It's been awhile since I looked at calculating electrical power with Ohm's Law. I believe, and could be wrong, but the problem is compairing 120v AC power supplied by UPS to the 5/12v DC supplied by the computer PS.The UPS power load display is what the computer PS is using to operate. That would be different from what the computer is demanding from the PS.If you want to determine how much load is on the PS, you would need a load meter on the PS not the UPS.Maybe someone current in Ohm's Law can better explain. From what I recall, power (watts) at 12 volts at a given amperage would be different than 120 volts at the same amperage.P=IxE or Power/Wattage (P) equals current/amperage (I) times voltage (E).120 volts times 20 amps = 2400 watts12 volts times 20 ampes = 240 wattsHope I didn't mess this up.Regards,Richard

>It's been awhile since I looked at calculating electrical>power with Ohm's Law. >>I believe, and could be wrong, but the problem is compairing>120v AC power supplied by UPS to the 5/12v DC supplied by the>computer PS.This sounds believeable Richard. Thanks for the insight. I DOES however, suggest a 1,500W UPS was . . . overkill :()

>My guess is that you wouldn't survive with the 4870 X2.Acknowledged! :-)Gary

>I

>It's been awhile since I looked at calculating electrical>power with Ohm's Law. >>I believe, and could be wrong, but the problem is compairing>120v AC power supplied by UPS to the 5/12v DC supplied by the>computer PS.>>The UPS power load display is what the computer PS is using to>operate. That would be different from what the computer is>demanding from the PS.>>If you want to determine how much load is on the PS, you would>need a load meter on the PS not the UPS.>>Maybe someone current in Ohm's Law can better explain. >>From what I recall, power (watts) at 12 volts at a given>amperage would be different than 120 volts at the same>amperage.>>P=IxE or Power/Wattage (P) equals current/amperage (I) times>voltage (E).>>120 volts times 20 amps = 2400 watts>>12 volts times 20 ampes = 240 watts>>Hope I didn't mess this up.>>Regards,>>RichardWatts are the true power ratings. So when you put a watt meter on the 12v hardware, it's measuring watts. On the 12v side, the amperage is increased to compensate for the low voltage, and it's reporting in watts. The UPS' watt meter is measuring how much power the PC draws, thru it's powersupply. I can imagine, due to various resistances involved between the components, there would be a slightly inexact representation as one over the other. I'm also seeing how the demand source (the pc CPU/GPU and other power users) has to *suck* hard, to pull in the requisite power. Therefore, the demand side will be HIGHER (in watt demand), while the source power can simply push along what is needed, which will display as less wattage. I wonder what the conversion factor would be? Have to tinker with that idea . . .NoelOr, inherent inefficiencies in the conductors involved will widen this disparity.QX9650 w/ Retail HSF|ASUS P5E3 Premium WiFi|4GB Muskin Ascent 7-6-6-18 1T DDR3-1600|EVGA 8800GT|Seagate SATA 2 x 2|Seagate Cheetah 15K.x|XP Pro SP2|Vista 64--maybe never to be installed

Thats because you are looking at the inverted load being supplied to the AC side of the PSU and NOT the DC demand the PSU is delivering to the internal systemHUGE difference... If I had time I would post the formulas for that, look them up on the net... check into VA and conversionsAnyone running less than 600watts on a quad core and overclocking is pushing their luck even with a mid-grade video card and that also depends on the PSU maker. A PC Power and Cooling PSU delivers where most would be lucky to deliver 65-75% of the advertised ability. I dont know how many times I have solved performance problems where no crash was seen, only poor performance by replacing the PSU with the right unit. Usually the system temps dropped significantly at the same time.EDIT: I see Richard already stepped in and somewhat corrected this complete misconception of watt draw. AC input and DC output are 2 different animals. Then there are the rails and how they are designed which also factors in. One would be glad they have that 750watt PCP&C unit after calcualting everything correctly because then HEAT and overhead current also play into a system for stability and performance.Goes hand in hand with many other misconceptions I still see posted around hereThe minimum UPS I would personally spec for a clocked quad core system and a 24inch LCD monitor with several hard drives and a decent video card would provide 658 W @ 1096 VA

There are raw formulas to estimate VA and convert it to power however there is much more to testing for efficiency between input and output draw than simply guessing and using basic formulas or even using meters at home. A bench test unit designed specifically for that purpose is required and PCP&C uses such a system when rating their PSU's and checking them for QCThe EFF rating is based on the full load ability of the PSU, therefore the ability is variable based on the load and HEATη (EFF) = Σ Po,ii------- X 100P inThat 750watt PSU is worth its weight in gold for a cool and stable system, and it overts future upgrade costsI do agree with a dual core processor and a typical mid range video card on 2GB of memory, 750watts is very overkill. I would still not run any system today with 3D hardware/use on less than a 500-600 TRUE watt PSU. Most el'cheap-o crap sold to be 500-600watts only delivers 65-75% of that

>There are raw formulas to estimate VA and convert it to power>however there is much more to testing for efficiency between>input and output draw than simply guessing and using basic>formulas or even using meters at home. A bench test unit>designed specifically for that purpose is required and PCP&C>uses such a system when rating their PSU's and checking them>for QC>>The EFF rating is based on the full load ability of the PSU,>therefore the ability is variable based on the load and HEAT>>η (EFF) = >>Σ Po,i>i>------- X 100>P in>>>>That 750watt PSU is worth its weight in gold for a cool and>stable system, and it overts future upgrade costs>>I do agree with a dual core processor and a typical mid range>video card on 2GB of memory, 750watts is very overkill. I>would still not run any system today with 3D hardware/use on>less than a 500-600 TRUE watt PSU. Most el'cheap-o crap sold>to be 500-600watts only delivers 65-75% of that Nick I am going to install a 4870 or 4870x2 I'm pretty sure. I'm thinking I'm still ok with that, and I base it on this utility as an estimator: http://www.extreme.outervision.com/PSUEngineThey have all my parts represented except the 4870x2. I used the 3870x2, and it comes out to 620W PS rating. I'm guessing I'll be ok still with the 750W from PC P&C. You think?

If its a PCP&C unit... yes, you will be fineThe trick with PSU's past their "Watt" and "Amp" rating game and "Multiple Rail" game is the Efficiency ratingSince that unit as I recall has a 82-85% EFF rating you are definitely good to go. It will actually peak at around 800-840 if I remember the specs on that one correctly. Its not duty cycle rated for that 800+watt load but it will peak under high stress without failureYou get what you pay forThe older PCP&C 750 does not come with the 6+8 pin ATI PCIe plugs but their newer version does. You should get the adapter with the x2 card. ..and do make sure they are both connected.That online calculator tends to display power needs by a 15-20% overhead margin. Make sure you did not miss anything in its list because it can thow off the estimate very easily the way that page is designed. It is a decent estimate since it does take into account overhead but they do not post that.On my systems, system A show a 748W PSU is needed and system B requires 925. Both of those are about 20% higher (including a 5% margin of overhead) than I actually need but I do use a 750 in system A and a 1000w unit in system B none the less. You never know when a upgrade may push that over the top and you always want to keep at least 10-15% in reserve for overhead. The system runs much cooler that way.I skip the capacitor age estimate on that page because the PCP&C units I use have duty rated caps for the life of the unit.you get what you pay for;-)

Thank you for your thoughts Nick. I have a nice UPS now too. Any predictions on whether ATI 4870 cards will do better in FSX with driver revisions, or do you think this is unlikely?QX9650 w/ Retail HSF|ASUS P5E3 Premium WiFi|4GB Muskin Ascent 7-6-6-18 1T DDR3-1600|EVGA 8800GT|Seagate SATA 2 x 2|Seagate Cheetah 15K.x|XP Pro SP2|Vista 64--maybe never to be installed

The 4870 is the best thing ATi has going for them and I do think it is a precursor to better engineering from ATi. Its the only ATi card I would consider over the 8800GTX/Ultra for FSXThere are some problems with their BIOS and and fan control however I am sure they will release a patch for that and in the mean time fan control is located in a new XML format so its easy to fix manually.As with any card they usually get better with driver updates. I wont be upgrading again until closer to mid Q4. I would prefer to upgrade as a unit than piece-meal it over time. Usually and its been the practice for a few years now, they will release upgrades after Xmas to play the 'come and get something better' game to try and score a double on the xmas season so i will be watching closer to November to see what may be getting released in January too before making any final decisions

>I wont be upgrading again until closer to mid Q4. I would>prefer to upgrade as a unit than piece-meal it over time.I do also, to a degree. I just figure this platform could handle considerably more GPU and plus I like ATI's IQ better. And I won't be doing the quantum upgrade for several years.

Nv and ATi have been much closer to each other on IQ in the last year however I do agree overall ATi still manages to squeak past Nv on the IQ. The image tends to have a much better feel to it which screenshots converted to JPEG can not capture. Converting images from the sim/BMP to JEPG does degrade the image a bit even if compression is not used and what most people look at are screenshots which convinces them there is no difference.. there is an IQ difference but by not leaps and bounds and it's best seen in-game and in real time.

>Nv and ATi have been much closer to each other on IQ in the>last year however I do agree overall ATi still manages to>squeak past Nv on the IQ.You know I think in terms of what the driver contributes to IQ, I would agree, they are not very far apart, if at all. What I'm talking about has to do with a hardware component somewhere on the card I'm assuming. Just take a look at your POST screen on an LCD between nV and ATI. This sharpness in IQ is present before the windows driver loads, and it translates into the windows environment as well. With nV, I see a certain softness/fuzziness of the POST screen characters, and I see a razor sharp edge with my two ATI cards used in the past. I think this is really what people are refering to when they almost universally say ATI has superior IQ.