What would happen if the power supply design only looked at the voltage drop and not the current density?

Infineon / Mitsubishi / Fuji / Semikron / Eupec / IXYS

What would happen if the power supply design only looked at the voltage drop and not the current density?

Posted Date: 2024-01-26

As we all know, the theory of DC power supply design is actually very simple. In the final analysis, it is Ohm's law. The power chip supplies current to the load. The current has a certain voltage drop through the transmission path. The voltage value that finally reaches the load end is the result of the attention of our receiving chip.

In PCB design, from the power chip to the load end reception, the factor that can cause voltage drop in the middle is our PCB design link, including copper, vias, wiring and other paths. As the main conductive material of PCB, copper itself is a conductor with a specific resistivity. Therefore, when current passes through, a voltage drop will occur, thereby reducing the voltage that finally reaches the load end.

Many people actually think this way. As long as they pay attention to the voltage drop on the link, and then see if the voltage drop reaching the load terminal meets the chip's reception requirements, such as 5%, if it meets it, then there will be no problem.

Below we use a simple copper sheet to analyze this situation. On this simple piece of copper, the left side is defined as the output terminal of the power supply, and the right side is defined as the load terminal.

Assume that the power supply output voltage is 1V, the output current is 2.5A, and the allowable voltage drop at the load end is 3%.

Then by doing a simple voltage drop simulation, you can know whether the voltage drop can meet the requirements when the voltage is received at the load end. The simulation results are like this.

It's very simple, right? This piece of copper will produce a voltage drop when transmitting current, so the voltage drop slowly decreases from the power output end on the left to the load receiving end on the right. When it finally reaches the load end, the power supply is 0.972V, which is okay. , meeting the pressure drop of 3%.

Yes, many engineers who are new to the industry do only focus on the voltage drop at the load end. If the voltage drop meets the requirements, they are done. So listening to what Mr. Gao Guo said, are there other indicators that we should pay attention to? Haven't you noticed that in Ohm's law, resistance is represented by copper via traces in the link? We focus on the voltage at the link and the final load end, but what about the current? Of course, what Mr. Gao Gao said is not just to pay attention to the 2.5A current written above, but the same as the distribution of voltage on the copper sheet. We also need to pay attention to the distribution of current on it. Give it a professional term called current density distribution. .

Still in the above case, the current density diagram during the current transmission process of this piece of copper is like this:

It is about 27A/mm2. Experienced colleagues know that this value is not large. So the question is, why should we pay attention to current density? This is the core content of this article. In addition to ohmic current, we need to know another law called Joule's law. To put it simply, current not only causes voltage drop, but also generates heat where the current density is high. That is to say, the current density also affects the temperature of the PCB!

Now that we know that current density will cause the temperature to rise, if conditions permit, our simulation actually needs to perform electrothermal simulation. Let's use this piece of copper for electrothermal simulation. With the same 1V output and 2.5A current, let's see if the temperature of this piece of copper changes. It is worth noting that the initial temperature for electrothermal simulation and electrothermal simulation here is set to 25 degrees, which means that the electrothermal simulation will be performed on the basis of 25 degrees.

It can be seen that under the same current conditions, the copper temperature obtained by electrothermal simulation will probably rise to about 38 degrees.

Just this 2.5A current can cause the temperature of this small copper plate to rise by more than 10 degrees. And do you think that it will be over when the temperature of the board rises? It's not that simple, which in turn affects current density and voltage drop. Let’s take a look at the voltage results at the load end of the copper skin at this temperature:

You will find that as the board temperature increases, the voltage drop at the load end also increases. From the previous voltage drop simulation of only 972.4mV to the electrothermal simulation of 969.9mV, a reduction of 2.5mV was achieved, and it was no longer able to meet the original 3% voltage drop requirement.

It seems that the copper current density itself after verification in this case is not very large, but the impact on temperature rise and voltage drop cannot be ignored! If you take a look at the PCB design you have done, you will even find that it is a luxury to have such an ideal complete plane for power flow.

For example, the plane used for flow becomes as follows. Can you imagine the final impact of current density and temperature rise?

Don't think this plane is exaggerated. In fact, in real PCB design, some places are not much better than this!

Or if the current is not 2.5A, but 25A, do you think the impact of this current density is still small? If the output voltage itself is lower or the reception requirements at the load end are higher, will the margin be smaller?

Here comes the question:

Where do you think the current density will be higher on the PCB, and what design methods should be used to optimize it?

Review Editor Huang Yu

#happen #power #supply #design #looked #voltage #drop #current #density