ATX Computer Power Supply Spark in Non-computer Application [closed]

Question

I would like to ask for your assistance with using a Silverstone DA1650 Gold computer power supply (you can find more about it here: Silverstone DA1650 Gold) for a non-computer application. In this specific setup, I intend to utilize only the 12V rail, which could potentially draw up to 110A. My plan involves cutting the connectors of 9 EPS 8-pin CPU +12 and GND power connectors and connecting them using solderless crimp terminal connectors (yellow 10-12 AWG). Each of the 3 application inputs will require 3 EPS connectors, each consisting of 4 wires.

Brief Problem Description: I encountered some issues and questions during my initial power supply testing. When I connected the Green PS-ON wire to GND, the power supply turned on successfully. I applied a 1A electronic load to measure the 12V rail, and it displayed a stable voltage. However, every time I turn on the power supply, I hear a loud noise resembling a spark. I initially thought it might be a relay, but after multiple attempts, I'm certain it's a spark. Despite this noise, the power supply continues to function. The spark noise is not consistently present, particularly if I turn the power supply on shortly after turning it off. However, when the power supply has been off for an extended period and is turned back on, the spark noise returns.

My Question: I'm seeking guidance on identifying and addressing this significant spike or spark issue. My primary concern is whether I can continue using this power supply for testing. Additionally, if there are any questionable choices in my setup or testing approach, I would greatly appreciate any feedback. Thank you in advance for your assistance, and please let me know if you need any additional information.

Detailed Explanation of My Steps (probably not necessary to read): Here, I'll outline the steps I took that led to these problems, as there may be some relevance. Initially, I connected the Green PS-ON wire to GND, successfully turning on the power supply. Next, I aimed to measure the 12V line using an electronic load. Initially, I applied a 100mA load to the single pins of the EPS 8-pin CPU +12 and GND, resulting in a measurement of only around 1V. I sought help online and learned that newer computer power supplies often require a larger load to activate the power rails. Consequently, I decided to gradually increase the load, starting with 1A. However, when I applied the load and turned on the power supply, I heard a large spark sound, hence, I turned it off. I suspected that I might have damaged the power supply, so I connected it to an old computer, where it successfully powered up the system, making me think that it is working properly. However, I still occasionally heard the spark noise when turning on the power supply, which concerned me. Why I did not get the 12V before, turned out to be because I did not connect the 6 additional sense wires (originally attached to the EPC 24-PIN motherboard connector) to the power supply. Now, I disconnected it from the computer to test it as before. When I applied a 1A load again, the 12V rail remained stable, but the spark noise persisted. As mentioned before, this spark noise doesn't consistently occur, especially when turning the power supply on shortly after turning it off. However, when the power supply has been off for an extended period and is turned back on, the spark noise reappears, and it's quite noticeable.

Answer 1

It is unlikely that what you hear is a spark. It is just likely due to abnormal operation of the power supply because it is not used correctly with the way it is assumed to be used.

If it does really spark, it's broken, damaged, and potetially dangerous so you should not use it anyway.

ATX supplies are not intended to power arbitrary loads.

The situation is even worse if you have no load, or only load one output rail. Loading only one rail may cause too high voltages on unused rails and the overvoltage protection circuits may trigger and try to shut down the supply or clamp the overvoltages to safe level, but as there is no load it will quickly be within limits again and the supply resumes normal operation until hitting the overvoltage limits again.

If you do a search here, most questions about ATX supplies being used for custom purposes is why they don't work properly.

You need to at least draw current from all voltage rails, the amount how much is unknown for any general supply or what your supply needs for proper operation. It will also likely help to add capacitive load, to help with certain scenarios you may subject to the power supply, like only drawinh current from 12V load and suddenly disconnecting the load or suddenly reducing the amount of current taken by load.

There are specifications for ATX supplies and they must work properly when used as per the specification requirements. If you don't use the supply as per the specification, the supply does not need to work in your usage properly.

Answer 2

I would look for loose connections. There may be a bad solder joint at one of the main filter capacitor leads. Loud sparking might result, while output current would go uninterrupted.