Scaling up is a reference to the ideal of making something work the same way, but on a larger scale. Normally, ratios between things remain constant as the scaling takes place. If VA is 25% more than watts on a particular design, then if you pick the 1200 watt scale, you have 1500 VA, and if you pick the 1800 watt scale, you have 2250 VA. Of course, if everything in the design is scaled up the same, that means everything including battery, wiring, etc. The controlling CPU probably would not need to be scaled up. In real life, it's probably not exactly that simple, especially if the chosen scales don't always match economical components.

The peak current rating the 1200W and 1350W model have would be the same, 12.5 amps for 120 volts. But, since the watts limit cannot be exceed, either, you might not be able to reach 12.5 amps. For example, a load with unity power factor would not be able to go above 10 amps or 11.25 amps depending on which of these is chosen.

One concern I have is efficiency, since it determines what is drawn from the mains. These different designs could be more or less efficient, and even have higher or lower power factor themselves, for what they draw from the mains. Since they are double conversion, I would expect the power factor at the UPSes load to not have a bearing on the power factor the UPS presents to the mains as a load. But, a different design could present a different load, as could a different amount of power being draw across the convert/inverter interface (e.g. the DC loop). These are things I want to know. I was hoping someone that knew the details of such designs might know if, in general, a different VA/watts ratio in the design would imply a different efficiency, or a different power factor as a load to the mains (both could determine the type of mains circuit truly suitable).

BTW, some specs given on some equipment is what I would term "silly". For example, consider a 208 volt 1000 VA 900 watt UPS. Why would it need a NEMA L6-20P plug, even while the 208 volt 3000 VA 2700 watt UPS in the same series only needs a NEMA L6-20P plug? That doesn't even make sense ... except that maybe they decided to use the 20 amp plug on everything to avoid having to buy 15 amp plugs and. Here they are ...

http://powerquality.eaton.com/PW9130G1000T-XL.aspx?CX=3
http://powerquality.eaton.com/PW9130G3000T-XL.aspx?CX=3

Of course. But does the difference in design that allows this difference in power at the same VA rating change other aspects, like efficiency, or the power factor the UPS presents to the upstream mains supply? Designs did change to make UPSes more efficient in recent years. Designs did change to make UPSes have a higher power factor in recent years. Design changes do present constraints on other design aspects. But does anyone know what those are in the case of UPSes with different VA/watts ratios? Apparently no one does.

It seems you both do understand Watts vs. VA, but aren't using the same terms.

There are two ways to multiply here, because AC electricity is a time series. Technically, AC is not just one voltage or one amperage, since it changes over time. The values we typically work with are integrations of absolute values over time. What matters in the case of calculating Watts vs. calculating Volt-Amps is whether we do the multiplication at each instance of time and then integrate (giving us Watts) or integrate both units separately then do the multiplication (giving us Volt-Amps).

Hi,

What is the difference between Voltage-Amps (VA) and watts ... should provide everyone the means to understand VA and watts. Therefore speak with the same terms and understanding.

Be sure to read the whole reference: What is the difference between Voltage-Amps (VA) and watts ...
You will garnish loads of information to add to your understanding.

Power Solution just happens to be helpful to users that have problems which their company can address. I have found that when one needs to communicate or get the facts it is best to speak to product/design engineers not the contact desk or salesman. It is your investment dollars so you should use the power to control the situation.

Disclaimer: I am not associate with or do I represent Power Solution

That's the same link that macemoneta gave in post #2.

I would say that it is obviously NOT the case that the watt rating is 60% of the VA rating. The one I saw from Eaton (1500 VA 1350 W) has a watts rating at 90% of the VA rating. The heater example in the linked page would work in this case if it were scaled down to 1000 VA since that would be 900 W. This page looks to me like it is a bit dated, maybe by 20 years. The other two I saw were 80% and 70% respective.

I suspect it is possible to design a UPS to handle even extremely low power factors. Maybe load them up with capacitors? But I suspect this increases cost. These days it seems PC SMPSUs have better power factors. Why make a UPS go overboard on VA if it increases costs and increases the circuit requirements.

Yes, I would agree speaking to the design engineer would be best. That seems to not be an option with companies like APC, Eaton, Emerson/Liebert, and Tripplite.

If the critical parts of this could become off the shelf components, maybe we can solve some of the various issues in data centers that current vendors seem to be loath to do.

http://opencompute.org/

The issues (solvable) I see with this are:

1. Lack of VNC remote control access (things like IPMI seem to be a big failure for open computing).
2. Does not (yet) scale DOWN very well (e.g. how about a single cabinet, batteries at the bottom).
3. Assumes generator as the secondary power. Facebook obviously needs that. Not everyone does.

ARRRGGGGHHHHHH!!!!

Hits head on wall.

That was my thought. My question was never answered. Not that I should have expected there to be a UPS design engineer here. The public readership probably got a decent dose of VA vs W. But at this point all I know about the 1350/1500 model compared to the 1200/1500 model or the 1050/1500 model is that it can handle more power if the load is closer to unity. But I knew that before this thread.

BTW, Watts and Volt-Amps are not the only limits. In certain circumstances Power Factor itself can be a limit (in the sense of not going too low) even if the load is low. And this can be at different points depending on the load's mix of harmonic current and reactive current. So get those PFC PSUs in your computers.

From my personal experience, companies list VA ratings only because they are higher than Watt ratings, and thus the consumer is more likely to buy it not knowing that the number doesn't really mean anything. Sure you can use it to estimate how long it might last, but personally I haven't found these accurate. I completely ignore the VA and ask for the wattage. Knowing the Watts that it provides and how many watts you need, as long as it provides more than you need, it will work. The more it provides the more it will last. Oh, and the other reason VA is there is to remind you that you can't chain-link UPSs and expect magic.

Skaperen,

I like you and your posts, so don't get me wrong. I welcome good and bad questions. They make me think.

The reason I posted that was, I get the feeling people have posted a reply to your question. I keep looking at it and I can't figure out where the difference in what you want and what was offered is lacking.

Now you get into power factors, this is a more complex issue. Yes, most AC is calculated with a notion that a simple ac motor is running. That the load is constant and never changes. The case of many VFD drives or some poorly or even good designed electronic devices is non-linear loads. For example we have thousands of VFD's all over running loads. As the systems start at stop and change speeds they tend to send back harmonics or can cause harmonics to appear on the line. Now what was 60 hz is a combination of 60 hz and an untold number of other frequencies. This is play havoc with traditional meters and load sensing devices. This is unlikely to be a condition for the design of any consumer device still. They still are usually designed for 50 to 60 hz and no harmonics. We use a special fluke meter to monitor these harmonics and have had a few issues. We have found very well made devices fail from what we believe is this harmonic issue. Now we have added in all sorts of line conditioning in places. Our first attempts have turned into 4 or 5 as the loads proved more than expected.

All of that has nothing to do with how a company rates VA. They can put any number they want on based on their own reasons for the most part. Don't use VA unless you consult the UPS company for more details. Use Watts and consider that Watts is the rating at FULL battery quality and again at RMS not true RMS.

1 members found this post helpful.

The first round of answers I got was was trying to explain to me what Watts vs. Volt-Amps is, and that I should be sure my load parameters, expressed in Volts, Amps, and Watts, needs to not exceed all of these parameters. These are the common parameters. And it is the case that lots of people just don't understand the difference. Some understand the important and still don't understand what or why. I understand it all, so all these explanations were not what I was looking for.

They should not (and I don't believe engineers would) just pick arbitrary numbers. Sales people? Well, let's just say I don't trust them.

The actual designs of UPSes vary. There are basic principles for the double conversion type I mostly look for. But they vary in many ways. The reasons they choose these designs can be varied. One is the desire to meet specific product targets. They want a unit that can deliver certain levels and still meet many other requirements (reactance as a load, harmonic current, efficiency, reliability, cost).

If you have a UPS design that can deliver a specific VA and W level, in general the same design can be rescaled to various levels, within some range. A 1000W 1200VA design can be a 2000W 2400VA design by just doubling most things. But if the goal is to double the VA and not the watts, certain parts can be left not doubled to keep a lower cost. You can get the 2400VA at maybe 1440W in a new design that has a cost somewhere between the 1000W/1200VA design and the 2000W/2400VA scaled up design.

Finding a 1050W/1500VA design from one manufacturer (APC), a 1200W/1500VA design from another manufacturer (Tripplite), and a 1350W/1500VA design from yet another (Eaton, apparently formerly Powerware), I have to think these are not identical designs, even if scaled up or down a bit differently.

My question was about what kinds of other effects and tradeoffs would exist in these different designs (more different than just being scaled up or down in level). Could one design be more efficient than another? Not all the manufacturers state efficiency (I'd vote for a law that mandates stating efficiency). And efficiency is not the only factor if any importance (a big one, for sure, but there are others). Fault current is one, and so far none of them state this. And yet it should be figured in, even though UPSes tend to have lower fault currents than corresponding transformers (of the same VA and W rating). I'd mandate stating this spec if I could mandate it, so I better understand my choices. Other things include how the design handles varying load levels, circuit break trip points and curves, battery recharging rates, and the harmonics and/or reactance it presents to the upstream mains. For example I want to know how much current the UPS draws in so I know if I can use it on a 15A circuit, or a 20A circuit, or a 30A circuit. This can effect the choice of UPS, and my layout (1, 2, or 3 UPSes per cabinet, or maybe where available, three phase). And 30A is not always a usable choice since you can't put 20A outlets on a 30A circuit (you can put 15A outlets on a 20A circuit, and you can plug 15A plugs into a 20A outlet if it's not one of those special ones required for some cases in Canada).

There really is a whole lot to consider, and sadly, most people never do, and have a suboptimal, possibly even dangerous, situation as a result.

Maybe an answer might have been "those UPS designs where the watts is a closer to the VA, are more efficient because ..." or "those UPS designs where the watts is a closer to the VA, present more harmonics upstream because ...". I don't know what the specific answers might be (if I did, then I would not need to ask). Things like that can be important. For example if one design has other benefits, but has more triplen harmonics, I might prefer the 208V/240V L-L connected models over the 120V or 230V L-N connected models.

If that doesn't explain what it is I was looking for, then I guess I will just not be able to.

Hi,
Missed the earlier reference. Not to date the link but I've had it for a long time.
I believe their(Power Solution) example is valid within their context. Still valid for the approach. Differences would be newer physical components and design circuitry. Theoretical rules have not changed but some are modified within the circuits design for the loads & conditions. Just like other competitors, design criteria revelation would be limited so as to prevent industrial design theft. Sure, a good engineer or tech should be able to reverse engineer but obvious theft of critical designs would win in court suits.
Yes, doable when needed but at greater cost and environmental situations will dictate the design. Every manufacture must be FCC certified for emissions within limits for use in the USA. Consumer electronics is unlike commercial since the environmental conditions are not the same. Do not forget that most large commercial units are designed to run using 3 phase systems. Where as consumer grade will be single phase balanced input (3 wire, single-phase, midpoint neutral system). So the cost in design and circuitry would be simpler for a 3 wire consumer design unit thus easier to reverse engineer.
That is one of the reasons to communicate with a third party consultant(s) for larger designs.

Engineering Academia opened doors for me that most would not be able to open. Trade shows help develop contacts/networking with design engineers. Sure some of the reps would be a PE with little time in the field but one can use that contact to move up the ladder.

Most designs do have similarities with each having enhancements peculiar to their manufacture design. Just so many ways to cook things!

I don't go for the big giant UPS that serves a whole room. I've had to deal with those in 4 places I worked. The first was for mainframes, so no real choice. But in all cases, the big room sized UPSes were always trouble. That, and various problems not specifically in the UPS were not covered by the UPS, such as circuit trips. I will never put those big ones in. The best I approach I see is 1 to 3 UPSes per cabinet.

So some people can get to the design engineers. Not everyone can. Unfortunately, the companies overly shield them to the point where even the sales "engineers" are not allowed to ask the design engineers any questions, and have to just make up something (often plain wrong) when the customer asks a question that goes over their head.

Here's one: A 208V/240V UPS uses/has no neutral, in or out. So what is the voltage RELATIVE TO GROUND? For 208V derived from 2 phases of a 3 phase system, the answer is 120V. But what about in a UPS. The most knowledgeable answer I ever got was someone that said not to use the ground as a substitute neutral. Well, that is correct. But he missed the point of my question and refused to discuss it beyond that (so I suspect he was just reciting stuff he didn't really understand). I only wanted to know the ground reference relationship. One other person said there was only a single inverter and its output was not grounded. So would they really make a floating output? He tried to sell me a $1000 transformer to connect to the output. I didn't want to pay $1000 to cover up a bad UPS design, and eat up 2U of rack space, too.

The Tripplite equivalent must have had two inverters connected in series and grounded in the middle ... because it used/had a neutral in and out. Some specs did suggest the output was 208/120 but could be changed to 240/120, so I suspect they just vary the sinusoidal phase relationship of the two inverters. The catch is I had to provide it with a L-N-L circuit. But all existing circuits were L-L only.

What I want to know from a design engineer is to consider what changes would need to be made to make a UPS provide more or less watts relative volt-amps, or more or less volt-amps relative to watts, and describe how those changes would affect other aspects of the design, such as efficiency.

On observation I've made over the years is that the USA power system just makes the whole industry much more difficult. It requires too many different models because of so many variations on power systems to work with. I'd really want to push for data centers to be powered at 416/240 volts and just feed the 240 L-N straight to the UPSes. It would simplify things because usually one circuit could handle one cabinet with one UPS.

The opencompute.org project just went ahead and specified the 480/277 volt system in their design. Do it in EMEA and you probably have to do it at 400/230, but that's not so bad. BTW, I found their project by googling for a PSU that ran on 277 volts.