why doesn't my PC run on 170V DC?, updated!

[Jerboy]

As far as I know, a typical PC power supply has a bridge-connected diodes and capacitors to create a DC bus with approximate voltage of AC line peak, which is 170V (120 x sqrt 2). So, why doe it refuse ot boot up when I power it up with 170V DC from the plug? The lamp comes on and hard drive spins, but it never boots.



Update:

Ok Mark R must have been right. The computer power supply has a voltage multiplier which gives peak-to-peak voltage. 120 x sqrt 2 x ~340V



It is very difficult to measure the power consumption of computers, because of its highly harmonic current waveform. First, you'll need a true RMS DMM to get the apparent power. Secondly, since its power factor isn't quite cos phaseshift angle=PF since it's not a plain inductive or capacitive load, you can't easily figure out it's PF on O'scope screen.

The only real solution in measuring AC input is using a professional grade wattometer which can calculate wattage even in complex current waveform, but this is prohibitively expensive. Since the input AC is rectified to DC right away in the front-end of power supply, a more practical way to measure power relatively accurately is feeding the computer with DC.

You'll feed it with smoothed 340V DC and measure current with DC ammeter. Using P=V I, the result should be quite accurate.
I fed a capacitor and rectifier with 240V AC and this yields ~340V DC(240 x sqrt2).

Actual voltage was 352V.

For this PIII machine(PIII 600 Katmai, ABit BE6, one CD-RW drive, one HDD, 230W ATX generic supply takes 82W in idle.

352V DCx 233mA=82watts

 

[Bozz]

Got the schematic diagram? Without it we are simply poking holes in the air...

A guess is the supply may require a 50Hz pulse from the AC side to get the switchmode circuit started...

But it still doesn't explain why your HD spins up, that indicates the supply is working...

 

[Jerboy]

Originally posted by: Bozz
Got the schematic diagram? Without it we are simply poking holes in the air...

A guess is the supply may require a 50Hz pulse from the AC side to get the switchmode circuit started...

But it still doesn't explain why your HD spins up, that indicates the supply is working...

Basically it's the same symptom I get when I run the PSU on 120V with the supply voltage selector switch set to 230V.

What if I simply start with 170V DC, then start cranking up the variac until it boots up 😀

 

[Mark R]

The problem is the rectifier:

When the PSU is set to 230 V, then the input stage is a conventional bridge rectifier and resevoir capacitors.

However, when the input switch is set to 115 V, the input stage is rearranged into a voltage doubling configuration. The voltage doubling fails to operate if a DC voltage is applied. If you apply 170 VDC, then all you get is 170 VDC whereas the PSU is designed to operate with 340 VDC.

The HDD may well spin up, because it has some power, albeit at a lower voltage than desirable. The rest of the system won't power up, because the PSU detects the inadequate voltage, and doesn't issue a power good signal, allowing the motherboard to power up.

 

[capybara]

ooops ////double posted

 

[capybara]

have u considered just transforming down to the voltages you need -
such as + and -12v, 5v, 3.3v.,,,,,,,,,,,,,,,,,,then yiou wont need a psu at all

 

[Jerboy]

Originally posted by: Mark R
The problem is the rectifier:

When the PSU is set to 230 V, then the input stage is a conventional bridge rectifier and resevoir capacitors.

However, when the input switch is set to 115 V, the input stage is rearranged into a voltage doubling configuration. The voltage doubling fails to operate if a DC voltage is applied. If you apply 170 VDC, then all you get is 170 VDC whereas the PSU is designed to operate with 340 VDC.

The HDD may well spin up, because it has some power, albeit at a lower voltage than desirable. The rest of the system won't power up, because the PSU detects the inadequate voltage, and doesn't issue a power good signal, allowing the motherboard to power up.

Can I feed it with 240V rectified to 340V DC then? I need to power computer from DC in order to measure power consumption.

It has such harmonic content that measuring it with standard AC V and A meter is not easy.

So if I feed DC, I can simply do V x I = W

 

[Mark R]

This should work - but I'd try and either get the schematics, or reverse engineer the input stage just to be sure.

Would hiring a power analyser be out of the question?

 

[Bozz]

Originally posted by: Mark R
The problem is the rectifier:

When the PSU is set to 230 V, then the input stage is a conventional bridge rectifier and resevoir capacitors.

However, when the input switch is set to 115 V, the input stage is rearranged into a voltage doubling configuration. The voltage doubling fails to operate if a DC voltage is applied. If you apply 170 VDC, then all you get is 170 VDC whereas the PSU is designed to operate with 340 VDC.

The HDD may well spin up, because it has some power, albeit at a lower voltage than desirable. The rest of the system won't power up, because the PSU detects the inadequate voltage, and doesn't issue a power good signal, allowing the motherboard to power up.

Of course!! I must have had a brain fart while trying to work out what could be causing it and not realising the obvious! 🙂

Cheers

 

[Demon-Xanth]

Power supplies typically have step down transformers to get them at a voltage that they can work with, no AC in==no power out.

 

[Bozz]

Originally posted by: Demon-Xanth
Power supplies typically have step down transformers to get them at a voltage that they can work with, no AC in==no power out.

You're thinking of linear power supplies where a huge iron core transformer's primary winding is connected straight to the mains power.

Switchmode power supplies operate on a different principle- they take the mains 240v (or 120v), rectify it, filter it so you have a clean 160-170v per capacitor (two caps for dual voltage power supplies), add the two voltages up and you have about 330 volts DC. The switchmode name refers to what happens next, a high speed "chopper" circuit oscillates this high voltage DC through a small transformer which is usually ferrite core (ferrite has a higher permiability if I recall correctly). The switchmode speed is approximately 20-80khz in most power supplies. Why do they do it this way, you may ask? Efficiency - A transformer can only pass AC, or a changing voltage. It cannot pass a direct current. Imagine the 240 volt mains power passing through a regular linear power transformer at 100 times/sec (120 times/sec for 60hz locations), that is a certain amount of energy that can be converted. But if the same voltage is passed through the transformer 20,000-80,000 times per second then MUCH more energy can be converted. By doing it this way you reduce wasted energy, reduce heat, improve efficiency but slightly reduce reliability (more components), reduce regulation ability and increase the amount of noise on the DC lines. Thats why amplifiers use linear power supplies (apart from car audio amplifiers).

Cheers

 

[earthman]

There is the issue of weight too. Switchmode supplies are fairly light, linear ones can be very heavy, not too mention the huge magnetic fields they produce.

 

[Jerboy]

You're thinking of linear power supplies where a huge iron core transformer's primary winding is connected straight to the mains power.

Switchmode power supplies operate on a different principle- they take the mains 240v (or 120v), rectify it, filter it so you have a clean 160-170v per capacitor (two caps for dual voltage power supplies), add the two voltages up and you have about 330 volts DC. The switchmode name refers to what happens next, a high speed "chopper" circuit oscillates this high voltage DC through a small transformer which is usually ferrite core (ferrite has a higher permiability if I recall correctly). The switchmode speed is approximately 20-80khz in most power supplies. Why do they do it this way, you may ask? Efficiency - A transformer can only pass AC, or a changing voltage. It cannot pass a direct current. Imagine the 240 volt mains power passing through a regular linear power transformer at 100 times/sec (120 times/sec for 60hz locations), that is a certain amount of energy that can be converted. But if the same voltage is passed through the transformer 20,000-80,000 times per second then MUCH more energy can be converted. By doing it this way you reduce wasted energy, reduce heat, improve efficiency but slightly reduce reliability (more components), reduce regulation ability and increase the amount of noise on the DC lines. Thats why amplifiers use linear power supplies (apart from car audio amplifiers).

Cheers

Bozz, you sure it's safe to run my power supply on 240V rectified to ~330V ?

If the internal voltage isn't this high, and I feed it, it will KILL IT instantly.

I won't be using those wussy 100VA travel 120V-->240V converter.

It will be fed from a 2KVA transformer, 10,000µF capacitor bank and 50A bridge rectifier. Because of this, impedence is really low and can easily destroy components if it's not the proper voltage.

 

[Jerboy]

Bu/\/\|>ag3

 

[Bozz]

I cant add too much more, like you say, Mark R hit the nail right on the head

Search for a power supply schematic that is dual voltage, the voltage doubler circuit only adds three components, from memory, to a primary switchmode circuit design. It really is an ingenious method of doubling the voltage. All you need to do is pull out the bridge rectifier, hardwire the 110/240v switch to the 240v position and put 320-340VDC across the +ve and -ve outputs from the bridge rectifier. Naturally install a fuse for protection, go half the size of the fuse that's already in it since it's designed for 110v use (double the current compared to 240v)

Cheers