Typical mains power for mid-range PC?

Most supplies are better at outtputing their max steady-state capacity than they are at a real dynamic load such as CPU that can vary by a couple dozen watts instantaneously.

Of course they are, and switchers have transformers.

What does Everest say about the SMART temp ?

Dont get anything like that effect with my Samsungs, they are well above ambient with no extra airflow.

Likely HUTIL is having a brain fart or something.

kony wrote in news: snipped-for-privacy@4ax.com:

Exactly. I've just experienced this on a customer system. Max chip temperature reading was 62C, whilst SMART reported 37C with ambient of 31.5C. Temporarily rigged up a cooling fan pointing directly at the PCB which reduced the max chip temp to around 40C: rather than several system lock-ups per day, customer reports no lock-ups in the 5 days since I added the lash-up cooling.

On Tue, 01 Aug 2006 04:36:35 -0400, kony Gave us:

The quality of the supply is its ripple spec at max rated load. Since it will be at a lower load, and even lower ripple at any other time, that is the max that would be presented to it and is the standard by which such supplies get qualified.

Ripple and regulation.

Dynamic, in operating range loading will not challenge that.

A shitty, noisy supply will likely be so throughout its operating span, including you transient response notions. A good supply will breeze through it... quietly. That is what computers like, remember?

Not fan noise! Good PC operation is about ELECTRICAL noise abatement!

Another reason to move toward optical computing.

On 1 Aug 2006 06:02:24 -0700, "Terry" Gave us:

Didn't run it (the calc), but did read the page.

Our supplies had 100% proper continuous load ratings.

We would never make a supply operation declaration label that declared peak wattage, unless the supply was custom, like an ionizer supply or such.

Our LVDC supplies were a rated rail at a given wattage with a voltage variance (regulation) of a declared amount, and a ripple of a declared amount at that full load (ripple spec).

It doesn't get much simpler, but reminds me of the false declarations made by car stereo folks in the 70's. You get what you pay for, and you have to research what you buy.

Life is simple. We build things for our use. One must not over-complicate the techno devices in life. Go with the flow.

As a student at the end of the (insert decade here), I was involved in d&d work on a fluid logic control system - it was immune to electrical noise too (and nuclear radiation..). Great fun and the "wet blouse effect" when it sprung a leak..

The next door lab was working on holographic projection...which also seemed a good idea at the time.

??

It's one parameter, but hardly the definition of quality.

Not necessarily true, because switchers dynamically adjust frequency or duration of switching. The longer the interval between "on", the higher the ripple.

Generally, no. The ATX spec exists to address the issue and most PSU are spec'd against that, the datasheets do not list the actual measured ripple at full load. Further, supplies are qualifed by several factors, ripple is but one parameter as was mentioned above re: quality.

Yes it does. Some call it recovery time, to describe this particular type of ripple.

All of these global feedback types of supplies have a minor lag between load and adjustment which is not a persisting ripple but because the load is so, EXTREMELY DYNAMIC, each singular event is constantly reoccuring. It can easily be seen on most PSU reviews, the larger peeks and valleys on oscope are common from the dynamic load.

Yes, except it still depends on what "shitty" means as you arbitrarily try to suggest ripple means quality.

Digital circuits do NOT need the lowest ripple possible. It is a good goal, and too much ripple will put wear on system components, BUT beyond a certain point, lower ripple has no useful purpose, makes a supply no more "quality" per se. If lowest ripple were the issue they'd not have used fairly typical switching supply circuits in the first place.

You've taken an area where there "could" be problems with crap supplies, and tried to overextend the argument for more than it's worth. Analog circuits are far more picky about power than digital, especially digital where there's still multiple stages of further regulation and LC/C filters.

Within parameters, a little better than the ATX spec is not a bad goal. After that point, other factors matter as much such as efficiency, discrete component lifespan (and fan as well).

The main problem is merely one of budget. The typical computer supply is a commodity grade item, even considering the better brands, where $5 more would go a long way but to get that $5 put into one a buyer typically spents multiple times as much... more than the profit margin was previously... more of today's disposible society in action.

Smart temp is what HUTIL reported.

That what effect? Any that are designed the same should have near identical response... in the same environment, upside down on a cardboard box with modest side-flow from a remote fan. It was not meant to simulate conditions in a system case in any way, I merely had the drive out and decided to compare the smart temp vs actual measurements elsewhere.

Nope, similar occurs with many HDD temp reports, they are unrealistically lower than the higher heat producing parts.

This is the same situation as with any single-point temp report. One cannot look at their chassis temp and know about their CPU or GPU either. Smaller system (on a HDD), but same issue with any parts not thermally coupled.

kony wrote

The SMART temp/ambient temp difference.

I did however miss your bit about the fan on first reading.

Sure, I was just commenting on the difference between the SMART temp and ambient and missed the fan comment.

It isnt even supposed to be displaying the temperature of those other bits.

Its got nothing to do with 'not thermally coupled', just a result of heat sources and significant thermal gradients.

Awesome experiment!

"Phat Bytestard" wrote

We must have the best of both, and all worlds.

Quiet psu's with excellent electrical characteristics.

"xModem" wrote

Aww, that's a tiny little 2 1/2 incher. Quiet computing folks often use little notebook drives.

It doesn't take into account the quality of the noise, something silentpcreview always mentions.

Not surprising that the Samsung P80 is the quietest desktop drive. A Seagate

silentpcreview wasn't impressed with the sounds of the 7200.9 and 7200.10, though they thought the 7200.7 was pretty good and loved the Barracuda IV. The Samsung P80's took over their love. The P120's are a bit loud for them. Recently they found a WD 500gig that is just great.

They're just showing idle noise. Seek noise is important.

On Tue, 01 Aug 2006 14:39:45 +0100, Palindr?me Gave us:

Used a LOT in automotive car washes.

Ever heard of a company called "Clipper" out of Cincinnati?

Expensive pneumatic and liquid logic devices.

Holo data storage is on the rise.

On Tue, 01 Aug 2006 11:03:43 -0400, kony Gave us:

If a DC supply has high ripple, it doesn't properly qualify as a DC supply, and yes, is also of poor quality.

On Tue, 01 Aug 2006 11:03:43 -0400, kony Gave us:

Nope. Most PC switchers generate a 400V rail, and then take the other rails from that, and some even have linear subsections in them.

They operate at a single frequency, and ripple is practically ALWAYS lower when the loading is lower. Most have a lot of stored energy and do not exhibit higher ripple throughout the load range.

One thing that will cause a higher noise figure is a high line condition.

Our supplies auto-switched from 90 VAC in to 265 VAC in.

They operated best at low line.