UPS (Uninteruptable Power Supply) Question

[rboerdijk]

so I'm looking for an Uninteruptable Power Supply.

I'm especially interested in the mustek powermust ups's (because
of the price around 100-140 euro), more specifically the 1400 or
2000AV - but what does the 'AV' mean?

How much do I actually need?

I want to connect two computers to it, one with a 450W power-supply
and one with a 350W power supply. Furthermore two tft's will be
connected.

Furthermore, many ups-devices mention they can be used
with telephone/modem... I assume this is not the standard network
cable which goes into my network-card?

Thanks for any information!

Regards

 

[fishmahn]

AV? On my UPS's I see VA, which stands for Volt-Amps. But I'm in the U.S., so maybe they use Amps first over there where you are...

How much runtime do you need when power goes out? Usually you need just 5-10 min so you can shut your systems down properly. If you need more, then more battery is needed...

Next you need to find out the load in watts for each device. That isn't necessarily the PSU rating, although you could use that as long as the systems aren't heavily loaded with devices.

Then add a couple amps (240w or 480w depending on your area's voltage - since you're mentioning euro's, I'm guessing you have 240v) for each TFT (My 17" TFT says 2a).

Anyways, lets say you end up with 1000 watts. The UPS information should tell you an estimated runtime at partial or full load, something like this:

Partial Load: (480w) 17min.
Full load: (960w) 7min.

So, you do the math to see if that UPS has enough for your needs, or you need to go up/down in size.

Regarding Telephone/modem, you are correct - its surge supression for your old-style dial-up modem.

Mike.

 

[rboerdijk]

Ok, so assuming the following setup

Tft #1 says on the external power-supply
- input: 100-240Vac, 1.2A, 50-60Hz
- output: 12Vdc, 4.2A, 50W (max)
Tft #2 says 1.2A (internal power supply, doesn't mention anything else)

Computer #1 PSU has the following text:
200-200Vac 6A 50Hz (actually the PSU has a gfx with much more digits
on it), see http://members.home.nl/gnarly/gfx/PSU_400W_50proc.jpg
The computer has 4 harddisks and 3 cd/dvd-writers, gf6800 (so pretty
much loaded).

Computer #2 - assume it's identical to computer #1

Which UPS would be suited for that:

1. http://www.apc.com/resource/include/techspec_index.cfm?base_sku=BR1000I with 1000VA/600W
2. http://www.mustek.de/ger_/html/produkte/ups1400usb.htm with 1400VA/850W

And how is it with energy usage of the UPS itself - if it's only 600W
instead of 850W, will it then use less energy? Or does the UPS only
'use' what the connected devices need?

Regards

 

[Guest]

It sure looks like that power supply you posted wants 240V*6A=1440W input. That seems awfully high and I'm guessing it isn't continuous or there would be some serious heat issues with that thing. Also it would have to be under absolute maximum loading conditions, which who knows where you are actually sitting. It is most likely due to the inrush requirements or this so called "PFC" deal they have going on plus max loading conditions.

If you wantr to play it by the book you pretty much have to get a UPS that'll do the 1440W*2 = 2880 plus what your 2 TFT's need 2(240*1.2)=576W for a grand total of 2880+576=3456W. Note this is not VA's.

As I said this seems ridiculously high to me and you may be able to get away with much less especially if you power things up serialy rather than simultaneously. Honestly I think you would be just fine with something in the 1600W range if you power each thing up serialy with 5-10 seconds between starting each device. Go much lower than about 1600W and I predict that you will have trouble with things especially as the UPS batteries start degrading.

The problem you are going to face here is that unfortunately the power supply doesn't mention if the 6A off the mains is continuous, start, peak or what not to mention the UPS guys don't give any mention of what the UPS peak rating and it's peak sustaining abilities are either. So as I mentioned if you want to play it strictly by the book with ammo to use on one of the manufacturers in case something isn't working you pretty much have to go with rated which is the 3456W number.

 

[fishmahn]

I think Knewton is taking the max possible draw of the PSU and that's not what we're looking for. We're looking for average draw.

I would take the watt rating of the PSU (350w, 435w, whatever that is) as the estimated draw, knowing draw is usually below that. That allows for a system loaded to 75% of max PSU, with 75% power efficiency. (400w x 75% load = 300w x 75% efficiency (1.33) = 400w actual draw)

So, each PC draws up to 400w during normal operation using your 400w PSU and the above guesstimated formula. Each TFT draws about 66w (50w max output x 75% efficiency = 66.5) Assuming since the Amp rating on both is the same that the max output will be the same.

That totals up to 933w of expected draw. What that means is that #2 may give you enough juice (pretty close to our estimated draw and I estimated high for a safety factor), but I doubt #1 will.

I doubt you'll get 25 min like the website said at full draw - maybe 12. If 12min is enough, then that's a good choice.

I just remembered some power consumption reviews of AMD & Intel PCs. Intel Pentium4E (prescott) used about 225w (varies of course so this is very rough) with 1 HDD and a 6800-type video card, and you add 12w per extra HDD you get about 260. Times 75% efficiency of the PSU = 345. If we use that as our consumption, then its just over 800 (345+345+66+66=822), just inside #1's specs.

There you go. In my opinion, #2 isn't powerful enough to handle the load of both PCs (though it may do it, runtime will go down very fast, so you will not get the 6min they show and I think that's too short for safety. #1 will be fine, and you should get over 10 min to shut everything down.

Once the battery is charged, a UPS uses minimal power itself (some for the lights and internal circuitry - a few watts - that's about it).

Just out of curiosity, since my German is rusty (LOL I took 1 year of conversational German in 9th grade - 30 yrs ago...) On the Mustek page, one of the words Google couldn't translate was Akkubetrieb. Its in reference to battery run time in the chart near the bottom of the page. Could you give me a translation or breakdown of its root words? Thanks.

Mike.

 

[Guest]

You are wrong when you say:

I think Knewton is taking the max possible draw of the PSU

If you don't understand where the numbers I gave came from just ask. Please don't assume anything about what I say as it is clear you don't understand.

You are giving bad advise to people about items which cost several hundred dollars to purchase. This wouldn't bother me except that on top of this you are telling people that advise which is sound, namely my advise which I will stake my proffesional reputation on, is not good advise. Please keep your opinions to yourself about things you are clearly not qualified to comment on.

While it seems you are somewhat knowledgable in this area there are a few critical concepts here which you seem to be overlooking.

For your benefit:

Do you think the power supply manufacturer was joking when they put 6A at 240V, or that the TFT manufacturer was joking when they said 1.2A @240V?? Note that both of these powers add up to way more than the continuous draw rating of each device. That is because they are accounting for inrush current commonly called startup current. It is what happens when a power supply switch is turned on and the capacitors are charging. An engineer knows that a capacitor will look very nearly like a short circuit to ground and will require imense "inrush currents" to charge up to working voltage.

These are numbers given by the manufacturer which include inefficiencies, worst case loading, and inrush current. The numbers are an indication of what they feel is a safe condition to gaurantee proper operation. I notice you made no account for inrush current in your calculations. This is where you went wrong in your calculations. I can tell you from experience as a professional electrical engineer working in the industry who is paid to make these decisions every day that this "inrush current" must be accounted for if you expect your system to function.

Furthermore you go on to say that a UPS which is rated for lower than your own continuous draw calculations is sufficient. WTF??????

Now go back and read my post again. You stand to learn something. If you have any questions I will answer them. BTW if you don't believe what I am saying then go ahead and give the UPS manufacturer a call. I think you will find that if anything the 1600W rating I said you could probably get away with will be the bare minimum they will recomend.

 

[fishmahn]

I don't have time to explain this to you so you'll understand, so I'll keep it short. Especially since I don't waste time with people who take offense when someone mentions their OPINION (I did say 'I think' didn't I? 8O ).

Then again... what you are saying is that ANY current-generation single PC could never run on either of those UPSs because by your calculations just to run 1 computer you need BOTH of those UPSs in parallel to make your 1440w, and don't plug a monitor into that - you'll overload it.

Now, if you take a half a second and look at that statement, you may see how ridiculous it is. Then again, maybe not.

But, since you're an engineer, and I'm only someone who's been using UPSs in production environments for 15+ years (including dedicated graphics workstations - and I don't mean Macs), I'll bow out and let you waste this man's money.

Mike.

PS: How long does inrush current last? (btw - I'm not so stupid that I didn't understand you). Ooooh, half second? Wow! That'll pop the circuit breaker easily...

 

[Guest]

The only part I took offense to was when you started predicting what my thinking was. That isn't for you to do as I am capable of speaking for myself.

I am confident that what I have said is completely reasonable given what is known about this problem. You are the one asking questions here. If you don't understand and don't want to reread what I have said then why don't you call the manufacturer. What do you have to lose? If it turns out you find information which contradicts what I am saying it would give you something useful to say to me in your next response.

I don't have time to explain this to you so you'll understand

Don't bother, I understand what you wrote. It was bad advise. I understand how to rate a power supply.

Then again... what you are saying is that ANY current-generation single PC could never run on either of those UPSs because by your calculations just to run 1 computer you need BOTH of those UPSs in parallel to make your 1440w, and don't plug a monitor into that - you'll overload it.

Again, try reading what I said again. That isn't what I said.

Now, if you take a half a second and look at that statement, you may see how ridiculous it is. Then again, maybe not.

Yeah that misinerpretation of what I can only assume is your understanding of what I said is pretty ridiculous. I would never say something like that.

Somehow I not seeing how being a UPS user for 15 years gives you any qualifications for spec'ing them in. I note that you don't mention that you have ever spec'd any in on a professional level only that you've used them.

How long does inrush current last?

That depends on the size of the capacitor the exact representation of the voltage applied with respect to time and the parasitics present in and around it. Different equipment will have different inrush profiles and the inrush profile will vary among even the same piece of equipment depending on at which point on the input voltage waveform the power grids are at when the switch is thrown. Fortunately since we live in an imperfect world the parasitics keep the inrush current to less than infinite. Without knowing the exact details of the circuits present I (or anybody) can only say that it will be somewhere between zero and infinity. That is why I think the best idea is to keep what the equipment manufacturer says about it in mind, and thus my suggestion.

Ooooh, half second? Wow!

Regardless of how long it lasts, the UPS better be able to handle the requirement. The technical explanation for why this is true is that a gross impedance mismatch can create conditions likely to cause instability in the power supplies switcher regulators or overheating in the power supplies switcher regulators.

Instable regulators can cause PC damage or a never start mode. Too much heat can cause the regulators to fail.

Opening breakers is another subject and obviously the wall circuit better be sufficient, but that is OT for this discussion.

 

[Fastboatslowcomp]

I notice you made no account for inrush current in your calculations. This is where you went wrong in your calculations. I can tell you from experience as a professional electrical engineer working in the industry who is paid to make these decisions every day that this "inrush current" must be accounted for if you expect your system to function.

Just for those of us looking to learn from these wonderfully spirited discussions, why would you account for inrush current in the calculations for sizing the UPS? Unless you are trying to start up during a power outage you are not taxing the UPS because you are using line voltage/current. The UPS battery kicks in when the line voltage drops. If you are considering powering up the system using the UPS battery rather than powering down the system after power failure then I understand your accounting for inrush current. Is that common practice for sizing a UPS?

 

[Guest]

you are using line voltage/current.

I don't believe this is true. This is how I understand the flow: The mains power is converted to DC which charges the battery. Another circuit takes power from the battery to create a second sine wave or pseudo sin wave as the case may be. The secondary sine wave is what the PC gets its power from. In this way the PC is isolated from tha mains and the baddies present there, and is always running off the UPS power.

Believing this explanation leads to the answers to your questions.

Is that common practice for sizing a UPS?

It is common practice for sizing any supply which will be supplying startup power.

 

[fishmahn]

you are using line voltage/current.

I don't believe this is true. This is how I understand the flow: The mains power is converted to DC which charges the battery. Another circuit takes power from the battery to create a second sine wave or pseudo sin wave as the case may be. The secondary sine wave is what the PC gets its power from. In this way the PC is isolated from tha mains and the baddies present there, and is always running off the UPS power.

Believing this explanation leads to the answers to your questions.
One problem. That's a continuous flow UPS (whatever the exact term is - I don't care, every mfg had different terminology) These don't use that method because PC's use switching power supplies, They bypass the battery until the power cuts out, then within 3 ms (according to the website he linked, IIRC) the battery cuts in. That is plenty of time for the PSU to maintain stable power. Since you aren't going to be starting up the system on batteries (at least, only a stupid person would - not that stupid people don't come here, but they don't do homework like the OP appears to have done), you don't need startup current capability. Thus the demand it much lower.

Is that common practice for sizing a UPS?

It is common practice for sizing any supply which will be supplying startup power.

See above.

Umm, now if you reread my statement that upset you and take it from a 'layman's' perspective, you may find that it was pretty accurate. You don't have to take the max possible draw of the PSU (which is exactly what you were doing - 6Amps X 240Volts, so what are you complaining about :?: ). bad wording here, upset should be offended and max rated draw would be a little better than possible draw. 😳

Mike.

PS: Edits are to correct some grammer - got ahead of myself and left out parts of sentences - edits italicized for clarity.

 

[Guest]

They bypass the battery until the power cuts out, then within 3 ms

By god you are right. I should have paid closer attention. I admit that was a major flub and explains why you got so bent. Hopefully you can accept my apology for that. When I read back through this thread I am feeling like an ass. I never even clearly spelled out my incorrect assumption before puting my foot in my mouth.

As I mentioned early on I understood what you said the first time I read it. It is only now in hindsight do I see that it was indeed an acceptable analysis.

I still don't get why you didn't point this out earlier. It seems like this final reply is what would have made sense in the beginning when I was going on about inrush current. After all everything I said is true if my assumption about the type of UPS's these were was good. Why would I concede to your way of thinking when I hadn't yet realized I had made an incorrect assumption?

BTW why would you say this?

These don't use that method because PC's use switching power supplies

That just isn't true.

I'm starting to think fastboatslowcomp knew what was going on all along and was really just a devils advocate in disguise.

 

[fishmahn]

Hopefully you can accept my apology for that.

Absolutely.

I still don't get why you didn't point this out earlier.

Like I said in my first reply to you, I didn't have time. By the time I got back to it in the late eve., Fastboatslowcomp had beat me to it, so I could take the short method instead of having to explain both types of UPSs and why you want either in different situations. You sounded like an engineer who had no idea about the realities of UPSs. I apologize for assuming you were 'one of them'.

BTW why would you say this?

These don't use that method because PC's use switching power supplies

That just isn't true.
Ok, I think you got me there. Why isn't it true? :?:

While a continuous flow is better (that's what we used on the Scitex, Hell, and other graphics equipment in color seperations - those things were huge!), a PC's switching PSU means you don't need continuous current as long as switchover is fast enough. Maybe its not cause & effect like my statement implies (is that what you meant?), but its close enough for layman's work. 😀 More accurate would be they don't have to use that method...

I'm starting to think fastboatslowcomp knew what was going on all along and was really just a devils advocate in disguise.

Mike.

 

[Guest]

having to explain both types of UPSs

I guess I wasn't clear. There hasn't been a time during this post or any time in the last couple of decades for that matter when I didn't know what the differences were, it is that I hastily jumped to the wrong conclusions about architecture in this case. In other words I didn't look very closely at the data sheets at first. My bad.

I fell into the trap of the whole problem with making assumptions, and look what happened. I guess I get into the mindset that: obviously everyone will do things my way when there are still those who don't buy the best equipment regardless of cost. Clearly this is a problem in my way of thinking.

I didn't have time

The part that I guess I don't understand is that the explanation was one line and the part that wasn't took multiple paragraphs. I don't see how that was easier. When you said "They bypass the battery until the power cuts out, then within 3 ms " it was all that was necessary for me to suddenly see the light and know exactly what was going on.

Why isn't it true?

It comes down to the power supply ideally not caring or knowing the difference between power coming from a battery backed supply or the power grid. This will be true in a properly sized unit. This of course assumes equal power quality, which is not always the case with these pseudo sine wave output UPS's. As you have pointed out there is a certain non ideal aspect of the switchover type in that they take a few mS to make the transition. Not a problem if the switchover time is kept small enough. There is also the trade off that often times they are undersized for reasons of cost so they can't handle startup current, because of this they aren't burly enough to offer the more ideal protection from grid transients offered by the power flow architecture I have been talking about. For example these switch over type UPS's that I'm now aware we are talking about are able to succesfully power the PC when the grid dies and the PC power supplies are fine with the power. If there is bit more oomf such that they can handle the system startup current there is no reason not to move to the superior protection offered by isolating from the grid and use the type that I had made the assumption we were discussing.

 

[fishmahn]

I guess I wasn't clear.

You were clear enough, I just originally suffered from a similar assumption problem. I assumed *back then* you needed a thesis on UPS technologies, which would have meant me either blathering out BS as I gathered my thoughts because its not something I deal with on a daily basis (and there's too much other stuff rolling around in my head) or spending lots of time looking up the details again so I could explain it technically and accurately. Obviously that's not the case, and its resolved in my mind.

That would also explain why I said I didn't have time - If I would have made the connection that you didn't need the long version, I could have made it the short version and saved lots of misunderstanding between us.

Most if not all the switch-over UPSs I know of offer surge protection as well to protect against variations coming from the grid. Don't remember specifically on the Mustek's the OP asked about but the 2 Tripp-Lite switch-overs I just bought for the office have surge protection (and that $100k or something warranty :roll: ).

Anyways, all's good, and it seems we're both on the same page as far as UPSs. You know it from the technical side much better than I ever could, but I still have enough experience and have done enough research on it in the past to know it from the user's side plenty well to make informed decisions.

Mike.

 

[Guest]

OK cool I'm glad that's over with. Although it never hurts to have someone point out the error of my ways to keep me on my toes.

At $100K Tripp Lite certainly must have some confidence in their surge suppression hardware. I think that is even better insurance than that offered by my old UPS which is of the type you have been calling continuous flow.

 

[RichPLS]

CyberPower 1100

I see this one as being good bang for the buck, and all around good choice.[/url]