The 4-leg one's dropout voltage is just a third of the 3-leg. That sounds like better efficiency and less heat.

TELVM wrote:

The 4-leg one's dropout voltage is just a third of the 3-leg. That sounds like better efficiency and less heat.

No, not by itself better efficiency/heat but it sort of makes it possible. Both regulators can make 3.3V output with a 5V input and they have same efficiency (on a large scale when you don't look at the few milliamps the regulator itself needs in order to work). The dropout voltage means how much volts more the input needs at minimum to be able to supply the output voltage in a stable fashion. So basically, the other with 1V dropout needs at least 4.3V input to make 3.3V output, and the other with 0.5V dropout needs at least 3.8V input to make 3.3V output. Thus largely irrelevant for an AT motherboard where minimum you have available is 5V input to make 3.3V.

The difference between 3-leg and 4-leg devices is how the regulator itself gets power and how it affects voltage setting resistors.

In the 3-leg device, the regulator's own supply current flows through the voltage setting resistors. Thus you need very stiff feedback voltage divider so a lot of current flows through the divider, so that the regulator's own supply current also flowing through the voltage setting resistors makes very little difference in the output voltage. This wastes power, although very little compared to how much a CPU needs current. Also if different manufacture batch or regulators from different manufacturer consume different amounts of power supply current, you see that as a bit higher or lower output voltage, so it's less accurate. The divider resistor values are made so low that so high current flows that the contribution of the regulator power supply current is kept minimal and the output voltage is still within some tolerance specification.
Actually, for some regulators, this divider is just made to consume at least some amount of current, to make sure the specification for minimum load current is achieved, if there is very light load that does not consume enough by itself, or it is possible that at sometimes there is no load at all.

In the 4-leg device, there is an extra ground pin for the regulator's supply current to flow out, and the voltage feedback pin can sense the voltage with very little current flowing in or out of the feedback pin. Therefore you can use much larger resistance values for the feedback divider, and much less current flows through the divider, and less power is wasted there. Also, since there is a bit of voltage drop in copper wires, it is possible to arrange the feedback from the point of load (say, at the CPU socket), rather than take the feedback from the regulator output.

Maybe this would be handy to:

http://www.youtube.com/watch?v=GSzVs7_aW-Y

My knowledge has improved a litte bit more.. So i guess on those motherboard is always better to use low-drop out voltage regulators.. Most you have to install an 5V on te get 4V, 3.45V or 3.3volt.
Because the differents of unregulated input voltage(i think this is AC =active current) is very close to the DC= Direct current) output voltage. So normally voltage regulator would normally should have a to high drop out voltage..

Now some questions.. I tried to repair the VRM part of an ADI motherboard i have.. But its missing de vrm part orginally. That board uses the UM8498F chipset.. There are 4 jumpers to set.. The first three jumpers determine to set the voltage to 3.45 volts. On that part there is only one light brown usual resistor.. Its an 4 band one with the colors: `orange, black, black, gold` I put this information in a 4 band resistor calculator..
It seems to be an 30 Ohms micro farrad resistor 5% tollerance. Iam not really sure both are black on that resistor, because those lines are tiny written on the resistors.. Maybe i think of buying a tiny loop (magnifier glass) and inspect it a little bit better.. But iam about 85% procent sure that the color code is `orange, black, black, gold` But the point here is, that iam missing a resistor on that motherboard to set it to 3.3 volts.
Iam not really good in calculations, so i really like to know which value resistor i need to have installed to get 3.3 volts on that motherboard.. Is seems the board doesnt support 4 volts..

Anyone having sugestions??

Can anyone explain me why the trace / electric path on the motherboard is bigger for the Vout, compared with the electric part for the Vin side (that the tiny side).

Robin4 wrote:

... Can anyone explain me why the trace / electric path on the motherboard is bigger for the Vout, compared with the electric part for the Vin side (that the tiny side).

Might be because the lower volts downstream of the VRMs mean higher amps on the Vout, hence the beefier traces.

Couple interesting links:

Everything You Need to Know About The Motherboard Voltage Regulator Circuit - Hardware Secrets

Graphics Cards Voltage Regulator Modules (VRM) Explained - Geeks3D

Robin4 wrote:

So i guess on those motherboard is always better to use low-drop out voltage regulators.. Most you have to install an 5V on te get 4V, 3.45V or 3.3volt.
Because the differents of unregulated input voltage(i think this is AC =active current) is very close to the DC= Direct current) output voltage. So normally voltage regulator would normally should have a to high drop out voltage..

Yes, that's why a low drop-out regulator is needed. Normal regulators need at least 2 to 2.5 volts over it to keep the output voltage stable, many times 3 volts is a good approximation. But the input voltage to the regulator is already regulated DC as it comes from the DC power supply. It is not unregulated, and it is not AC either (alternating current), as regulators do not handle AC.

Robin4 wrote:

Now some questions.. I tried to repair the VRM part of an ADI motherboard i have.. But its missing de vrm part orginally. That b […]
Show full quote

Now some questions.. I tried to repair the VRM part of an ADI motherboard i have.. But its missing de vrm part orginally. That board uses the UM8498F chipset.. There are 4 jumpers to set.. The first three jumpers determine to set the voltage to 3.45 volts. On that part there is only one light brown usual resistor.. Its an 4 band one with the colors: `orange, black, black, gold` I put this information in a 4 band resistor calculator..
It seems to be an 30 Ohms micro farrad resistor 5% tollerance. Iam not really sure both are black on that resistor, because those lines are tiny written on the resistors.. Maybe i think of buying a tiny loop (magnifier glass) and inspect it a little bit better.. But iam about 85% procent sure that the color code is `orange, black, black, gold` But the point here is, that iam missing a resistor on that motherboard to set it to 3.3 volts.
Iam not really good in calculations, so i really like to know which value resistor i need to have installed to get 3.3 volts on that motherboard.. Is seems the board doesnt support 4 volts..

Anyone having sugestions??

I am sorry but with this information, I cannot know how the resistor is connected and where. At least a picture would be required, but schematics would tell the absolute truth. The circuit board can have many layers and you can only see the top and bottom layers how components are connected.

Robin4 wrote:

Can anyone explain me why the trace / electric path on the motherboard is bigger for the Vout, compared with the electric part for the Vin side (that the tiny side).

Maybe.

TELVM wrote:

Might be because the lower volts downstream of the VRMs mean higher amps on the Vout, hence the beefier traces.

No this is definitely not it in this case, because that and the links provided apply to switch mode power supplies, not linear regulators. This motherboard has linear regulator, and in linear regulators, current going in is the current going out (plus whatever the regulator itself requires to operate).

But the trace on the output might be bigger because it means less resistive losses in the copper between regulator and CPU, so voltage does not drop too much while flowing. The input side does not matter so much how much voltage drops at that point, as long as the trace is big enough to carry the current without heating too much and still meeting the minimum voltage requirement on regulator input.

Can somebody explain me what the differents is of `voltage regulator` and `transistor`.. I thought that it would be the same to people only give that one `thingy` two different names.. But i guess there should be other differencens..

Is it true that not always an voltage regulator is used on 486 board, but should also been possible to be a transistor?? (because an board i have uses the characters TR1) on the pcb.. So i guess there need to be placed an Transistor instead..

It lies, moreover, in the type of circuit that is used, whether there is a transistor or a voltage regulator is to be applied?

Robin4 wrote:

Can somebody explain me what the differents is of `voltage regulator` and `transistor`.. I thought that it would be the same to people only give that one `thingy` two different names.. But i guess there should be other differencens..

They both come might have three legs and plastic cases, but otherwise they are completely different animals. A transistor is a transistor, it can amplify or switch currents/voltages depending on how you connect it. A voltage regulator is a integrated microchip that is designed to keep output voltage stable, internally it may contain a circuit with tens of transistors and other components to perform the regulation.

Robin4 wrote:

Is it true that not always an voltage regulator is used on 486 board, but should also been possible to be a transistor?? (because an board i have uses the characters TR1) on the pcb.. So i guess there need to be placed an Transistor instead..

It lies, moreover, in the type of circuit that is used, whether there is a transistor or a voltage regulator is to be applied?

Yes, it depends on the circuitry. There might be a circuitry on motherboard that selects voltage and this voltage is buffered with a transistor. After all, before there were any voltage regulator chips, they were built with separate discrete components.