The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
From: Ian Walker
Subject: Re: How to check for a 50 or 75 ohm connector
Date: Mon, 23 Dec 2002 19:24:45 +0000
NNTP-Posting-Date: Mon, 23 Dec 2002 20:01:08 +0000 (UTC)
User-Agent: Turnpike/6.02-U ()
In article , DarkMatter
>On Sun, 22 Dec 2002 23:48:54 +0000, Ian Walker Gave
>>In article , DarkMatter
>>>On Sun, 22 Dec 2002 21:31:08 +0000, Ian Walker Gave
>>>>In article <email@example.com>, DarkMatter
>>>>>On Sun, 22 Dec 2002 15:15:52 +0000, Ian Walker Gave
>>>>>>Wrong! Whilst 50 and 70 ohm N (BNC) type connectors are both N (BNC)
>>>>>>type they are incompletely described by calling them N (or BNC) type.
>>>>>>You must also specify the impedance for correct electrical (and
>>>>>>mechanical in the case of N type) performance, the type of cable
>>>>>>providing it is suitable for the connector is irrelevant (i.e. you can
>>>>>>have RG-174 on one side of an N system and RG-203 on the other).
>>>>> You prove in your own statement that it IS indeed the cable that is
>>>>>the determining factor.
>>>>Only in that the properties of the system determine the type of cable
>>>>that must used (RG-174, RG-58, RG-203, Su-104, etc.), both the cable and
>>>>the system determine the connector that must be used (BNC, N, C, K,
>>>>PC3.5, PC7, 2.4, 1.0, etc.). If the connector side is not designed to
>>>>have the same impedance as the cable then you will get a reflection from
>>> Wrong. SOME BNC connectors have differences in the media between
>>>the grounding shroud, and the center pin. That media is usually air.
>>>Some use a hard dielectric.
>>They might be mostly air as in the case of precession low loss N type,
>>but they must have some hard dielectric otherwise you would not be able
>>to maintain concentricity with a consequent degradation in return loss
>>and an increased risk of damage which would further degrade
From concentric // adj. (often foll. by with)
(esp. of circles) having a common centre
>Of what? The grounding shroud?
Yes, the centre inner must be on the line formed by the centre of the
outer in order to produce a non-reactive transmission line.
> Are you joking?
> AGAIN, we are NOT talking about "N" type BNCs, we are talking about
Again you introduce some mythical connector called '"N" type BNC' maybe
you are getting confused with "C" type which is ay least a bayonet like
the BNC but it is not called '"N" type BNC'. Am regret to say this only
typifies your ignorance of co-ax connectors
>>>The differences in losses or impedance
>>>will only show up at high GHz level frequencies and are all but
>>>negligible in many if not most commercial applications for such
>>Perfectly true that the effect is generally negligible in many
>>commercial applications; but not in all.
>>> Note that the current industry uses newer, smaller connectors for
>>>today's modern Ghz applications as BNC is too bulky in today's
>>Not only are the smaller connectors less bulky but they provide a better
>>performance above 1GHz.
> Which makes your previous statement all but moot.
There are several factors determining a designers choice of cable:
voltage stress, I squared R heating, loss due to voltage drop, cut-off
frequency, flexibility, and last size constraints. The physics of the
materials determine the cable which must be used, the equipment must
accommodate the size of the cable.
>>>>If you have only got a bag of unidentified plugs you will have to
>>>>measure the inner and outer diameters and, along with a guess at the
>>>>dielectric calculate the impedance.
>>> All you have to do is determine the cable type that it was meant to
>>>be terminated with. That cable's impedance determines the impedance
>>>of the jumper, and the correct connector to use for it. The connector
>>>itself MUST be identical on the PLUG/JACK side so that anyone can plug
>>>into any jack when both are declared as BNC. Sheesh, this is so
>>>simple, I can't believe you don't get it.
>>There are too many types of cable to make this practical,
> I'm sorry, but AMPHENOL lists over a hundred coax types that mate to
>BNC front ends.
Which only serves to confirm my point that there are too many to make
measuring the rear of the connector and then comparing against all the
possible cables practical. It is much easier to measure the front where
all you need do is measure the thickness of any solid dielectric between
the inner and outer conductors.
>> it would even
>>be possible to have cables with identical inner and outer diameters but
>>different impedances by choice of dielectric. I have not looked for
>>cables of this type, but then again that is how 50 and 75 ohm BNC
>>connectors are implemented. Is it sooo difficult to understand, if the
>>inner and outer must be the same diameter to allow non-destructive
>>mating then the only way of making the impedance different is to change
>>the dielectric between them. I can not believe you do not understand
> What? That is what I said. The pin, and the shroud are the same
>for all. The media between is all that changes. That IS exactly what
It is you who have insisted that the impedance is determined by the
cable side and not the mating side of the connector
>>> IT IS the CABLE that determines the jumper's impedance, and the
>>>connector is of very little sway in that regard. That FACT that one
>>>cannot use a RG-6 BNC connector on an RG-174 coax proves that. You
>>>can make either impedance jumper, and BOTH will still plug into any
>>>BNC connector. If someone wants to claim "Oh no. It won't plug into
>>>an "N" type BNC connector." then the idiot is too blind to grasp the
>>>fact that he is referring to a different connector entirely.
>>> BNC is BNC is BNC. Impedance is cable specific. Pretty simple.
>>> Another proof is in simple numbers. A 100 foot length of cable is a
>>>capacitor, for sure. A half inch connector is a negligible amount of
>>>that run length. If you remember ANY of your training, you will know
>>>that such tiny additions are typically rounded off.
>>We are not talking about capacitors but transmission lines.
> Impedance is a reactive term. That means capacitance and
>inductance. Guess which coaxial cable exhibit that determines their
>impedance the most?
A transmission line is best represented as an infinite lumped LCR
network and can be drawn as follows
¦ ¦ ¦
C C C
¦ ¦ ¦
Where there are an infinite number of LCR elements in a finite length of
cable (you will need your calculus for this), XL = -XC (thus its
impedance is not frequency dependent) and R increases with frequency
(this is inevitable).
> Of course we are talking about capacitance. Have you ever seen an
>HV transmission coax? It looks JUST LIKE a large diameter coax for an
>old CB radio set.
It only looks capacitive at DC, not at AC where it is a transmission
line and we are talking about transmission lines. Even at DC the applied
step voltage requires that it be analysed as a transmission line.
> When one applies the juice through it, however, the thing jumps.
>That is a physical contraction due to capacitive charging.
The charge is not instant along the length of the cable but propagates
along it at the characteristic velocity (VP), until the charge reaches
the far end, and any reflections have died out it must be analysed as a
> We have always been talking about a capacitor.
We are talking about a transmission line
>If one looks at a
>circuit where there is a jumper between two circuit element, and that
>jumper is coaxial, it appears as a capacitor to the circuit.
It appears as a transmission line.
>Every time. Unless used to pass DC, of course. Even in that case, it
>does charge up. I can make various HV capacitive loads for testing HV
>supplies by simply changing the cable length.
Of course you can, but the normal use of a co-axial line, and the
function for which it was designed is as a transmission line
> It IS basic electronics. Capacitive reactance is called impedance.
As can inductive reactance, as also can any combination of inductance
and capacitance. They should always be written as either "R+jX" or "R
theta degrees" (of course you know that the two R in these expressions
re not identical but are related by converting between Cartesian and
polar co-ordinates). For the frequency at which XL = -XC the impedance
will be indistinguishable from a resistor and can be safely written as R
(you might have to specify the frequency). In the special case of a
transmission line XL = -XC for all frequencies of use.
It IS basic transmission line theory, which is perhaps a bit more
advanced than basic electronics; but then most electronics has to be
interconnected and whilst you might ignore transmission line theory at
Dc you can not do so at AC.
>>If I connect
>>two 50 ohm BNC cables via a 75 ohm coupler it will exhibit a worse
>>return loss than if I use a 50 ohm coupler. If I really feel bored and
>>have nothing useful to do tomorrow then I might perform this experiment
>>with an SNA, I won't confuse you with a VNA and a sliding termination.
> Hahahah... whatever.
I conducted the experiment using the following HP (Agilent) apparatus:
85032B, 8757A, 8350B, 83595A and 85027A. I also used the following
adapters: two off Suhner 33 BNC-N-50-1, a 50ohm BNC female coupler, and
one 75 ohm female BNC coupler as the item to be investigated.
I connected one of the 33 BNC-N-50-1 to the 85027A via an APC7 - N(m)
adapter from the 85032B and fitted a 909F from the 85032A to the second
33 BNC-N-50-1 and coupled it to the first 33 BNC-N-50-1 using the 50 ohm
BNC coupler. I then normalised this and replaced the 50 ohm coupler with
the 75 ohm and plotted the results. This produced a rising response from
~odB at 10MHz to +18dB at 1GHz. Since there was no cable in the
transmission line under test, only the coupler, this indicates that the
properties of the connectors is determined by more than the cable they
Had I used an SNA I would have obtained readings in the form R+jX or R
Using a step driven TDR produces the following:
----------- ----------- >
A B C D
A-B is the 50 ohm line from the step generator and any adapters
required to allow connecting the test line,
B-C is the 75 ohm test line
C-D is the 50 ohm termination after the test line, and in fact
continues to infinity.
If normalised to 50 ohm then the line B-C represents an SWR of 1.5:1.
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup