My XFX Core Edition Pro 450W power supply started making a weird ticking noise some time ago. Eventually, I figured out what the issue was, and managed to fix it. I purchased this unit back in 2012, so it's an old model, but I'm still posting this in case someone finds it helpful or interesting.
THE PROBLEM
The power supply started making an irregular ticking noise when running, but would otherwise work perfectly fine. The issue would sometimes come and go. A Google search for "xfx core ticking" found multiple other people reporting the same issue, so I assume this is a design flaw. Here is a random YouTube video (not mine) demonstrating the issue: https://www.youtube.com/watch?v=je6ugQAqt8g
My first thought was that this is clearly a fan issue, but this is not the case. The sound does not physically originate from the fan; in fact, the sound persists even with the fan disconnected.
I poked around with my oscilloscope, and turns out that each "tick" corresponds with a fast increase in the PFC output capacitor (main tank capacitor) voltage. In other words, the PFC circuit attempts to start for 1-2 mains half cycles, drawing a short current surge which makes a "tick" sound in the filter inductors, but then immediately shuts down again. This cycle then repeats irregularly. The rest of the power supply works fine on my 230V mains despite the non-functioning PFC.
Eventually, I figured out that I could reliably make the problem start and stop by poking the feedback circuitry with my oscilloscope probes in a specific way. The primary side controller is an Infineon ICE1CS02. This chip has separate "PFC VSENSE" and "PFC OVP" inputs (datasheet attached). It turns out that on this unit, the VSENSE (output operating voltage) was set to 382V and the OVP (output overvoltage) was set to 390V. These voltages are probably too close together for reliable operation, so the overvoltage protection can randomly trigger during startup/normal operation. (Also cutting close to the 400V rating of the PFC output capacitor.)
THE FIX
I replaced a 1.1M surface mount resistor (R322) in the PFC VSENSE divider with a 1.0M resistor. The resistor in question is shown in the top right in the attached photo. I barely managed to squeeze in a 1/8W through-hole resistor in the space provided (didn't have any surface mount components at hand). This reduced the PFC output voltage to 372V, which provides enough headroom to the overvoltage protection level. The power supply has worked fine since this replacement.
THE PROBLEM
The power supply started making an irregular ticking noise when running, but would otherwise work perfectly fine. The issue would sometimes come and go. A Google search for "xfx core ticking" found multiple other people reporting the same issue, so I assume this is a design flaw. Here is a random YouTube video (not mine) demonstrating the issue: https://www.youtube.com/watch?v=je6ugQAqt8g
My first thought was that this is clearly a fan issue, but this is not the case. The sound does not physically originate from the fan; in fact, the sound persists even with the fan disconnected.
I poked around with my oscilloscope, and turns out that each "tick" corresponds with a fast increase in the PFC output capacitor (main tank capacitor) voltage. In other words, the PFC circuit attempts to start for 1-2 mains half cycles, drawing a short current surge which makes a "tick" sound in the filter inductors, but then immediately shuts down again. This cycle then repeats irregularly. The rest of the power supply works fine on my 230V mains despite the non-functioning PFC.
Eventually, I figured out that I could reliably make the problem start and stop by poking the feedback circuitry with my oscilloscope probes in a specific way. The primary side controller is an Infineon ICE1CS02. This chip has separate "PFC VSENSE" and "PFC OVP" inputs (datasheet attached). It turns out that on this unit, the VSENSE (output operating voltage) was set to 382V and the OVP (output overvoltage) was set to 390V. These voltages are probably too close together for reliable operation, so the overvoltage protection can randomly trigger during startup/normal operation. (Also cutting close to the 400V rating of the PFC output capacitor.)
THE FIX
I replaced a 1.1M surface mount resistor (R322) in the PFC VSENSE divider with a 1.0M resistor. The resistor in question is shown in the top right in the attached photo. I barely managed to squeeze in a 1/8W through-hole resistor in the space provided (didn't have any surface mount components at hand). This reduced the PFC output voltage to 372V, which provides enough headroom to the overvoltage protection level. The power supply has worked fine since this replacement.
Attached Files
CAPs POOF
Comment