Power Question

[lboucher]

Excuse me if this question is silly for some reason.

So I just ordered a few smart string controllers to play with this year, but thinking forward.
I (like many) tend to collect stuff over time. Due to this I have a few lower grade computer power supplies laying around.
Is it possible to cut and solder and connect these in parallel to get the required amps to drive a large number of light?

Or would this be bad on the power supply circuitry for some reason?


Just looking for the cheapest way forward.

[Corey872]

I was told basically "You could, but it would be a bit of a hassle"

You are not allowed to view links. Register or Login

I guess there are issues with load balancing and such which add cost to the proposal.  I've searched around for 'cheaper' solutions as well, and surprisingly, the PC power supplies seem to be about the cheapest on a watt-to-watt basis.

Though who knows - maybe some experimentation is in order in the 'off season'.  I was also told it was impossible to run 20 cheap diode bridges in parallel to replace the burned out rectifier in my welder.  They have been happily cranking away for nearly a decade.

[lboucher]

Thanks for the info, i figured the answer was probably buried somewhere in a posting.
Figured it was worth its own thread as lots will be thinking it to.

Also for the ME in the room.
Can someone explain what "Diode back every power source on each supply" means?
Just a link to something talking about this concept would be great. Thanks

[rrowan]

Basically diodes are a one way flow switch.

Wikipedia page
You are not allowed to view links. Register or Login

Cheers

Rick R.

[Corey872]

I think you probably want an EE, but my simplistic understanding:

If you have multiple power supplies on a common rail, there will be on supply which is the 'weakest' and one which is the 'strongest' - ie one which is putting out the least power and one the most.  You would have this condition even if the power supplies are same manufacturer, same model, same output, etc - and things would get even worse if you had different make/model/output supplies.  As the load on the common rail fluctuates, the weak supply may have a tendency to 'absorb' or sink some of the power from the strong supply.  This is even worse if one supply happens to be turned on before the other.  So there is some risk in burning up the power supply.  Plus, when that happens - depending on how it burns up, there may be a risk of releasing the magic smoke from down stream components...hubs, controllers, a string of nodes, etc.

Diodes on the output would prevent this reverse current 'sinking' - but add another level of components and complexity - and another mode of failure.

[chrisatpsu]

plus with atx power supplies, they require a load to function, so there will be more fitting of diodes besides just the +12v wires.

[lboucher]

I think you probably want an EE, but my simplistic understanding:

If you have multiple power supplies on a common rail, there will be on supply which is the 'weakest' and one which is the 'strongest' - ie one which is putting out the least power and one the most.  You would have this condition even if the power supplies are same manufacturer, same model, same output, etc - and things would get even worse if you had different make/model/output supplies.  As the load on the common rail fluctuates, the weak supply may have a tendency to 'absorb' or sink some of the power from the strong supply.  This is even worse if one supply happens to be turned on before the other.  So there is some risk in burning up the power supply.  Plus, when that happens - depending on how it burns up, there may be a risk of releasing the magic smoke from down stream components...hubs, controllers, a string of nodes, etc.

Diodes on the output would prevent this reverse current 'sinking' - but add another level of components and complexity - and another mode of failure.

Thanks that all makes sense.