Question Seasonic Focus platform high current tripping resolved?

[Pextaxmx]

I see some nice discounts on Seasonic Focus/Prime PSUs these days.
Then I remember the high current tripping problem reported early this year (mainly with 3090 GPUs).
Then I see some discussions about how RMA's Seasonic PSU's manufactured after 2020/12 don't trip anymore with 3090.

Need replacement for Seasonic TX-850 with better OCP for 3090

Bottom line: this PSU (Seasonic Prime TX-850 Titanium) just can handle the 3090. OCP keeps tripping randomly regardless of load and I have to reset the PSU at the switch to get my system back. This is a known issue with certain PSUs and the 3090. Should I go to a Seasonic Prime 1000 watt...

hardforum.com

Can anyone confirm this is the case?

 

[--SID--]

Focus GX/PX should be fine. Prime still has issues, that's why Corsair AX Titanium is discontinued.

 

[rcald2000]

I can't speak to the model, but I would agree with the opinion that you should purchase at least a 1,000 watt unit, if your budget allows.

 

[jonnyguru]

I see some nice discounts on Seasonic Focus/Prime PSUs these days.
Then I remember the high current tripping problem reported early this year (mainly with 3090 GPUs).
Then I see some discussions about how RMA's Seasonic PSU's manufactured after 2020/12 don't trip anymore with 3090.

Need replacement for Seasonic TX-850 with better OCP for 3090

Bottom line: this PSU (Seasonic Prime TX-850 Titanium) just can handle the 3090. OCP keeps tripping randomly regardless of load and I have to reset the PSU at the switch to get my system back. This is a known issue with certain PSUs and the 3090. Should I go to a Seasonic Prime 1000 watt...

hardforum.com

Can anyone confirm this is the case?

It's not OCP/OPP.

Not sure why people CONTINUE to report it as such.

 

[Pextaxmx]

It's not OCP/OPP.

Not sure why people CONTINUE to report it as such.

can you please elaborate? do they turn off regardless of current spike?

 

[jonnyguru]

can you please elaborate? do they turn off regardless of current spike?

It's excessive noise on the +12V sense line being fedback from the graphics card.

 

[Pextaxmx]

It's excessive noise on the +12V sense line being fedback from the graphics card.

so it is rather a compatibility issue with certain graphics card group?
or there is certain noise allowed and PSUs are supposed to be capable of filtering them out? then these Seasonic units fail to comply?

Who to blame?? GPU or PSU?

 

[jonnyguru]

Who to blame?? GPU or PSU?

Both.

Some people have fixed the problem by removing the +12V sense wire from the 24-pin. Some people have fixed the problem by putting I ferrite bead on the +12V sense wire.

 

[Pextaxmx]

Some people have fixed the problem by putting I ferrite bead on the +12V sense wire.

if that can solve the problem, why not Seasonic just send out replacement cable with ferrite bead installed on that wire? Seems to be a sound solution to save their reputation.

 

[Zerk2012]

if that can solve the problem, why not Seasonic just send out replacement cable with ferrite bead installed on that wire? Seems to be a sound solution to save their reputation.

It's not really a Seasonic problem it's the huge sudden spikes the video cards can have way over their TDP.

EDIT but yes I would think they would address it.

 

[Pextaxmx]

It's excessive noise on the +12V sense line being fedback from the graphics card.

A question came to my mind.
Does your statement imply:
(1) RTX 3000 GPUs are sending same noisy signal to any PSU that is powering the GPU
(2) or it is combination of SEASONIC PSU and RTX 3000 that mutually generate the noisy signal.

if it is case #1, actually Seasonic is doing the right thing to protect equipments that is being powered by it. I would rather be concerned about other PSU's that is ignoring warning sign from the GPU.

If it is case #2 then yes, we should avoid Seasonic until they fix the problem

Can you please elaborate?

Thank you for your contributions by the way - I learn a lot from your posts.

 

[jonnyguru]

A question came to my mind.
Does your statement imply:
(1) RTX 3000 GPUs are sending same noisy signal to any PSU that is powering the GPU
(2) or it is combination of SEASONIC PSU and RTX 3000 that mutually generate the noisy signal.

if it is case #1, actually Seasonic is doing the right thing to protect equipments that is being powered by it. I would rather be concerned about other PSU's that is ignoring warning sign from the GPU.

It's #1. And the PSU shouldn't protecting if the circuit could just filter out the noise. It's a signal wire for measuring voltage . It's job isn't to report that the GPU produces too much noise.

 

[Pextaxmx]

It's #1. And the PSU shouldn't protecting if the circuit could just filter out the noise. It's a signal wire for measuring voltage . It's job isn't to report that the GPU produces too much noise.

unless the noise is very transient and easily filtered out by setting proper threshold, how exactly PSU distinguishes noise from genuine distress signal? maybe noise has distinct frequency contents?

 

[jonnyguru]

unless the noise is very transient and easily filtered out by setting proper threshold, how exactly PSU distinguishes noise from genuine distress signal? maybe noise has distinct frequency contents?

Yes.

Where are we going with this?

 

[Pextaxmx]

Where are we going with this?

I am just trying to learn. If I am making you repeat already discussed topic, please direct me to the existing discussion.

(1) is the Seasonic trip mechanism the noisy voltage spike hitting the overpower threshold? which is tighter than others? (it must be awfully tight, based on the magnitude of noise shown above)
(2) PSUs without tripping problem simply has higher wattage threshold? or they somehow predict the transient event and hold the tripping?

 

[jonnyguru]

I am just trying to learn. If I am making you repeat already discussed topic, please direct me to the existing discussion.

(1) is the Seasonic trip mechanism the noisy voltage spike hitting the overpower threshold? which is tighter than others? (it must be awfully tight, based on the magnitude of noise shown above)
(2) PSUs without tripping problem simply has higher wattage threshold? or they somehow predict the transient event and hold the tripping?

1) No. It's not overpower. OPP is not measured on the sense line. It's noise. Introduced by the GPU. Most PSUs can either filter out or differentiate the noise from the correct signal. But every PSU's circuit is different depending on who/how it's engineered. Apparently Seasonic's engineers didn't take this into consideration.

2) NO! Again. It's not over power protection. The IC sees the high frequency noise and freaks out and shuts down the PSU. It has absolutely NOTHING to do with OPP. AGAIN: Many people solve this problem by removing the +12V sense lead or putting a ferrite bead on the +12V sense.

 

[Juular]

if that can solve the problem, why not Seasonic just send out replacement cable with ferrite bead installed on that wire? Seems to be a sound solution to save their reputation.

They apparently do if you contact their support with this problem. But for whatever reason they chose not to inform the public about this problem working case by case with people who already bought it, not the way you should handle this sort of problem IMO. Corsair for example, has made announcements about issues of smaller magnitude.

 

[Pextaxmx]

They apparently do if you contact their support with this problem. But for whatever reason they chose not to inform the public about this problem working case by case with people who already bought it, not the way you should handle this sort of problem IMO. Corsair for example, has made announcements about issues of smaller magnitude.

Interesting.. are you saying they send out exactly a replacement cable with ferrite bead installed on V sensing line? So that is their official cure for the tripping?

 

[emitfudd]

I thought this issue was resolved with a GPU driver update to limit the amount of boost?

I am running a Seasonic Prime Titanium 850 with a 3080 and no problems.

 

[sizzling]

I thought this issue was resolved with a GPU driver update to limit the amount of boost?

I am running a Seasonic Prime Titanium 850 with a 3080 and no problems.

The boost problem was an entirely different issue with the boost design, nothing to do with the PSU.

Im running a Seasonic 650w Prime Ultra Platinum and a Gigabyte 3080 Gaming OC for over a year with no issue

 

[--SID--]

Im running a Seasonic 650w Prime Ultra Platinum

That's a Focus platform, not Prime.

 

[Juular]

Interesting.. are you saying they send out exactly a replacement cable with ferrite bead installed on V sensing line? So that is their official cure for the tripping?

I don't know really, but i've heard reports that they send out some 'special' replacement cables which solve the issue, if i have to guess then it's probably just the 24-pin cable lacking 12V V-sense line or whatever 'shielded' PCIe cables they were sending out as a remedy for prior Seasonic Focus problems (even though in this case, with 12V V-sense on 24-pin it shouldn't help).

I am running a Seasonic Prime Titanium 850 with a 3080 and no problems.

It depends on the amount of factors. In some cases there are no issues, but we know for fact that Jon 'JonnyGURU' Gerow has tested this platform on his own, being the PSU R&D head at Corsair who has employed Seasonic TX platform for Corsair AX, and he confirmed the issue even for 1kW units, which was partly the reason why Corsair AX is now discontinued.

Im running a Seasonic 650w Prime Ultra Platinum and a Gigabyte 3080 Gaming OC for over a year with no issue

Surprised to see a 650W unit to pull off RTX3080 (not downclocked/power limited ?) but that unit is actually based on Focus platform, not Prime. Judging by the single review on Prime Gold / GX we have (because Seasonic doesn't send review samples on it so people wouldn't figure it out apparently), it's even exactly the same (components wise, that it relatively cheaper ones than you usually see in this price range) as Focus Gold / GX, just with single-ended cables and 2 years longer warranty.

 

[sizzling]

Surprised to see a 650W unit to pull off RTX3080 (not downclocked/power limited ?)

From what I can make out people compensate for the quality of the psu by increasing the wattage rating. I am not running a cpu that requires 200W or more, a 3700x. While stress testing the whole system peaked at 546w at the wall using a smart plug. Ok not the most accurate way of measuring wattage but gives a ball park figure. While gaming it pulls around 450w average. I even did some overclocking of the gpu (didn’t take any power readings) and had no problem getting an extra 100mhz before hitting stability issues at anything over 2100mhz. In the end I decided the tiny gain in performance wasn’t worth the extra heat and noise, it gave better scores in benchmarks but added nothing to actual gameplay. I have seen a YouTube build of a SFF system using a FE 3080, 9900k and Corsair SF600. Now that did make me cringe give the CPU’s power needs but the system past all the benchmarks and games they threw at it with average scores.

 

[Zerk2012]

From what I can make out people compensate for the quality of the psu by increasing the wattage rating. I am not running a cpu that requires 200W or more, a 3700x. While stress testing the whole system peaked at 546w at the wall using a smart plug. Ok not the most accurate way of measuring wattage but gives a ball park figure. While gaming it pulls around 450w average. I even did some overclocking of the gpu (didn’t take any power readings) and had no problem getting an extra 100mhz before hitting stability issues at anything over 2100mhz. In the end I decided the tiny gain in performance wasn’t worth the extra heat and noise, it gave better scores in benchmarks but added nothing to actual gameplay. I have seen a YouTube build of a SFF system using a FE 3080, 9900k and Corsair SF600. Now that did make me cringe give the CPU’s power needs but the system past all the benchmarks and games they threw at it with average scores.

I have a Kill-A-Watt meter but I don't think it can react fast enough to catch very short power spikes.

 

[Pextaxmx]

I am seeing enough evidences no reason to walk into this mess by buying any of Seasonic PSUs at the moment. I don't understand why Seasonic would be so discrete about their acknowledgement of this issue and make users wonder whether the issue is resolved or not. Seasonic used to be #1 PSU manufacturer for long which is quite a title to have. Something that a company only can achieve over generations of excellent track record. Now they became the #1 premium PSU manufacturer that we should avoid as much as we can if RTX3000 is in your system, or is in your future plan.