# Is a 3.3V-5V MCU runnable at “nominal” 3.3V?

I've got some 5V MCUs whose rating is 3.3V Min to 5.5V Max. My question is whether running them at a "nominal" 3.3V is OK.

I can imagine that a 3.3V source runs typically at ~3.35V, so no problem. And I know that if the voltage drops to 3.29V the MCU won't suddenly quit - probably.

The reason I ask is that I have some 3.3V components that I'd like to interface to it, and they can't handle 5V as well as it can handle 3.3V. Given that most 3.3V components max out at 3.6V, maybe I should run them all at 3.45V (if I can design the non-standard power supply)?

I guess I'm asking whether 'tis better to use a "standard" voltage, or up it a little to satisfy both sides of the equation.

### Edit:

As requested: the datasheet. Please see page 423, under "DC Specification (5V MCU)"

• Link the MCU datasheet. Sometimes the notes on min voltage and tolerance take some "interpreting" and without it, we can only guess. – Brian Drummond Sep 15 '16 at 11:05
• Given that datasheet I would take 3.3V as an absolute minimum for guaranteed operation, and arrange for a higher nominal voltage. – Brian Drummond Sep 15 '16 at 11:28

Not really, at least not in your case.

This MCU comes in two variants: One designed to operate on a nominal 5 volt, and one designed to operate on a nominal 3.3 volt.

The datasheet is just a little bit odd, but you can draw some conclusions from a few places. Chapter 1, section 1.1: Features:

Operating voltage range: 5.5V ~ 3.5V or 2.2V ~ 3.6

Clearly, they think that the 5 volt variant is good between 3.5 to 5.5 volts. If they knew that it worked well for a lower voltage, they would have included it in the features. It's a selling point.

Let's look at page 423 which you linked to:

DC Specification (5V MCU)

VDD Operating Voltage Min: 3.3 Typ 5.0 Max: 5.5

Now, here the range is wider. Why? Not sure. There are usually some test conditions. The common way to interpret these figures is that they guarantee by design and/or testing that the MCU will operate correctly for the whole temperature range and clock frequencies if you keep the voltage between 3.3 and 5.5 volts. Anything else and you're operating on luck.

Now, look at page 54:

2.3.5 Power-On Reset (POR)

When VCC drops below the detection threshold of POR circuit, all of the logic circuits are reset. The nominal POR detection threshold is around 1.9V for 3V device and 3.3V for 5V device.

There you have it. If you try to operate this on a nominal 3.3 volt, and it goes slightly below that for various reasons, your MCU will likely reset.

• The 3.3v reset threshold seems to be a likely candidate for why they're recommending at least 3.5v at the start of the document; giving a reasonable safety margin up front for anyone who doesn't read the document exhaustively (or know exactly what gotchas need to be searched for). – Dan Neely Sep 15 '16 at 15:29

The specification states 3.3 - 5.5 V so that means that the manufacturer guarantees operation at 3.3 V.

Assuming the 5 V device (which you should not use, you should use the 3.3 V version):

Will it work at 3.29 V ?

Yes, very likely it will. In order to guarantee that 3.3 V there can be no hard limit just below 3.3 V so in practice most 3.3 V minimum ICs can work under typical conditions down to 2.5 V or so. Of course the maximum clock rate will be lower then especially at high temperatures.

Will it work at 3.2 V ?

Yes, very likely it will.

In practice if your 3.3 V does not vary a lot (it stays above 3.0 V) then it will just work.

But if you're going to make 10 million products with this CPU at 3.3 V then you might want to consider increasing that to 3.5 V just to be safe.

• Except maybe for the POR threshold at nominal 3.3V – Sean Houlihane Sep 15 '16 at 11:33
• Sure for the 5 V device but not for the 3.3 V one as it has a lower POR threshold. – Bimpelrekkie Sep 15 '16 at 11:35