Repurposing a PC power supply

[MaxPower]

There's an article in an old Everyday Practical Electronics issue about how to go about doing this. I was wondering that since most pedals use very little current if it would be okay to disconnect the power supply's cooling fan if one kept the current demand low (below one amp for example)?

[Prehistoricman]

Using a PC power supply for audio applications is asking for noise and whine. They're optimised for heavier loads and may perform "pulsing" for a low load (depending on quality).

However, if you still wish to go ahead, try it! It will probably passively cool itself fine but it will also not be able to cool itself if it's loaded heavily for some reason.

[MaxPower]

Son of a.... nothing is ever straight forward and simple. Would a voltage regulator and it's associated filtering help combat possible pulsing?

I take it this (noise, etc.) would also be an issue with a gaming console power supply? I was given one of those and it has 5v and 12v outputs with a high current rating. Thought it might make a good power supply for a 12v power amp project. I did figure it might need some filtering.

Sigh, need to study up on power supplies some more.

[GibsonGM]

Yes, you can (and should...) use a regulator with this kind of power supply.  Not difficult to incorporate, and should take care of the noise as well as give you a nice stable voltage!     Probably the same issue w/the gaming console unit...12V will regulate down to 9V well, tho. 

Just for curiosity's sake, can you like to the article from EPE?

[Rob Strand]

I was wondering that since most pedals use very little current if it would be okay to disconnect the power supply's cooling fan if one kept the current demand low (below one amp for example)?

It's a small risk.

Check out this for PC power supply efficiency,
https://en.wikipedia.org/wiki/80_Plus

Assuming your power supply complies to this standard and is on the edge of complying. So if you have a 400W PSU which is 80% efficient at 20% load and you assume that the loss levels off below 20% load then you have to dissipate 400W * ( 1 - 0.8 ) * 0.2 = 16W.

If that 16W is only being dissipated by a small number of parts then it is possible the heat removal for those parts is not sufficient without a fan.   You just don't know.

Another option is to run the fan as low as possible.  Most fans need  > 7V.   Some PSU's use thermal control of the FAN which complicates things.  However, if you want to reverse-engineer the FAN circuit you could set the minimum FAN voltage to just above 7V.  It *is* a pain.  It also means modding the PSU which can be dangerous if you don't know what you are doing.

[PRR]

Put a resistor series with the fan. 12V 0.15A is typical, so acts-like 80 Ohms. Try 40 or 50 Ohms 1 Watt.

[MaxPower]

Thanks everyone. Haven't had the chance to be around lately, sorry for the late reply.

As for a link,  I own the issue. I'll post what issue it's in when I get a chance.

Edit: okay, the project is in the May 2013 issue of EPE.

[printer2]

A PC power supply is so bulky. A laptop supply isn't. Most are 19V now, good for some pedals that like it, can probably still find 12V ones in thrift stores fore a couple of bucks.

[Sooner Boomer]

A PC power supply is so bulky. A laptop supply isn't. Most are 19V now, good for some pedals that like it, can probably still find 12V ones in thrift stores fore a couple of bucks.

This is a great way to get a head start building a power supply.  You can then use a 3-pin regulator to get a lower voltage.  There are a lot of these power supply "bricks" available from the usual surplus dealers on-line, at various output voltages
s and currents (some are even universal AC input!).

[amptramp]

A PC power supply already has the UL, CSA, VDE and possibly any other necessary approvals and since it uses a standard detachable input AC line, it can be adapted to any plug/socket in the world.  Some automatically switch between 120 VAC and 220 VAC, so all your problems are solved except switching noise.

Since the switching frequencies are relatively high (above audio), the necessary filtering is fairly simple.  The switching noise matters because harmonics of the audio can beat with the switching frequency to produce heterodyne signals in the audio band and some devices like delays have oscillators on board that may be affected by the switching frequency.  Linear regulators may limit some of the noise but they are not necessarily designed with a sufficiently high frequency response to do so.