5
\$\begingroup\$

My problem is that I have a RF chip which is supplied 3V3 and interfaces with my MCU over 1V8/2V8.

I have an external source of 3V3 which will supply the RF chip. I'm space restrained on my PCB and as such would like to have the least amount of external extra components. This means I prefer not to LLC the UART lines between the MCU and RF chip.

My MCU can run on between 1V8 and 3V3. As such it would be possible to run the MCU at 1V8 and bypass the LLC question.

Would it be possible to run this MCU using a resistor divider circuit wherein the 3V3 is converted into 1V8?

My MCU sinks 142 µA/MHz at 3V and runs at 24MHz making 3,408 mA at 3V. (5,68mA @ 1V8) Which is a relatively small amount of current which makes me think the normal problems with resistor circuits as supply source (e.g. variable and highly subjective voltages) would be possible.

Is this line of thought correct or would I be better of using an 3V3->1V8 LDO?

| improve this question | |
\$\endgroup\$
  • 5
    \$\begingroup\$ This is never a good idea, so the answer to when you should switch is "right now". That calculation on current consumption is just a typical sort of average, the actual figure will vary wildly and produce massive changes in supply voltage. \$\endgroup\$ – Finbarr Nov 18 '19 at 10:24
  • 1
    \$\begingroup\$ You said your MCU can run on voltages between 1V8 and 3V3? In that case, why do you need to step it down to 1V8? Why not just keep it at 3V3? \$\endgroup\$ – Curious Nov 18 '19 at 10:41
  • 2
    \$\begingroup\$ "and interfaces with my MCU over 1V8/2V8." So it can't be supplied 3V3 over the UART lines or I'll blow it up as Absolute Maximum rating is: General digital Interfaces: min: -0.2, max: 3.0. (normal operation) VCC is defined by: min: 3.2, typical: 3.8, max: 4.2 \$\endgroup\$ – Tarick Welling Nov 18 '19 at 11:02
  • \$\begingroup\$ As others have said, don't use the divider :) Please see this question and its answers. \$\endgroup\$ – bitsmack Nov 18 '19 at 21:16
9
\$\begingroup\$

It is never a good idea to power anything from a resistor divider. Any variation on the load will change the voltage, and your MCU will very likely vary its current consumption based on what it is doing at any time. This means your supply voltage will be affected, and as you are wanting to go to 1V8 (the lowest acceptable voltage) then this is an extremely bad idea.

You can get LDOs in very small packages if it is space constraints you are worried about. Getting a fixed version means no external components except capacitors.

To sum up though, while it is theoretically possible to do it, it just isn't worth the hassle if/when problems start to occur, which is quite likely.

| improve this answer | |
\$\endgroup\$
4
\$\begingroup\$

Maybe? If the amount of power wasted in the divider is much more than that of the MCU then it will be more stable. You could try a 30ma divider. Make sure you have enough capacitance near the MCU.

But why? Saving the pennies at cost of reliability? LDOs are a simple three-terminal drop-in component and can be extremely small for that amount of current.

| improve this answer | |
\$\endgroup\$

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.