Transmitting video signals over long distances can be tricky. Cheap co-ax cables won’t do the job. You either need amplifiers along the path, or need to use expensive, high quality shielded co-ax cables – both of which can end up costing a lot. [Maurizio] built a low cost solution to transmit video over long distances using twisted pair cables.
The system is cheap and uses readily available parts. The idea is to convert the video signal into a differential output using a pair of op amps and transmit them over a pair of twisted pair wires, then extract the signal at the receiving end using another amplifier.
A differential amplifier usually requires a dual-polarity power supply, which may not be available when adding this upgrade to an existing system. To over come this limitation, [Maurizio] uses a bias voltage equal to half of the power supply value. This bias voltage is added to the non-inverting amplifier signal, and subtracted from the inverting amplifier signal. The resultant differential signal is then fed into the twisted pair cable through impedance matching resistors. At the receiving end, a single amplifier receives the differential signals and outputs a signal that corresponds to the original video signal.
This symmetrical configuration renders the system immune to external noise. The design was tested for transmitting video on inexpensive CAT-3 twisted pair wire. According to him, when transmitting on 300m of wire, good quality color video was displayed on a monitor with an NTSC input. He used LMH6643 op-amps for this experiment, but other devices with similar characteristics can be used. Here’s a useful PDF document that discusses signals, cables and connections.
If you want to check out more of [Maurizio]’s work, see how he figured out how to send serial data from Excel.
Kinda would be nice to at least use a picture of CAT-3, not what looks like CAT-7…
That’s the picture from the article, which I can only assume you didn’t read.
sick burn bro
hasn’t this been known and done even before the invention of the Ethernet cable?
There are plenty of commercial products, even passive baluns designed specially for this not to mention 10s of app notes and circuits on hobby sites. I have seen them on security cameras in quite a few places.
Yup, work for a company that makes these. That’s basically how our VGA extenders work.
From experience: You can use Analog Devices AD8123 as a triple-EQ device for CATx cable. If you use standard CATx cable (not “minimized skew”), then you will likely also need an Intersil EL9115 skew-comp chip. With high-quality diff-amps (AD8130, AD8133, and similar), you should be able to extend 1920×1200 VGA to 1200ft (400m).
A friend of mine installed a twisted pair video distribution all over my home town. The interesting bit is that in order to get the signal to an adjacent town he took a leased line from the telephone company and sent the video down that. It worked for years until BT put low pass filters on.
From experience: You can use Analog Devices AD8123 as a triple-EQ device for CATx cable. If you use standard CATx cable (not “minimized skew”), then you will likely also need an Intersil EL9115 skew-comp chip. With high-quality diff-amps (AD8130, AD8133, and similar), you should be able to extend 1920×1200 VGA to 1200ft (400m).
And, yes, I agree… VGA is effectively obsolete… but still more than adequate for many applications.
It’s more than adequate, because getting HDMI/DVI through any length of cable beyond 3 ft can be tricky as hell.
Did you mean 3 meters? Because HDMI 2.0 goes over a 2m (6.6ft) cable without too much trouble. And 1080p goes over HDMI or DVI cables about 10 meters long without issue most of the time, too (unless you buy some garbage 30AWG cables). I’m just saying… the vast majority of HDMI cables are of the 2 meter/6 foot variety. Seems that wouldn’t be the case if it was really as bad as you say.
(Also, there’s HDBaseT for getting much longer cable lengths… but that’s spendy.)
Uh, I have HDMI going over 50 ft from a single receiver at one TV to another all the way across the house… Never had a single problem with it.
Analog composite using balanced control lines. Yes, not a new concept, it was used on old style VGA KVM systems. problem there was if you did not get the wires within a certain spec on length, the monitor looked awful.
I note in the write up he didn’t tackle the subcarrier pulse (NTSC only) to show how it fared through this process. In long coax runs the first indication of an issue is the subcarrier integrity. Equalizing distribution amplifiers corrected for this issue on long runs by overcompensating the subcarrier to allow it to arrive intact at the distant end.
With HD-SDI and digital protocols this form of video transport is nearly gone from the world of broadcasting. I do remember a time when any copper pair was fair game in an emergency though.
This hack is more about cost. A video signal will actually go a long way down coax. It’s TV antenna signals that don’t go far because they’re very low level and you soon run into singal to niose ration trouble.
Nice hack. I would have thought Twisted pair cable would have about 300+ ohms impedence.
I think the characteristic impedance of Cat5/6/7 is 100 ohms, like a lot of differential cabling tends to be nowadays… I can’t speak to Cat3 though (nor do I see much use in going with Cat3 for such an application…)
I still have a buddy-pc host card and reciever that used a standard cat5 cable to transmit the vga signal + keyboard / mouse. The Trident chip says 97. http://imgur.com/a/7SY96
I’m an ROV designer, and we send send standard definition video over twisted pair over 500m all the time. You just need to add some frequency compensation (at either the bottom or top end) if the line gets really long. If you put your frequency compensation at the top end, you also need to follow it with a low-pass video filter so that you’re high-frequency compensation doesn’t extend out of the baseband video spectrum. Otherwise you end up boosting your noise more than anything else!
Bring back souvenirs of an 4 chanels microwave real time video transmitter i have used in a work setup, up to 20 miles of range in year 2000, never new the cost of that thing.
I thought co-axial cable was inherently shielded. The outside lead acting as a shield for the inside lead?
The required signal properties of the outer conductor are to cancel mutual inductance with the inner conductor and present a consistent impedance to the driving signal.
These requirements can be achieved with as little as 5% shielding.
To shield out unwanted external signals a greater percentage of shielding is used – between 40% and 100% normally.
Most cheap cable is about 60% – 75%, crappy cable that comes from your least favourite Asian country is as low as 5% for audio. Good stuff is 80% – 100% and then there is double shielded as well.
I thought using resistors to do impedance matching was generally bad practice when designing transmission lines.
But if it works, it works. Nice hack!
There are working baluns for most video-ports that use ethernet-cables. Cheap too. Like cheap-cheap.
Or am I missing something here?
You can get quad shield RG-6 in quantities of 100’s of feet for pretty cheap and off the shelf video amps pretty cheap as well, if you even need them. Sounds more straight forward than this.