Hi, Nice site, I'm a first time poster and hope this is correct forum.

We are modeling 20Ghz and 50 Ghz oscope bandwidth tests that use source, spitter and power meter/sensor so a level signal is presented to the oscope input. We have uncorrected residual errors, primarily power measurement accuracy, splitter tracking, and correction errors on cal kit/ VNA as uncertainty contributors.

Since we are using E8257D, 11667B,C, 8487A-H84, E8361A w/85056A 85052C 85058B already and have to connect oscope to get complex input parameter we'd like to consider calibrating VNA source power and s11 using mechanical cal kits then using the source output power to establish our reference level.

To do this we need source power to be known and constant with varied load ( 8487A source cal check and S11 of oscope input) so reference level is the same at low freq and observed 3 db point on oscope.

Seems like we eliminate the splitter tracking and source match errors and add in the variation in output power due to VNA source match with different reflections from sensor to oscope input.

Any idea of how constant calibrated VNA source power is over rho 0.1 to 0.4 range?

Any obvious misses on uncertainty contributors?

Thanks,

Karl

We are modeling 20Ghz and 50 Ghz oscope bandwidth tests that use source, spitter and power meter/sensor so a level signal is presented to the oscope input. We have uncorrected residual errors, primarily power measurement accuracy, splitter tracking, and correction errors on cal kit/ VNA as uncertainty contributors.

Since we are using E8257D, 11667B,C, 8487A-H84, E8361A w/85056A 85052C 85058B already and have to connect oscope to get complex input parameter we'd like to consider calibrating VNA source power and s11 using mechanical cal kits then using the source output power to establish our reference level.

To do this we need source power to be known and constant with varied load ( 8487A source cal check and S11 of oscope input) so reference level is the same at low freq and observed 3 db point on oscope.

Seems like we eliminate the splitter tracking and source match errors and add in the variation in output power due to VNA source match with different reflections from sensor to oscope input.

Any idea of how constant calibrated VNA source power is over rho 0.1 to 0.4 range?

Any obvious misses on uncertainty contributors?

Thanks,

Karl

You might consider using the external leveling capability of the the signal generator and the power meter to level the power at the splitter. The effective source impedance of a splitter is very good in this mode

see this note:

http://cp.literature.agilent.com/litweb ... 6699EN.pdf

or here:

http://www.home.agilent.com/agilent/faq ... &id=524913

you can characterize the actual equivalent output Z of the splitter using the VNA, then ade this number together with the scopes measured input impedance to calculate the actual voltage at the input.

If you want to use the VNA source, the actual source impedance of the VNA is available, after you have perfromed a 1port or 2port calibration - the is the Esf error term.