I am designing a DCDC Buck regulator for 12V to 3.3V system power supply. My controller; bluetooth & some other peripherals have following nominal currents on 3.3V rail as per datasheets (Details given below).

There is a required parameter of Io(max) for DCDC regulator. How to calculate maximum required current for circuit?

CAN Transciever: Nom: 40mA Max: 70mA Bluetooth LE Chip (+5dbm Transmission): Nom: 21mA Max: 32mA Microcontroller(@ 16MHz): Nom: 10mA Max: 28mA GPS: Nom: 46mA Max: 100mA

I would like to know which value should be taken to calculate maximum load current on this rail? Is it nominal current or peak current?

We generally follow: Io(max) = I(nom)*1.5. Is it OK? Because calculating with Max current ratings in datasheet will generate very high & way off current value for regulator current calculations as number of peripherals on line increases.

E.g. Here in this case: Nom (Total) = 117mA & Max(Total: 230mA)

**It is also observed that system current never reach to Max (total) current with full load on system.

Hence I would like to know which current value from datasheet should be generally considered for power supply rail current calculation & How to calculate Maximum rail current based on it?

  • \$\begingroup\$ Bear in mind that microcontroller figures do not include switching losses (which although small, will still have an effect). If you identify all the modules, we can look at what the datasheets may say about why a particular maximum current has been specified. \$\endgroup\$ – Peter Smith Sep 7 '18 at 14:35
  • \$\begingroup\$ Hi @PeterSmith, thanks for your valuable suggestion. Currently these are only main current consuming devices in my system. Others are very small devices such as LEDs. So I am not considering them as a part of this question. I would like to know "How in general; maximum power supply rail current is calculated? Based on nominal current or max currents in datasheet? & How much buffer to be kept for nominal current and max current calculation for design. \$\endgroup\$ – Akky Sep 7 '18 at 14:42
  • 1
    \$\begingroup\$ You write "calculating with Max current ratings in datasheet will generate very high & way off current value", but the difference is not even a factor 2. That's hardly way off. \$\endgroup\$ – pipe Sep 7 '18 at 14:42
  • \$\begingroup\$ @PeterSmith how should I get value for switching losses in microcontroller? \$\endgroup\$ – Akky Sep 7 '18 at 14:52
  • \$\begingroup\$ You should test it at maximum internal ambient temp for max CMOS current consumption as CMOS runs slower and draws more current. Then consider load regulation error and Tcase to not exceed 65'C at worst case ambient for a reliable design. or whatever is recommended for RF tuning error. \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Sep 7 '18 at 15:00

I think you're overthinking this way too much.

1) what are you hoping to gain by designing the DCDC converter to exactly match the current consumption of the system?

2) If you dimension the DCDC converter such that it cannot deliver the current required you risk system instability (it might reset). So the safest choice in my opinion is to use the Maximum current and add some margin. So here that's 230 mA with some margin so 300 - 400 mA. From my experience, that number sounds about right for a BT + GPS system.

3) The system's current consumption is NOT constant, it will vary with activity. When everything is active it will consume more current compared to an idle situation. The software in the microcontroller has a huge influence on current (power) consumption. Imagine the difference between updating the location every second or once per hour.

4) Since power consumption varies so much, for battery powered applications it can be useful to select a DCDC converter that has an efficient low-current mode. Some DCDC converters use a PWM mode for high currents and a PFM mode low currents. In PWM it switches continously, in PFM it switches in short bursts.

5) If you are not concerned with efficiency you could use almost any DCDC converter capable of delivering at least 300 mA.

6) Only after you have your system up and running will you be able to measure the actual current consumption and then you might be able to optimize the value of the inductor in the DCDC converter to gain a little bit of efficiency.

Oh a, final tip: DCDC converters can be tricky beasts. There is no shame in just using a ready-made module. There are cheap ones from China but also direct drop-in replacement modules (for the 78xx series of linear regulators) from reputable manufacturers. That will save you the hassle of trying to get the DCDC converter working when it doesn't work for some reason. In some teardowns of very expensive (professional) equipment I have seen ready-made DCDC modules being used, the designers of that equipment wanted something that "just works" so they used a module designed by DCDC experts. It might be a little more expensive but when in a hurry, it might save you b#tt!

  • \$\begingroup\$ thanks for valuable input. I would like to know in general; should I consider Max. values given in datasheet for Max. system current calculation? \$\endgroup\$ – Akky Sep 7 '18 at 14:50
  • \$\begingroup\$ Uhm, it depends. You need to look at your system and think about where the power is used. What lights up (LED, light bulb) emits heat, activates a (relay) coil, emits an RF signal. Many ICs do not use much power (like a microcontroller) but if you make that uC light up a LED (without using a transistor for switching) then the LED's current goes into the uC's supply input. That current cannot exceed a certain value and that's the rating listed. The uC will only hit it's max. current rating if you use it in a particular (a bit wrong) way. \$\endgroup\$ – Bimpelrekkie Sep 7 '18 at 14:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.