On a sunny day (Sat, 9 Jan 2010 22:26:36 -0800 (PST)) it happened " snipped-for-privacy@sushi.com" wrote in :
Video over twisted pair does NOT use transformers. Composite video 'FBAS', is unsuitable to pass through a transformer. Unless you modulate it first on some carrier. Of have many meters of video here via CAT5, I use 1 twisted pair like this:
1 video
2 video ground and an other pair like this;
1 +12V
2 power ground.
The +12V is locally, at the camera side, way out there, regulated down to +5V with a decoupled 7805.
A wire of a third pair is used to send RS232 data back from my temp_pic temperature sensor. The ground side of that uses power ground, it is also powered from the +5V locally.
Power and video ground are connected together on the PC side, where the video enters the digitiser.
Video has very low frequency components in it, and even a DC component, you could clamp each line, at 15 kHz, to get hat back, but transformers would severely distort the signal, be to bulky because they would need to be able to pass 50Hz or 6 0Hz the V component, and then would not be able to handle up to 5MHz or so for the smaller details.
Didn't find your answer? Ask the community — no account required.
J
Joerg
That's good medicine, as Phil Hobbs would have put it. And it sort of answers your own question below :-)
Every little discharge can degrade a transformer winding, piece by piece.
A CAT5 pair does not have a ground, it floats. If the whole cable floats (no shield) and is, for example, strung along poles it can be even worse. When I was a kid and started ham radio I woke up one night. A loud *POCK* noise at irregular intervals right in my bedroom. Did I leave my rig turned on and some tube was arcing? Then I noticed blue flashes. They came from a coax connector that was laying on the table. I always unplugged that before going to bed. Unfortunately by the time I noticed all this the center pin had pitted quite badly, had to snip and mount a new connector. Which back then caused a little financial hurt.
Next day I placed a 100K up across the antenna connection, never happened again.
Regards, Joerg
http://www.analogconsultants.com/
"gmail" domain blocked because of excessive spam.
Use another domain or send PM.
M
miso
s:
+5V with a decoupled 7805.
emperature sensor.
V locally.
ideo enters the digitiser.
ou could clamp each line, at 15 kHz,
o bulky because they would need
ble to handle up to 5MHz or so
Ok, then what do you call this:
formatting link
%20recorder.html
I see these video over cat all the time in surplus shops.
P
Paul Keinanen
I have seen that at the end of the feeder for a 3.5 MHz antenna, with one 20 m uninsulated wire in one direction connected to the coaxial center wire and an other 20 m uninsulated wire in the opposite direction connected to the coax shield. In suitable windy conditions, "static rain" caused a large charge build up, causing flash over between the "UHF" connector center pin and shield.
However, I doubt that there would be that much voltage in a CAT5 cable hanging on poles. The differential voltage would be nearly zero, due to the twisted pair structure. The CAT5 also has insulation on each pair and there is also a common shield, so the charged particles can not hit directly the conductor.
If there are unused pairs, these could be grounded to drain any static charge from the cable, while the actual data pairs could be floating.
J
Jan Panteltje
On a sunny day (Sun, 10 Jan 2010 22:17:31 -0800 (PST)) it happened " snipped-for-privacy@sushi.com" wrote in :
That is not a transformer, as in the sense of transformer, see:
formatting link
It is a BALUN. Just 2 coupled coils to get impedance change, while the DC is *not* separated. It is also not a very good solution, as I sincerely doubt it will behave flat from 15kHz to 5 MHz. and for lower frequencies it will not do anything, think DC, this is how it is connected, see that page:
composite in --------L----------- ===== balanced to CAT5 twisted pair ground ---------L-----------
couple coils You can make that yourself from the right core material pot core. How much do you think the coupling is at 100Hz? And yes, there can be 10Hz and lower components in the composite video signal, think white frame to dark frame, that is a DC jump, as slow as you can imagine. So they solely rely on there being a line clamp at the receiving side. And that is NOT a guarantee. Bad product!
I have seen a lot of crap solutions, even in TVs on the market, I have even seen AC coupling without any line clamp in TVs, if the picture is black it goes grey (average value). Consumers buy the stuff. It is not professional, it sucks.
For a composite 75 Ohm to long CAT5 cable the right way, you need a DC coupled amplifer x2 (at least compensate for cable losses) with differential output of 100 Ohm. And a C coupled 100 Ohm differential amp at the other side, and that side terminated with 100 Ohm. You can use an AC coupled amplifier, if you KNOW there is a line clamp on the other side. That driving amplifier should also have a frequency / phase correction to compensate for any cable losses.
If you do not want all that, and the distance is not very long, you can stick the 75 Ohm output on one pair, and terminate with 75 Ohm on the other side, that will work too, will have some reflections perhaps, and attenuate the signal somewhat, but works OK that way here, but maybe is only
10 or 20 meters, and a low quality security PAL camera, the 4.43 MHz chroma comes out fine though
I would not do it that way for anybody else, I have an other camera, further away, that actually uses an correction amplifier and audio cable.... And finally I have an other camera that is an IP camera and uses the CAT5 for Ethernet. That is the right way, the picture is perfect, but that camera is way more expensive:
formatting link
So, that is the professional way these days.
J
JosephKK
Yes. See below.
It is cheap and highly advisable.
There is considerable slowing of the signal in the transmission media=20 and considerable attenuation.
After reading the thread, i still think my base proposal is worth=20 considering. CAT5 is a reasonable choice. Get some old 10BASE-T=20 network cards to get the transformers off of. Put a continuous=20
1MHz carrier on it and give it a 20 cycle burst of alternate 1s=20 encoded Manchester (biphase) style for the detection. Use other pairs=20 to transmit power.
More information if you want it.
M
miso
t
thi=3D
to=3D
c t=3D
+5=3D
e v=3D
, y=3D
e t=3D
e a=3D
ated.
flat from 15kHz to 5 MHz.
it is connected, see that page:
=A0 balanced to CAT5 twisted pair
gnal,
agine.
en seen AC coupling without
e).
upled amplifer x2
m.
terminated with 100 Ohm.
the other side.
compensate for any cable losses.
tick the 75 Ohm output
too, will have some reflections perhaps,
is only 10 or 20 meters, and
ugh
her away, that actually uses an correction amplifier and audio cable....
for Ethernet.
e expensive:
I agree is it not a good scheme. I suppose if you nitpick, a balun is not a transformer, though there are video transformers too:
formatting link
Generally you want a tilt in the frequency response, so active performance is better.
J
Jan Panteltje
On a sunny day (Mon, 11 Jan 2010 23:45:23 -0800 (PST)) it happened " snipped-for-privacy@sushi.com" wrote in :
Well, I am amazed, but then again it is specced from 10Hz to 8MHz, not from DC. Still pretty broadband for a transformer, and it should work with a line clamp at the other end, not without... And I wonder if they used more then one transformer to get that frequency range covered.
Also I see an other BIG problem using this product, a typical video output stage in cheap equipment looked like this: + | V c composite ---- b NPN Q1 e | +2.5V 75 Ohm 75 Ohm coax C |------------0=========================0------||----- amp - H clamp ----- | | | 1kOhm | 75 Ohm | | | |------------ ----------- ///
This works OK without transformer. I wonder what it does with transformer, I mean now transistor Q1 sees only 75 Ohm to ground, resulting in 2x the Ic....
I would stay clear of those video transformer products,
Indeed, and there are very good chips for all this these days.
M
Michael A. Terrell
from 15kHz to 5 MHz.
connected, see that page:
Then how do you get vertical sync? Video Isolation Transformers are common in TV production. The keep the differences in ground potential from inducing hum into the video.
Not all baluns are DC coupled.
Greed is the root of all eBay.
J
Jim Thompson
from 15kHz to 5 MHz.
is connected, see that page:
DC restorer/sync separator, as I posted here December 4...
...Jim Thompson
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
M
Michael A. Terrell
separated.
flat from 15kHz to 5 MHz.
is connected, see that page:
Yeah, but Jeorge can do it with two used, rusty 16D nails, a rubber band and an old condom for under two cents per unit. ;-)
Greed is the root of all eBay.
M
miso
om DC.
clamp at the other end,
range covered.
Ohm coax =A0 =A0 =A0 =A0 =A0 =A0 =A0 C =A0
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D0------||----- amp - H c= lamp -----
But video really is an AC signal since you have the porch for DC restore.
J
Jan Panteltje
On a sunny day (Tue, 12 Jan 2010 15:22:27 -0800 (PST)) it happened " snipped-for-privacy@sushi.com" wrote in :
Well, only if you clamp can you treat it as an AC signal. Can you guarantee every thing you connect to clamps?
M
Michael A. Terrell
Only if you know what you are doing with Video.
Greed is the root of all eBay.
J
Jim Thompson
separated.
flat from 15kHz to 5 MHz.
it is connected, see that page:
Probably! "Poor Joerg..." ;-) ...Jim Thompson
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
K
krw
The same pairs can be used to transmit power. The usual method is attach the power and ground to the center taps of the transmit and receive pairs. More pairs in parallel for longer distance or more power.
J
JosephKK
If you use SMPTE standard compliant equipment sure. That is part of=20 those standards. These standards are ALL about interoperability.
J
JosephKK
Absolutely correct. PoE is standardized.
K
krw
It works even when you don't use the standard. Most don't.
J
JosephKK
=20
pairs=20
Yes it does. Most recent equipment is compliant now. There is also=20 some combinations of (usually older) equipment that do not work together.
Join the Discussion
Have something to add? Share your thoughts — no account required.
Didn't find your answer?
Ask the community — no account required
Report Content
You are reporting this content to the moderators. They will look at it
ASAP.