Cyberpower UPS problems: Causes PC to restart on power loss when on high load

[antony209494]

I have the Cyberpower BR1000ELCD

BR1000ELCD - Backup UPS Systems | CyberPower

CyberPower BR1000ELCD offers home and office users a reliable battery backup and safeguards office PCs, network communication equipment, and other electronic devices from surges, spikes, brownouts and other power incidents. Designed with GreenPower UPS™ Technology to improve operating efficiency...

www.cyberpower.com

3 months old - under warranty.

When the power goes out if the load on the UPS is small (< 350 W) everything is fine and keeps the PC along with 1 router, 1 modem and one monitor running. If I am gaming or mining the other 3 devices transfer to battery correctly and PC always restarts (but doesn't turn off). Mining total consumption from the wall is 470W. Gaming on a resource heavy game like GRW brings it to 500W. The UPS runtime chart which you can check on the link above claims that at 510W it gives 1 min of runtime. The LCD on the UPS wildy claims 4-5 mins under that load which of course is BS. I don't need 5 mins. The power drops for less than a second! I barely need it to stay on and not restart and disrupt my mining while I am away or not drop me out of a game for less than a second of power loss! And its advertised as a 600W UPS....
PSU is the Xilence Performance X 750W.

Xilence - XN073 XP750MR9 750 Watt Power supply, Performance X | XP750MR9

Komponenten zur Kühlung und Geräuschminimierung von Desktop-PCs. Leise Netzteile, Kühler und Gehäuselüfter

www.xilence.net

Now here come the questions:
1) If its a transfer time issue: This is the most common thing that comes up while googling for these scenarios. Cyberpower claims a typicall transfer time of 4ms. I cannot find any info on the hold up time of the PSU but it would be ridiculous if it was less than that right? Why when the load is low it transfers fine then? Does increasing the load also increase transfer times? (Doesn't make much sense to me)

2) If its simulated sine wave - Active PFC compatibility issuse: I am a bit confused by this. My PSU is Active PFC. Some people say it's fine for home PCs and pure sine wave is only a must for servers. Others say it is incompatible. Has me thinking if I sould buy something like the much more expensive CP1300EPFCLCD that is pure sine wave. Again, how does this explain that it works in low wattage conditions?

3) If its the PSU: Worst case scenario. I really don't want to have to replace the PSU.

Any thoughs/commments/suggestions?

 

[Deleted member 205776]

I have a friend with that exact UPS you mentioned (CP1300EPFCLCD), but with a Thermaltake Smart 500W RGB. He experiences the exact same issues as you. If he increases the load on the GPU (by increasing detail settings), and a power cut happens (which is quite often for him), it immediately restarts. Just fine on low load.

It could very well be the PSU. Xilence isn't my go to brand when someone asks about quality power supplies.

I have a Pure Sine Wave CP1500EPFCLCD with a Seasonic Prime PX-750 and I'm experiencing no issues like this (my PSU does have about 30ms of hold-up time, which is well over the ATX standard of 16ms). I doubt Active PFC (which every modern PSU has) is the problem, or simulated sine wave. Pure Sine Wave isn't necessary for a desktop system but nice to have, PSUs like it more.

 

[thesmokingman]

With modern psu you should be looking at Pure Sine Wave and not the alternative. Google it. Pure Siine Wave ups deliver power in even arching waves where the modified/simulated delivers its power in chunks. You can imagine the havoc the chunks could play on power delivery.

 

[antony209494]

Damn...The problem is you are both potentially right. Idk if I should get new a new PSU or a new UPS (can't do both). I held Xsilence to a decent standard when it comes to power supplies given that it is owned by the same company that owns be quiet and it has proven solid so far (about 2 years).
If only I could test the UPS with another PC to see what happens with a different PSU.

 

[thesmokingman]

Damn...The problem is you are both potentially right. Idk if I should get new a new PSU or a new UPS (can't do both). I held Xsilence to a decent standard when it comes to power supplies given that it is owned by the same company that owns be quiet and it has proven solid so far (about 2 years).
If only I could test the UPS with another PC to see what happens with a different PSU.

You don't wanna use that ups on modern psu especially with modern gpus that have very fast transient responses, not to mention Ryzen cpus that can change power states in 1ms. As shown in the graph, with a mod/simulated sine wave, the psu is always in a too much power or not enough power situation. Google it. You'll see modern devices using sim sine wave running hotter, buzzing, having issues.

 

[antony209494]

I have a friend with that exact UPS you mentioned (CP1300EPFCLCD), but with a Thermaltake Smart 500W RGB. He experiences the exact same issues as you. If he increases the load on the GPU (by increasing detail settings), and a power cut happens (which is quite often for him), it immediately restarts. Just fine on low load.

It could very well be the PSU. Xilence isn't my go to brand when someone asks about quality power supplies.

I have a Pure Sine Wave CP1500EPFCLCD with a Seasonic Prime PX-750 and I'm experiencing no issues like this (my PSU does have about 30ms of hold-up time, which is well over the ATX standard of 16ms). I doubt Active PFC (which every modern PSU has) is the problem, or simulated sine wave. Pure Sine Wave isn't necessary for a desktop system but nice to have, PSUs like it more.

You don't wanna use that ups on modern psu especially with modern gpus that have very fast transient responses, not to mention Ryzen cpus that can change power states in 1ms. As shown in the graph, with a mod/simulated sine wave, the psu is always in a too much power or not enough power situation. Google it. You'll see modern devices using sim sine wave running hotter, buzzing, having issues.

I forgot to mention I e-mailed cyberpower about this and still waiting on their response.
Right now I am leaning more towards getting the CP1300EPFCLCD and maybe transfering this one to my home theater/TV setup (although it would be a waste of 5 plugs ). I'd rather have a better quality UPS for potentially more power hungry components in the future. If it ends up solving my problem even better! If it turns out to be the PSU I will try to replace it maybe later this year.

 

[newtekie1]

With modern psu you should be looking at Pure Sine Wave and not the alternative. Google it. Pure Siine Wave ups deliver power in even arching waves where the modified/simulated delivers its power in chunks. You can imagine the havoc the chunks could play on power delivery.

I never understand why people always spread this misinformation. A modified sine wave is perfectly fine for modern PSUs. The AC to DC circuitry can handle it just fine. Jonnyguru himself has said modified sinewave UPSes are fine for modern PSUs. They are designed to handle impure input power and smooth it to DC current, it is part of a good PSUs design.

 

[antony209494]

I never understand why people always spread this misinformation. A modified sine wave is perfectly fine for modern PSUs. The AC to DC circuitry can handle it just fine. Jonnyguru himself has said modified sinewave UPSes are fine for modern PSUs. They are designed to handle impure input power and smooth it to DC current, it is part of a good PSUs design.

So I guess you are team "replace PSU" too?

 

[newtekie1]

So I guess you are team "replace PSU" too?

Yes. If the computer is rebooting when the UPS switches to battery, but stays on on boots back up on battery, then the PSU should be replaced because the hold-up time isn't long enough. The ATX spec requires power supplies to have a hold-up time of 17ms, but it isn't uncommon for cheap units to have way less than that.

 

[antony209494]

Yes. If the computer is rebooting when the UPS switches to battery, but stays on on boots back up on battery, then the PSU should be replaced because the hold-up time isn't long enough. The ATX spec requires power supplies to have a hold-up time of 17ms, but it isn't uncommon for cheap units to have way less than that.

I mean its not that cheap...I paid 80 euros for it. More like good value.
I am still stuck on one thing though. How does it transfer normally below 350W? Is PSU hold up time or UPS transfer time dependent on load? Can't explain it any other way.

 

[Hardcore Games]

Old UPS designs with obsolete inverters are all too common. Modern inverters are clean sine wave unlike 20 year old designs.

 

[RJARRRPCGP]

You'll see modern devices using sim sine wave running hotter, buzzing, having issues.

With my earlier UPS that I bought in 2009, one of my CRT monitors that I had, would buzz noticeably.

 

[newtekie1]

I am still stuck on one thing though. How does it transfer normally below 350W? Is PSU hold up time or UPS transfer time dependent on load? Can't explain it any other way.

Yes, the load definitely affects the hold-up time of a PSU.

When the power drops out, the PSU is still supplying power from the charge left in the capacitors. The higher the load on the PSU, the quicker the capacitors will be discharged. A good PSU should have capacitors that are capable of keeping the unit running at full load for 17ms without any input power. But capacitors are often the first cost saving area manufacturers use. They switch to lower capacity caps and the hold up time drops. They bet on people's computer being idle most of the time, so the hold-up time at idle might still be long enough, but not under load.

Old UPS designs with obsolete inverters are all too common. Modern inverters are clean sine wave unlike 20 year old designs.

That's not really accurate or good advise. Pure sine wave UPSes existed 20 years ago and there are a lot of modified sine wave units on the market today. It isn't about the age of the design, it's about building the unit to a cost. Most inexpensive UPSes sold today are still modified sine wave. Heck, even a lot of the more expensive ones are modified sine wave. However, the UPS on my main system, is at least 15 years old and is a pure-sine wave unit.

 

[RJARRRPCGP]

But capacitors are often the first cost saving area manufacturers use. They switch to lower capacity caps and the hold up time drops.

Reminds me that with PC PSUs, you need to watch out for less expensive models using cheap caps, (nothing to do with capacitance as much as the caps being cheap stuff that shall be perm-banned by the CPSC!) because you may be happy in the short term, but likely will regret it in 6+ years, if not as little as 3 years!

 

[antony209494]

ΟΚ....I have also e-mailed Xsilence to ask an official spec on the hold up time. Just to see what the response will be. I shall look to replace the PSU when I can.

 

[thesmokingman]

I forgot to mention I e-mailed cyberpower about this and still waiting on their response.
Right now I am leaning more towards getting the CP1300EPFCLCD and maybe transfering this one to my home theater/TV setup (although it would be a waste of 5 plugs ). I'd rather have a better quality UPS for potentially more power hungry components in the future. If it ends up solving my problem even better! If it turns out to be the PSU I will try to replace it maybe later this year.

You don't wanna use sim/mod sine wave ups on audio equipment either. And as I mentioned pure sine wave is best for 80 plus psu and active pfc psu. The sim/mod is fine for psu pre 80 plus etc.

 

[newtekie1]

And as I mentioned pure sine wave is best for 80 plus psu and active pfc psu. The sim/mod is fine for psu pre 80 plus etc.

That is wrong. Active PFC and 80+ PSUs are just fine with a modified sinewave UPS.

 

[Shrek]

How do active power factor correction power supplies deal with a modified sine wave input? they are trying to keep voltage and current in phase, which must be quite challenging. there is a report that seems to claim the primary capacitor can be exposed to some rather high voltages.

PSU_and_UPS_Compatibility_Test_Report.pdf (libero.it)

 

[newtekie1]

How do active power factor correction power supplies deal with a modified sine wave input? they are trying to keep voltage and current in phase, which must be quite challenging. there is a report that seems to claim the primary capacitor can be exposed to some rather high voltages.

PSU_and_UPS_Compatibility_Test_Report.pdf (libero.it)

A good PSU will handle it.

If you need more confirmation that simulated sine-wave UPSes are perfectly fine to use with active PFC power supplies, see Jonnyguru's post from just a few months ago here:

https://forum.corsair.com/v3/showpost.php?p=1071428&postcount=6

The man designs power supplies for Corsair, he knows what he is talking about.

 

[antony209494]

OK responses from Xsilence and Cyberpower came in:

"Die Hold Up Time of the XP750MR9 as per INTEL Specs is 16.6 ms; however, the the capacity of the main caps can vary it can also be slightly more, upto 20 or 22 ms, but most probably not more.
Unfortunately weh ave very little testing with the set up you mentioned, or generally with the use of UPS Systems ( as mainly private users dont use them ) so we cannot say really what can cause the problem"

Cyberpower claims it is because of simulated sine wave (response attached below).

At this point I think replacing either will fix it: if the PSU was better it would handle the simulated sine wave like a lot of you said. And if it was pure sine wave this PSU would have no problem.

 

[DeathtoGnomes]

I mean its not that cheap...I paid 80 euros for it. More like good value.

You get what you pay for. It is a good value for mid-range unit.

btw is the Cyberpower management software being used?

 

[antony209494]

Final response from CP representative who has been extremely helpful:

Hello dear Sir,

Many thanks for your quick reply.

The issue is NOT switching timing…. The BR1000ELCD has about 8ms switching time. The comment in the forum is inaccurate.

The issue has to do with wave output distortion on BAT mode…. This is the real problem.

In the past many a-PFC PSUs could still work if distortion was not too high, but nowadays they are much-more sensitive.

Indeed the CP1300EPFCLCD will not have an issue with a-PFC power supplies so a good choice.

PR Tower Series are also Great and virtually tool-free in maintenance and the next step over the CP-PFC series; (eg the PR1000ELCD -> https://www.cyberpower.com/eu/en/product/series/professional_tower )

BTW all Apple PCs are also a-PFC and Apple’s favorite UPS is the CP-PFC (especially the CP900EPFCLCD).



It is really sad nowadays… The “low-cost” UPS will be bye-bye since really most computers use a-PFC. The BR are great models for older PCs, or for SOHO items that are impervious to a-PFC -> routers, phones, lights, laptops, small PCs, NAS systems, etc.

Take care and Kind Regards,

 

[micropage7]

Final response from CP representative who has been extremely helpful:

Hello dear Sir,

Many thanks for your quick reply.

The issue is NOT switching timing…. The BR1000ELCD has about 8ms switching time. The comment in the forum is inaccurate.

The issue has to do with wave output distortion on BAT mode…. This is the real problem.

In the past many a-PFC PSUs could still work if distortion was not too high, but nowadays they are much-more sensitive.

Indeed the CP1300EPFCLCD will not have an issue with a-PFC power supplies so a good choice.

PR Tower Series are also Great and virtually tool-free in maintenance and the next step over the CP-PFC series; (eg the PR1000ELCD -> https://www.cyberpower.com/eu/en/product/series/professional_tower )

BTW all Apple PCs are also a-PFC and Apple’s favorite UPS is the CP-PFC (especially the CP900EPFCLCD).



It is really sad nowadays… The “low-cost” UPS will be bye-bye since really most computers use a-PFC. The BR are great models for older PCs, or for SOHO items that are impervious to a-PFC -> routers, phones, lights, laptops, small PCs, NAS systems, etc.

Take care and Kind Regards,

oh they replied it well, so we know the exact problem

 

[newtekie1]

Final response from CP representative who has been extremely helpful:

Hello dear Sir,

Many thanks for your quick reply.

The issue is NOT switching timing…. The BR1000ELCD has about 8ms switching time. The comment in the forum is inaccurate.

The issue has to do with wave output distortion on BAT mode…. This is the real problem.

In the past many a-PFC PSUs could still work if distortion was not too high, but nowadays they are much-more sensitive.

Indeed the CP1300EPFCLCD will not have an issue with a-PFC power supplies so a good choice.

PR Tower Series are also Great and virtually tool-free in maintenance and the next step over the CP-PFC series; (eg the PR1000ELCD -> https://www.cyberpower.com/eu/en/product/series/professional_tower )

BTW all Apple PCs are also a-PFC and Apple’s favorite UPS is the CP-PFC (especially the CP900EPFCLCD).



It is really sad nowadays… The “low-cost” UPS will be bye-bye since really most computers use a-PFC. The BR are great models for older PCs, or for SOHO items that are impervious to a-PFC -> routers, phones, lights, laptops, small PCs, NAS systems, etc.

Take care and Kind Regards,

So the Cyberpower representative recommend you buy more cyberpower equipment. Are we surprised by this response?

Again we have a response from a man that designs modern power supplies for a living. I'd take his response that the power supplies can handle modified sine wave more than the response of a company rep saying "buy more of our product". Just saying.

 

[micropage7]

So the Cyberpower representative recommend you buy more cyberpower equipment. Are we surprised by this response?

Again we have a response from a man that designs modern power supplies for a living. I'd take his response that the power supplies can handle modified sine wave more than the response of a company rep saying "buy more of our product". Just saying.

as usual, marketing trick. like when you wanna buy motherboard then the seller said why don't you consider this one, it has bla bla bla and much better than it
for me thet reply they show that they know about their product, not like the other who clean their hands by saying our product is fine, it has good quality control, and in your case it's not from our product and defend it without checking any further