Power supply install

[unrox]

I am tring to install a 500W Power supply, replacing a 250W supply, all hookups fit perfectly, as does the unit, upon rebooting I get a message""Reboot and select proper device" and "Insert boot media in selected boot device and press any key". Never installed a PS ,have installed numerous other devices with not too many problems, followed all instructions, and here I am. Help!

[Hipster Doofus]

Go into your bios & reconfigure the boot order. That might do it.

[setarip_old]

@unrox

Your question has nothing to do with the purpose of the Doom9 Forum.

You'd undoubtedly have more success posting this question at a hardware-specific forum...

[hello_hello]

Go into the BIOS and check to see if the hard drive shows up there. Changing the power supply shouldn't change anything, so it sounds like the PC can't find the OS to boot from because the hard drive's not working. Either the power cable or the SATA/IDE cable isn't connected properly at one end.

Just in case it's not one of those unfortunate co-incidences, make sure you don't have any external devices such as a USB thumb drive connected to the PC when you boot. If you do, the PC may be trying to boot from it and changing the PSU didn't the cause the problem. My old PC wouldn't boot with an external drive connected. It'd look for something bootable on it then give me the same error message.

[Ghitulescu]

or maybe he simply forgot to connect the power connectors to the drives ;)

[Wilbert]

Moved to appropriate forum.

[Ghitulescu]

I don't know why did you installed a PSU of twice the capacity, but I know there might be problems when the source is too strong and therefore it works at low loads (good quality sources should not have this problem, but cheaper ones may provide unstable voltages/amperes).

[ramicio]

There is no such thing as too strong of a source. The power supply is only supplying the current that the computer is drawing. Voltage is kept constant regardless of current, until you overload the thing, then it dips. Ideally it should shut off if overloaded. If your computer draws 200 watts on a 350-watt power supply, then the same 200 watts is going to be roughly the same draw at the outlet on a 700-watt supply. I said roughly simply because every PSU has a different efficiency curve. The tolerance will be within a few watts. With a 1200-watt power supply you could stably power a single LED, or power a crazy computer with 4 video cards at full load.

You either forgot to plug in a drive or you need to go into the BIOS and make sure the OS drive is selected as a boot device, and no other bootable devices are set before it, except removable storage.

[Ghitulescu]

No, this is the common sense, but it's not true at all, at least for cheap PSUs. The switching regulator cannot be modulated that much to light a LED and a fan-heater, and provide stable voltages on both ends of the power spectrum. Like a radio, it cannot tune from 153kHz to 108000kHz, it needs frequency banks. Some components (like your LED) are less sensible to voltage variations, others not (CPU, memory, GPU).

[ramicio]

I've had "cheap" PSUs and expensive PSUs all provide me with stable voltage from little load to high load... A motherboard has its own power regulators, so your analogy is useless. A faulty PSU will not behave properly, but any one that is operating as it should will be able to handle the full spectrum.

[hello_hello]

Of course it's highly probable the original 250w PSU was a cheapy, and the new 500w PSU is a more reputable model.

I've come across a plethora of PCs with a wide variety of PSUs (quality wise) and I'm yet to see a PC which started up minus a single component (for instance a hard drive) because of the PSU. It's pretty much all or nothing.

The only issue I've ever come across which might relate to some "source being too strong" theory is the PSU not seeing enough of a minimum load and switching itself off as a consequence, but of course that'd effect the whole PC, not just the hard drive. http://www.playtool.com/pages/psuunderload/mincurrent.html
I owned one of the Antec PSUs mentioned in that article. It never skipped a beat while running in the PC I originally bought it for, but after I upgraded the MB it took a bit of work to get the PC to boot.

[Ghitulescu]

Thanks for the article, I was prepared to get a bit in details on how a switched PSU works.

[ramicio]

There's no minimum load.

[Groucho2004]

There is no spoon.

[StainlessS]

Very early PC's needed a dummy load if a hard drive were not installed,
although this has not been the case for a long time, I think.
(Dummy load being an array of 'Sandbar resistors', I think).

[Ghitulescu]

There's no minimum load.

So I'm forced back to theory.

A switched power supply functions (very very basic) like this: it has a high-frequency oscillator that commands a gate - the duration of the impulse (that coincide with the gate being closed then opened) sets the amount of the electricity that passes through that gate. A filter made of a coil and some condensers keeps the current sort of constant. As one can understands, the more the gate is opened the higher the current that passes through.

However, both the filter (to clean the peaks) and the switching driver (the oscillator) have their own preferences (it's called resonance) and work extremely bad outside a certain range (square of 2 divided by 2, some 70%), exactly like a radio.

Now, if one needs to extend the range from 0.001 mA to 500 Amperes, that would be done only at the expense of power efficiency, or by employing parallel automatically selected drivers and filters.



The longer explanation you have to read it in any electronics book.

[ramicio]

i'm sorry you need to revert back to theory. I've been experienced with electronics since I was of a single digit age. All modern power supplies have voltage regulators after this PWM circuit. The gate you are referring to simply fires a transformer. This design is for efficiency. If you are just blindly powering a transformer with full power, you aren't going to be as efficient as you can get. If you can lower the pulse width to it, and hence the overall energy going to it, when you don't need it, and then ramp up the duty cycle when there is higher load, you are squeezing out all the efficiency you can. This is all based on current. The voltage regulators keep the voltage at a maximum no matter if you load 0.001 A or 500 A.

You could easily make a simple power supply out of a transformer and special high current regulators and power a computer. It just won't be very efficient across various loads.

[Ghitulescu]

i'm sorry you need to revert back to theory. I've been experienced with electronics since I was of a single digit age. All modern power supplies have voltage regulators after this PWM circuit. The gate you are referring to simply fires a transformer. This design is for efficiency. If you are just blindly powering a transformer with full power, you aren't going to be as efficient as you can get. If you can lower the pulse width to it, and hence the overall energy going to it, when you don't need it, and then ramp up the duty cycle when there is higher load, you are squeezing out all the efficiency you can. This is all based on current. The voltage regulators keep the voltage at a maximum no matter if you load 0.001 A or 500 A.

You could easily make a simple power supply out of a transformer and special high current regulators and power a computer. It just won't be very efficient across various loads.

I also am experienced with electronics from the same age as you, so I wonder what you don't get that you can't have the pulse width so narrow that a filter cannot clean it afterwards from all ripples. Everything there works on resonance, otherwise one would need trafos 3 times bigger than one found in a Pioneer AVR (or Denon for that matter or HK or or or ...) instead of trafos having matchbox sizes, and still be dependent on the frequency. Also because it's difficult to squeeze such a large range into the linear curve of a transistor, most switching PSUs use ICs as drivers, and they still cannot work for the remotest ends of the range. It may work, but it's not granted.

I know that most electronic boards have additional power supplies on-board, but this doesn't solve the issue of a bad source.

[ramicio]

I don't buy that you are experienced in this field.

Maybe because the frequencies we are talking about are in the megahertz range... That's how "filters" are able to level the pulses.

I really don't care what you think about this. A computer power supply is not going to fry low-current devices.

[Ghitulescu]

That was really my impression about you, too. Everyone interested in PSUs knows this issue.
How can one (maybe you) modulate the pulses to get a constant current of 0.01A (say a LED) when the source must output a 100% load of say 45 A? The only way of having a range so broad is to use additional variable dissipative loads, which will lower the efficiency, increase the heat production and the complexity of the drivers - which questions the use of a switched PSU instead of a classical one in the first place.

[ramicio]

How is an LED going to draw more current than it does? It's only going to draw as much as it draws. Plus, they always have a current-limiting resistor in series with them. I don't get how you can't get this basic principle. I think you are overthinking this. There is no problem.

[Ghitulescu]

Case 1:
There is a load (pure resistive) that draws 0.01A by 3.3V.

Case 2:
There is again a pure resitive load that draws 45A by 3.3V.

A sPSU can supply 3.3V constant at a current of max. 45A.
My statement was that the PSU can be used for case 2 only. Yours that the PSU can be used for both.

Not only my experience, but others' as well, says that case 1 has a too low consumption and it's not sensed as a load, and even if, then the system cannot reliably/stably deliver the tiny fraction of the Pmax. For data comunication it will bring noise, for CPU will probably force the reset etc etc.

PS. pure resistive, to simplify the reasoning. Inductive loads complicate the design.

[Guest]

Closing thread as it is way off topic now and has become personal.