UPS for FreeBSD
Paul Pathiakis
pathiaki2 at yahoo.com
Sun Nov 30 16:11:47 UTC 2014
To all that responded on this thread:
Thank you. I feel I have a better understanding of all this now. :-)
The time spent to better educate others and myself is welcome and so
"FreeBSD Community".
P.
On 11/30/2014 00:33, John Johnstone wrote:
> I agree this thread is extremely long so I'll consolidate.
>
> On 11/28/14 10:55 AM, Paul Pathiakis via freebsd-questions wrote:
>> Again, these UPSes are not double sine wave, they are stepped. You can
>> get one for cheap, but, again, how much value do you place on your
>> server and the information it has? If you're looking at APC, you have
>> to look at their 'server class' UPS. They start (I think) at 1500VA and
>> about $850.
>
> I think things could be clearer about the terms "double sine wave" and
> "stepped". I think it should be made clear that although a UPS that
> outputs a 60 Hz square wave is producing a poor approximation of a
> sine wave, better designs still produce a sine wave in a stepped
> fashion, just with a higher number of steps. One that produces an
> infinite number of steps would be indistinguishable from a pure sine
> wave.
>
> On 11/28/14 5:04 PM, Paul Pathiakis via freebsd-questions wrote:
>> I found this which seems to be pretty solid. I believe it someone
>> mentioned UltraUPS.
>>
>> http://www.controlledpwr.com/whitepapers/uutopla1.pdf
>
> This whitepaper refers to the stepped approximation of a pure sine
> wave in their design:
>
> "The inverter reconstructs a sinewave from the DC using PWM (pulse width
> modulation); a method that “digitizes” the DC into various width pulses
> to make an AC sinewave."
>
> The number of steps needs to be high enough to not cause problems.
>
> On 11/27/14 2:25 PM, Darren Pilgrim wrote:
>> Sine-wave approximating inverters do bad things to any power supply
>> with a regulator cap (which is everything that won't catch fire on
>> its own). The issue is the high frequency components and the
>> hundreds of under- and over-voltage events per second inherent to the
>> stepped square waveforms used (every step is a spike or sag).
>>
>> UPS manufactures know this is bad, so they try to hide it by calling
>> it "modified sine wave", "quasi sine wave", "simulated sine wave",
>> "PWM sinewave", etc., and hope you're dumb enough to fall for it. I
>> have yet to see a consumer UPS that doesn't do this.
>
> I wouldn't say it fair to interpret UPS manufacturer's descriptions of
> their methods of simulation as "hiding".
>
>> You need to buy a server-grade UPS to get something that won't damage
>> your electronics. APC SmartUPS, Cyberpower PFC Sinewave or Smart
>> App, Eaton 5P/PX or 9 series, Tripp Lite SmartOnline, etc.
>
> On 11/28/14 1:22 PM, Paul Pathiakis via freebsd-questions wrote:
>> "Simulates" utility power. This is the sneaky little bit of
>> marketing. Saying it's sine output.... It is, but it's simulated.
>> I've been roped in by the APC marketing machine previously.
>
> As far as "simulated" being a bad thing, by definition all UPS's have
> to simulate a sine wave once the AC power goes away. It is just a
> question of how close an approximation is it. There's no question
> that a 60 Hz square wave would be a poor approximation.
>
> As the number of steps increases, the magnitude of the "high frequency
> components" and "under- and over-voltage events" becomes so small as
> to be negligible. There are plenty of "server-grade" UPS's that
> produce a step approximated simulated sine wave as the Controlled
> Power whitepaper states.
>
> One aspect to consider is that if a UPS isn't an on-line design, any
> negative effects experienced by the load due to the quality of the
> output sine wave will only occur when the output is being produced by
> the UPS during a power failure. It's reasonable to expect that this
> is going to be a short interval which in many cases may not cause any
> power supply problems.
>
> -
> John J.
> _______________________________________________
> freebsd-questions at freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> To unsubscribe, send any mail to
> "freebsd-questions-unsubscribe at freebsd.org"
More information about the freebsd-questions
mailing list