CC2340R5: How to reduce the power consumption of cc2340

Hello Yi,

Your code is matching pretty closely to the rfPacketTx example, so it is pretty close to the minimal current consumption for the given radio configuration, assuming all other HW peripherals (e.g. UART, SPI) are not used in this test.

yi feng said:

1：Is it necessary to close the rf and then open it when switching channels, i.e. call RCL_close() and then RCL_open()? This may consume more time increasing power consumption.

No, it is not necessary to call RCL_open/close. The RCL will only be active if there is a current command being handled.

yi feng said:

Can there be a way to increase this percentage?

I assume you want to minimize overall energy consumption. Since voltage typically stays constant, you can just look at current consumption.
Also, to reduce percent of energy_tx/energy_tot, either reduce energy_tx or increase energy_tot. Increasing energy_tot is trivial (just keep the device in active), but also consumes more current... I assume the main goal is to reduce overall current consumption (which reduces overall energy consumption).

To reduce current consumption, based on info provided, some of my suggestions are:

1. As discussed in first point, you don't need to call RCL_close and RCL_open. No tradeoffs.
2. Reduce TX power (tradeoff is range)
3. Reduce TX packet size (tradeoff is data throughput)
4. Reduce amount of code leading up to the RCL_Command_submit: Have 3 separate txData (or txPacket) with the channel field already set (tradeoff is memory consumption)
5. Increase the interval between the TX sequences. Each TX sequence is assumed to be 3xTX as you show (tradeoff is throughput)
6. Reduce amount of time spent in Active mode

Thanks,
Toby

Hi Toby,

For the first point, I would like to know why you say it is possible to do without calling RCL_close and RCL_open?Don't I need to modify PBE_GENERIC_RAM.WHITEINIT after I change the frequency? Then call RCL_open in order to pass in the value of PBE_GENERIC_RAM.WHITEINIT.I changed the frequency and PBE_GENERIC_RAM.WHITEINIT but without calling RCL_close and RCL_open and then using wireless sniffer, I can't receive data from the changed channel.

2、3、5：Our request is that txpower and txpacket not be lowered any further.The frequency of transmission cannot be reduced either。

4：I tried to reduce the code before RCL_Command_submit (no more updating the data, sending the same data for each channel), but that doesn't seem to help either!

Then I tried other methods and created three rclHandles corresponding to three different channels and PBE_GENERIC_RAM.WHITEINIT. as in Figure 1

Then, I noticed that the time interval between sending the two channels did become shorter, as shown in Figure 2,But the startup time before sending the first channel becomes more. As in Figure 3, then the total time to send does not change.as in Figure 4

The following figure shows the measurements of the very first code

Hello Yi,

Understood there are application requirements of what techniques can be used to lower overall energy consumption.

Can you share the code for switch_white?

It may be that this is changing something in the radio at a level requiring a re-load, which is expected to take more time.

Thanks,
Toby

This is to modify the PBE_GENERIC_RAM.WHITEINIT in the configuration.And then loaded in via RCL_open

There is also a problem that when using this program, when the chip is near a conductor (e.g., a small magnet, a hand, an iron cabinet) an abnormal current waveform occurs, resulting in increased power consumption.figure below，This anomalous waveform has a frequency of 50 Hz。

It looks like this when it's not near normal

Hi Toby,

This abnormal current waveform is caused by the fact that the voltage source we are powering the chip with is not grounded.

Thanks,
Yi

yi feng said:

This abnormal current waveform is caused by the fact that the voltage source we are powering the chip with is not grounded.

Does this go away if the voltage source is grounded?

yi feng said:

This is to modify the PBE_GENERIC_RAM.WHITEINIT in the configuration.And then loaded in via RCL_open

Ok, in this case then it is normal to have more time spent in the the RCL_open, because the registers would have to be re-loaded. (This will happen even if you only call RCL_close then RCL_open).

Note that LRF_commonRegConfig is meant to be constant... meaning that RCL expects this value to not change. But if you do change it, then yes, a RCL_close and open is required.

No anomalous waveforms when the voltage source is grounded

Thank you for confirming.