|
Combat Pretzel posted:Not sure if this is the best thread for this, but I guess the use case overlap is pretty high. Make sure they have actual UPSs and not just "passthrough charging" Without the former they'll just keep running the system through the battery while also simultaneously charging it and you'll wear that thing out really quickly. You want one with a proper UPS relay switch that'll connect your stuff to the mains when there's power and swap you back to the battery (quickly) when they cut out. I think Bluetti only has two models with proper UPS relays, https://www.bluettipower.com/products/bluetti-eb3a-portable-power-station and https://www.bluettipower.com/products/bluetti-ep500pro-home-battery-backup Though I'd imagine they will be rolling them out to their other products as they update them since they add a lot of value for very little cost to them. This guy does a lot of reviews for battery backup/solar gen systems. I'd recommend checking out whatever you might be looking for on there first. He also usually has coupon codes for the models he reviews so that could save you some money. https://www.youtube.com/c/HOBOTECH/search Nitrousoxide fucked around with this message at 15:45 on Sep 18, 2022 |
#
?
Sep 18, 2022 15:41
|
|
|
|
| # ? May 15, 2024 05:41 |
|
|
A UPS is designed to sit on the mains 24x7 and cutover without impact to any of the connected devices - can those large power banks operate in that mode? You can get Li-ion batteries in a UPS now, APC have done it for a few years.
|
|
#
?
Sep 18, 2022 15:41
|
|
|
Combat Pretzel posted:Eh, according to various tests on the interwebs, it can deliver the battery runtime promised. Also, "only" 6.6 amps. On 230V, that's like 1500W. My current Cyberpower UPS only delivers 900W max, beyond that it'll shutdown. 6.6 amps at 120v. The limiting factor is the wattage. So it will only provide 3.3A at 240. Also, your Cyberpower UPS is in the same class as this Bluetti thing, maybe a smidge better. I'm comparing these things to actual enterprise gear, because one of the things you specifically mentioned was true sine output. More "endurance" requires things that consumer devices does not require, like the ability to rapidly charge off mains power and dissipate the heat that creates. It also requires safely storing that amount of fault current while attached to mains power and safely providing rapid discharge of that amount of power, which sounds like a nightmare for portable, plug-in residential gear (which battery solar installations are absolutely not).
|
|
#
?
Sep 18, 2022 15:59
|
|
|
I fully expect there to be at least one "I turned a battery pack from a crashed Tesla into a UPS" video on youtube. But yeah, you'd think putting a bunch of 4680 cells and a new charge controller into an existing UPS design, and selling it for a non--silly price, would be within the capabilities of most companies.
|
|
#
?
Sep 18, 2022 17:54
|
|
|
My Cyberpower unit has true sine.  Actually, I'm pretty hot and bothered that there aren't any more affordable consumer units with it, especially since PSUs with active PFC hate the stepped sinewave when under load. After that yellow-glue-fire-drama around Cyberpower UPSes, I've been casually looking for an 1500VA-ish APC unit with true sine, and the thing costs double of my Cyberpower unit. --edit: Anyway, the point was, this camping module ships with quite a few Wh of capacity for an acceptable price, a "consumer" UPS unit costs a multiple of that, and I don't see why. For the price of that Bluetti unit with 716Wh I get an APC consumer unit with around 220ish Wh. That's just bleh. Computer viking posted:I fully expect there to be at least one "I turned a battery pack from a crashed Tesla into a UPS" video on youtube. Combat Pretzel fucked around with this message at 18:06 on Sep 18, 2022 |
|
#
?
Sep 18, 2022 17:57
|
|
|
Yeah I have a cyberpower too and I’m concerned after that video, if a new thread favourite arises I would be interested in moving over.
|
|
#
?
Sep 18, 2022 18:01
|
|
|
priznat posted:Yeah I have a cyberpower too and I’m concerned after that video, if a new thread favourite arises I would be interested in moving over.
|
|
#
?
Sep 18, 2022 18:03
|
|
|
I picked up some eatons after the video and already had great customer support and have been super simple to hook up to linux. Lost power a few times already and my house is still standing so no complaints.
|
|
#
?
Sep 18, 2022 18:29
|
|
|
Combat Pretzel posted:My Cyberpower unit has true sine. So now you understand the trade off, right? Combat Pretzel posted:My current Cyberpower UPS only delivers 900W max, beyond that it'll shutdown. Combat Pretzel posted:After that yellow-glue-fire-drama around Cyberpower UPSes, I've been casually looking for an 1500VA-ish APC unit with true sine, and the thing costs double of my Cyberpower unit. You're not seeing how all of these things are related? What you are asking is why aren't things that are well built out of expensive parts cheaper because there are other things made out of cheap parts with bad/exaggerated specs that kinda look like the thing you want. The reason is that 99% of the market for the thing you want is enterprise and enterprise demands even higher specs (particularly around output) than you want or need. There is not enough of a market for what you need to make something just for it, so you either run home gamer hardware or enterprise hardware. Those are the two choices. And it sounds like what you want is enterprise gear, you just don't want to pay enterprise pricing for it. Edit: I also don't understand why you or anyone else is worried about true sinewave output. Your 90 to 250 volt 50 to 60 hz switching power supply does not give one poo poo what you throw at it. Motronic fucked around with this message at 18:45 on Sep 18, 2022 |
|
#
?
Sep 18, 2022 18:42
|
|
|
Motronic posted:Edit: I also don't understand why you or anyone else is worried about true sinewave output. Your 90 to 250 volt 50 to 60 hz switching power supply does not give one poo poo what you throw at it. My understanding is that rectifying out the harmonics places more load on your power supply, in addition to potentially affecting output quality in less high quality units. Source: https://www.anandtech.com/show/17398/the-amazonbasics-aurora-vista-1500-ups-review-passable-power/2 Anandtech posted:The peak of the wave reaches over 400 Volts and, since it is not an actual sine wave, throwing it as an input to any switching power supply will generate harmonics. As a matter of fact, the square wave is the worst-case scenario in such designs and will generate harmonics ludicrous in both amplitude and number. These two parameters are a disaster for the PFC circuitry of any PSU, no matter how well made it is. It will (probably) not cause the PSU to immediately blow up but does burden the unit severely, significantly reducing its reliability. Several reputable PSU manufacturers had to revise entire platforms to resist the massive stress that such a UPS could place on them. Do you have something to support the claim that this is not a problem?
|
|
#
?
Sep 18, 2022 19:08
|
|
|
Motronic posted:Edit: I also don't understand why you or anyone else is worried about true sinewave output. Your 90 to 250 volt 50 to 60 hz switching power supply does not give one poo poo what you throw at it. It matters if your PSU has power factor correction, which almost all modern switch mode power supplies do.
|
|
#
?
Sep 18, 2022 19:16
|
|
|
Just hack a 12V/5V DC supercap based power storage in to PCs imo 
|
|
#
?
Sep 18, 2022 19:19
|
|
|
I guess I don't know how much power is lost in converting DC from the UPS battery to AC for its outputs but between that, the cost to do full sine wave instead of a square wave, and the conversion back to DC inside the PSU, it feels like we're missing a trick by not having combination UPS-PSUs (UPSUs?). They could do direct DC-DC conversion for standard ATX power supply outputs and like 2A 110VAC out for monitors or other accessories. There probably isn't enough demand for UPSes specifically for desktop computers to make that even close to commercially viable though. Eletriarnation fucked around with this message at 20:06 on Sep 18, 2022 |
|
#
?
Sep 18, 2022 20:01
|
|
|
Eletriarnation posted:Do you have something to support the claim that this is not a problem? There is a difference between "literally the shittiest inverter on the market" and what your average quality UPS puts out. Also, the last sentence you quoted. I've spent literally decades in this industry, in and out of datacenters and in house server rooms. Plenty of places have been running non-true-sine APC and other quality manufacturer gear for a very, very long time with no ill effects even in the some of the worst places for power quality and reliability. If you personally feel it is important to you and worth it then nobody's going to tell you it's not better. It's just not enough better to justify the price in most cases. You're better off stocking spares.
|
|
#
?
Sep 18, 2022 20:08
|
|
|
My understanding was pure sine wave is more for things that run a motor or for medical devices. PCs should be fine going through their power supply especially if it’s a decent quality one.
|
|
#
?
Sep 18, 2022 20:14
|
|
|
Motronic posted:There is a difference between "literally the shittiest inverter on the market" and what your average quality UPS puts out. Also, the last sentence you quoted. Alright, but if you're admitting that a lovely consumer UPS can have bad results then implicitly it's important to be able to distinguish which consumer UPSes won't have those bad results. Many simulated-sine models probably aren't going to list the various relevant factors of their output waveform, especially at various load levels. How is anyone to know if a random Cyberpower or APC that doesn't say true sine is going to be just as bad without trying it on their own oscilloscope? Your saying "well, I've run lots of commercial-grade equipment and never seen an issue" is of questionable relevance because it's not only anecdotal, but you're not talking about the same kind of kit. I would also assume that whatever random piece of gear you are running inside a datacenter gives a better output than an AmazonBasics UPS. Eletriarnation fucked around with this message at 20:19 on Sep 18, 2022 |
|
#
?
Sep 18, 2022 20:15
|
|
|
Eletriarnation posted:Alright, but if you're admitting that a lovely consumer UPS can have bad results then implicitly it's important to be able to distinguish which consumer UPSes won't have those bad results. No, I'm not. I'm admitting the article that was linked says so. Eletriarnation posted:Your saying "well, I've run lots of commercial-grade equipment and never seen an issue" is of questionable relevance because it's not only anecdotal, but you're not talking about the same kind of kit. Why is it of questionable relevance in a thread where a significant portion of people are running actual, real enterprise grade gear? Eletriarnation posted:I would also assume that whatever random piece of gear you are running inside a datacenter gives a better output than an AmazonBasics UPS. Oh, are you ADMITTING this? I mean, come on man. I don't even know what you're on about here but you seem to be really mad about it. You wanna go pull out the test gear and start documenting things, feel free. But I already know what works in the reality of running gear professionally. If you feel that experience isn't relevant to you then don't listen to it. I'm done with this. You home gamers can go on spec sheet racing your UPSes.
|
|
#
?
Sep 18, 2022 20:46
|
|
|
Motronic posted:Why is it of questionable relevance in a thread where a significant portion of people are running actual, real enterprise grade gear? It's of questionable relevance because (1) we were talking about the characteristics of consumer grade UPSes (2) the people running the actual, real enterprise grade gear do not include the actual person you initially responded to who was asking about a replacement for his 1500VA Cyberpower unit that wouldn't cost twice as much. The thread's not about my emotions, but if you want to go in that direction I'm less annoyed about anything you claim and more about you doing some self-appeal-to-authority bullshit when someone says that something you claimed is controversial, provides a source claiming that you're wrong, and notes that some sources would help support your side of the argument. If your response to being called out on that poo poo is to dismissively go "I'm done with this", I don't think you'll be sorely missed. Eletriarnation fucked around with this message at 22:19 on Sep 18, 2022 |
|
#
?
Sep 18, 2022 20:55
|
|
|
My NAS (QNAP 2 bay, forget the model but its like 7 years old) is making GBS threads the bed. It was a good run but now its time to upgrade. I'm looking for a 4bay drive to mostly just run plex (no transcoding) and back up files for work. My NAS is extremely slow are they faster now? Is there a go to pick for 4 bays?
|
|
#
?
Sep 19, 2022 02:09
|
|
|
lol what the hell happened here. Either way, I've seen teardowns on Youtube of the relevant Cyberpower and APC units, and other complaining about that cheap glue poo poo (which may or may not be the cause for previous drama) and once a while questionable soldering in Cyberpower units, the designs were deemed largely the same. Even the batteries in both came from RBC. So yeah, the very least I ought to be allowed to naively question why the APC unit is considered worth twice the price (more so considering the relatively cheap price of the batteries, at least what my replacements cost). "Schneider Electric" seems to mean gently caress-all, when I listen to the complaints of the electricians and electronics engineers at work. Personally, I'd love to have a double conversion one, considering the poo poo power lines in my town. But no passive cooling (at lower loads) and $$$.
|
|
#
?
Sep 19, 2022 17:32
|
|
|
Just go full industrial and install a motor-generator set in the basement; with a suitable flywheel it should be able to smooth out a lot of noise and transient events. (Note: Do not do this)
|
|
#
?
Sep 19, 2022 17:55
|
|
|
Combat Pretzel posted:lol what the hell happened here. Is APC actually double the price? It might just be a temporary market/supply issue or specific to your region. Here's a 1500W Sine wave APC I bought from newegg which is currently $222: https://www.newegg.com/apc-br1500ms2/p/N82E16842301736 Cyberpower 1500VA sine wave is $219 on amazon: https://smile.amazon.com/gp/product/B00429N19W/ The last time I bought that cyberpower it was $199 in 2019, and when I got that APC from newegg it was $199 because it was list priced at $239.99 but had a $40 discount at the time of order. That was in March of 2022.
|
|
#
?
Sep 19, 2022 22:36
|
|
|
Combat Pretzel posted:The complaint is more, why do "budget" UPS systems (or any other) not have more endurance for the price. If a power outage lasts more than five minutes it's probably going to last a few hours, so there's a significant value gap between 15-30 minutes of runtime and 3-4 hours of runtime where you've multiplied your costs as well as the size and weight of the batteries but you're still probably going to have to shut down before the grid comes back.
|
|
#
?
Sep 19, 2022 23:58
|
|
|
I've never understood why twice as much of a thing costs more money. Just an amazingly hosed up state of affairs when you think about it. High capacity batteries are expensive. You can use lithium which is expensive materials and needs good engineering for fire risk. Or SLA which is expensive to ship and in most places has big extra taxes when you go above a fairly small size to cover the disposal & environmental costs. So that's why consumer units cap out at a rather low watt-hour endurance, besides just the "big money from enterprise" reason.
|
|
#
?
Sep 20, 2022 00:38
|
|
|
Klyith posted:I've never understood why twice as much of a thing costs more money. Just an amazingly hosed up state of affairs when you think about it. That kind of made me curious about how much of the cost of a consumer grade UPS is the batteries. Looking at cyberpower, and rounding a bit, their $200 UPS has $100 replacement batteries. Which suggests that twice the capacity would cost $300 plus whatever it costs to ship the more expensive unit, build the larger chassis, design and support another variant, and whatever modifications are needed to actually charge and use the extra cells. Plus the risk of running into the regulations you mention, of course. Basically, it seems like it's not "twice as much", it's "twice as much of something that makes up about half the price of the unit, plus a complicated overhead". Which I guess in a way is the answer to "why does it cost twice as much to double the capacity".
|
|
#
?
Sep 20, 2022 01:37
|
|
|
FWIW, as an example, I believe the classic Mac Pro line was actually pretty sensitive to simulated sine wave UPS' with its Active PFC PSU, and Apple still recommends that users typically go with a pure sine wave UPS if possible when using one, but I could be misremebering.
|
|
#
?
Sep 20, 2022 08:24
|
|
|
e: ^ They're all Active PFC now unless you buy a really cheap model. Anything with an 80+ rating is likely (certain?) to have it. ^ Well, I know it's a challenge to prove a negative so I'm not expecting some kind of exhaustive demonstration that pure sine vs. simulated-sine is never going to make a difference to a PC. That said, if you go look at a thorough review of a power supply one thing that basically always gets evaluated is ripple in the output. For those who aren't aware, ripple refers to the AC harmonics and noise which are still present in a DC signal. Obviously, AC harmonics in the output must originate from those in the input that aren't entirely filtered out. Some small amount is OK, but you want it to be as low as possible. Now, maybe this is just a gut feeling, but I think it's a reasonable guess that if you introduce all kinds of new harmonics into the input (say, by using a square wave instead of a sine wave) you are going to get one or both of two things. The first is a lot more harmonics in your output, which will impact the stability and lifetime of all the expensive components inside the computer. The second is that the components inside the PSU responsible for smoothing the waveform will work harder, run hotter, and potentially decrease its lifetime. In the worst case you might get a catastrophic PSU failure which in addition to forcing you to eat some downtime and buy a new one, is potentially very bad for expensive and sensitive components connected to the outputs. Also, I know this is the NAS thread but some folks might also have overclocked gaming computers on UPS and those would be even more sensitive to stability issues from dirty input power. So, with all of that in mind, I go to Amazon and search "cyberpower ups" and I see that the difference between a simulated sine wave model and a true sine wave model at 1500VA is $35. The box I'm plugging into this has a $100 PSU and $1000 of hard drives connected to it, with data that would require a lot more than $1000 of work to replace. Now, if we had some kind of credible demonstration that output power quality (and maybe component lifetime, though I know that's harder) isn't noticeably impacted for a current PSU plugged into a cheap UPS, I might be willing to take that bet. I searched and I couldn't find one, though. Eletriarnation fucked around with this message at 14:56 on Sep 20, 2022 |
|
#
?
Sep 20, 2022 14:03
|
|
|
Eletriarnation posted:Now, if we had some kind of credible demonstration that output power quality (and maybe component lifetime, though I know that's harder) isn't noticeably impacted for a current PSU plugged into a cheap UPS, I might be willing to take that bet. How harmful the bad not-true-sine power is to the PSU & downstream components will depend a lot on how heavily the PSU itself is loaded. That's why you're not going to find a lot of demonstrations -- to show anything happening you have to push some mildly unusual scenarios. (Also you're kinda asking for an article proving the baseline assumption that an average cheap UPS doesn't kill PCs: nobody is going to do that. The proof is that PCs aren't dying all the time from cheap UPSs.) A home server NAS that's got 4-6 drives, an average CPU, and no / minimal GPU will be using like 50-100 watts. That will be way below the PSUs spec power, even with a small PSU. The active PFC will not be working hard and a good quality PSU can deal with lovely power just fine at that load. I would feel comfortable plugging that type of NAS server into an average-quality UPS with simulated-sine, because I think the risk of anything bad happening in normal operation is extremely low. If you have a really massive number of drives there might be some caveats. For example, lets say you have the UPS data hooked up & set to turn the server off after 5 minutes of no power. As part of the shutdown, the server spins up 20 HDDs to flush cache and properly unmount. Drives need a way more power when spinning up, and a whole bunch together is non-trivial load. I would not feel comfortable plugging a crazy datacenter-size server into a meh-quality UPS. Your home OC'ed gaming machine, if you want to keep playing when the power goes out in the middle of a heated gamer moment, yeah buy an expensive UPS. If you are ok just hitting alt-f4 when the lights go out in the middle of a game, you'll be fine. It's not an instakill. (And if you are worried that your barely-stable OC will crash from dirty power, I don't know what to tell you other than stop hitting yourself.)
|
|
#
?
Sep 20, 2022 16:19
|
|
|
Klyith posted:(Also you're kinda asking for an article proving the baseline assumption that an average cheap UPS doesn't kill PCs: nobody is going to do that. The proof is that PCs aren't dying all the time from cheap UPSs.) quote:A home server NAS that's got 4-6 drives, an average CPU, and no / minimal GPU will be using like 50-100 watts. That will be way below the PSUs spec power, even with a small PSU. The active PFC will not be working hard and a good quality PSU can deal with lovely power just fine at that load. I would feel comfortable plugging that type of NAS server into an average-quality UPS with simulated-sine, because I think the risk of anything bad happening in normal operation is extremely low. quote:(And if you are worried that your barely-stable OC will crash from dirty power, I don't know what to tell you other than stop hitting yourself.) Eletriarnation fucked around with this message at 17:00 on Sep 20, 2022 |
|
#
?
Sep 20, 2022 16:30
|
|
|
I was going to upgrade my NAS a long with the drives, but for now I've decided to eek out some more time with this old Synology 413. The 1522 looks good but I don't really want to shell out for the thing and then still spend another few hundred getting the 10Gbit upgrade stuff too. And without a network speed increase my current one will be just as fast as shoving files across cat y cables.
|
|
#
?
Sep 20, 2022 17:34
|
|
|
Reasier posted:My NAS (QNAP 2 bay, forget the model but its like 7 years old) is making GBS threads the bed. It was a good run but now its time to upgrade. I'm looking for a 4bay drive to mostly just run plex (no transcoding) and back up files for work. My NAS is extremely slow are they faster now? QNAP TS-453D or Synology 920+ are both great but might be a little overkill for you. I'd also check out Synology DS418 and QNAP TS-431k. They're all gonna be faster than something from 2015, but it probably won't be night and day.
|
|
#
?
Sep 20, 2022 20:59
|
|
|
wolrah posted:Budget or not, generally a "plug-in" class UPS system is generally intended to maintain operation during short outages or voltage excursions. For a longer outage they give you time to recognize this and either start the generator and warm it up before connecting the load or cleanly shut down.. This scales to the enterprise level as well. The solution to "unlimited runtime without the grid" is not "more battery", it's "generators". I'd argue that the type of environment where all of these are true, is exceptionally rare: *Uptime is critical *The grid is expected to be offline for hours and that cannot cause downtime *You have sufficient room / safety measures in place to install enough batteries to support your critical load for X hours plus Y margin *You do not have any means to install a generator and store enough fuel to handle supplying the critical load Consider how price-per-kWh stored scales with batteries, versus "more diesel fuel". As far as power quality is concerned, anything designed to ITIC standards is supposed to handle some pretty massive fluctuations in power. There's a whole designed curve and anything in between the two lines (one for overvoltage, one for undervoltage) is supposed to be something the power supply handles without interruption or damage.
|
|
#
?
Sep 21, 2022 22:24
|
|
|
Generators can also be refuelled if they run low.
|
|
#
?
Sep 21, 2022 23:10
|
|
|
Klyith posted:Crypto in general is terrible, but this storj thing is barely crypto. Their token isn't burning resources for proof of work, it's running a cloud storage biz by renting space on home user drives. Yeah, it's a dumb thing they can fart out so that crypto shitheads with too much (or alarmingly little) money hear about TrueNAS. On the rare chance it leads to any impact on anything and isn't just quietly dropped in a year then I will get mad.
|
|
#
?
Sep 22, 2022 00:14
|
|
|
ilkhan posted:Generators can also be refuelled if they run low. I've been wondering if a high power inverter would be more practical alternative to a generator. A generator is big and requires quite a bit of maintenance if you want to rely on it working when you need it once a year or every other year. A car is in constant use and can be expected to work. Running a car just to produce electricity would be inefficient, but for rare needs it wouldn't be a too big an expense.
|
|
#
?
Sep 22, 2022 03:14
|
|
|
ilkhan posted:Generators can also be refuelled if they run low. You'd be surprised how often they can't. Not because they physically can't, but you can't get the fuel - even if you have guaranteed delivery contracts. In large outages, governments have a habit of overriding business concerns and claiming all available fuel for things like hospitals and their own needs.
|
|
#
?
Sep 22, 2022 03:32
|
|
|
unknown posted:You'd be surprised how often they can't. I guess that would be irrelevant for the question, since any battery backup will run out way before generator fuel. If your fuel is running low you either acquire more or you relocate your operation somewhere that has electricity. Trying to acquire electricity by trucking charged UPS batteries isn't a practical option, unless you hire a fleet of F-150 Lightnings to come an plug in.
|
|
#
?
Sep 22, 2022 04:02
|
|
|
Saukkis posted:I've been wondering if a high power inverter would be more practical alternative to a generator. A generator is big and requires quite a bit of maintenance if you want to rely on it working when you need it once a year or every other year. The maintenance is pretty low and I have never worked any where that had diesel generator backup where the generator wasn't run at least 20 minutes every 2 weeks for testing. They are pretty loving reliable and cost effective.
|
|
#
?
Sep 22, 2022 05:00
|
|
|
Can someone explain what the deal is with rackmount *lengths* is in a practical sense? Is there a standard depth for racks? (possibly 31"?) It sounds like there is also a short depth standard (24"?) for network gear and audio and some other similar stuff. And it sounds like on the long end it goes up to 36", but I assume that extra long might cause problems for some colos and some racks? Can you put an over-long server into a rack (especially near the bottom)? It'll pop out the back a long way, ofc. Can you put a short server into a full-size rack, or does it depend on the rail kit, or what? I will keep an eye open for things popping up locally (tbh I don't think I have the cooling to let noise damped work very well, so it'll probably just be a standard rack too) but otoh is there anything wrong with a basic 4-post unit from raising? https://www.ebay.com/itm/154760929653
|
|
#
?
Sep 22, 2022 05:08
|
|
|
|
| # ? May 15, 2024 05:41 |
|
|
Lowen SoDium posted:The maintenance is pretty low and I have never worked any where that had diesel generator backup where the generator wasn't run at least 20 minutes every 2 weeks for testing. They are pretty loving reliable and cost effective. That certainly applies for the large static generators, I was mostly thinking the small two-stroke gasoline generators meant for home use. Does anyone bother to run them even monthly.
|
|
#
?
Sep 22, 2022 05:42
|
|