Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[MD Corie]

A newbie question:

Background info:
I have 16 SC23A (wired) cameras connected together using the "Y" technique specified on the X10 site.  Several of these cameras are located further away than the ~60' of supplied cable, so I have added the Cat5e four-twisted-pair cable (also as specified by X10) to extend the distance.  I assumed that I should use the standard 8P8C modular connectors as are used when the Cat5e cable is used for is normal LAN applications, so I terminated the Cat5e cables with those connectors and then used in-line modular adapter jacks to connect the cables to the respective camera cables.  This all "sort of" works, but has some performance problems - hence my questions:

With the cameras all wired together this way (and with only one camera activated at any given time), the image quality is seriously degraded:  It appears "foggy" and also has multiple "ghost" images (about five or six that are apparent) that are slightly offset in a cascade to the right of the primary image.  This appears to be video signal "reflections", most likely caused by mis-matched impedances somewhere in the system - like one might get when using an RF coax TV cable with unbalanced connections - but I'm a little surprised to see this sort of effect at the relatively low frequency of the line-level video signal.  So, my question is whether my theory of reflections caused by mis-matched impedances is valid, or if there is some other explanation for the problem?

By the way, if it IS caused by signal reflections, then what do I need to do to correct the problem?  Do I need some sort of baluns?  If so, what, and where?  Where can I find out the impedance of the cameras' video outputs and/or the TV's video input, so that I can put in appropriate components to match those impedances?

Thanks in advance...

[Brian H]

"Y" Technique?
You have more than one connected to the same video input?

I will see if I can find this data on the X10 site.

[dave w]

Brian H and I are probably going down the same road. You have 16 cameras "Y"ed down to one output? I'm surprised you have any picture at all. Each camera has an output impedance of around 72 ohms which is present whether the camera is ON or OFF ("standard" video camera output impedance is usually 72 - 75 ohm, but who knows with X10). So essentially you have 16, 72 ohm resistors all in parallel configuration, or an output impedance of about one ohm. "Y"ing 2 or 4 cameras will work, but 16...no work.

You need find a 16 input video switcher.

[Brian H]

Well the wiki does show the 'Y' method but with 16 cameras you are asking for interference.
http://kbase.x10.com/wiki/Connecting_Multiple_Wired_Cameras

[MD Corie]

Brian H and I are probably going down the same road. You have 16 cameras "Y"ed down to one output? I'm surprised you have any picture at all. Each camera has an output impedance of around 72 ohms which is present whether the camera is ON or OFF ("standard" video camera output impedance is usually 72 - 75 ohm, but who knows with X10). So essentially you have 16, 72 ohm resistors all in parallel configuration, or an output impedance of about one ohm. "Y"ing 2 or 4 cameras will work, but 16...no work.

You need find a 16 input video switcher.

Thanks for the reply.  Actually, there is a somewhat usable picture, but it is rather poor - even without the "ghosts".  I didn't think too much about the poor image quality, though, because quite frankly, the image from my SC23As isn't all that great even when they are connected individually to the monitor.  So, I hadn't really noticed that it was further degraded until I started disconnecting segments of the "network" and noticed some improvement when certain portions were isolated.  Maybe the output of these cameras is designed to work with UTP, considering the 60' non-coax cable that they come with... so maybe that's why they work as well as they do in this case.  ()

Anyway, is there any way that I can use passive components to work-around the impedance problem?  (If I have to go with an active mixer, that will mean extra cabling because all lines would need to be fed independently back to the mixer, rather than being combined (in some cases) at remote locations, with the combined signal being sent back along the long cable run).

[dave w]

Anyway, is there any way that I can use passive components to work-around the impedance problem?  

Passively? I don't think so. But maybe someone else knows of a way. How does your system work? How are you switching from one camera to another now?

If you are using the X10 addressable power supplies to switch only one camera on at a time, you could install a 12V relay at each camera so the camera output is open circuit through the NO contacts contacts of the relay. When the power supply turns on to supply power to the camera, the 12V  will also close the relay and complete the video path. You could use a DPST (or DPDT which is more common) relay to open the audio line also.
http://www.allelectronics.com/make-a-store/item/RLY-521/12-VDC-REED-RELAY/1.html
http://www.allelectronics.com/make-a-store/item/RLY-212/12-VDC-DPDT-MINIATURE-RELAY/1.html
http://www.allelectronics.com/make-a-store/category/500/Relays/1.html

[Knightrider]

HA Dave has used UM506's to switch camera feeds as well.

[MD Corie]

Yes, camera switching is accomplished via the X10 addressible power supplies, located at the "central office", with the individual power sent out over lines in the Cat-5 "extension" cables in the case of the remote cameras.

Relays at the remote end would ensure that only one camera output gets put on the cable at a time, but I'm not so sure that would eliminate the impedence/reflections/ringing issues unless impedance-matched relays are available.  (I'll have to look into that).  Meanwhile, another concern with relays driven from the unregulated X10 addressible power supplies at the far end of the cable would be how much power they would sap from the cameras - and whether the cameras would still work OK with the reduced supply voltage.  With some degradation already apparent from losses on the long cable runs, adding a substantial load with a relay seems like it might only trade one problem for another - that is, trading degradation from impedence mis-match for degradation from below-spec power to the cameras.  Relays also raise the questions of reliability and even operability at the extremes of temperature that the junction boxes are exposed to.  (Probably the only way to find out any of this for sure will be to try it, but it seems like there are several likely problems involved).  I'll have to see if there are any solid-state relay chips available that may avoid some of the issues with mechanical relays... but I'm also thinking there may be some buffer or compensator chips out there that would take care of any transmission line problems.  I would be nice to know whether that is the actual problem, though;  I hate to put money into "solving" things that are non-problems.

[Brian H]

Here is the signal pin out for the SC23A cameras.
http://kbase.x10.com/wiki/Pin-out_for_Wired_Camera_Connectors

If I did my RJ45 8C Network cable wiring to RJ14 6C signal conversion properly. Both the Video and Audio signals are not twisted in a pair with a common or ground.
Only the raw +12 volts and common {X10 calls it Neutral   } are a twisted pair.
I have no idea what having the video and audio on a single wire of a separate pair will do to the performance. Though I doubt it helps.
 

[dave w]

Relays at the remote end would ensure that only one camera output gets put on the cable at a time, but I'm not so sure that would eliminate the impedence/reflections/ringing issues unless impedance-matched relays are available.  (I'll have to look into that).  Meanwhile, another concern with relays driven from the unregulated X10 addressible power supplies at the far end of the cable would be how much power they would sap from the cameras - and whether the cameras would still work OK with the reduced supply voltage.  With some degradation already apparent from losses on the long cable runs, adding a substantial load with a relay seems like it might only trade one problem for another - that is, trading degradation from impedence mis-match for degradation from below-spec power to the cameras.  Relays also raise the questions of reliability and even operability at the extremes of temperature that the junction boxes are exposed to.  (Probably the only way to find out any of this for sure will be to try it, but it seems like there are several likely problems involved).  I'll have to see if there are any solid-state relay chips available that may avoid some of the issues with mechanical relays... but I'm also thinking there may be some buffer or compensator chips out there that would take care of any transmission line problems.  I would be nice to know whether that is the actual problem, though;  I hate to put money into "solving" things that are non-problems.

The 12V relay URLs I posted had coil resistances of 1400 or so ohms, or about 9 ma coil current, and a price of a buck each. Since these same power supplies are used for the wireless cameras and your cameras do not have the wireless transmitter circuitry, I would bet they have enough reserve to power the additional nine millamps. I don't see why you could not wire the relays back at the respective camera power supplies where the RCA video line couples to the CAT 5 cable, with the N.O. contacts in series with the video feed. I don't think you want to run video through a solid state relay unless they are made for 600 MHz. Admittedly using off the shelf relays won't be ideal for coupling video, but "Y"ing 16 cameras down to one is considerably less ideal. Since 15 of the video lines will be open circuited, only one video line at a time will be conneced to your monitor or computer, I am confused about you impedance / reflected signal concern. Maybe you did not understand the original suggestion.

[MD Corie]

If I did my RJ45 8C Network cable wiring to RJ14 6C signal conversion properly. Both the Video and Audio signals are not twisted in a pair with a common or ground.
Only the raw +12 volts and common {X10 calls it Neutral   } are a twisted pair.
I have no idea what having the video and audio on a single wire of a separate pair will do to the performance. Though I doubt it helps.
 

Not sure that I'm understanding correctly:  Are you saying that the video (and audio) signals should NOT share a twisted pair with their respective signal commons?  (If so, then I have a big problem right off, because mine do... and I'm puzzled why this would be the case).  Please clarify how the wires in the Cat-5 cable should be allocated.  Thanks!

[MD Corie]

The 12V relay URLs I posted had coil resistances of 1400 or so ohms, or about 9 ma coil current, and a price of a buck each. Since these same power supplies are used for the wireless cameras and your cameras do not have the wireless transmitter circuitry, I would bet they have enough reserve to power the additional nine millamps. I don't see why you could not wire the relays back at the respective camera power supplies where the RCA video line couples to the CAT 5 cable, with the N.O. contacts in series with the video feed. I don't think you want to run video through a solid state relay unless they are made for 600 MHz. Admittedly using off the shelf relays won't be ideal for coupling video, but "Y"ing 16 cameras down to one is considerably less ideal. Since 15 of the video lines will be open circuited, only one video line at a time will be conneced to your monitor or computer, I am confused about you impedance / reflected signal concern. Maybe you did not understand the original suggestion.

Possibly I did not understand the original suggestion.  My experience with mechanical relays suggests the issues that I mentioned earlier, but I'd have to try them to be sure of anything.  To compensate for the multiple "Y" connections at the remote junction boxes, the relays would have to be inserted there - not at the monitor end - in order to accomplish what I think you are suggesting.

Probably solid-state relays was a poor choice of wording;  what I was thinking of are some line driver chips that have a high-Z output state.  Something of this nature - if it also had the appropriate "on-state" impedance - might be a better choice than mechanical relays, both in terms of environmental tolerance and long-term reliability.  (Just saying...)

Regarding the reflections issue, I'm assuming that reflections are occuring and are what cause the "ringing" in the video images.  Presumably, impedance mis-match is the cause of the reflections, hence my concern over that aspect.  The typical solution for mis-matches is to use baluns or some such, although it looks like ones that would be suitable -and would physically fit - for this particular application may be hard to find.  Anyway, it may be that some sort of twisted-pair transmitter/receiver chip set or Cat-5 compensator chip may be a better way to correct any mismatches that actually exist.  I'm just not yet sure whether mis-matches are the actual problem, or if something else is going on.  Unfortunately, it looks like any determination via trial and error is going to be rather expensive...

[Brian H]

No I am saying the way I traced it out. The video and audio seemed to be 1/2 of a pair with an NC for the other wire in the pair.
With the SC23A only having one common pin {#3}. I don't see how more than one twisted pair could have a common in it.

Did you maybe do a custom wiring of the cable or have an older camera where all the pins on the RJ14 camera connector is used?

Now I don't have a clue as to what the interface splitter box does with the power, audio and video signals at the receiver end.

[MD Corie]

No I am saying the way I traced it out. The video and audio seemed to be 1/2 of a pair with an NC for the other wire in the pair.
With the SC23A only having one common pin {#3}. I don't see how more than one twisted pair could have a common in it.

OK, now I see:  You were referring to the camera's provided cable, not to the Cat-5 extension cable.  (Sorry for the confusion).

I have never "dissected" one of the camera cables - other than to replace a modular connector - so I don't know whether that camera cable uses twisted pairs or not... but when I stripped back the insulation at the end in order to replace the modular (6P6C, but effectively 6P4C) connector, the wires did NOT appear to be in twisted pairs.  In any case, there is only the one common, as you mentioned.  (I believe the documentation mentions that the older cameras used all six conductors, providing a common with each of the three signal lines).  All of my cameras use the single common wire arrangement, leaving two n/c cavities in the connector.

Did you maybe do a custom wiring of the cable or have an older camera where all the pins on the RJ14 camera connector is used?

I don't have any custom-wired cables from the cameras themselves, but the Cat-5 extension cables ARE custom-wired - at least in the sense that I rather arbitrarily assigned wires in the cable to particular signals.  (I did make a feeble attempt to follow the pin ordering as used in the camera cables, by laying out the center six conductors in the same way as used in the camera cables, but then assigned the resulting n/c cavities to the wires from additional power supplies.

[Brian H]

Thanks for the added data.
I will digest the information and maybe see if I can think of something.