ethernet frequency and power on ethernet

I'm not sure you can, and still be legal. The ampacity of the wire is going to limit your current, the various codes for low-voltage wiring are going to limit your voltage, and the intersection of the two is going to give you your maximum power (minus line losses). 802.3af peaks out at about 20 watts input to the cable (15.4W per device), though you might be able to cheat and use all four pairs to double your power. The standards are a nightmare to parse, but it might be possible. Dunno what standards you have in India, or if this is for commercial operation or not. Can you pull another cable?

In article , praveen wrote: :I am want know the frequency of the ethernet data so that i can decide :on the value of the inductor through which i can feed DC power. :I have heard the ethernet can range from DC to 10 Mhz (10TBase), in :this case i cannot feed power on ethernet.

I don't know about DC, and I don't recall the details about 10BaseT, but at the higher bit rates, the frequency is not just

1/(megabits per second): instead, they use a slower carrier and more bits per symbol. Some error correction is used, so the raw number of data points sampled exceeds the nominal bandwidth.

10/100/1000 BaseT are all async, so when there is no data going through, there are no pulses on the line. I never looked deeply enough to find out whether it uses a DC carrier or if the line floats free.

If you are using 10BaseT, then several of the wires on a typical Cat5 8-wire RJ45 setup are not used at all (not even as grounds): if you need to carry power, could you perhaps use those wires? This might not help if you are running Cat3 instead of Cat5.

do some hunting for 802.3af - standard for power over ethernet - it can provide up to 13W or so to your device.

even if you dont use the standard, it explains how to do this....

[Walter Roberson]

| 10/100/1000 BaseT are all async, so when there is no data going through, | there are no pulses on the line. I never looked deeply enough to | find out whether it uses a DC carrier or if the line floats free.

As far as I know, 100TX and 1000T are both synchronous - there is a signal on the wire at all times.

Steinar Haug, Nethelp consulting, snipped-for-privacy@nethelp.no

Well the maximum transition rate on 10baseT is 10 MHz, on

100baseTX is 100 MHz, and 1000baseT is 125 MHz. But these are essentially encoded digital signals without carrier. So actual lines in use see many "missed" transitions and even longish runs (6+) without.

I think inductors would soak up too much of the signal.

It doesn't matter. Both 10baseT and 100baseTX only use two pairs for signalling. Both Cat3 & Cat5 cables typically have four pair. Only 100baseT4 (obsolete) and 1000baseT use all four pairs. Watch-out, though. Some hubs/cards ground unused pairs. And you'll be stuck with NEC low-voltage limitations.

-- Robert

If it is 15W per pair then he should be able to get close to 60W with four pair. (If it is limited by cable heating then it will be a little less.)

I would have tried center tapped transformers before using inductors to isolate it.

It is supposed to be that three phase AC makes the most efficient use of its wires, so three phase between three pairs might do it. That is, if you are limited by average current squared and peak to peak voltage. It shouldn't be much harder to keep 60Hz away from ethernet, though it could run at a higher frequency if you have to generate it anyway.

-- glen

It's actually pairs of pairs, so he could double the power. Google for the PoE spec...

Well, obviously there is no DC in 10BASE-T protocol as DC can carry no signal. The actual spectrum analysis of a 10BASE-T signal would probably show max energy at or around 5 to 7 MHz. Even then, I don't see how it affects your power-over-CAT5 calculations. All you need to worry about is max current, and not the frequency. You can run 1A of current on every pair of a AWG24 cable, such as CAT5. Since two pairs are used already for

10BASE-T, you are left with two pairs or 2A, which would make it 30V for the operating voltage. You will have to add few volts for the voltage drop, depending on the length of your cable.

The frequencies used by 10baseT is about 10 - 20 MHz. However, you should look at the PoE gear. Some of it may be suitable for what you need. My biggest concern, is that you don't exceed the current capability of the wire.

10baseT, uses Manchester encoding, which means the frequencies used will range between between 10 & 20 MHz.

???

I think you'll find there's always some signalling on 100baseT and faster. No DC though.

They'd isolate the DC circuit from the data. A properly selected inductor will have a high impedance, at the desired frequencies.

Actually, for 10 Mb/s Ethernet (using Manchester encoding) the fundamental frequency will be between 5 MHz (alternating ones/zeros) and

10 MHz (all ones/all zeros). The energy spectrum of a packetized Ethernet signal using Manchester encoding at 10 Mb/s is concentrated under 30 MHz, with signal energy down to (but not including) DC.

-- Rich Seifert Networks and Communications Consulting 21885 Bear Creek Way (408) 395-5700 Los Gatos, CA 95033 (408) 228-0803 FAX

Send replies to: usenet at richseifert dot com

And including DC for coaxial ethernet. For 10baseT, though, the energy should be pretty low at lower frequencies. They still have to get through the transformer.

I still didn't get around to testing 10baseT with a common mode 240V AC 60Hz voltage. (I will be sure to test it on cheap parts, just in case.) If it works, I will believe it will work with just about everything else that has been suggested here before.

-- glen

I realized my error after posting. I had used info that I'd found in the O'Reilly Ethernet book, by Charles Spurgeon. On page 118, it says "While Manchester encoding makes it easy for a receiver to synchronize with the incoming signal and to extract data from it, a drawback of the scheme is that the worst-case signaling rate is twice the data rate. In other words, a 10 Mbps stream of all ones or all zeroes results in a Manchester encoded signaling rate of 20 MHz on the cable.". He was correct in the first sentence, but wrong in the 2nd, where he confuses signaling rate with frequency. I have worked with other systems that used Manchester encoding, but that was almost 30 years ago.