Hello guys,

I recently picked up an old vector network analyser in working order
but minus the 3 interconnects between it and the
transmission/reflection bridge. These are specified in the manual to
be exactly 24" long, 50 ohms and terminated by N-type plugs.
I've been told the length of these patch leads is quite critical to
getting accurate measurements with this VNA, but am at a loss to work
out why 24" is specified when the frequency range of this device is
4Mhz to 1300Mhz. If it were only capable of measuring at one fixed
frequency, I could understand the need for a specifically cut length
of some fraction of a wavelength.
Can anyone explain the relevance of 24" in this context?
Also, will *any* 50 ohm coax suffice for this purpose or has it got to
be something special?
Thanks,
P.
--

"What is now proved was once only imagin'd" - William Blake

On Sat, 03 Dec 2005 16:52:30 +0100, Paul Burridge
k wrote:

Hello guys,

I recently picked up an old vector network analyser in working order
but minus the 3 interconnects between it and the
transmission/reflection bridge. These are specified in the manual to
be exactly 24" long, 50 ohms and terminated by N-type plugs.
I've been told the length of these patch leads is quite critical to
getting accurate measurements with this VNA, but am at a loss to work
out why 24" is specified when the frequency range of this device is
4Mhz to 1300Mhz. If it were only capable of measuring at one fixed
frequency, I could understand the need for a specifically cut length
of some fraction of a wavelength.
Can anyone explain the relevance of 24" in this context?
Also, will *any* 50 ohm coax suffice for this purpose or has it got to
be something special?

Maybe a mention of the model number would be helpful. My best guess
is that the absolute length is not very important as long as all three
are the same length.

The stability of the cable is likely more important than the exact
type.

On Sat, 03 Dec 2005 14:50:02 -0700, Wes Stewart
wrote:

Maybe a mention of the model number would be helpful.

Thanks Wes (and others - I'm building up a picture of what's what from
your contributions...)

The VNA in question is an HP 8754A made in about 1979.

My best guess
is that the absolute length is not very important as long as all three
are the same length.

That would explain a lot.

The stability of the cable is likely more important than the exact
type.

Yup. So I've been given to understand. But is the loss factor relevant
over just 2 feet at 1300Mhz?
BTW, the guy that asked about the impedance of RG8: it *is* 50 ohms as
far as I can ascertain and if it *were* 52, it surely *would* screw up
the measurement, IMHE!

--

"What is now proved was once only imagin'd" - William Blake

On Sun, 04 Dec 2005 00:45:00 +0100, Paul Burridge
k wrote:

On Sat, 03 Dec 2005 14:50:02 -0700, Wes Stewart
wrote:

Maybe a mention of the model number would be helpful.

Thanks Wes (and others - I'm building up a picture of what's what from
your contributions...)

The VNA in question is an HP 8754A made in about 1979.

My best guess
is that the absolute length is not very important as long as all three
are the same length.

That would explain a lot.

The stability of the cable is likely more important than the exact
type.

Yup. So I've been given to understand. But is the loss factor relevant
over just 2 feet at 1300Mhz?

Seriously doubt it.

BTW, the guy that asked about the impedance of RG8: it *is* 50 ohms as
far as I can ascertain and if it *were* 52, it surely *would* screw up
the measurement, IMHE!

Doubt that too.

I can't find any data on this box, but it's likely that the
interconnects are carrying 1) the rf excitation out of the mainframe,
2) a signal from the referance arm of a bridge and 3) the
reflected/transmitted signal.

When calibrated, as I believe I described in another thread, errors
(if any) caused by the interconnects should wash out.