Power + DSL signal in same 700ft cable length

I really do want to try to using the existing spare pair in my drop wire, to keep costs to a minimum. After more research, plus people's suggestions here, I think I'm going to try going the AC tranformer route:

At the pedestal: Put a 120VAC (primary) to 48VAC @3A max (secondary) transformer.

Connect one of the two pairs of 22AWG wire to it Connect the other pair to the telco's NI to carry the DSL signal.

Assuming my equipment will draw max 1.25 amps (safe assumption?), V = 1400ft * 1.25amps * 16ohms / 1000ft = 28 volts loss 48VAC - 28VAC = 20VAC @ 1.25A 20VAC * 1.25A = 25 Watts available...?

At the peak: 20VAC available assuming 1.25A draw. Put a 120VAC (primary) to 24VAC @ 3A max (secondary) tranformer, but in reverse. (My understanding is that these tranformer's windings are optimized to compensate for losses, so using the tranformer in reverse won't yield the same ratio as the other direction. But maybe it's close enough?) Use the wallwarts from the devices (or maybe higher-efficient ones, available cheap).

I really don't know if this is going to work, but I get the impression that as long as my numbers are close, the only real way I'll know for sure is to try it out. When the budget is so tight to begin with, it's a bitter pill to swallow; spending the money just to determine whether it will work. DLR echoes my sentiments perfectly:

Lets do some math here....

1) Assuming the battery is rated at one kilowatt of total stored power, you can use only half that. You never want to use more than half their rated capacity if you expect them to last more than one season. 2) Assuming you are using deep discharge batteries. Car batteries are designed for short (five seconds) or high amperage output, not for a steady low power drain. 3) Assuming you need 10 watts of power, that's 240 watts per 24 hours - or in your case 120 watts for a 12 hour period per day.

Therefore, one deep discharge battery will be good for EIGHT days.

Use five of them in parallel for forty days then recharge them. Depending on ambient temperatures, they may only hold a charge about that long anyway. Also consider the fractional output at colder temperatures.

• Mount all the batteries on a wagon that you can pull up and down the hill.

• Keep a fire extinguisher handy.

• Wear coveralls and safety glasses.

• Keeps lots of baking soda around.

You forgot the mule to haul the wagon. :)

I assumed he was gonna have his wife...oh, nevermind. [snicker]

I haven't finished reading through all the post yet so forgive me if this has been mentioned. And I'm just trying to come up with a non-conventional solution.

What if; You put the DSL modem and a Wireless AP (maybe a single box) at the pedestal. Powered by the 120vAC You run coax from the AP to a wireless repeater on the peak. Kind of like cable TV. This would be powered by solar and batteries.

Another idea, which may not be possible given your area, is either satellite internet, or commercial wireless internet from the nearest city.

Coax loss would be astronomical at 2.4 GHz. Solar power would cost in round ad hoc numbers - over $3,000.

I've taken a few days away from the project, and have picked up a couple of books on Electricity from the library. I haven't been to the library for a while -- I'd gotten accustomed to just looking everything up on the web. But, geez... I would have had to visit hundreds of sites to find as much information as these couple of books offer. Seems the Internet can't replace the library, yet.

I'm learning a lot. It's kind of fun.

Thanks for the suggestions, RC. I looked into satellite internet. These new services that don't require phone line for upload make it a feasable option for me. But, I'm really hoping to be able to do VoIP and gaming. Latency is really high, with ping times around 1000ms!

I checked commercial wireless as well. I have line of sight to a sliver of the city/valley, 18 miles away, but there are no commercial WISP towers in the area. I *did* contact a guy in that "sliver" who was willing to let me put a parabolic on his roof, in exchange for me paying for his internet access. And, I'm please to say, I actually managed to connect 18 miles away! Latency was great (3-5ms), but I didn't get around to testing transfer speeds. My understanding is that at those distances, I would have needed to use expensive commercial equipment that let me adjust ACK timings. And I feared stability of an

18mile link, so I didn't pursue it further.

Finally, I also checked high-speed cellular plans like Verizon/Sprint EVDO and Cingular's UMTS. They've actually been deployed in my city, but the signal ends a couple miles from my house. Plus, I understand that the carriers artificially increase latency times -- to several hundred milliseconds -- to discourage their subscribers from trying to use the connection for VoIP.

I think if I'd realized how challenging this 700ft distance was going to be, I might have opted for the 18mile wireless link instead. : (

A 25 dBi dish at your end will be very stable with a 15 dB fade margin at

18 miles. Subscriber unit and dish would run you about $500.

Again, I suggest Avalan for your application, Up to 40 miles on 900 MHz. I've got one in and working great!

Bob

decaturtxcowboy wrote:

In the last 10 days or so, I've made a *lot* of progress on this project. Here's a brief update. When this project is complete and I have internet access at home, I'll probably spend a few evenings home and blog about the entire effort, in case it's useful to others.

I ordered 2 transformers: 120V to 48V @ 3A, and 120V to 24V @3A. Action Electronics accidentally shipped me two of the 48V model. When I informed Jim(?), he immediately shipped out the 24V one, free of charge, and let me keep the extra 48V one. I'm very impressed with their service. (No, I don't work there. :)

I also bought a multimeter from Radio Shack.

I did a few tests. First, I tested resistance of my 750ft 22AWG solid copper phone drop wire. It measured exactly as expected: ~12ohms/750ft, which equals the rated 16ohms/1000ft.

Then I moved my testing to the front yard. Here are some numbers:

- Wall outlet measures 121.7VAC

- I connect outlet to primary of 48V transformer, and measure secondary: 54V (a bit higher than 48V expected).

- I attach drop wire to secondary of 48V transformer, and measure output:51.8V. Only 2.2V loss. So my multimeter isn't drawing much amperage. I = E/R = .092 amps draw. Not sure if this tells me anything helpful, though.

- I attach the other end of drop wire to secondary of 24V transformer (meaning that I've placed it in reverse). Primary measures 170V! I don't dare attach any equipment to this...

- I remove 24V transformer, and attach end of drop wire to secondary on spare 48V transformer (again, in reverse). In other words, I have a

48V transformer on either end of the drop wire. I measure primary output: 115.8V. At least I dare to put a higher load on this... I shouldn't fry my equipment. If anything, the increased amperage draw will increase my voltage loss and I won't have enough power.

- I go ahead and plug in my cheap ActionTec modem. Its rated 12V-15V @

500mA minimum, so I don't expect it to get enough power in this configuration. And yet, it boots right up! I measure voltage again with load. 108V. Now I'm really curious.

- I unplug the modem and plug in my WET11, which is rated 5V @ 2A! (Though I found a website that says it needs more like 1.5A). I'm surprised to find that it, too, boots right up. I measure voltage with load: 113.7V.

- I plug in the modem and the WET11. Both boot up. Voltage with load:

105.2V.

- This means that from 121.7VAC at the wall, across 750ft and 2 transformers, voltage with expected load drops to 105.2V. For a total loss of 16.5V. Another calculation tells me that average draw from the two devices is .7 amps...

By the way, in all of these tests, I'm only using one wire conductor for hot and one for neutral. The other pair will carry the DSL signal.

So, now I'm mystified. I expected far more voltage loss than what I'm seeing. Or maybe there's some trick to getting accurate voltage measurements with load... something I'm doing wrong? But there's no arguing with the fact that my devices both booted up. So what gives???

Finally, I went to a friend's house who has DSL. I brought along my drop wire and two 48V transformers, and did speed tests. The 750 ft wire didn't affect speeds at all (which I expected to be the case), and neither did 750ft with ~48VAC on the second wire pair (which I was relieved to see).

In the last few days, I've laid the wire out to the peak, have set up the parabolic antennas, have put 1 amp fuse near the pedestal, have put a few splices on the line to break away if deer trip on the wire, and have ordered Qwest installation which should happen next Tuesday. I'm getting pretty excited about all of this.

But, I still don't understand why there is so little voltage drop across the wire. It kind of makes me nervous, like maybe I'm not understanding all of the theory I've been reading. Any thoughts on this?

(snip)

Usual for cheap meters is 1k ohm/volt, or 1ma. I don't know any that are worse than that.

(snip) > Then I moved my testing to the front yard. Here are some numbers: > - Wall outlet measures 121.7VAC > - I connect outlet to primary of 48V transformer, and measure > secondary: 54V (a bit higher than 48V expected).

The rated 48V is with a 3A load. It will be a little higher unloaded. That includes the loss in the transformer winding and magnetic losses in the core.

There are some things that will fry with too low of input voltage, but that is rare. Motors will easily, if they don't start. Switching power supplies, like the usual PC power supply, may try to pull full power at a lower voltage and higher current.

Looking back, the calculation I see was 26.6ohms round trip. Now you say 16ohms, is that round trip?

Which calculation gets 0.7A?

Some is the difference between 16 ohms and 26.6 ohms, and some I think is that the devices take less than rated power. You lose about 1/7th of the power in the wire, or about 2.7W. That leaves about 2.7*7 or

19W for the load. That sounds about right. Note that your load power should decrease with input voltage with a linear regulator.

With a 16 ohm wire and 24V out you could get 36W out.

With 48V into a 16 ohm wire and wire current I the voltage drop is 16*I, the output power is (48-16*I)*I, with a peak at 1.5A and 36W. For 19W out it is 0.44A in the wire, 0.176A in the load at 108V.

-- glen

[snip]

[snip]

Unless your Radio Shack carries better stuff than mine, I would be suspect of your multi-meter.

Ohm's law again. According to my approximate calculations, your two power supplies at the far end are using 24W, drawing 0.2A. The turns ratio of the transformers (120:48 = 2.5) means that the current in the 48V line is 2.5 * 0.2 = 0.5A. The voltage drop in the 12 ohm conductor is 6V. The transformed voltage drop is 6 * 2.5 = 15V.

Sanity check, 121.7 - 105.2 = 16.5 which corresponds to my approximate value of 15V. I have ignored losses in the transformers because they are insignificant compared to the loss in the cable. Besides that I don't have enough data to calculate it.

carl

Oh, you're right. I was measuring just one way. Wire is rated 16ohms per 1000 ft. Across 750 ft, I measured 12ohms.

The calculation before of 26.6ohms... hmmm, not sure where I got it. Round trip on 700ft. 16ohms/1000ft x 1400ft = 22.4 ohms. Fat fingers?

I = E/R == 16.5V / (16ohms * 1500ft/1000ft) = .69Amps

I'm using 1500ft here because I ended up running 750ft of cable instead of the expected 700ft.

Strange. There are all of these sites out there saying that the equipment I'm using should draw much more current than I'm seeing drawn. Guess I should be happy. : )

Most equipment specs list only 1 power number and that's the max. Typical is usually less.

I wonder how this is coming along?

I had thought before that you could easily use a shunt regulator to keep the voltage down to 24, independent of load.

Well, 24V RMS has a peak of 24*sqrt(2), or about 34V, though I would probably set the regulator at about 30.

Across the line just before the transformer you put a bridge rectifier, and then a shunt regulator made from something like an LM323. It conducts just enough current to keep the voltage down to a set amount.

With a heat sink, I think the LM323 can keep it down to 22 or 28 (Plus 1.5 for the rectifier) even with no load. That, into the 24V transformer, should be fine at the output.

-- glen

Sorry to be away so long. I'm pleased to say that I now have high- speed Internet at my cabin. I really appreciate all of the ideas and suggestions and (patient) explanations from the members in this group.

Glen, I hit the library for books on electricity, and read quite a bit, but must not have made it to the chapter about shunt regulators and rectifiers. :)

To make this work, I went ahead with AC over one of two-pair 22-gauge drop wire. I put a 48volt transformer on one end (at the pedestal), and a second 48volt transformer in reverse on the other end, 750 feet away, at the peak. Then I connnected my cheapo ActionTec modem/ wireless router with a wallwart.

I was worried the draw would be too much (the ActionTec says Max

1.5A), but it's been fine. It's been running without a single reboot for nearly 45 days!

I was also worried about sub-freezing temperatures (the ActionTec is only rated down to 32 F), but again, no problems. Last week we got down to -8 F overnight, but the wallwart + transformer + modem in a single enclosure must generate enough heat. Or maybe I didn't need to worry about it.

Finally, despite comforting assurances from posters above, I had a lingering worry about loss across the wire, running power and DSL signal in parallel for 750 feet. But, sure enough, it's a non-issue. I'm getting 1.5 mbps (down) at the pedestal and 1.5mbps at the peak. In fact, 2 1/2 miles further up the canyon, with a wi-fi relay antenna pair in the middle, passing data through 5 APs/repeaters along the way, I'm consistently getting ~1.4mbps at my home!

I have started a blog to document the entire project. I'll probably focus more time on the wi-fi aspect than the electrical, but this part will get coverage too. I'm hesitant to reveal the address yet (since there's not much there), but here it is: http:// habitat11.blogspot.com .

Thanks again for all of your help.