# How Do You Power Cameras? PoE / 12V DC / 24V AC

1. What percentage of cameras you install have PoE built-in?

2. What percentage of PoE cameras you install will use PoE to power the camera (I don't want to assume) vs AC or DC (if either is supported or available)?

3. If you cannot use PoE for some reason, what is your preferred power source? 12V DC or 24V AC (assuming you could get a camera and choose the power option) OR does it depend on the price level of the camera since low cost cameras typically offer 12V DC only whereas more expensive cameras offer 24V DC or both

I have poe available most of the time, but i dont ever use it.

Why is that? Distance limitation? no PoE switch?

I need to power the ir as well. Since it is all battery backed up, poe just means more devices to power and less efficient. I usually use 12v if distance istn a problem, otherwise i convert to 24vac.

...poe just means more devices to power and less efficient.

If battery back-up is a concern, wouldn't using a POE switch as the power source for all devices make it easier to co-locate and consolidate the backup systems at the switch?

As for efficiency, 48v-55v transmitted with less power loss than 12v... Where do you see the ineffiency?

If you look at it from the perspective of the output of a UPS (for this exercise, lets ignore the efficiency of the UPS) powering a POE power supply which drives a transmission line (Cat5) to the POE device. You have two inefficiencies:(i) the conversion of the power supply and (ii) the losses on the transmission line. Compared to a UPS feeding say a torroid transformer 120V-24VAC and transmitting power over #10 wire.

For a 48 v POE system, imax is 350mA, Lmax is 100M, Rmax is 20 ohm the voltage drop is 7VDC and power loss is 2.45W to deliver 12.95W (IEEE standard). You need to inject 15.4W to the transmission line to get 12.95W at the POE device at the maximums.

Now the POE powersupply:

http://www.phihong.com/assets/pdf/PSA16U-480-POE.pdf

the spec says at 90VAC input it needs 0.5A input or 45W to deliver the 15.4W or 34%, at 240VAC it draws 0.35A or 84W to deliver 15.4W or 18%. Take the average at 26%. You will need to input 59W. All in, the efficiency of the wiring and the powersupply you need 59W to get 12.95W at the POE device or 22% efficienty overall.

For the torroid and #10 wire at 24VAC:

#10 wire is 3.277 ohms/km, I = 12.95W/24VAC =0.54A,

Transmission Ploss = R x I**2 = (3.277 x 2 x 0.1km) x 0.54**2 = 0.19W

Torroid Txfmr efficiency is 90% for a 500VA size. Lower for smaller VA

http://www.tortran.com/transformer_design_guide.html

The power input to the transformer is:

(12.95+0.19W)/0.9 = 14.6W

Overall Efficiency is 12.95/14.6 = 89%

I am kind of surprised it is this different. Did I miss something?

I am kind of surprised it is this different. Did I miss something?

Only my actual objection, which was based on what you said you 'usually' use, 12v, not 24vac, as I stated:

As for efficiency, 48v-55v transmitted with less power loss than 12v...

(DC I assumed, since you specify 24vac later).

Anyway I'll admit that my comment was shortsighted and that you are correct that the DC stepdown must be accounted for in any real model. Even I know the awesome effeciency of the simple transformer at stepping up/down voltages AC voltages, one the holds no worthy counterpart in the DC world.

But I'm up for learning something regardless and I would like to understand your equations fully and see if I can witness that dramatic inefficiency 89% vs 22% that you are predicting, for 24vac.

I would like to ask you if you think that this simple test will be of any value:

i have a 5mp stardot camera which has poe, 24vac and 12vdc on the back. I don't have 100 M of zip handy, but I might have 20 for a first test.

I would just plugin either the 48, 24, or 12v adapters into a watt meter, power up the camera using a 48v passive adapter that will put the power on cat 5, verify that its working and record the kw/h usage of the whole system over 1 hr. Then do the same with 24ac and 12dc with zip cord... i will get you the exact specs of everything if you think its worth doing.

Will it show the difference?

"Even I know the awesome effeciency of the simple transformer at stepping up/down voltages AC voltages, one the holds no worthy counterpart in the DC world."

Actually a good inverter will take 12VDC and convert to 120VAC with 85% efficiency. The problem is that the POE power adapters are a very simple design without thought to efficiency. These devices are under \$50, if efficiency was important they might be \$200 devices. Frankly, I just have had little experience in dealing with them because I needed lots of power for the IR and it seemed redundant to use another power device when I have already got the power there. There are also high efficient DC power supplies (85~90% efficient). With respect to transformers, typically control transformers (not torroidal) their efficiency is more like 50%.

For your test if you want to confirm the power losses in the POE adapter I would use just a multimeter and measure the AC current and AC voltage to calculate your power input. I would assume that the AC current and AC voltage is in phase with each other, otherwise you will need an oscilloscope to confirm this. Similarily on your output measure the DC current and DC voltage to calculate your power output. Compare the product of this with your watthour meter measurement (current needs to be steady). What is the accuracy of the Watthour meter? I would trust the product of the two measured quantities with a multimeter.

If you want to measure the losses in your transmission line it will need to be 100M to have the voltage drop to get the power losses. Hope that helps.

"Even I know the awesome effeciency[sic] of the simple transformer at stepping up/down voltages AC voltages, one the holds no worthy counterpart in the DC world."

Actually a good inverter will take 12VDC and convert to 120VAC with 85% efficiency

To be sure, the DC analogue to the transformer would be just a DC-DC converter, not an inverter. Though like your inverter, the DC-DC efficiency has improved steadily over the years, and 85% is common enough now.

But ac-ac converters, (transformers) will run at >99%, so for every watt of waste heat generated by transformer, a dc-dc converter would generate ten. And then add in the simplicity of the transformer (its a wire), vs the semiconductor (its not) and maybe you will agree that its high time we invent the DC-DC transformer. :)

If you want to measure the losses in your transmission line it will need to be 100M to have the voltage drop to get the power losses. Hope that helps.

Helps alot, thanks. For the record how do you think the three options wil compare? Do you think that short cables (10M) vs long (100M) will actually have a different power type winner?

"But ac-ac converters, (transformers) will run at >99%, so for every watt of waste heat generated by transformer, a dc-dc converter would generate ten"

Transformers are not that efficient unless they are at very high frequency. They have a lot of losses due to the conversion of electrical to magnetic and back to electrical energy at 60hz. As I indicated earlier, unless you use torroids the efficiency is 50% at 60hz. As a matter of fact, in switching power supplies they convert 120VAC to 24VDC in a highly efficient method >90% by converting the 120VAC to dc, then creating a high frequency AC (>20K Hz) and then doing the voltage level conversion and DC conversion at the higher frequency where high frequency transformers are more efficient.

"Do you think that short cables (10M) vs long (100M) will actually have a different power type winner?"

I am not sure if I understand your question. Power loss in the cable (where the delivered power is constant) is effected by voltage, wire size and cable length, but in different ways. If you keep the power constant and keep the cable length and copper cross sectional area constant; if you double the voltage it will cut the power loss by a quarter. Keeping the power and voltage constant and either doubling the cable length or doubling the cross section area of the copper will cut the power loss in half.

My original 'nonpareil transformer' statement was in reference to the one used in your calculation, sans rectification, so torroid it is: (i wont defend it anymore.)

"Do you think that short cables (10M) vs long (100M) will actually have a different power type winner?"

Apologies for the goofy wording there :) What I was trying to ask was:

Do you think that the power supply that is most efficient at 10M will be the most efficient at 100M? For example, if the 24AC setup is determined to have a higher efficiency (than the 12vdc or 48 POE) at 10M, would you expect it to still be the efficieny winner at 100M?

Taking your calculations into account I'd say that at

• 100M - 24VAC should clearly win, POE second, 12vdc out of juice.
• 10M - 12VDC may win, 24VAC second, POE last

Agree?

I don't think I would look at this way. For example, take the 100M example you have different options available. If you use POE you are using Cat5 cable, if you are going with 12VDC or 24VAC for power with a seperate Cat5 for communications. You can make it work with either 12VDC or 24VAC you just need to use bigger wire than Cat5. 24VAC will have the advantage of using smaller wire than 12VDC. At some point it isn't economical to use big wire that will be dependent on the individual situation and something else will need to be tried. The efficiency of the transmission line is dependent on voltage and wire size.

It is not uncommon for us to use #6 wire because we may have 150W at 24VAC of load 300' away.

It is not uncommon for us to use #6 wire because we may have 150W at 24VAC of load 300' away.

In that case I had no business questioning your transmission efficiency since even without a calculator I'm pretty sure 24VAC over 100M of #6 wire will be many times as efficient as the same length cat 5 POE. Is it hard to work with that thick stuff?

you can get highly stranded cable as opposed to building wire and it is very flexible. You need to use ferrules to make solid connections.

98% of our cameras installed are POE and we use it. I don't know what the difference would really be between POE and non-POE powered devices in efficiency.....

Only time we use AC is if it's aspecial use camera like a tag camera, or one where AC/DC is required for a heater, and then 12V is usually fine.

100% POE. All of our companies POE switches are already generator backed up. So I do not have to worry about verifying generator back up at the camera location like I used to. POE has made install and troubleshooting a breeze.

All fixed IP cameras (so far) are powered by PoE. All PTZ IP cameras (as well as 99% of non-IP cameras) are powered by 24VAC.

We prefer 24VAC when PoE is not an option.

BTW, some IP cameras (specifically Axis, if I recall) don't have an option to use non-PoE power.