Slashdot Mirror
IBM Bringing Powerline Broadband Back?
KindMind writes "IBM, in partnership with International Broadband Electric Communications, appears to be bringing back
powerline broadband back from the dead.
This time, the idea is to build out in rural areas not currently serviced by broadband, and isn't for competing with other broadband solutions.
From the article: 'Their strategy is to sign up electric cooperatives that provide power to sparsely populated areas across the eastern United States. Rather than compete toe-to-toe with large, entrenched cable or DSL providers, IBEC is looking for customers that have been largely left out of the shift to high-speed Internet.'"
All of the other efforts failed because it caused interference to ham radios and to emergency broadcast channels.
It had nothing to do with lack of backing, and large corporate backing doesn't necessarily translate to instant success.
Some of the powerline broadband manufacturers were able to produce systems that didn't interfere with public safety and amateur radio.
This is necessary, since even a distant powerline broadband system can interfere with emergency communications - the signals skip off the ionosphere and around the whole world, and sometimes contacts by legitimate radio operators can be made at astonishingly low power - meaning that the power line carriers probably have the potential for worldwide interference.
Earlier this year, ARRL won a suit against FCC that will lead to more realistic parameters for interference. The previous ones applied a single-point interference specification made for consumer electronic devices to any point on a wire, and of course over the total length of the wire the interference power was much higher than the spec.
The problem is that power lines are not like telephone lines or coaxial cable. Telephone lines are carefully balanced so that they cancel out much of the interference they would otherwise generate. Coaxial cables have their own shield. Power lines are driven in unbalanced mode when RF is injected into them, and thus act just like long antenna wires, and they radiate a great deal of any RF sent down them. No amount of signal processing can fix that.
Why not use WiMax? It's higher bandwidth, requires less infrastructure overall to install (since you don't have to bypass transformers, etc.) and works for mobiles. Pretty much every business that has invested in BPL for home internet delivery has failed.
The broadband competition in those areas will end up being between WiMax and cellular.
Bruce
Bruce Perens.
Okay, there's two problems with this, there always have been, and they still aren't practical to solve. The first is transformers. The second is interference.
Transformers: They have a resonant coupling frequency. Try to pass high frequency RF through a power conversion transformer and you get scrambled eggs on the other side. So at every point along the line where you meet a transformer, you'll need an RF pass-thru. These aren't cheap; They need to be lightning resistant, fail safe no matter what (otherwise people die -- no joke here), and in general very well designed. A typical loop is going to see maybe 2-4 step-downs from the plant to your house. At least one RF bypass will need to be installed for each customer, along with whatever CPE is required to get the signal.
Interference: High frequency RF tends to degrade quickly. And above 800 MHz (someone who's an EE, correct me if I'm wrong on the threshold for skin effect) it won't even "stick" to the lines. Because these lines are unshielded aerial lines running in one direction for miles, they make awesome antennas. Which would be great, except... FCC regulations dictate no harmful interference. So any signal being sent down those lines is going to have to be very low power to avoid becoming an omelette with another signal... like say, emergency services. Shannon's law people -- you've got 800 MHz to deal with, a low power signal, and it needs to travel along an antenna some tens of miles along, sucking up every stray RF in the neighborhood. Can you say signal degregation? Any signal you push over that line had better have a helluva lot of error correction. Given it tops out at 3 megabits per second, on a shared link... with 800 MHz of bandwidth to work with... That should give you an idea of just how much the Suck factor is (Low Q for you techies)
So, great article, I applaud IBM for making the effort, but unless you've got some really nifty new electronics, like a DSP from hell, I don't see this being anything but a money sinkhole. Comcast may suck, but they've got a few gigahertz to work with and no FCC restrictions... Just really bad management, which is the only thing making this even remotely practical.
#fuckbeta #iamslashdot #dicemustdie
I am an amateur radio operator, so I've heard a lot of pros and cons against BPL. Aside from the obvious and well worn HF interference issue, it was my understanding that BPL actually isn't great for rural areas because the distances over which it will work well are way too small. In other words, it needs a fiber connection to feed the powerline grid for a small area. Each area of distribution has to be fed by another fiber run. Seems to me like WiFi or WiMax are much better alternatives for rural areas. And what about the new whitespace frequencies that the FCC recently approved? Wouldn't this make wireless even more attractive?