Maximum Power Supply Draw For 1360 Nodes White?

[dearvbguru]

I am trying to get a close estimate or actual of how many amps the power supply would draw for a fully addressed hub of 1360 nodes (16 x 85 per strand) all white.

From the power calculations cheat sheet by RJ, it should be close to 41.6 Amps @ 12V DC.

If I "guess" I would say 41.6 * 12 = 499 Watts so depending on actual voltage (110 to 120) would be 4.5 to 4.1 Amps assuming no losses in power supply.

Just trying to figure out how many power supplies I can safely put on the same circuit.  If I have 20 amp circuit then I should be able to put four on the same circuit.  The max would only be pulled when all the strings were on all the time and all white.  Normally I would only have a brief burst of all white in the whole display.

I know it will vary based on power supplies, but just curious if anyone has actually put a kill-o-watt meter on it and confirmed?

[zwiller]

This is great question.  Hopefully someone has used their killawatt to test this or has experience with this, first year for smart strings for me.

I follow your logic and your numbers look good to me in theory.  However, you might want to consider that odds are your "20 amp" receptacle is most likely only 15a and not a dedicated circuit unless you wired it yourself and verified.  If you are planning to run 4 fully addressed hubs (5,440 nodes) on a single circuit I think you are pushing it and should give yourself some leeway.  I would suggest 2 per circuit.  Also, you could run each separately from a non dedicated circuit. 

[dearvbguru]

This is great question.  Hopefully someone has used their killawatt to test this or has experience with this, first year for smart strings for me.

I follow your logic and your numbers look good to me in theory.  However, you might want to consider that odds are your "20 amp" receptacle is most likely only 15a and not a dedicated circuit unless you wired it yourself and verified.  If you are planning to run 4 fully addressed hubs (5,440 nodes) on a single circuit I think you are pushing it and should give yourself some leeway.  I would suggest 2 per circuit.  Also, you could run each separately from a non dedicated circuit.

I am in control of the circuits.  I have been doing this since 2003 before LEDs and would draw about 90 Amps full on power.  I had a special 50Amp 240V range socket installed right next to my 200 Amp panel in my garage.  I had a 50 Amp 240 V breaker installed in main panel that feeds the plug.  I then had an extra extra heavy duty cord to feed a subpanel mounted on plywood.  Then the sub panel has individual breakers in it that each breaker feeds one outlet mounted on the plywood. Currently each breaker is a 15 Amp breaker but I have spare slots in it that I will put 20 Amp breakers and 10 AWG wire to 20 Amp outlets on the board.  I already have 2 100' 10 AWG extension cords so I can run 20 Amps. 

I am trying to minimize the number of cables I have to send out of my garage for power because I will have 8 power supplies to run for pixelnet in addition to all my DMX stuff and regular LOR stuff including a 12 CCR tree.  In previous years, the bundle of 12, 14 and 16 AWG  cords going out of my garage was about 4 inches in diameter.  Just trying to consolidate some things this year since I will have several pixelnet hubs in close proximity to each other.

[tbone321]

10 guage is a bit of overkill for 20A, especially if your outlets are mounted to the same board as your pannel.  The 80% rule for current load is a good one to follow in these situations.  Also remember that there is more current being drawn than what the nodes are drawing.  The 5V and the 3.8V circuits are also being used as well as the load resistor on the 12V side used to keep the regulators on the PC supply stable.  If you go above 80%, then the breaker will blow eventually if you hold it there and the closer you get to the rated current above 80%, the less time you have. 

[zwiller]

You have a power inventory?  How many nodes and other lights total? 

If it were me, I would try to run your show just off of the 2 - 10AWG lines.  That gives you 60A but limited to the 50A breaker.  You could run the 8 pixelnet power supplies off of 4 - 15A receptacles on the panel knowing they pull a total of around 32A max and have another 20A receptacle for dmx lights, etc which I would probably split to 2 - 15amp receptacles.  Panel has 6 - 15A receptacles. 

This is very down in the dirty and probably not ideal but you're talking to a guy that played in a band where I hooked up an amp rack requiring 100A to 2 old receptacles the bar provided for power and we played near top volume for 4 hours at a time.  Also the same guy who added a 200A service to his house and was told by power company that they will only feed 120A to it. 

To me, with your setup there is enough protection that if you overdo it you will pop a breaker.  Any electrician which I am NOT would never advocate such and that is not necessarily bad.  Keep running your numbers and be careful.  Hopefully a solution will present itself. 

[chrisatpsu]

1360 nodes = 42.5 Amps of 12VDC

as long as you find a power supply that outputs 50+ amps of 12v power, you can put this all on one supply/hub.

judging strictly from the 12v output, you can't determine the power draw from an ATX power supply, since it also has different rails.
The sticker should tell you what it's input draw is in Amps. Use that and go from there in your planning.

[dearvbguru]

In response to a couple previous posts..

I was curious for one hub so I could multiply by number of hubs (. I don't just have one hub.

50 amp breaker in main panel is 240 v so I have 100 amps at 120. If you double voltage you half the current. This is why power companies actually transmit the power at thousands and thousands of volts. Power = current squared * resistance.  By going up to thousands of volts they reduce the actual current to minimal to reduce losses on the lines.

I am planning on using 10 awg wire to match extension cords 10 awg rated for 20 amps. The 80% of capacity rule is for constant current draw.  Breakers operate on heat and magnetic. Spikes in the current draw can easily exceed the breaker if they are brief.  This occurs any time you were to start an inductive load there are large inrushes of current. Motors and incandescent lights are inductors. Compressors on air conditioners and refrigerators and freezers are the worst culprits.  The larger the gauge wire the less resistance it has and you can somewhat longer runs and have more power available at the end because you have less voltage drop because of resistance.  I plan in using the extra heavy duty extension cords to somewhat central points and can brand out to the hubs and each branch won't use as much do they can be regular sized extension cords.

I agree the calculation is not accurate because of the small amount of draw on the 3.3 and 5 volt rails but this is negligible in comparison to the 12v rail current draw and there is also an efficiency factor for the power supply itself. Typical power supplies are 80% efficient at converting the 120 into the dc voltages rest is lost to heat.

You can't take the current draw listed on the label unless you are drawing max on all the dc rails. The power supplies I am using hardly draw anything while just sitting there powered on and connected to a hub. The auto adjusting fan sometimes stops and starts as needed but if you put your hand on the power supply you can't even tell its on by heat.

I was primarily curious as to the accruals if anyone had actually measured. I guess I will have to put my meter on it and turn on individual strands and get my actual draw and record it do I will know how many I can put on each.

So far in 9 years I haven't popped any breakers yet because I make sure my loads are balanced as best I can over each leg of the 240 and limit power on each to known limits. Even when I was almost running fully loaded I could run 90 to 100 amps for brief pops and not trip any breakers.

Now gfci outlets tripping that's another story all together and have had years fighting it to find on loose or broken mini bulb on a wireframe and when the rain allowed it to go to ground through the wireframe... But that's what they are supposed to do...

[dearvbguru]

Hubs = 8 not (

[chrisatpsu]

just making a point.

I have 2 power supplies sitting on my desk

Power supply #1  36A @ 12V   input 115VAC 7A
Power supply #2  30A @ 12V   input 115VAC 16A

as you can see, they're both providing almost the same output on a 12V rail, yet the draw on household current is considerably different.

[dearvbguru]

just making a point.

I have 2 power supplies sitting on my desk

Power supply #1  36A @ 12V   input 115VAC 7A
Power supply #2  30A @ 12V   input 115VAC 16A

as you can see, they're both providing almost the same output on a 12V rail, yet the draw on household current is considerably different.

I understand that each is different and guess the best way to determine is to actually measure.

My power supplies say max AC input current which is if everything is loaded down.  One type I have is capable of 24A @3.3 24A @5 and 50A at 12 if you add the watts up you get 800W but it says total power of 630W (which would be 79% efficient on the DC side) but also says peak power 730W (for 60 seconds which would be 91% efficient).  If you look at the AC side of the label it says  100V-240V - 10/5 Amps.  At listed 100V which is low voltage and more current, thats 1200 Watts of power used on the input side and that would mean the overall efficiency of the power supply is between 53% and 61% even though its listed at "80 Plus Certified" efficiency.  So my 12V is the primary current driver and adding some more in for 3.3 and 5V I would only be using abour 52% max on the DC side of its capacity so I should be using less current then 1200 W on AC side.

This is why I was asking the question because I guess I dont understand the power ratings on the power supplies and how they come up with actual watts used on input and output side to try and get in the ballpark of the actual current and see if anyone had a better idea of what to expect.

I got half my large order of pixels in from Ray today and my connectors so I maybe I can finish putting the pieces together on my 85 node strings and run some tests with my meter and report the results.  It will take a bit and will run on the second power supply type as well to see what the differences are.

[zwiller]

Yep 2nd leg gives you another 50A for a total of 100A.    

[tbone321]

I am planning on using 10 awg wire to match extension cords 10 awg rated for 20 amps. The 80% of capacity rule is for constant current draw.  Breakers operate on heat and magnetic. Spikes in the current draw can easily exceed the breaker if they are brief.  This occurs any time you were to start an inductive load there are large inrushes of current. Motors and incandescent lights are inductors. Compressors on air conditioners and refrigerators and freezers are the worst culprits.  The larger the gauge wire the less resistance it has and you can somewhat longer runs and have more power available at the end because you have less voltage drop because of resistance.  I plan in using the extra heavy duty extension cords to somewhat central points and can brand out to the hubs and each branch won't use as much do they can be regular sized extension cords.

As I said, 10 awg is still overkill on the panel outlets and will really offer no benefit in the few feet between the panel an the outlet but if it makes you feel better to use it, go for it.  The 10 gauge extension cords may be helpful with the much longer runs.  I am aware of how breakers work and what the 80% rule means and if you get a freeze in the show with all lights on and you are exceeding it and pushing the rating of the breaker, you will have lights out sooner than you expect (although that may be a good thing then  ).  As for incandescent bulbs, they are NOT inductive loads.  They are resistive loads and have NO kickback voltage.  What they do have is a variable resistance (much like a toaster) dependent on the temp of the filament but the variable resistance has nothing to do with inductance.  Flurecent fixtures OTOH, are inductive loads as are relay coils and the motors that you described. 

[dearvbguru]

As for incandescent bulbs, they are NOT inductive loads.  They are resistive loads and have NO kickback voltage.  What they do have is a variable resistance (much like a toaster) dependent on the temp of the filament but the variable resistance has nothing to do with inductance.
 

It may depend on the light bulb but most filaments in regular household incandescent light bulbs (not miniature or specialty lights) are actually a coil of wire.  Any wire has inductance but when you coil a wire you increase its inductance.

Look at the attached photo.  Its a magnified element from an incandescent bulb.  The wire is self is coiled and then the coil is coiled.

But yes the elements in them do increase their resistance as they heat up.  They appear as a resistive load when warm and stable but they do behave like an inductive load at start up with a high inrush current.

Back EMF is produced by a changing electrical field.  This can occur when you have a motor that takes time to start spinning because of its inertia or size or load.  Back EMF is also produced in relays with the energizing coil.  If you drive a relay with a transistor and turn off the relay through the transistor you will create a high back EMF which is a "voltage" which will want to flow through your transistor in reverse.  This is why you put a diode backwards across the coil so when you abruptly turn it off the back emf will be able to flow through the diode and the coil until dissipated saving your transistor.

[tbone321]

We can go around and around with this but a light bulb is NOT an inductive load.  The reason that the filament is coiled is to get more of it into a small space for more light.  You tend to see that filament type in high powered halogen automotive bulbs.  There is no core there and the winding is not tight enough to consider it an inductor.  Electrolytic capacitors also have a large inrush of current when they first get power and are discharged.  Do you call that an "inductive load"?   As you said, any wire or coiled shape of wire may have some inductive properties but that does not by definition make them inductive loads.

One final thing about inductors, their current draw and passing capability is is dependent on frequency.  An incandecent light bulb doesn't care if it is operating at 0 Hz (DC) 60 Hz (typical AC in US) or even 1000 Hz, the current draw remains pretty much the same.  Try that on a relay coil and you will see a difference.  Motors have a high inrush current because the coils acts as close to a dead short until the armature begins to spin and manipulates the magnetic field.  This manipulation of the magnetic field INDUCES counter EMF voltages which impede the flow of current and reduce the current draw.  If something slows down or stops that armature, the manipulation of that magnetic field also reduces or stops and the current draw of the motor increases or shoots back to max (onlt controlled by the wires actual resistance).  The fact that the primary current flow is manipulated and controlled by INDUCED counter EMF voltages makes the device an inductive load.

A light bulb has a high inrush of current because the filament (regardless of shape) when cold has a very low RESISTANCE.  As the filament increases in temp, the resistance of the filament also increases which is what is controlling the flow of current.  This increase of resistance also impeeds the building of any magnetic fields that may be trying to form due to he shape of the filament.  Since the primary current is being controlled by the RESISTANCE of the filament, it is called a resistive load. 

[rrowan]

maybe this discussion needs to stop.

How about a Gentleman's agreement to agree to disagree.

Rick R.