting 30 At: NPL National Physical Laboratory, Teddington, UK Volume: Session III, Measurement uncertainty and
    traceability

    Abstract

    It is impossible to analytically estimate real uncertainties directly
    from complex differential errors - Real rectangular uncertainties are estimated from the orthogonal circumscribed about the complex DER -
    Real polar uncertainties are estimated from the annular sector
    circumscribed about the complex DER - The proposed method of
    uncertainty estimation, by building complex DERs, can be applied to
    any VNA calibration technique.

    https://www.researchgate.net/publication/230625680_Building_Complex_Differential_Error_Regions

    2/3
    Complex DERs in Non-Zero Length Thru VNA Measurements

    October 2009

    Conference: ANAMET Meeting 32, At: National Physical Laboratory NPL, Teddington, UK Volume: Session III

    Abstract

    After "Building Complex Differential Error Regions" for two-port
    calibration VNA measurements, in 30th ANAMET meeting, we are
    considering now the case of Non-Zero Length Thru or indirect Thru with
    SLOT calibration for either case of same type/sex connectors or
    different type connectors, on both ports.

    https://www.researchgate.net/publication/230625718_Complex_DERs_in_Non-Zero_Length_Thru_VNA_Measurements

    3/3
    Measurement Uncertainty in Network Analyzers: Differential Error
    Analysis of Error Models Part 4: Non-Zero Length Through in Full
    Two-Port SLOT Calibration

    September 2016

    Journal: FunkTechnikPlus # Journal, Issue: 11, Year: 4, Pages: V1-7 to V1-29

    The most accurate full two-port calibration of a VNA Vector Network
    Analyzer requires a Direct or Zero–Length Through connection. However,
    it is not uncommon at all to have one or two cables and a DUT Device
    Under Test with incompatible connectors, either of different type or
    of the same type/sex, which enforce then the use of some kind of
    barrel or adapter. Thus, in this paper, we study these cases of
    Indirect or Non-Zero Length Through, we estimate the effects of such connections on the measurement uncertainty by using our theory of Differential Error Regions and Intervals DERs/DEIs, and we evaluate
    our resulting method by applying it in practice to a built two–port network, which was measured against frequency with a SLOT calibrated
    VNA extended by two lengthy cables.

    https://www.researchgate.net/publication/358164808_Measurement_Uncertainty_in_Network_Analyzers_Differential_Error_Analysis_of_Error_Models_Part_4_Non-Zero_Length_Through_in_Full_Two-Port_SLOT_Calibration

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jos Bergervoet@21:1/5 to gin-pez arg on Fri Feb 4 23:50:32 2022
    On 22/02/04 10:26 AM, gin-pez arg wrote:
    ...
    The most accurate full two-port calibration of a VNA Vector Network
    Analyzer requires a Direct or Zero–Length Through connection. However,
    it is not uncommon at all to have one or two cables and a DUT Device
    Under Test with incompatible connectors, either of different type or
    of the same type/sex, which enforce then the use of some kind of
    barrel or adapter.

    You need a kind of "dummy DUT". I usually take something with a more
    or less known electrical length and not too much loss.

    Thus, in this paper, we study these cases of
    Indirect or Non-Zero Length Through, we estimate the effects of such connections on the measurement uncertainty by using our theory of Differential Error Regions and Intervals DERs/DEIs,

    Why don't you just stick to the matrix equations that Helton and
    Speciale already showed us in 1983? <https://www.semanticscholar.org/paper/A-Complete-and-Unambiguous-Solution-to-the-Problem-Helton-Speciale/6ca55d0258bbd7a1470a125ebd0aa1f2dfec4e7f>

    --
    Jos

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From gin-pez arg@21:1/5 to Jos Bergervoet on Sat Feb 5 06:42:16 2022
    On Friday, February 4, 2022 at 11:50:37 PM UTC+1, Jos Bergervoet wrote:
    On 22/02/04 10:26 AM, gin-pez arg wrote:
    ...
    The most accurate full two-port calibration of a VNA Vector Network Analyzer requires a Direct or Zero–Length Through connection. However, it is not uncommon at all to have one or two cables and a DUT Device
    Under Test with incompatible connectors, either of different type or
    of the same type/sex, which enforce then the use of some kind of
    barrel or adapter.
    You need a kind of "dummy DUT". I usually take something with a more
    or less known electrical length and not too much loss.
    Thus, in this paper, we study these cases of
    Indirect or Non-Zero Length Through, we estimate the effects of such connections on the measurement uncertainty by using our theory of Differential Error Regions and Intervals DERs/DEIs,
    Why don't you just stick to the matrix equations that Helton and
    Speciale already showed us in 1983? <https://www.semanticscholar.org/paper/A-Complete-and-Unambiguous-Solution-to-the-Problem-Helton-Speciale/6ca55d0258bbd7a1470a125ebd0aa1f2dfec4e7f>

    --
    Jos

    =

    Thank you for asking these and here are my replies:

    = "You need a kind of "dummy DUT". I usually take something with a more or less known electrical length and not too much loss."
    - I don't think so. If you are happy with it, then you really don't need what we said.

    "Why don't you just stick to the matrix equations that Helton and Speciale already showed us in 1983?"
    - Because, we were concerned ourselves with the uncertainties introduced by similar complete and unambiguous solutions. But, once again: If you are happy with it, then you really don't need what we said.

    pez

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jos Bergervoet@21:1/5 to gin-pez arg on Sat Feb 5 23:01:37 2022
    On 22/02/05 3:42 PM, gin-pez arg wrote:
    On Friday, February 4, 2022 at 11:50:37 PM UTC+1, Jos Bergervoet wrote:
    On 22/02/04 10:26 AM, gin-pez arg wrote:
    ...
    The most accurate full two-port calibration of a VNA Vector Network
    Analyzer requires a Direct or Zero–Length Through connection. However, >>> it is not uncommon at all to have one or two cables and a DUT Device
    Under Test with incompatible connectors, either of different type or
    of the same type/sex, which enforce then the use of some kind of
    barrel or adapter.
    You need a kind of "dummy DUT". I usually take something with a more
    or less known electrical length and not too much loss.
    Thus, in this paper, we study these cases of
    Indirect or Non-Zero Length Through, we estimate the effects of such
    connections on the measurement uncertainty by using our theory of
    Differential Error Regions and Intervals DERs/DEIs,
    Why don't you just stick to the matrix equations that Helton and
    Speciale already showed us in 1983?
    <https://www.semanticscholar.org/paper/A-Complete-and-Unambiguous-Solution-to-the-Problem-Helton-Speciale/6ca55d0258bbd7a1470a125ebd0aa1f2dfec4e7f> >
    Thank you for asking these and here are my replies:

    = "You need a kind of "dummy DUT". I usually take something with a more or less known electrical length and not too much loss."
    - I don't think so.

    If you don't think so, you don't understand the calibration principles.
    You need at least one standard which is a through (among the five
    standards that you need as a minimum).

    If you are happy with it, then you really don't need what we said.

    That remark does not have any convincing (or even relevant) technical
    or scientific content.

    ...
    "Why don't you just stick to the matrix equations that Helton and Speciale already showed us in 1983?"
    - Because, we were concerned ourselves with the uncertainties introduced by similar complete and unambiguous solutions.

    None are "introduced" by the equations. Uncertainty is introduced by
    the calibration "standards" used (you need at least 5 of them and more
    than five can help of course). The mathematical relations connecting
    them to the error-network T-matrix do not introduce uncertainties,
    they are only showing you how those errors propagate.

    But, once again: If you are happy with it, then you really don't need what we said.

    You may have something to offer, but by reverting to advertisement
    jargon you will not convince many readers here. You still have not
    justified why anyone would be better served by what you "said", then
    by the mathematically exact relations that have been well-known for
    decades now.

    All n-port calibration is basically just group theory of Sp(2n, C).

    --
    Jos

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jos Bergervoet@21:1/5 to gin-pez arg on Sun Feb 6 15:37:37 2022
    On 22/02/06 3:34 PM, gin-pez arg wrote:
    On Saturday, February 5, 2022 at 11:01:43 PM UTC+1, Jos Bergervoet wrote:
    On 22/02/05 3:42 PM, gin-pez arg wrote:
    On Friday, February 4, 2022 at 11:50:37 PM UTC+1, Jos Bergervoet wrote: >>>> On 22/02/04 10:26 AM, gin-pez arg wrote:
    ...
    The most accurate full two-port calibration of a VNA Vector Network
    Analyzer requires a Direct or Zero–Length Through connection. However, >>>>> it is not uncommon at all to have one or two cables and a DUT Device >>>>> Under Test with incompatible connectors, either of different type or >>>>> of the same type/sex, which enforce then the use of some kind of
    barrel or adapter.
    You need a kind of "dummy DUT". I usually take something with a more
    or less known electrical length and not too much loss.
    Thus, in this paper, we study these cases of
    Indirect or Non-Zero Length Through, we estimate the effects of such >>>>> connections on the measurement uncertainty by using our theory of
    Differential Error Regions and Intervals DERs/DEIs,
    Why don't you just stick to the matrix equations that Helton and
    Speciale already showed us in 1983?
    <https://www.semanticscholar.org/paper/A-Complete-and-Unambiguous-Solution-to-the-Problem-Helton-Speciale/6ca55d0258bbd7a1470a125ebd0aa1f2dfec4e7f> >
    Thank you for asking these and here are my replies:

    = "You need a kind of "dummy DUT". I usually take something with a more or less known electrical length and not too much loss."
    - I don't think so.
    If you don't think so, you don't understand the calibration principles.
    You need at least one standard which is a through (among the five
    standards that you need as a minimum).
    If you are happy with it, then you really don't need what we said.
    That remark does not have any convincing (or even relevant) technical
    or scientific content.

    ...
    "Why don't you just stick to the matrix equations that Helton and Speciale already showed us in 1983?"
    - Because, we were concerned ourselves with the uncertainties introduced by similar complete and unambiguous solutions.
    None are "introduced" by the equations. Uncertainty is introduced by
    the calibration "standards" used (you need at least 5 of them and more
    than five can help of course). The mathematical relations connecting
    them to the error-network T-matrix do not introduce uncertainties,
    they are only showing you how those errors propagate.
    But, once again: If you are happy with it, then you really don't need what we said.
    You may have something to offer, but by reverting to advertisement
    jargon you will not convince many readers here. You still have not
    justified why anyone would be better served by what you "said", then
    by the mathematically exact relations that have been well-known for
    decades now.

    All n-port calibration is basically just group theory of Sp(2n, C).

    --
    Jos

    - - - - -

    : )

    Well, the judgment to the readers then!

    There are still three of them, I believe..

    --
    Jos

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From gin-pez arg@21:1/5 to Jos Bergervoet on Sun Feb 6 06:34:27 2022
    On Saturday, February 5, 2022 at 11:01:43 PM UTC+1, Jos Bergervoet wrote:
    On 22/02/05 3:42 PM, gin-pez arg wrote:
    On Friday, February 4, 2022 at 11:50:37 PM UTC+1, Jos Bergervoet wrote:
    On 22/02/04 10:26 AM, gin-pez arg wrote:
    ...
    The most accurate full two-port calibration of a VNA Vector Network
    Analyzer requires a Direct or Zero–Length Through connection. However, >>> it is not uncommon at all to have one or two cables and a DUT Device
    Under Test with incompatible connectors, either of different type or
    of the same type/sex, which enforce then the use of some kind of
    barrel or adapter.
    You need a kind of "dummy DUT". I usually take something with a more
    or less known electrical length and not too much loss.
    Thus, in this paper, we study these cases of
    Indirect or Non-Zero Length Through, we estimate the effects of such
    connections on the measurement uncertainty by using our theory of
    Differential Error Regions and Intervals DERs/DEIs,
    Why don't you just stick to the matrix equations that Helton and
    Speciale already showed us in 1983?
    <https://www.semanticscholar.org/paper/A-Complete-and-Unambiguous-Solution-to-the-Problem-Helton-Speciale/6ca55d0258bbd7a1470a125ebd0aa1f2dfec4e7f> >
    Thank you for asking these and here are my replies:

    = "You need a kind of "dummy DUT". I usually take something with a more or less known electrical length and not too much loss."
    - I don't think so.
    If you don't think so, you don't understand the calibration principles.
    You need at least one standard which is a through (among the five
    standards that you need as a minimum).
    If you are happy with it, then you really don't need what we said.
    That remark does not have any convincing (or even relevant) technical
    or scientific content.

    ...
    "Why don't you just stick to the matrix equations that Helton and Speciale already showed us in 1983?"
    - Because, we were concerned ourselves with the uncertainties introduced by similar complete and unambiguous solutions.
    None are "introduced" by the equations. Uncertainty is introduced by
    the calibration "standards" used (you need at least 5 of them and more
    than five can help of course). The mathematical relations connecting
    them to the error-network T-matrix do not introduce uncertainties,
    they are only showing you how those errors propagate.
    But, once again: If you are happy with it, then you really don't need what we said.
    You may have something to offer, but by reverting to advertisement
    jargon you will not convince many readers here. You still have not
    justified why anyone would be better served by what you "said", then
    by the mathematically exact relations that have been well-known for
    decades now.

    All n-port calibration is basically just group theory of Sp(2n, C).

    --
    Jos

    - - - - -

    : )

    Well, the judgment to the readers then!

    pez

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)