Merging / splitting DC/AC signals sharing a conductor...

[Externet]

Hi all.
Looking for recommendations to connect a NTSC color surveillance camera of the garden variety, which has 12VDCin and video out connections, to a single coaxial submarine extension; it will be less than 50 feet.

I believe should be like this:

CV>----------||--------------J........................................J------------||-------------->Video
C+<---------uuuu--------J-----------------coax----------------J-----------uuuu---------<+12V

Ignore the dots above the word coax.
All cables are shielded, (-) ground to shield
J are junctions, nodes, connection
Capacitors --||-- intended to allow video pass and block DC
Coils ----uuuu---- intended to allow DC pass and block video
Between splices 'J' is a ~40 feet long RG-59 or suitable coaxial cable
C is camera end
Arrows show signal 'flow'
V is video
+ is DC supply

I believe the NTSC video signal takes 4.5 MHZ bandwidth, the video impedance about 75 Ohm.
How should I calculate optimal values for the capacitors and inductors IF my schematic contraption makes sense ? If any other solution, welcome!

Thanks,
Miguel

[MrAl]

Hi there,


Just wondering, how much current does your camera draw from the
12v power supply?

Just to note... strictly speaking, the NTSC signal isnt really an AC
signal is it? I believe it has ac components but also a dc component.
This doesnt prevent you from using this method though, as i think
we can come up with something that works.
For example, replace the two coils with resistors of the proper value,
raise the 12v power supply to 15 volts (or something like that),
then install a 12v regulator with caps at the camera end, then modulate the
signal on the coax center lead with an amplifier from the video,
then use an amp/level converter at the receive end to restore the
video to proper levels.
We need to know how much current your camera draws however.

[Robert Reed]

This set up is very common in microwave front end circuits and is called a "Bias Tee". The inductors should be of a value that have a self resonant frequency well above your signal frequency (near 0 - 4.5 Mhz). or are they riding on a carrier. The capacitors should be chosen to have a low reactance at the lowest operating frequency which may make them fairly large to handle your low end, especially since you are working with 75 ohm source/load impedance. The inductors should be rated to carry the circuits specified current and can be seriesed with some low value resistors (5-25 ohm) if needed to reduce circuit 'Q' in respect to the DC path of the cable. After recovery on the load end DC can be restored with the proper Op -Amp circuitry. The newer SMD Op-Amps will handle this especilly in view of the fact that you need only unity gain at that point.

[MrAl]

Hi again,


I dont believe inductors are really required, but yes i guess it
doesnt matter that much if you want to do it that way.
In any case the dc level has to be restored at the receive end.
I would also use a nice cap at the camera end too to help remove
any left over ac.

[haklesup]

I have a splitter that does that, it came with my cable amplifier.

One cable input goes to a wall wart and another goes to the cable(comcast). The output of this splitter goes to the RF+power input of my amplifier. In other words, there is an off the shelf part out there somewhere to do this.

http://cgi.ebay.com/ELECTROLINE-8-PORT- ... dZViewItem

Note the Remote power input and small splitter in the photos in this auction.

I noticed it was called a power inserter so searching on that I found

http://mjsales.net/items.asp?FamilyID=7 ... &Cat2ID=35

and

http://www.lectrosonics.com/manuals/biastman.pdf

[Externet]

Nice help, MrAl, Robert and haklesup.

Thanks. Did not know the inserters were available commercially.

As even the small ones are somewhat bulky for my application, I still prefer to calculate my own. The consumption of the camera is 150mA. The wire gauge of the coil should not drop more than a volt in insertion (~10%). Am fine there.
Actually, I believe the cameras are 5V and a regulator is built-in them, so a higher drop is permissible.

The larger the 'choke' inductance, the higher the rejection to video pass. The higher the capacitance, the lower the insertion loss to video. But for real figures, am not sure how to calculate this one. Empirically, yes, I could try and evaluate.
If the NTSC signal has any DC levels that I should be aware of, is another chapter.

Miguel

[rshayes]

The NTSC signal does have a substantial DC component. Separating this from the the DC supply mignt be tricky, since DC variations in the supply voltage would be interpreted as changes in brightness.

The choke would have to have substantial reactance at 60 Hz, since the NTSC signal has strong spectral components in the 60 Hz range. The high end of the NTSC spectrum is in the megahertz range, and winding a choke that is effective at 60 Hz with a self resonant frequency of several Mhz would be difficult, if not impossible.

If the video signal is modulated onto a high frequency carrier, either with AM or FM, the chokes and capacitors needed to combine and separate the power and signal become much smaller and easier to build. If the carrier is in the 50 MHz range, a transformer can be used to couple thw signal onto the coax line. This could be as simple as a few turns of wire on a ferrite bead.

[Lenp]

I don't think there are DC levels associated with NTSC transmission. We use balun transformers to send NTSC over Cat-5 cable, and they provide complete isolation from the unbalanced coax to the balanced Cat-5. It really is just a transformer in a box, available anyplace video equipment is sold. They also have audio/video versions as well.

Just droppig in a thought.

[ecerfoglio]

Thanks. Did not know the inserters were available commercially.

As even the small ones are somewhat bulky for my application, I still prefer to calculate my own. The consumption of the camera is 150mA. The wire gauge of the coil should not drop more than a volt in insertion (~10%). Am fine there.
Actually, I believe the cameras are 5V and a regulator is built-in them, so a higher drop is permissible.

The larger the 'choke' inductance, the higher the rejection to video pass. The higher the capacitance, the lower the insertion loss to video. But for real figures, am not sure how to calculate this one. Empirically, yes, I could try and evaluate.
If the NTSC signal has any DC levels that I should be aware of, is another chapter.

Miguel

The "DC + signal" systems I have seen (like the power inserters) don't use NTSC video signal but RF modulated by the video signal.

They were (are?) used to power TV boosters (amplifiers) that were mounted on the antena.

You may add an RF modulator along with your camera and "transmit" on cannel 2 or 3 inside your coax. At the receiving end add an old plain TV tuner. The higher frecuencies and no dc component of the RF signal will make it easier to separate them.

[MrAl]

Hi again,

Lemp, that's an intestesting view. Not all baluns are wired the
same however, so maybe a simple test is in order.

I propose a test:
Connect a capacitor of say 1uf or 10uf between the 'normal' transmitter
and receiver, in the center line, and see if the video transmits
unaltered (ie the picutre looks exactly the same with or without the cap).
If this test passes than the NTSC signal being dealt with is not going to
be affected by the dc levels in the signal, but if it fails then the dc level
has to be restored.

The only balun i have on hand to test is a TV type balun, which might
not be built the same, and that will not pass dc because one winding
would short it out completely so it would have to be ac, but that
is for the higher frequency TV signals, not composite video which i
think we are talking about here.
I worked with these some time ago but cant remember how those
camera baluns are wired now. Oh yeah, and they have to be the
passive type too not the active type to understand if they pass dc.

Of course having to pass 60Hz is also a bit of a task too as mentioned
by rshayes, so perhaps we are better off using resistors instead of
inductors? Any other input here?

[MrAl]

Hi again,

I found some specs on a balun used for CATV transmission line
matching, and it is clearly spec'd for DC to 8MHz, unlike the TV balun
which because of the way it is wired, would present a short circuit
to any DC signal while still passing very high frequency signals.

[Robert Reed]

Hello Exter
In my first reply, I was not sure whether your base band was riding on a carrier or not and I sort of assumed it was. If you are truly handling only the base band (DC to 4.5 Mhz) and as Steven pointed out, it probably would be impossible to select an RF choke to cover that bandwidth ratio- Millions to one.
You would also need this to have a reactance of at least 10 tlmes the transmission line characteristic impedance - maybe approaching 1000 ohms. Then a thought occurred me (now assuming you are only carrying that base band) that this could simply be accomplished by carrying your DC power thru a single strand of #22 wire taped or nylon tied to your cable.
Admittedly, not very pretty or professional, but you can now DC couple your video to it's termination and eliminate a whole slew of problems that would be involved in designing a bias "T" for this system and the DC restoration circuit that would have to follow.

[Externet]

Thanks fellows.
Robert, if I use Belden 9451 plain 100% shielded 2 conductor + shield instead of a 75 Ohm "RG-59" coaxial, using one conductor for video and one for power, I could comply with your suggestion with no splitting/merging filtering at all. Simple direct wiring.
The 9451 is 45 Ohm impedance, (not that far); vulgar microphone cable of widely common use.
Perhaps I can aim for another cable with closer impedance match, but unsure if chosing a -dual- 75 Ohm cable will present other loading/capacitive effect against the conductor carrying +12V.
Suspect the loading will not be significant for the length, but is just a guess.
(Any externally attached wire would be more prone to tangle in a submarine application),

bwccat.belden.com/ecat/pdf/9451.pdf

Miguel

[MrAl]

Hi again,

Here's a basic layout drawing...

[Robert Reed]

Exter
The 9451 cable has 34 Pf per foot and a nom. impedance of 45 ohms between conductors. It has 67 pf per foot and an unknown impedance from either conductor to ground (probaly lower than 45 ohms). This was basically designed for as a balanced line. In the range of frequencys you have to handle, I think the line would just look like a lumped capacitance at the very low end - about 2500 Pf for your length. I have a hunch that the output stage driving this line would probably just use a line driver type of chip and therefore not be so fussy as to the type of line. If you have the equiptment still set up on the bench, you could note the effect the cable has on that. Maybe terminate it with 100 ohms or higher for experimentation. One word of caution if you do go this route, is to make sure the DC power feed is very clean as it will have 40 feet of coupling to the signal conductor,
There are a line of "siamese" cables made by Belden that you might want to look at and one that comes to mind is type 5499X5. It is a 75 ohm coax (16.3 Pf per foot) and a pair of #18 conductors all in one cable. These are called siamese as their actually joined side by side and can be separated at their ends so as to allow type 'F' or BNC connectors to the coax portion. They have a whole family of this type that you may want to further investigate on their E-Catolog. Have no idea what the pricing is thogh. Good luck and let us know what you come up with.