The datasheet of CC2652R1 gives clear information in table 5.3 
"Recommended Operating Conditions". It runs fine with 1,8 - 3,8 Volt.

Lower voltage brings lower energy consumption. Don't forget that the 
nominal voltage of a battery depends on the technology and charge state.

Two 1,5 V alcaline cells have 1,8 - 3,2 V.

A reachargeable Lithium cell has 3 - 4,3 V, which is too much, so that 
requires a voltage regulator.

VV V. schrieb:
> Thanks a lot.
>
>> A reachargeable Lithium cell has 3 - 4,3 V, which is too much, so that
>> requires a voltage regulator.
>
> So to what voltage should I regulate it down?

Something between 1,8 and 3,8 Volt - obviously.

I assume that there is more than this single microcontroller, so you 
should check the requirements of the other parts as well. In case of 
doubt, use 3 Volt. Because that's a common standard within the allowed 
range.

Be careful with the selection of the voltage regulator. This device has 
an RF interface, which usually means that all they consume short peaks 
of much more current than written in the datasheet. You need a voltage 
regulator that can follow these peaks quickly enough. Also you possibly 
want to reduce the standby consumption to a minimum. Only some few 
voltage regulators support both together.

The HT7830 is one of them.

VV V. wrote:
> What is the aspect in the datasheet which tells me, that it can follow
> quickly enough?
> During sending, my device consumes 11mA and usually an average of 2mA.

You wont find that in the datasheet. You need to make your own 
measurements in the target circuit. Some chip manufacturers provide 
separate documents that tell you how to design PCP and power supplies 
properly.

To give you an impression what I mean with peaks in the power consumtion 
take a look at this website: 
https://www.ondrovo.com/a/20170207-esp-consumption/

This WiFi microcontroller is specified for average 70 - 80 mA in active 
mode but it consumes up to 435 mA while sending. You 11 mA are surely 
also an average value which does cover these peaks.

> You 11 mA are surely
> also an average value which does cover these peaks.

I did measurments and this is the maximum current it reaches. See the 
attached image. Average value 2mA, peak 11mA.
The specification in the datasheet says it takes 9.6mA when 
transmitting.
I only have these measurments and I thought I could trust them.

So what qould you recommend?

I assume that you did not measure properly. The peaks are so short that 
you cannot even see their waveform. You need to measure them with a much 
shorter time base.

I saw in the datasheet that the RF output power of that chip is only 5 
dBm which is about 3.2 mW. This is far less than I assumed, so your 
measurement might be Ok.

Anyway you still need a voltage regulator that can handle the peaks 
quickly enough, or the voltage would become unstable. Many Low-Drop 
regulators are slow, keep that in mind.

As I wrote, the HT7833 can be used.