Instructables
Picture of Power Over Ethernet (PoE) Adapter
Power over Ethernet or PoE, is the technology used for power transmission in network equipment, via network UTP cable, together with data. PoE is useful in situations when we want to connect network devices that are far away from a power source. In this situation we take advantage of the Ethernet cable, because there are 4 pairs of wires, but for the majority of networks, only 2 pairs are used for data transmission. PoE uses 2 unused twisted pairs, for electrical transmission needed to power the device.

There are several architectures of PoE and many of them existed before the technology being standardized, due this fact not all types of PoE are compatible. PoE devices receive power directly from the network cable, but should be noted that according the manufacturers, specs like nominal values of voltage and current, and even pairs used to transport electrical energy may be different.

Source here.
 
Remove these adsRemove these ads by Signing Up

Step 1: Description

Picture of Description
Applying a voltage across terminals not dimensioned for that, may damage the network device. To avoid this, the best way is to make a PoE adapter, separating the two input signals, via two simple circuits, an injector and a splitter. The injector will feed the network cable with power and this circuit is placed near the origin of the installation, where we have a modem or a switch, and of course a power plug. Then we have as outputs of the Ethernet cable, data and power that will connect to the device. Now we connect the splitter circuit, which receives data and power. From this circuit leaves an Ethernet cable and a DC plug avoiding applying unknown voltages to the devices.

As example consider the eLab case, which have a switch on the back of the room and we want to feed a 12V wireless router which is at the opposite end of the room, to obtain greater coverage by departmental area, in a place near a windows without power sources nearby. The network diagram is the one above.
1-40 of 60Next »
Good Afternoon From UK, Linda. I have a query, although I am no technical wizard, but is there no danger of this device generating Radio frequency interference, given that the cables over which the power travels may act as an antenna. This has happened many times here in UK, Europe and North and South America. I am not saying this WILL happen, only quoting the experience of what is called Power Line Transmission devices, which uses power cables to pass data between distant points in the same building. You are to be praised for your efforts posting, and this is a genuine concerned enquiry. Thanks for youre time, Best Regards, Gary
eLab (author)  whitewolf4991 year ago
Thank you! Well, that is a common concern about PoE adapters but Ethernet communication is "kind of immune" to this type of electromagnetic radiation, besides the power lines being used in these cases are very low power. Therefore using some wires for power and some wires for data in the same cable, should not affect the transmission of data packets or the speed itself.
GregS2 eLab2 months ago

The power on the ethernet wires is not your power company's grid power, but the power coming from a power adapter. The power adapters are usually a DC output and as such have filter capacitors to stabilize the power they provide. Due to this, the power adapter would have the side-effect of low-pass filtering out any back-fed RFI and preventing it from entering the power grid.

I'd add two more reasons - twisted pairs used in ethernet cable (and phone lines) are paired & twisted in an effort to minimise radiation/induction with an AC / RF signal on the wire.

PoE is DC and tends towards embedded devices with a steady current draw, therefore limited scope for radiation even from a straight wire conductor.
That is a good point. Also note that the use of STP (shielded twisted pair) network cables rather than UTP (unshielded twisted pair) network cables will reduce the likelihood of interference.
whitewolf, I believe that the problems with Power Line Transmission devices that you are referring to have to do with sending (analog or digital) signals over building mains wiring—AC powerlines—as with X10 and other devices or protocols.
It is sending low voltage, low amperage voltage over a data line. It is NOT the same as data over power line, PLT, as the cable is made for data.

I use a similar set up in my shack with no RF problems at all.

Very!! informative article on Ethernet Adapter but can you suggests me one of best adapter which i could use for my computer, I have already visited some of the online store to like sfcable.com, amazon etc... but still confuse!! Do different adapter support different net speed or all are the same ??, please help me!!!!

eLab (author)  davidjohnson876 months ago

Thank you!

Well we can't really suggest any of them in particular because we haven't tried them, that's why we made this one on the first place. But they're probably very similar, there isn't much to choose. Adapters like this will only work for 10/100Mbps networks, which is usually the standard speed on a domestic Ethernet LAN. If your network supports 1Gbps then you shouldn't use these kind of adapters.

restiandan7 months ago
sir can we control the output voltage? lets say the main power is 48V then the output must be always 12v? max utp used is 100m, if i have 5~10m UTP the output voltage is 12v, if i have 50~100m UTP the output voltage still 12v? can you teach me? thank you in advance..
eLab (author)  restiandan7 months ago

Hello!

These adapters are just a pathway for power and data. The voltage output is supposed to be approximately the same as the input (obviously there is a voltage drop depending on the distance and current). If you want the output to be 12V our advice is just to use a 12V power input. If you want to convert 48V to 12V, that is a different thing, it has nothing to do with PoE adapters. Still if that's what you want to do, there are a few ways to do it, like DC/DC step down converters, voltage regulators, voltage dividers and other electronic circuits. The choice depends on some factors like efficiency and power consumption. For more efficiency, probably the best thing would be to buy a 48V to 12V DC/DC step down converter, you can find them on ebay for example.

Gadget931 year ago
This is a really cool and useful idea. As a CCNA and a CCDA I must mention though that due to the power and data running next to one another, there will be interference and noisy data. Alot of packets will be dropped and other problems will cause problems with data. Although if you don't have any other choice, go ahead...it is better than nothing.
stajp Gadget9310 months ago
That's not the case if UTP is used. Even in CCNA learning literature there is an explanation why the UTP is twisted: to combat crosstalk. DC power, which doesn't change in regards to voltage or current, has minimal crosstalk to other twisted wires. The whole IEEE PoE standard takes a lot of consideration for optimizing this idea (and Cisco has PoE devices done according to standard), but there are similar projects to this which are used for years without any noticeable loss in data.

The only problem here is the quality of work when it's made and used. If it's done badly, then don't expect miracles, and crosstalk and other problems would arise :D
charles5431 year ago
How much current can the ethernet cable handle?
eLab (author)  charles5431 year ago
It's around 0,577A for each conductor.
pcooper2 eLab1 year ago
To avoid ambiguity, you should write "577 mA", as the comma is a thousands separator in the U.S.A. and many other places, not the decimal point. The way you wrote it, someone unfamiliar with such conventions might interpret your answer as 577 A.
eLab (author)  pcooper21 year ago
you're right! It's less confusing if used 577mA.
mahrouch1 year ago
Verry nice instructable, but i wonder if we can use this structure to power a POE device (which have not a plug of power but take it straight from the RJ45 cable). I have a "StraightCore GWP-217VE Universal Outdoor Wireless Client" without its PoE system. do you think this DIY can work with it?
eLab (author)  mahrouch1 year ago
It might not work, and even worst, it may damage your device. Different devices and brands might use different technologies, and they may or not be compatible with this particular one.
miguipda1 year ago
Hi,

here is a cheaper way : http://www.seeedstudio.com/depot/passive-poe-cable-set-p-1175.html?cPath=207

Bye
It's only cheaper if the power plug and jack are compatible with one's equipment. This Instructable shows you in a general way how to built injectors and splitters for virtually any situation.
charles5431 year ago
Why not use 4 and 8 for positive and 5 and 7 for negative or the other way around? That way the twist will cancel out any interference
eLab (author)  charles5431 year ago
When we made this, we didn't care that much which conductor to use for power, but that's good idea!
kevinfodor1 year ago
I would like to make a comment that this really should not use the label "PoE". To be PoE is something very specific outlined in the IEEE specs. This is of course a good hack, clever idea if you will, but it is NOT PoE. Yes, it does send power over what happens to be an Ethernet cable, but this alone does not make it a PoE device and I think some people may be confused by that.

There are very specific requirements which must be met by a PoE device. One is that the polarity can be reversed (using a bridge). Another is the load capacitance seen by the PSE (power supplying equipment) and input protection, etc. There are many links on the internet about what specifically PoE is.

Anyway I don't mean to knock the instructible itself, just the fact that it is calling it self "PoE" which is not accurate.
eLab (author)  kevinfodor1 year ago
Yes you're right, it's not PoE IEEE standard, which is exactly why the 2 adapters must be used together. Otherwise you could just plug the cable into a device. Either way, the technique used is still transmitting power over Ethernet cabling, so no harm done :P
lbolhuis1 year ago
I have been using this technique successfully for several years at distances up to 100 ft to power my access points. Even where there is power nearby this way the power comes from a central point where a UPS is available. I've also powered cordless phone base units this way and thus they don't beep when the power is out. Very nice description on how to to this!
eLab (author)  lbolhuis1 year ago
Yes, it's a very useful technique for many reasons. Thank you!
laserlad1 year ago
Excellent Instructable and excellent community comments! I hope I am not muddying the water, but I would like to offer one caution concerning using the signal transmission side of Ethernet cabling rather than the power transmission capabilities that eLab has done such a great job of communicating. Because of the very reasons already mentioned here, such as the use of twisted pairs within the cable to reduce interference, ethernet cables are sometimes used as an inexpensive and easily available method of sending control signals from point A to point B that are not monitored anywhere in the system. Sorry - I realize that statement isn't very clear - but as an example, that could be something like running a handy ethernet cable from a custom-made (home-made) switch panel out to relays. In this case, I'm talking about a "switch panel" in its lowest form; simple mechanical switches mounted in a convenient location with only the electrical wiring necessary to connect them to the circuits they are to energize/de-energize. The cable will certainly work, and telephone cable with RJ11 connectors has also been used satisfactorily for that. But... it should be noted that under those circumstances there is no cognitive processing of those signals - be it through Arduino, laptop, or whatever - and also no hardware/firmware processing of the signals on either end such as limit checks, voltage and/or current regulation, or other safety checks. That is not critical in many applications any more than it is on each light switch in your home, office, lab, or whatever. BUT, if whatever is being controlled by the switches or other simple mechanical devices on the circuit has a non-FailSafe potential, you should not rely on Cat 5e, or 6 or any other simple twisted pair configuration to carry those signals. Again an example: if the control signal you are sending when that switch is thrown to energize the circuit is intended to activate the relay that turns on a laser system or a furnace or any number of pieces of equipment or systems that could cause harm if their activation (or termination) were to go unnoticed. It may sound intuitive, but I know of more than one building burnt to the ground in which the root cause failure analysis determined this to be the culprit. I also know of a person with 3rd degree burns over about 5% and 2nd degree burns over about 30% of their body that are attributable to the same cause. Please don't misunderstand me - what eLab has put in this Instructable is very useful and great information, especially as augmented. He is talking about transmitting low levels of power through ethernet cables, that you would presumably be running anyway, in order to conveniently drive components down line. As I followed the comments it occurred to me that some of those, especially concerning the inherent characteristic of twisted pair lines, such as found inside ethernet cable, to have some level of resistance to EMI, might cause some readers to infer (correctly for some specific instances) that they could be used to transmit control signals without interference. So, I'm not trying to make anyone afraid to touch a Cat 5e cable; I'm only trying to warn against anyone taking the accurate statements of contributors, who seem both well-informed and well-intentioned, beyond their intended scope. If in doubt, I would highly recommend an appropriate coaxial solution (there are many), wireless system using proper EMW spectrum and security, or best of all.. ask for help from a known, knowledgable source. I am very sorry if this is off-base or detracts unnecessarily from the great information submitted. When you have actually seen more than one occurrence of reasonable people being devastated by taking a certain action it seems mandatory to at least flag it for others.
MikB1 year ago
If you are worried about voltage losses in longer cable runs (and you should be!) then certainly feeding a higher voltage and converting down remotely is a good idea.

Simple: 7805 or 7812 regulator to drop from a higher voltage. Gets warm, wastes power, current in (at say 15v) is same as current out to device.

Slightly Harder: Something like this http://www.ebay.co.uk/itm/140906034058 fitted at the remote end will take whatever's available from about 7v upward (this module goes to 24v) and transform it down to 5v.

It's a switching converter (buck converter) meaning it draws less current than the load. This is good. It reduces the losses in the wiring, and is even more efficient.

It's not exactly PoE standards compliant, it's not 48v, but it works.

I'm using some of those modules (and similar) to power remote IP cameras (draw 400-500ma, at 5v) from a centralised "injector" -- actually a hacked 8 port router with an extra PSU to give 15v on pins 4,5 and 7,8 on 7 of the ports (not the UPLINK port), just as shown here, but internal to the router

Each port can source 500ma (limited by it's own polyfuse), the load from a single camera is now down to around 280ma.

The voltage does drop too much to just feed 5v and hope for the best, especially on long runs.

darman121 year ago
Great Instructable :) I really liked your explanation of PoE. It was clear and concise. I personally would buy an injector rather than go the DIY route, just to rule out any errors I could make; but, this would be a cool project t accomplish. Good job :)
eLab (author)  darman121 year ago
Thank you very much ;)
AndyGadget1 year ago
 
Nice write-up with plenty of information and the appropriate warnings. I'll be using something similar to run a remote IP webcam. Be aware that in some countries 568A is the norm so check the connections carefully. (I love standards - There's always so many to choose from #;¬)
Curious, does the standard you choose really matter?
So long as you use the same pin out on both adapters it seems like the wire in the middle wouldn't matter all that much.
I took a short Leviton certification course 12 years ago, and if I remember correctly when making your cables you can choose either A or B, so long as you use the same one on both ends of the cable. (if you don't, you end up with a cross cable)

If this is incorrect then let me know, as I'd love to try this to free up an extension cord. :)
You are correct, you can go either way, but following the standard in your environment will make troubleshooting much easier. According to Comptia, two B ends would be used for a straight through cable. While two A ends would work as well, two B ends is the standard. If I were to find an A on one end, I would assume the cable is crossover; that is why following the standard is important.
Plus, it is a good idea to stay in good practice whether it is in a department with other IT staff or it is in your own office/home.
One reason to be concerned with this is if you are working in a facility that requires a second party inspection of all work prior to being used. Granted in that case, this POE solution would not be expected to be used, but it's good to keep it in mind. It also needs to be kept in mind if you are going to be running this solution through a patch panel for facilities. Make sure that any ports that will be carrying voltage that is unusual for your facility is clearly marked.
eLab (author)  adillbeck1 year ago
As long as you use the same norm in both adapters, it's all fine ;)
You're absolutely right - As long as you're connecting the same both ends the standard won't matter. My comment was more of a 'heads up' in case anyone was using a mix of off-the-shelf components which could be marked differently.
(If you use anything other than one of the recognised standards, ensure the signal pairs are consistent otherwise the transmission characteristics would be affected.)
eLab (author)  AndyGadget1 year ago
Thank you! We wanted to keep it simple but with all the necessary info.
rusty01011 year ago
For those concerned about the fact that this solution is only workable for 10 and 100 mbps networks, which only use 2 pair rather than all 4 pair, and wouldn't work for gig-e, that's accurate as far as it goes. The solution to add support for gig-e is to isolate the power on the pairs from the connection to the equipment, and to block the signal on the pairs from the power side. You can run both DC and AC on the same wires. Capacitors will pass the signal, but not the DC, and you'll need to rectify and use capacitors on the DC side of the rectifier to filter out the signal on the power side of the injector/tap hardware. All the while making sure that you're not attenuating the signal to the point that the equipment can't negotiate a reliable 1-gig connection.

Also check the equipment you are planning on doing this with. If it already supports POE you would be better off getting an Injector designed to support it.
1-40 of 60Next »