Why does my Seasonic SSR-500GB3 whine??

[Pimpom]

Re: Why does my Seasonic SSR-500GB3 whine??

A stethoscope is great for locating the source of a noise. Even the cheapest ones will do. Just remove the chestpiece and point the open end of the tube around. I've been using the technique for 50+ years in my work.

[kaboom]

Re: Why does my Seasonic SSR-500GB3 whine??

Yes but I already attempted to warranty it and they said no. In the mean time I've voided the warranty by removing the cover and I've also swapped the fan with a different one because I wanted something with more airflow for a project.

I'd replace the standby output caps with either FR or PW, using the biggest that physically fit. Eg- if it takes a D8mm, go for the tallest that fits, and of the highest value. Try for something like 2200u 6.3 or 10v. If it's got a pi filter, the first may be D10, the second D8.

The tertiary aux supply, derived from the feedback winding, likely also supplies that PFC/PWM riser board. Tracing & proving this, to find other caps that need to be changed, is easier said than done. Black PCBs matter, lol...

The only two things I can think of to be causing that noise: either the 5VSB circuit (could be the transformer, or it could be from "sensitive" feedback loop on the 5VSB - though I suspect the latter does not apply, as that usually only happens when very low ESR caps are used on the 5VSB output filter)...

The typical 431 feedback/opto arrangement sometimes has a small ceramic cap (Cff in example) from the cathode of the opto LED to the gate of the 431. If the first filter (C1 in example) is questionable, noise triggers the 431 via Cff & Rff (Rff is not always present) and the series ckt back to C1. At C2, things are calmer (after the pi section). "Feed forward" is attempting to dominate thru Cff & Rff while closer-to-DC feedback is taken after the pi-section; C1 no longer "dominates" and feedback/regulation hunts. This is usually low-frequency enough to be audible. The fact that the standby transformer is noisy in no way condemns it.

Example:


While the IC used in the above may be different, the feedback action remains the same.

Shovey, try the following with the PSU removed or disconnected from the system in question. You'll be applying some loads to 5VSB.

Note the noise with it unloaded. Now try either small light bulbs of increasing size (you can use 12v ones here) or various resistors. 18-20awg wire should work to probe the pins of the ATX connector.

What you're looking, or listening , for is a change in that noise vs 5VSB load- I'd assume it'll be quite noticeable.

[shovenose]

Re: Why does my Seasonic SSR-500GB3 whine??

I'd replace the standby output caps with either FR or PW, using the biggest that physically fit. Eg- if it takes a D8mm, go for the tallest that fits, and of the highest value. Try for something like 2200u 6.3 or 10v. If it's got a pi filter, the first may be D10, the second D8.

The tertiary aux supply, derived from the feedback winding, likely also supplies that PFC/PWM riser board. Tracing & proving this, to find other caps that need to be changed, is easier said than done. Black PCBs matter, lol...



The typical 431 feedback/opto arrangement sometimes has a small ceramic cap (Cff in example) from the cathode of the opto LED to the gate of the 431. If the first filter (C1 in example) is questionable, noise triggers the 431 via Cff & Rff (Rff is not always present) and the series ckt back to C1. At C2, things are calmer (after the pi section). "Feed forward" is attempting to dominate thru Cff & Rff while closer-to-DC feedback is taken after the pi-section; C1 no longer "dominates" and feedback/regulation hunts. This is usually low-frequency enough to be audible. The fact that the standby transformer is noisy in no way condemns it.

Example:


While the IC used in the above may be different, the feedback action remains the same.

Shovey, try the following with the PSU removed or disconnected from the system in question. You'll be applying some loads to 5VSB.

Note the noise with it unloaded. Now try either small light bulbs of increasing size (you can use 12v ones here) or various resistors. 18-20awg wire should work to probe the pins of the ATX connector.

What you're looking, or listening , for is a change in that noise vs 5VSB load- I'd assume it'll be quite noticeable.

Alright I am currently using it in a system but I'll be re-casing that system sometime next week as well as using a different PSU so I will try these things. We'll make it shut up somehow