Re: Insertion delay needed for BNC-BNC couplers UG491A/U (M-M) and UG914/U (F-F)

Kurt Poulsen
 

Hi John
You are doing some calculations which is not reflecting the physics. The BNC adaptors characteristics dependent on the mating part. When both are fitted with isolation (PTFE) then the velocity factor is 0.695. Bur the SDR Kits male short adaptor is fitted with isolation but the open is not. You are trying to involve the female short into the calculation but that is so the delay thru the FF adaptor terminated with the male short is not just the summed delays. E.G. the calculated value 77.11ps is not the same value as the S21 thru delay which is 72.45ps. It is far more complex and these delays you use are simple delays which is the average delay over the frequency span. In the arbitrary settings you will see that there is a long expression which is the frequency dependent delays for the male and female short and open. Load the female kit and create a custom trace with the expression s_open as well as s_short and right click on the custom trace, select copy and paste the copy to S11. Then used extension port delay and trim delay to a horizontal phase line and you will se it is not possible as it is very frequency dependent. So you cannot claim a VF=0.67 due to this, sorry to say it is not that simple to find the male short delay based on mechanical dimensions. The internal structure of the female female adaptor is also not ideal as there sare sections not filled with PTFE that is why I have so implemented so complicated expressions in the cal kit definitions.
Kind regards
Kurt


-----Oprindelig meddelelse-----
Fra: VNWA@groups.io <VNWA@groups.io> På vegne af John KN5L
Sendt: 23. maj 2020 01:23
Til: VNWA@groups.io
Emne: Re: [VNWA] Insertion delay needed for BNC-BNC couplers UG491A/U (M-M) and UG914/U (F-F)

Hi Kurt,

Yes, 3.333... is correct value. Normally I use VNWA port extension window to compute physical length from time.

SDR-Kit 4-pc cal kit, which uses a F-F adapter with dimensions, length = 32.5mm and includes Amphenol marking. Looks very much like:
https://www.amphenolrf.com/downloads/dl/file/id/3542/product/5386/031_219_rfp1_customer_drawing.pdf

Based on SDR-Kit calibration parameters, Short difference between Male and Female is 77.11pS one way.

Measured external surface to plane = 8.5mm (both ends)

Plane to plane = 32.5 - 8.5 - 8.5 = 15.5mm

15.5 * 3.3333 = 51.66pS

Vf = 51.66 / 77.11 = 0.670

Correction for Short cal device calculated Plane to short condition using 3.3333ps/mm is 11.4mm

Can't tell where the short condition plane is within the device, but 11.4mm between ref and short seems plausible.

John KN5L

On 5/22/20 5:23 PM, Kurt Poulsen wrote:
Hi John and David
Sorry for interfering I this debate. The calculation of the delay is
not correct. You state 3ps/mm it is 1/3 = 3.3333333333ps/mm so better to divide by 3. The velocity factor is indicated to 0.67 and for ptfe (Teflon) is 0.695 (some manufacturer state 0.7 The question not yet responded to asked by Dave is why a negative delay and even twice the delay. It is the convention used by the VNWA and to be understood as the time travel from the calibration plane to the reflection point and back again to the calibration plane. A negative because that is the travel time for moving the reflection point back to the calibration plane. In the VNWA the Extension port delay demonstrate clearly this phenomena as adding a positive one way delay, is like moving the calibration plane forward e.g. to the end of a short where the phase and ImageZ (=0) trace then is straight line (Phase reference 180/-180degree).
This is contradiction to the HP and other manufacturer which uses the term offset delay which for calibration standards are a positive number and is just a term describing the one way travel time form calibration plane to reflection point/plane. The calculation inside the HP VNA used this to calculate the double offset for a reflection, so it is a matter of the formulas used.
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: VNWA@groups.io <VNWA@groups.io> På vegne af John KN5L
Sendt: 22. maj 2020 23:01
Til: VNWA@groups.io
Emne: Re: [VNWA] Insertion delay needed for BNC-BNC couplers UG491A/U
(M-M) and UG914/U (F-F)

Hi Dave,

On 5/22/20 3:11 PM, David Feldman via groups.io wrote:
trying to understand, so please bear with (and correct) me --- I
gather that at ~3 pSec/mm, the physical distance between the BNC
short
BNC Male Short:
https://www.amphenolrf.com/downloads/dl/file/id/7341/product/5273/2021
14_customer_drawing.pdf

Male ref plane is outside edge. Using drawing above, full length is 15mm. ref plane is recessed a little and short is also recessed a little from end. Lets call ref plane to short as ~13mm. Velocity factor of 0.67.

13mm * 3p/mm * 1/0.67 * 2 = 116.4ps

Measured value is 113.5ps. Physical length for 113.5ps = 12.7mm

John KN5L
--
Best 73 de Kurt de OZ7OU

Join VNWA@groups.io to automatically receive all group messages.