Archived from groups: alt.comp.periphs.mainboard.asus (
More info?)
"Paul" <nospam@needed.com> wrote in message
news:nospam-1002051352280001@192.168.1.177...
> In article <371kmgF58bkr0U1@individual.net>, "Arifi Koseoglu"
> <someone@somewhere.com> wrote:
>
>> "Paul" <nospam@needed.com> wrote in message
>> news:nospam-1002050806470001@192.168.1.177...
>> > In article <1107996840.065613.250950@c13g2000cwb.googlegroups.com>,
>> > "arifi" <arifi@turk.net> wrote:
>> >
>> >> Hello everyone,
>> >>
>> >> At work we wish to build a small-footprint demo computer using one of
>> >> those thin (10cm / 4 inches high) desktop cases, which will be easy to
>> >> carry to customer's sites, and such. We prefer this way rather than
>> >> buying a strong laptop since we believe such a system will perform
>> >> better and be more flexible.
>> >>
>> >> The case we have found has a power suply capable of producing 230
>> >> Watts, and they claim it easily drives P4 systems. The case can house
>> >> microATX size MoBos. The power supply is a thin and long one custom
>> >> for
>> >> this case, so there is no easy way to upgrade the PSU.
>> >>
>> >> Being primarily an ASUS house, we searched for "microatx" and "matx"
>> >> on
>> >> their web site, and the following motherboards were listed:
>> >>
>> >> P5GD1-VM
>> >> P4GE-MX
>> >> P4R800-VM
>> >> P4P800-VM
>> >> P4BP-MX
>> >> P4S800-MX
>> >> P4SP-MX
>> >> P4SP-MX SE
>> >> P4V533-MX
>> >> P4VP-MX
>> >> A7N8X-VM
>> >> A7N8X-VM/400
>> >> P5S800-VM
>> >> K8S-MX
>> >>
>> >> I would assume all to be nice boards, but we have not been able to
>> >> find
>> >> conclusive information about how much Wattage the complete system
>> >> built
>> >> around one ot them would require.
>> >>
>> >> The system will contain the Motherboard, 2x512MB DDR RAM, 1x P4 CPU,
>> >> 1x
>> >> SATA drive, CDROM Player, cabled keyboard and cabled mouse. No other
>> >> attachments except for a monitor or a projection device. The amount of
>> >> computing power we need is at the order of a P4-2800 with 1Meg Cache,
>> >> but we do know that we will have to settle with the max. that the
>> >> power
>> >> supply can drive.
>> >>
>> >> There are Power Supply calculators out there, but the ones I have
>> >> looked into do not consider all-in-one boards like the ones above
>> >> where
>> >> the graphics card, lan etc are all on the motherboard.
>> >>
>> >> So the Question is: How can I find out how much power these
>> >> motherboards will draw? What is the strongest CPU/MoBo combination I
>> >> can use with the 230W PS we have? (he system will be run at rated
>> >> speeds - no overclocking)
>> >>
>> >> I will appreciate any information and any pointers deeply.
>> >>
>> >> Many thanks in advance,
>> >> -arifi
>> >
>> > For power estimation purposes, you could go to the Intel motherboard
>> > web page, and look at some boards there. Each board has a manual,
>> > and there is a power estimate section.
>> >
>> >
http://developer.intel.com/design/motherbd/products.htm
>> >
>> > This is the manual for the D915GAG (microATX builtin graphics)
>> > These numbers are maximums, in every sense of the word.
>> > I use these numbers only to get some idea of what the
>> > non-CPU rails will be drawing.
>> >
ftp://download.intel.com/design/motherbd/ag/C6860002.pdf
>> >
>> > Mode DC Power +3.3V +5V +12V -12V +5VSB
>> > Minimum loading 200.00W 3.30A 10.00A 900mA 0.03A 0.80A
>> > Maximum loading 300.00W 6.00A 14.00A 16.00A 0.10A 1.40A
>> >
>> > Based on their description, start with the minimum loading
>> > spec. Your two DIMMs will add about 5W each. The +12V
>> > draw, exclusive of the processor, would be for fans.
>> > A Prescott P4 2.8/FSB800/1MB LGA775 has a TDP of 84 watts
>> > (see page 74 table 5-1).
>> >
>> >
ftp://download.intel.com/design/Pentium4/datashts/30235101.pdf
>> >
>> > The 84W comes from +12V. The Vcore converter could be
>> > considered to be 90% efficient (based on the fact that the
>> > MOSFETs and toroids don't get warm to the touch). The
>> > +12V current required is thus (84/12)*(1/0.90)=7.8A .
>> > (A 2.8GHz/FSB800/512KB Northwood is 69.7W ==> 6.45A)
>> >
>> > Adding this together, and assuming the DIMM power comes
>> > from +3.3V (3 amps to give the requires 10 watts), gives
>> > a Prescott based 2.8Ghz system with power of:
>> >
>> > Mode DC Power +3.3V +5V +12V -12V +5VSB
>> > Minimum loading --- 3.30A 10.00A 0.9A 0.03A 0.80A
>> > DIMMs --- 3.0A --- --- --- ---
>> > Processor --- --- --- 7.8A --- ---
>> > Disk drive --- --- 1.0A 0.5A --- ---
>> > CD/RW Drive --- --- 1.0A 1.5A --- ---
>> > Total 212.6W 6.30A 12.0A 10.7A 0.03A 0.80A
>> > (Fan current included in the 0.9A minimum)
>> >
>> > This online calculator will also work out powers for you.
>> >
http://takaman.jp/D/?english
>> >
>> > Other details. The hard drive will draw 2 amps from +12V
>> > while spinning up. (Since the processor doesn't run at
>> > full power in the BIOS, this is not an issue. On a
>> > system I measured with my ammeter, CPU power might be
>> > about 50% of max, while sitting in the BIOS as the drive
>> > spins up.)
>> >
>> > The CD/RW will only draw the 1.5 amp number while spinning
>> > as well. Optical drive currents drop substantially when they
>> > are not in use (no CD in drive). The listed hard drive
>> > current might be representative of a seek operation.
>> >
>> > You can find replacement supplies. But it won't be
>> > easy.
>> >
>> > http://www.pcpowerandcooling.com/prices/
>> >
>> > There are a couple of 1U power supplies here. Due to the
>> > small fan size, these will likely make a lot of noise
>> > if you draw decent power from them.
>> >
>> >
>
http://www.pcpowerandcooling.com/products/power_supplies/highperformance/turbocools/index_hp_1u.htm
>> >
>> > As Rob pointed out, there is a whole spectrum of power
>> > saving solutions out there. But you'll need time to experiment
>> > with them, to find the right one.
>> >
>> > As for actual measured systems - these measurements are
>> > exclusive of video card power. (My 9800pro video card
>> > can draw 5V@5.5A 12V@0.9A, but when the video card is
>> > working that hard, the processor generally cannot run
>> > at full power, due to waiting for the video card to
>> > complete operations. A FX5200 low performance type video
>> > card, would eliminate that extra power.) There is one hard
>> > drive and one CD on these systems.
>> >
>> > These two systems are roughly equal in gaming computing power.
>> > A fast video card can add 38W to the power numbers. This shows
>> > you how close a real system would get to your 230W total rating.
>> > With a FX5200 class video card, you would be in good shape.
>> >
>> > (Processor current comes from +5V...)
>> > A7N8X-E 3200+ 2x512MB dual channel
>> > 3.3V@5.2A 5V@16.6A 12V@0.53A (running Prime95 = 106W)
>> >
>> > (Processor current comes from +12V...)
>> > P4C800-E 2.8G/FSB800/512K 2x512MB dual channel
>> > 3.3V@11.6A 5V@0.55A 12V@6.0A (running memtest86 = 113W)
>> >
>> > The only thing that seems to be amiss, is Intel's user
>> > manual estimates a large +5V consumption, and at least
>> > with my P4C800-E, the power seems to come more from
>> > +3.3V . This means, depending on which company makes the
>> > motherboard, the same amount of power will be used, but
>> > it could either be drawn heavily from +3.3V or +5V rail.
>> >
>> > In the examples above, the AthlonXP board uses the +5V supply
>> > to power the processor. Pentium4 and Athlon64 boards use
>> > +12V to power the processor (that is why they both use the
>> > 2x2 power connector, for extra +12V current). Depending on
>> > how the amps pan out on the various outputs of your power
>> > supply, may help you decide whether a +5V powered processor
>> > or a +12V powered processor, is the right answer.
>> >
>> > As for Athlon64, the TDPs are listed here. Some of the
>> > Athlon64's have very nice power numbers. Whether they will
>> > work for you, really depends on whether your application mix
>> > works well with Athlon64's strengths and weaknesses. If you
>> > were building a gaming box, the Athlon64 would be the easy
>> > choice.
>> >
>> >
>
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/30430.pdf
>> >
>> > HTH,
>> > Paul
>>
>>
>> Dear Paul,
>>
>> Hello again, and of course MANY MANY THANKS for the (once again) very
>> much
>> detailed information.
>>
>> After reading your post, the case I mentioned actually arrived and I had
>> the
>> chance to look at the actual label on the power supply: First of all, the
>> PSU has a label stating that it is a 350W PSU and not a 230W as we were
>> told
>> by the supplier. HOWEVER, the ratings on the label are as follows:
>>
>> -----------------------------------------------
>> +3.3V | +5V | +12V | -5V | -12V | +5VSB
>> 7A | 13A | 6A | 0.5A | 0.8A | 1A
>> -----------------------------------------------
>> 75W Max |-------| 9.6 W Max |-------
>> -----------------------------------------------
>> 3 5 0 W
>> -----------------------------------------------
>>
>> Now, here is how I interpret the label:
>>
>> 1. The 75W Max. under the +3.3V & +5V columns probably means that the
>> combined simultaneous load on these lines cannot exceed 75Watts, while
>> the
>> +3.3V line can carry 3.3x7=23.1W and the +5V line 5x13=65W.
>>
>> 2. The +12V line is capable of producing 12x6=72W.
>>
>> 3. The 9.6W Max. under the -5V & -12V columns probably means that the
>> combined simultaneous load on these lines cannot exceed 9.6Watts, while
>> the -5V line can carry 5x0.5=2.5W and the -12V line 12x0.8=9.6W. Prescott
>> based 2.8Ghz
>>
>> 4. The +5VSB line is capable of producing 5x1=5W. (By the way, what does
>> VSB
>> mean?)
>>
>> 5. Taking into account the 75W and 9.6W Maximums, and adding the +12V and
>> +5VSB powers, I calculate the actual total power to be 75 + 9.6 + (72) +
>> (5)
>> = 161.6W YUCK !!! Is this calculation correct?? Where is the claimed 350W
>> ??
>>
>> You mention that the P4 and Athlon64 get their CPU power from the +12V
>> line,
>> while the AthlonXP from the +5V line. The P42.8Prescott has a TDP of
>> 84Watts
>> (your post) and the Athlon64 3000+ a Max. TDP of 89Watts (Athlon64 Power
>> and
>> Thermal Data Sheet.)
>>
>> Does that now mean that I cannot run either of these CPUs using this PSU
>> or
>> am I missing something?
>>
>> Under what conditions are the maximum Power levels reached? Typical
>> loading
>> of CPU on the Demo system will be a sequence like [Short/Intermediate
>> Peak] - [Longer Idle] - [Short Peak] - [Longer Idle] -
>> [Short/Intermediate
>> Peak] - [Longer Idle] ...
>>
>> Here comes where I get confused again. In your post you mention:
>>
>> >(Processor current comes from +5V...)
>> >A7N8X-E 3200+ 2x512MB dual channel
>> >3.3V@5.2A 5V@16.6A 12V@0.53A (running Prime95 = 106W)
>>
>> >(Processor current comes from +12V...)
>> >P4C800-E 2.8G/FSB800/512K 2x512MB dual channel
>> >3.3V@11.6A 5V@0.55A 12V@6.0A (running memtest86 = 113W)
>>
>> Here you calculate the total Wattage by collecting values from all rails
>> (3.3, 5, 12). Then probably I should not expect all the CPU power from
>> the
>> 12V rail alone in the Athlon64 and Prescott cases... If so, I could still
>> use and Athlon64 or Prescott... But in your table you had put the 7.8A
>> for
>> the CPU completely on the 12V line of the Prescott based 2.8Ghz
>> system...Or
>> ??
>>
>> Confused... I hope I am not asking the obvious.
>>
>> Many Thanks,
>> Cheers,
>> -arifi
>
> That has got to be the worst power supply I've ever seen.
>
> Are you sure that isn't the *minimum* load ? Some supplies
> have two sets of number, some minimum numbers and some
> maximum numbers. The power supply will not regulate to
> within 5% of nominal output voltage, unless the minimum
> load is applied.
>
> Is it possible for you to give more info about the
> case and power supply you bought ? Do you have a
> URL for a website, with details about the product,
> and maybe a picture of the label on the power supply ?
>
> I think you understand the basic principles. The power
> supply has limits for the maximum current that can
> come from any individual output, and there are also
> limits for certain groups of outputs. The group limits
> are caused by using a multi-winding transformer in the
> output. If two windings are supplying current, the thermal
> load on the transformer will bear some relationship to that
> load. Thus, the label on the power supply will state a
> combined total power for those outputs. The total power for
> the power supply should also be based on some fundamental
> limit of the supply (maybe the rating of the primary side
> and switching components ?). In any case, all those limits
> apply simultaneously, so you cannot exceed any output's
> maximum current, neither can you exceed one of the group
> power ratings.
>
> If the ratings on that label are for real, you've got enough
> power *maybe* for a Pentium-M or a Via EPIA. There isn't
> enough current available on +5V or +12V to run any of the
> motherboards I've got here (my home collection).
>
> So, work on verifying that label.
>
> The +5VSB stands for +5 volts standby. The +5VSB supply
> continues to run when the computer is in S3 Standby. The
> suspend to RAM function requires power to keep the memory
> chips refreshed, and that comes from +5VSB. The +5VSB is
> also used for any Wake on LAN, Wake on Ring, or similar
> functions, that require portions of the other chips to be
> operational while the computer is in standby. These days,
> 2 amps is a good number for +5VSB. If you only have
> one amp to work with, then all wake functions and any
> USB/keyboard header settings must be set not to use +5VSB.
>
> My measurements for my A7N8X-E and P4C800-E were done with
> a clamp-on DC ammeter. This is a device based on a Hall
> probe, that converts the magnetic field around a conductor
> into a voltage. My meter will measure AC or DC current in
> a conductor, by simply clamping the meter around a
> conductor or a group of conductors (the magnetic fields add).
>
> This is how I measured my home systems. You grab all the +5V
> wires on the ATX 20 pin power cable and put them in the jaws
> of one of these, then measure the current flowing in the
> bundle of wires. Useful for determining how close to the
> limits you are. Mine cost $400 Canadian.
>
>
http://www.extechproducts.com/products/extech/380941_942_947.pdf
>
> Paul
Dear Paul,
The web site of the company producing (actually, importing from China) the
case does not have any details on the PSU, so I removed the PSU from the
case, scanned the three sides on which the label was wrapped, combined the
three images and put the file on our web site:
http://www.mind2biz.com/download/codegen_psu_label.jpg
The case's web page is:
http://www.codegen.com.tr/desktop/ms-2.htm
We have the MS2-G7, which is the same as the MS2-G10 except for the color.
Some translation for the Turkish terms on that page (not that you wouldn't
be able to figure it out): "Boyutlar" means "Dimensions", "Surucu Yuvalari"
means "Drive Bays" and "Ses & Mikrofon" means "Sound and Microphone"
The actual dimensions of the PSU, if relevant, are 45mm x 80mm x 220mm (1.8"
x 3.2" x 8.8") (WxHxD), and it has two 4cm (1.6") fans, one one each end.
Here we have 220V/50Hz mains.
Pentium-M is hard to find here and importing from outside Turkey has too
much paperwork and extra costs attached. Of course, another case could be an
option, but strangely, this brand is the only one supplying these "slim"
ones and portability is also a factor.
Now, assuming that the information on the label's picture will not bring any
good news (but hoping for the opposite), should ve resort to plain AthlonXPs
and older P4s? For example, AMD documentation
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/26237.PDF
states that an AthlonXP 2800+ draws a max. of 68.3W, which is just slightly
above the 65Watts that can be drawn (according to its label) from the +5V
line of this PSU. Similar options may be available for the P4 family too...
When is the "Max" power needed? When the CPU is running at 100% capacity
according to Windows Task Manager, is it also drawing the max. current /
power from the PSU? Is the duration a factor here?
Many thanks *again*,
Best,
-arifi
Facebook
Twitter
Instagram