Power Supply Question

[hamsterears]

Okay, I recently came across a couple ATX type computer power supplies that people didn't have any use for.<p>I also found a website (http://www.nfdc.net/home/cbdb/12%20V%20 ... Supply.htm) where somebody describes converting an ATX supply into a bench supply. (Note: he says it's not a how-to, so if you blow yourself up trying it, it's not his fault).<p>Now, part of his method is to put a load across the +5V/Ground wires, using resistors (about 20W) worth. I began playing with it, and it worked.<p>However, the other evening I was checking the voltages, and discovered that the ATX was still running even though the resistor had come disconnected.<p>So the question is, is this okay? I figure there are three possible answers.<p>A) It's fine, the guy didn't know what he was talking about.<p>B) I was hallucinating, the power supply actually shut down when the resistor came disconnected.<p>C) It'll work okay for a while, then self destruct.<p>Any help would be appreciated.

[jimandy]

I checked out the URL and noted that the writer said the PS needed a load to turn on. I'n no expert on switching PS's (or anything else for that matter) but when the web author further stated <blockquote><font size="1" face="Verdana, Helvetica, sans-serif">quote:</font><hr> I cut the red and black wire from this plug and attached 2   10 ohm, 10 watt resistors wired in parallel. This was to simulate a load on this circuit. Usually the red wires supply +5v DC.
<hr></blockquote>
That's essentially 5 ohms across the 5V, a draw of 1 amp, which seems like an undue waste of precious current.

[haklesup]

Two issues are at play here.<p>To activate an ATX power supply, you need ground the power On wire just like the motherboard would do when you push the power button. AT supplies just turn on with the AC switch. <p>
The second issue is regulation. Look at the sticker on the side of the PS. Note that (most likly) the 5V supply has a max and a min current. Max current is typical of all supplies but the presence of a min value means that the manufacturer will not guarantee that the supply is regulated at 5V unless this minimum current is flowing. In other words, an open load will be unregulated and may measure any voltage.<p>These load resistors are required if you want to measure the voltage with no other load but if you have it connected to a project, you most likly do not need the Load R. Many switching supplies have this min current requirement but not all of them, you just need to look at the specs to know.

[terri]

(Almost simultaneous posting with hacklesup's post):<p>For most of these power supplies, the "dummy load" is just to get them started and keep them oscillating properly. If you already start them with another load, you don't need the dummy load. As I mentioned, I used to keep an old hard drive on the bench just to plug in and get them going to test them out.<p>It looks like maybe what happened here is that when the dummy load disconnected itself, it kept running because it was already loaded from the circuit you were testing. (Moderately likey.)<p>Or, the power supply already had a dummy load wired into its internal circuitry to avoid this startup problem. (Most likely.)<p>Or, the supply was of such a design that it didn't need a dummy load to oscillate properly in the first place. (Least likely.)<p>Yes, it's a waste of power, but these were not intended as bench supplies. They are only handy as bench supplies. As I also mentioned, IBM used to ship PCs that did not have hard drives with a dummy load to simulate the load put on the power supply by the missing hard drive. This resistor was not needed when a hard drive was installed, so there was no "waste" of power in this case.<p>[ February 02, 2005: Message edited by: terri ]</p>

[Enzo]

I have worked with many switchers in my day, though not that one. And some of them require a minimum load on them to even function. If there is no load on these, then they will "zick" once and quit. The output voltage spikes and the protect shuts it down. A lot of supplies have an internal load for each supply voltage.<p>A 1 amp dummy may be more load than it needs, but it will insure operation. For a 15 amp supply, giving up a half an amp or so is no big deal.

[hamsterears]

Okay, I didn't tell everything, since I didn't want to be too long winded.<p>I have two different power supplies - different manufacturer, ratings, etc. When the first one continued working, I began testing.<p>Both supplies would start up without the resistor. Voltages seem all right. So, I'm assuming that I don't need the resistors.<p>Thanks a bunch for all the help.

[rshayes]

Some of the computer supplies that I have seen used a flyback configuration. This stores energy in an inductor and then transfers it to a filter capacitor, which supplies the output current for much of the cycle. Multiple output voltages are obtained by using windings with different numbers of turns on the same inductor core. The output voltages adjust themselves to correspond to the various turns ratios when the energy in the inductor is transferred to filter capacitors through diodes. In effect, each output is a peak rectifier and the turns ratios determine the peak values that each capacitor charges to. The filter capacitors are discharged by the load currrents until they are recharged during the next switching cycle.<p>Only one output can be regulated. This is usually the +5 volt output. The other outputs "track" this output due to the turns ratios of the inductor windings.<p>If the load on the regulated output is light, and the load on the tracking outputs is heavy, the tracking outputs will fall faster than the regulated output. The control loop only looks at the regulated output and fails to supply enough energy to maintain the other outputs. A minimum load is needed on the regulated output to guarentee regulation of the tracking outputs. Normally this happpens automatically, since the +5 volt load is usually the heaviest load on the power supply.<p>A dummy load would probably be needed in situations where you were drawing most of the power from the tracking outputs and very little from the regulated output.