PoE (Power Over Ethernet)Step By Step Hack Howto |
power over your ethernet cable to an access point
by: Terry Schmidt < terry at nycwireless dot net > |
|
The 12 volt Apple Airport and the UGate 3300 use a 2.5mm Inner Diameter and 5.5mm Outer Diameter Coax DC Power Jacks. Other Access Points may use different sizes. (Original Linksys WAP11 [5 Volt, 2 Amp] uses 2.1mm Inner Diameter, New Linksys WAP11 uses a 5.5mm Outer Diameter and 2.5mm Inner Diameter) By using less expensive Wall Mount Modules (not CAT5 spec), or not using them at all, you can reduce the parts cost of this project. |
2. Drill a hole in your two-port mount housing. Mount the Male DC plug in the housing.
Also refer to wiring diagram at bottom of page. [Note this is the Intel, Symbol, Orinoco standard for wiring, not the Cisco standard for wiring]
|
![]() UGate 3300 and 250' Feet of Ethernet Cable Using PoE ![]() Apple Airport Guts and 250' Feet of Ethernet Cable Using PoE |
Another PoE Module Setup by Mike Cambell.
![]() Using a Lucent 48 vdc phone power supply for the source power. The power supply came as a bundled pack of about 12 that came with the last switch we purchased. It just uses RJ11 to connect to the 4 port doghouse. We use Orinoco AP-1000's for our infrastructure so we buy the units that come with the PoE splitter. Another credit to making this was the fact that at the time, Lucent only offered the 6 and 12 port injectors. I've since noticed that PowerDsine has come up with a single unit. I imagine that Orinoco will be offering those in the future. -- Mike Campbell Mike Campbell's PoE injector Module Bruce Kozlik's PoE hub. Uses a Lambda 50 Watt 48VDC supply, Siemon CT coupler and pannel. They were able to supply 4 access points, moint it all in a 1 rack unit. |
![]() If anyone would like to send me a better diagram, please do and I will replace my two-year-old-esque drawing. < terry-poe @ nycwireless dot net > |
The following setup was tested with a UGate 3300 Access Point (Specs 12.0V/0.8A DC) and an Apple Airport Base Station (Specs 12.0V/1.2A DC). They both use the same DC plug. Power- After having been contacted by many people, my power testing requirements were inadequate (I tested the voltage on an open circuit). I no longer have the PoE modules to test since they have been in a live installation working for 4 months now. Ping Times - There was no difference in ping times. All were <10ms. Ping Floods - Ping flooding was done from a Linux host using the 'ping -f' command over a 10/100 switch. The UGate 3300 suffered 8% packet lost under both tests, while the Apple Airport Base Station suffered 1% packet lost under both tests (with and without PoE, and with and without 250' Feet of Cable). The PoE setup did not affect the amount of lost packets under ping flooding. Conclusion- The maximum segment distance for Ethernet 10BaseT is 330 feet (100 meters) (specified in IEEE 802.3), so this PoE setup should work at the maximum segment distance with the stock voltage outputs of the DC transformers (depending upon your access point and DC transformer). 500 Feet of Cable Testing Addendum (2001 May 05)- This setup was additionally tested with 500 feet of CAT 5 cabling, 170 feet more than is allowed for ethernet specifications. The ethernet signal worked over the 500 feet of cabling at both 10baseT speed (Apple Airport) and 100BaseT (Ugate 3300) with no problems. The DC power however did have problems. Both access points would power up (i.e. the lights would come on), but they would not fully initialize and function. I don't really consider this a problem with this hack since your not supposed to use 500 feet of cabling for ethernet anyway. Check the packaging, documentation, or the bottom of the device to find out how much current it draws. Usually it'll be between 50 and 90% of what the transformer can supply. (If the device ships with a 1500ma transformer, expect it to draw around 1200 at peak, and normal operating current should be around 1000.) These numbers are based on my experience alone. -- From email from "Myself" You can fix this by using a higher voltage transformer to overcome the drop. Be careful! This is why I said to find the actual power draw of the device. You want to do all your testing using a large power resistor which draws approximately as much power as the device. When you've found the correct transformer voltage, doublecheck your polarity and try it with the device again. The caveat here is that your new transformer rating assumes the resistance that you tested it with. If you find a transformer that works over 500 feet of cable, and you run it over 100 feet, you've changed the resistance. Current draw (the device) is the same, but resistance has gone down, so voltage drop goes down. Ergo, the actual voltage reaching the device will go UP, and you might cook something. -- From email from "Myself" |
Electrical Math (thanks to Ryan Cole for emailing this)
Wire gauge / Ohms per 1000ft The TIA-568 spec which defines Cat5 (cat 3 as well) specifies that either one shall be no more than 9.38 ohms per 100 meters. Our testing of name brand Cat 5 says more like seven ohms per 100 meters, so 3 ohms per hundred feet is safe and conservative. -- Scott Carter Lets do another situation... 28 Gauge over 50 ft at 2 amps.... 6 Volts x 2 Amps = 12 Watts dissipated over the wire I would not expect that to start a fire outright, but you could use it to keep your pipes from freezing in the winter. From Bill Cherry: Vendor/Model Pins4,5 Pins7,8 -------------------------------------------------------------- Intel 2011B + 12/24 VDC - 12/24 VDC * Symbol 4121 + 12/24 VDC - 12/24 VDC * Cisco 350 Pwr Injector - 48 VDC + 48 VDC ** * Intel is rebrand of Symbol, these use the Bias T Pwr Injector by Symbol, standard 12 V power supply for up to 100', 24 V power supply for up to 350' ** No Load conditions |
|
© 2001, 2002 NYCwireless. All rights reserved. NYCwireless is a registered trademark of NYCwireless |