Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 21

Thread: Seasonic Prime Ultra Titanium 1000W and stepped approximation sine wave UPS?

  1. #11
    Join Date
    Jan 2020
    Location
    EU
    Posts
    10
    Thanks Thanks Given 
    1
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by Jon Gerow View Post
    Worst case: The PSU isn't designed with the proper front end protection and the bulk cap blows up, but that doesn't kill your hardware on the DC side of things.
    [...]the Corsair AX Titanium has been out for some time now, is based on the Prime Titanium, and has had no bulk cap failures or complaints via tech support that the PSU doesn't work with a step sine wave UPS.
    Thanks Jon, for taking the time to reply! I take it that a company with the reputation like Seasonic will not allow a 'worst case' design to grace their line up. I checked your review on the site here, it shows pictures of 400V 105C caps, and if I read your comments from the LinuxTechTips forum correctly, that should be enough to not be 'worst case'?

    *edit* currently talking via PM to Seasonic rep, I'll summarize the results once finished, since he didn't reply to the thread itself.
    Last edited by Jonathan; 4 Weeks Ago at 11:03 AM.

  2. #12
    Join Date
    Dec 2011
    Posts
    27
    Thanks Thanks Given 
    6
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by Jonathan View Post
    Thanks Jon, for taking the time to reply! I take it that a company with the reputation like Seasonic will not allow a 'worst case' design to grace their line up. I checked your review on the site here, it shows pictures of 400V 105C caps, and if I read your comments from the LinuxTechTips forum correctly, that should be enough to not be 'worst case'?

    *edit* currently talking via PM to Seasonic rep, I'll summarize the results once finished, since he didn't reply to the thread itself.
    Did you get the summary back?

  3. #13
    Join Date
    Jan 2020
    Location
    EU
    Posts
    10
    Thanks Thanks Given 
    1
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Sorry, had to pull a few allnighters, too tired.

    As promised, a summary of what the Seasonic rep answered to my questions. And more of what my own research found.

    I tried to get him to tell me if they design their PSUs with square wave UPS power in mind and programmed the controller accordingly, but did not get a straightforward answer. So I am assuming that is a 'No' then.

    He said they designed PSUs so that - at least some of them - could work with simulated sine wave input. He agreed with Jon's post above, and said literally that it's 'quite the lottery to have it working or not'. In 99% of the time it works, but if you don't want any hassle use a sine wave UPS.

    My interpretation of this is as follows: don't use square wave/simulated sine wave UPS.

    While from my conversation, Jon's post as well as previous mails from Seasonic, I do think their PSUs won't get damaged from simulated sine wave UPSs, that's not because they planned for it, just that the components are sturdy enough to handle the spikes.

    They don't give any warranty it will work though and here's why I don't think using a simulated sine wave UPS is a good idea: I tried the Eaton Ellipse Eco on my Seasonic Prime Ultra Titanium. It worked, in the single test I did pulling the mains plug. I saw a big inrush peak, so I would need a big 1600VA/1000W model to handle this.

    But, from what I gather here and from various places on the internet, this is not a guarantee it will work the next time I pull the plug. It will depend on how much load there is at the time, at what moment in the sine wave the power is lost etc. It may just as well make my workstation shut down or reboot.

    That means the UPS will only give a false sense of security. Next time you're working on something really important, guess what might happpen... Ask Mr Murphy.

    Add to this that you're unlikely to run just the workstation on the UPS. You're also plugging in the monitor, maybe like me some external HDD case and a NUC. These also have power supplies and you have to verify these also will work 100% of the time with simulated sine wave power, without taking damage in the long or short term.

    Since I'd be hard pressed to find someone to repair my old but much beloved monitor, it's quite the risk. And buying a UPS is not about risk taking, but about mitigating risk.

    Just out of curiosty, I mailed Intel and APD about the power supply on the NUC, Silverstone about the HDD case built-in power supply and Philips about the monitor. Sofar, only Philips responded and they're actually making an effort looking into it, but they couldn't just tell right away.

    Now, I don't blame Seasonic or any of the other manufacturers. There's hundreds of types of UPS (although a lot seem to come from the same rather shitty Chinese OEM manufacturer) and it's impossible to test with all of them and give any 100% assurance of compatibility. After all, it's not their responsibility - mains power is sine wave, that's the standard. The fault lies with the UPS manufacturers cutting corners.

    And they aren't interested in changing their lucrative shenanigans. Every single one of them will tell you to buy the more expensive (way overpriced in many cases) sine wave UPS if you don't like the simulated sine wave ones. No guarantees those will work, if you want security from risk get the sine wave ones. Now, if only buying a UPS wasn't exactly about getting safety...

    From what I researched, CyberPower has a sine wave type that is about the same price as other brand's simulated sine wave types. It works by modifying a triangular wave and yields a pretty nice sine wave. This proves the extra cost is really peanuts. CyberPower as a very poor reliability reputation for their consumer models (at least here in Europe, maybe the 110V range is different), so it's not on my list of options.

    I did find a few older or used sine wave types from more reputable brands, but here's the main other reasons I don't want those: they're too big (allowing for a lot of time on full load, but I don't need that) and especially, very VERY noisy. Not something you'd want on your desk. New models cost twice or more that of non-sine wave models, but still have the noise and size issues.

    So unless you're willing to spend a lot of money on a UPS with simulated sine wave, and be happy with say 99% safety and 1% risk, there's not much you can do. Except save your work often.

    At least SSD manufacturers seem to take note, I asked Samsung about the NVME 970 range of theirs, and they said these are immune to power loss corruption (they have a tantal capacitor with enough power to finish any writes and handling what's in the SDram cache).

  4. #14
    Join Date
    Oct 2017
    Location
    Grand Junction, CO
    Posts
    2,192
    Thanks Thanks Given 
    125
    Thanks Thanks Received 
    72
    Thanked in
    65 Posts

    Default

    https://digilander.libero.it/hibone/...est_Report.pdf
    shows voltages up to 497V on the bulk capacitor

    Many PC power supplies have bulk capacitors rated for 400V while the above report shows over 400V for every square wave input.

    I wish Corsair would repeat these no-load tests on their own offerings.
    Last edited by ashiekh; 4 Weeks Ago at 05:21 PM.

  5. #15
    Join Date
    Jan 2020
    Location
    EU
    Posts
    10
    Thanks Thanks Given 
    1
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    That report is limited to just a few types of two brands, a bit too limited, but it does show the problem.

    Jon pointed out in his post to LTT forum:

    "If you're going to use 400V bulk caps in your PSU, they better be 105C rated. Otherwise, they should be 420V. Then I don't think you would have any issues. "

    So for an emergency situation now and then, when the device gets shut down pretty soon after engaging simulated sine wave battery power, I suppose Jon and the Seagate rep are right and these PSUs won't take damage.

    Nevertheless, like I said, I want to be absolutely 100% sure, that my PSU doesn't decide that in certain circumstances it suddenly no longer likes those squary waves and shuts down. Without proper testing, like only the manufacturer of the UPS or PSU can do, that certainty isn't going to happen.

    It's really unfortunate this issue isn't widely known, and consumers just trust that whatever is offered for sale will be OK. There's no incentive for UPS manufacturers to make cheaper small UPS for workplace or home use with sine wave output.

    Anyone know of a maker forum with knowledgeable people who might be interested in starting a home built design project for such a UPS? I'm not an electronics engineer, but I could try to create some enthusiasm for such a project.

  6. #16
    Join Date
    Oct 2017
    Location
    Grand Junction, CO
    Posts
    2,192
    Thanks Thanks Given 
    125
    Thanks Thanks Received 
    72
    Thanked in
    65 Posts

    Default

    Quote Originally Posted by Jonathan View Post
    So for an emergency situation now and then, when the device gets shut down pretty soon after engaging simulated sine wave battery power, I suppose Jon and the Seagate rep are right and these PSUs won't take damage.
    I don't know about you, but I leave my PC plugged in when I am away; the test (https://digilander.libero.it/hibone/...est_Report.pdf) was for "Power Supply OFF - 5Vsb Mode"


    Quote Originally Posted by Jonathan View Post
    Anyone know of a maker forum with knowledgeable people who might be interested in starting a home built design project for such a UPS?
    You want a cheap solution? replace the bulk capacitor (if needed) with a 500V unit; problem solved. Cost ~$5

    * Disclaimer: don't do this unless you know what you are doing; such capacitors can hold a lethal charge.
    Attached Images Attached Images
    Last edited by ashiekh; 4 Weeks Ago at 10:36 PM.

  7. #17
    Join Date
    Dec 2011
    Posts
    27
    Thanks Thanks Given 
    6
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by Jonathan View Post
    Sorry, had to pull a few allnighters, too tired.

    As promised, a summary of what the Seasonic rep answered to my questions. And more of what my own research found.

    I tried to get him to tell me if they design their PSUs with square wave UPS power in mind and programmed the controller accordingly, but did not get a straightforward answer. So I am assuming that is a 'No' then.

    He said they designed PSUs so that - at least some of them - could work with simulated sine wave input. He agreed with Jon's post above, and said literally that it's 'quite the lottery to have it working or not'. In 99% of the time it works, but if you don't want any hassle use a sine wave UPS.

    My interpretation of this is as follows: don't use square wave/simulated sine wave UPS.

    While from my conversation, Jon's post as well as previous mails from Seasonic, I do think their PSUs won't get damaged from simulated sine wave UPSs, that's not because they planned for it, just that the components are sturdy enough to handle the spikes.

    They don't give any warranty it will work though and here's why I don't think using a simulated sine wave UPS is a good idea: I tried the Eaton Ellipse Eco on my Seasonic Prime Ultra Titanium. It worked, in the single test I did pulling the mains plug. I saw a big inrush peak, so I would need a big 1600VA/1000W model to handle this.

    But, from what I gather here and from various places on the internet, this is not a guarantee it will work the next time I pull the plug. It will depend on how much load there is at the time, at what moment in the sine wave the power is lost etc. It may just as well make my workstation shut down or reboot.

    That means the UPS will only give a false sense of security. Next time you're working on something really important, guess what might happpen... Ask Mr Murphy.

    Add to this that you're unlikely to run just the workstation on the UPS. You're also plugging in the monitor, maybe like me some external HDD case and a NUC. These also have power supplies and you have to verify these also will work 100% of the time with simulated sine wave power, without taking damage in the long or short term.

    Since I'd be hard pressed to find someone to repair my old but much beloved monitor, it's quite the risk. And buying a UPS is not about risk taking, but about mitigating risk.

    Just out of curiosty, I mailed Intel and APD about the power supply on the NUC, Silverstone about the HDD case built-in power supply and Philips about the monitor. Sofar, only Philips responded and they're actually making an effort looking into it, but they couldn't just tell right away.

    Now, I don't blame Seasonic or any of the other manufacturers. There's hundreds of types of UPS (although a lot seem to come from the same rather shitty Chinese OEM manufacturer) and it's impossible to test with all of them and give any 100% assurance of compatibility. After all, it's not their responsibility - mains power is sine wave, that's the standard. The fault lies with the UPS manufacturers cutting corners.

    And they aren't interested in changing their lucrative shenanigans. Every single one of them will tell you to buy the more expensive (way overpriced in many cases) sine wave UPS if you don't like the simulated sine wave ones. No guarantees those will work, if you want security from risk get the sine wave ones. Now, if only buying a UPS wasn't exactly about getting safety...

    From what I researched, CyberPower has a sine wave type that is about the same price as other brand's simulated sine wave types. It works by modifying a triangular wave and yields a pretty nice sine wave. This proves the extra cost is really peanuts. CyberPower as a very poor reliability reputation for their consumer models (at least here in Europe, maybe the 110V range is different), so it's not on my list of options.

    I did find a few older or used sine wave types from more reputable brands, but here's the main other reasons I don't want those: they're too big (allowing for a lot of time on full load, but I don't need that) and especially, very VERY noisy. Not something you'd want on your desk. New models cost twice or more that of non-sine wave models, but still have the noise and size issues.

    So unless you're willing to spend a lot of money on a UPS with simulated sine wave, and be happy with say 99% safety and 1% risk, there's not much you can do. Except save your work often.

    At least SSD manufacturers seem to take note, I asked Samsung about the NVME 970 range of theirs, and they said these are immune to power loss corruption (they have a tantal capacitor with enough power to finish any writes and handling what's in the SDram cache).
    Thank you very much for your reply and work.

  8. #18
    Join Date
    Jan 2020
    Location
    EU
    Posts
    10
    Thanks Thanks Given 
    1
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by ashiekh View Post
    I don't know about you, but I leave my PC plugged in when I am away; the test (https://digilander.libero.it/hibone/...est_Report.pdf) was for "Power Supply OFF - 5Vsb Mode"
    But even when the PSU is off, it needs to be on battery power of the UPS, which should not be common. Still, I do switch off my computer entirely, including peripherals. Electricity is quite expensive over here. But I still would think a non-sine wave UPS isn't a good idea, and should be avoided if possible.

    You want a cheap solution? replace the bulk capacitor (if needed) with a 500V unit; problem solved. Cost ~$5
    * Disclaimer: don't do this unless you know what you are doing; such capacitors can hold a lethal charge.
    Sure, but you'll also need to look into the filter caps. And you'll lose your warranty - which in case of Seasonic is 12 years...

  9. #19
    Join Date
    Oct 2017
    Location
    Grand Junction, CO
    Posts
    2,192
    Thanks Thanks Given 
    125
    Thanks Thanks Received 
    72
    Thanked in
    65 Posts

    Default

    Quote Originally Posted by Jonathan View Post
    But even when the PSU is off, it needs to be on battery power of the UPS, which should not be common.
    One over-voltage is all it takes to destroy a capacitor


    Quote Originally Posted by Jonathan View Post
    Sure, but you'll also need to look into the filter caps.
    Why?
    Last edited by ashiekh; 4 Weeks Ago at 07:25 PM.

  10. #20
    Join Date
    Jan 2020
    Location
    EU
    Posts
    10
    Thanks Thanks Given 
    1
    Thanks Thanks Received 
    0
    Thanked in
    0 Posts

    Default

    Quote Originally Posted by ashiekh View Post
    One over-voltage is all it takes to destroy a capacitor
    Supposedly the protection circuits in a PSU should kick in? From Jon's post at LTT: "If the PWM controller lets the higher voltage peaks through and the bulk cap is undersized, then long term abuse could cause that cap to get too hot and vent [...] Some PSUs I've used would just shut off because the PWM would see the higher voltage and trip a safety" - throughI'm no electric engineer, don't ask me, but Jon is and if he says it should be OK if the caps can handle 105C, who am I to argue?

    Why?
    I read in one of the articles I've been researching that these too are stressed by the square wave shaped power of these UPSs. I'll try to find again where.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •