Power sensing circuit

It sounds like you need to start with a sensor that detects when the computer needs power and when it doesn't.

Many laptops have a switch that senses when the display/lid closes and generates a signal to turn off the display and go to sleep.

Many computers have a timer that senses that no activity has occurred for a preset period and when the timer times out the computer goes into screensaver mode or sleep mode. These often wake up when a keyboard key is pushed, or the mouse is moved.

The shutdown or sleep process is usually controlled by the CPU, so the shutdown signal needs to go to the CPU, which controls at least part of the power supply.

The wakeup signal could go to the CPU if it has a built-in capability to convert this into a wakeup sequence, or it could directly activate power to parts of the circuit that were turned off. In that scenario, the CPU would get involved after waking up to keep power applied until a sleep cycle is required.

There are lots of possible triggers and lots of potential power up and down sequences, including a simple power switch.

Hi Pujan,

I am not an expert but it seems to me that this may have been a trick question. If the electronics of the computer are making the decision to go into sleep or low power mode does the power supply really have to do anything? The power supply only reacts to load demand and if the electronics has lowered the demand and therefore the load the power supply will see a higher impedance and provide less power. Most computer power supplies are switch mode and if they are line powered saving energy may not be an issue. If I have misunderstood or if I am wrong about the computer controls actually changing something in a power supplied it would be great to find out. The engineers on this site will quickly set me straight if I am off and we can both get some new ideas.

John

Hi Pujan,

   I think there is a misconception here. Power modes are achieved not by limiting the supply current from the supply but by reducing the conception from the load. Let me elaborate it for you.

In your case you wanted to achieve sleep mode. Then CPU gives commands to most of the peripheral controllers to go to sleep mode after that CPU itself will go into a state (no clocking)

where it wait for an external interrupt to wakeup. So its doesn't work like, say, right now your system draws 3Amps and you wanted to achieve low power mode and you limit the current from the source to 0.1Amps "No"!

But power supplies in computers are very intelligent they are called PMIC ( Power Management Integrated Circuits) They will take one input and generate multiple rails of LDO(Low Dropout Regulators) and DCDC Converters.This chip will be hooked to an I2C bus or some other communication bus to the CPU. All the peripherals will be powered from this IC. CPU will have the absolute control over all the functionality of these ICs. CPU can shutdown any rails using I2C commands.

In your case you can use a MOSFET to turn on or off the rails to a peripheral and there by achieve low power.

Regards,

Aswin

Thank you Douglas Wong for your detailed reply.

I completely understood what you have mentioned related to signal which controlled through CPU and that drives power supply.
Now next question to follow on this is:

Let's see CPU is able to send signal to power supply circuit as per required scenario. Now my question (this is what I wanted to know in my first post) is what kind of circuitry helps power supply to go in active mode or non-active mode? In other words, what would be block diagram of circuit (between CPU and PS) that tells power supply to act?

I hope I am asking correct way.

Thank you once again for your inputs.

Pj

Thank you John for your valuable inputs.

Yes, I thought the same that power supply would react based on "load demand". So having said that, can you help me or suggest me how that circuit looks like which satisfy our (load demand) needs?

Thanks once again,

Pujan

Great explanation Aswin Venu. I really appreciated how you have clearly explained.

I am very much familiar with I2C, SPI etc protocols that makes communication between peripherals. I have never worked on PMIC so would like to understand more in detail, specially with circuit/block dia-gram and/or videos if possible.

Now let's say if I am using MOSFET then there is no I2C as MOSFET is an analog component. So in this case, how would you tell MOSFET to operate?

Thanks once again,

Pujan

Hi Pujan,

If the question is related to a MOSFET in a laptop power supply then it is about switching and efficiency.

A MOSFET in the power supply is switching ON or OFF at the switching frequency, lets say 100kHz. This will generate losses that are present even at low output current.

The MOSFET is not jut by itself, it is controlled by a controller, analog or digital. The controller will keep the voltage constant at the output when the current is changing. The controller can have (or not) an enable input. Also, the output of the power supply has a capacitor so that even if the MOSFET will stop switching, the output voltage will decrease slowly, depending on the current drawn from the laptop.

In order to increase the efficiency, when the laptop enters sleep and the supply current will drop below a small value, let say 10%, you can stop the MOSFET from switching to save power. Once this happens, the output voltage will start to decrease, but slow because the current is below 10%, so you can wait some time. When the voltage drops below a set value, let say 95% of nominal, then you start the MOSFET again, the voltage rises to the nominal value.

There are several ways to implement this circuit in a generic way, that is, if it was not included in the controller already.

The easy way is with a comparator with hysteresis that will generate a enable signal when the voltage is at 95% and disable when the voltage is at 100% (enable "0" or "1" logic).

This may not be desirable at high power because you will always have a 5% voltage ripple at the output.

To resolve this you need another comparator and a current sense resistor in the GND wire. The 2nd comparator will generate a signal when the output current is smaller than 10%. This signal can be an enable for the first comparator so that only when the current is smaller than 10% the voltage will be allowed to drop to 95% while the MOSFET is not switching.

If the controller does not have an Enable pin, then you can use the same principle but you can act directly on the feedback pin of the controller, basically you will change the nominal voltage from 95% to 100% with a comparator by adding resistance in the feedback divider of the controller.

This control can disable the MOSFET for long periods of time if the current in the laptop is very small. You can have also other means of communication between the laptop and power supply but this is about how to sense the power.

Regards,

Stefan

Thank you so much Stefan for an excellent information and your way of explaining. I would definitely like to try this out (at-least simulation).

I really appreciate your time. Do you have any good informative material on PMIC? If so, could you refer me?

Thank you and stay safe,

Pujan

Hi Pujan,

I think the best way to find up to date information about PMIC is to follow the integrated circuit

That is, you can explore the offering of Farnell for example, and see what type of PMIC they have available. You can see some classification, for example DC-DC, off-line, controller with or without integrated power transistors, etc. You find the datasheet and the manufacturer this way.

One other way is to go directly to the IC manufacturer sites (like Vishay, STMicroelectronics, Texas Instruments,  ONSemi) and browse the converter categories. You will find there also additional materials like application notes.

Regards,

Stefan