I've talked about my little blended power feed before so I thought I'd post
I made one this morning and I used a couple clear heat shrinks to try to show the insides. Didn't turn out like I wanted in the pics, too much glare.
This is an extremely simple design.
PoE>Shielded end>*1>feed line
*1
blue/white > 1ohm resistor > zener diode > blue/white
brown/white straight through
shielded end ESD > green straight through to feed line ESD dump
remove orange/white and the leftover white lead from green/white
repeat for as many power supplies as you'd like to blend.
why does this do?
the resistor sheds volts based on the amps. I use a 1ohm because it's the smallest I could get in a >=1W capacity.
so, ohm*amp=lost volts.
the zener diode is a reasonably efficient one way valve to keep power from supply 1 from pushing into supply 2. It still will a tiny bit, but the supplies don't have any issue with ~1.5v backfeed.
voltage is pressure, and the highest pressure wins. So if there are 2 supplies and they each have a slightly different voltage, they will not naturally balance. Also, these supplies don't drop many volts under load.
supply 1 24.1V1A
supply 2 24.05V1A
supply 3 24.04V1A
In this case, supply 1 will be used exclusively until the voltage drops to 24.05V, then it will still do about 90% of the work. not ideal.
add the resistor. Now, an 8W 24V RocketACLite draws .33A. 1ohm*.33A = .33V drop.
That would bring supply 1 down to 23.77V, but supply 2 and 3 are both above that so basically this load would get split between them in an almost equal manner.
why 3? why not 2 or 5?
redundancy! Lets say I have 4 RocketAC drawing 8W each. 32W is 24V1.3A. That's more than a single 24V1A=24W supply can handle. 2 supplies is ~23.5V2A=47W. But what if one fails? 24W isn't enough. 3 supplies will basically feed 23.7V3A and 71W, -or- will deliver 47W if one fails.
If this were only 2, then I wouldn't need the resistor because that's really, at least for me sensibilities, a fail-over supply. That's great, but I'm almost always needing more than 24W.
Also, ubiquiti POE supplies are -over powered- for a reason. Why does an 8W radio ship with a 12W supply? so that it's only run at 2/3 capacity. If I need 24W, a 24W supply is a bad idea because it runs a full capacity. You might hope that the 24W supply was really a 32W supply with a lower rating to handle this...but it's not.
but what about more power? yeah, no problem. Add as many as you like. Your limitation is your feed wire and your input on the other end. It's usually pretty easy to run CAT6, where running 'power wire' can be troublesome with building management and inspectors etc. A CAT6 wire as dedicated power can do 100ft and lose about 1.3v per amp.
The cost of this is like $3, and that includes the shielded ends.
This does sap a little voltage off. At full draw on a 24V1A supply, you could lose 1V. That's 1W that needs dissipated by the resistor per supply (max). Realistically, that 4 Radio setup (I'm ignoring a switch etc...) is drawing 1.33A, so you would lose 1.33V aggregate across all the supplies, or about .45V each, and have to dissipate 1/2W off the resistors. I buy 1ohm3W resistors.
I do heat shrink these up, so there is an insulation barrier there, but it's very low output and I've never seen these get even warm to the touch.
The last thing to add here is that I put this at the base, before the lead to the tower. *IF* I were feeding each supply to the tower on it's own CAT5e/6 line, then I could skip the resistor and just do the zener diode. CAT6 24AWGx8 conductors at 80' is about 1ohm of resistance. I don't want to run extra wires though.
DC power blender...pics
-
rebelwireless - Experienced Member
- Posts: 607
- Joined: Mon Sep 01, 2014 1:46 pm
- Has thanked: 31 times
- Been thanked: 136 times
1 post
Page 1 of 1
Who is online
Users browsing this forum: No registered users and 15 guests