I was curious about the actual current draw and duration of an advertisement on the WICED Sense board with the 20737S chip so I did some testing and thought I'd share since I couldn't find any similar information on the forums. Tests were done with only advertising, no scanning.
Current is calculated by measuring the voltage over a known resistance of 10 ohms in the power path.
I found that given a normal ADV packet size (shown below) and an advertising interval of 100ms, the device drew approximately 25mA of current during the advertisement event for a duration of approximately 2.5ms. It drew roughly 4.5mA on average.
// total length should be less than 31 bytes
BLE_ADV_FIELD adv; // advertisement
adv.len = 1 + 1;
adv.val = ADV_FLAGS;
adv.data = LE_LIMITED_DISCOVERABLE | BR_EDR_NOT_SUPPORTED;
adv.len = 16 + 1;
adv.val = ADV_SERVICE_UUID128_COMP;
memcpy(adv.data, &test_uuid_main_vsc_service, 16);
// Tx power level
adv.len = TX_POWER_LEN + 1;
adv.val = ADV_TX_POWER_LEVEL;
adv.data = bleprofile_p_cfg->tx_power_level;
adv.len = strlen(bleprofile_p_cfg->local_name) + 1;
adv.val = ADV_LOCAL_NAME_COMP;
memcpy(adv.data, bleprofile_p_cfg->local_name, adv.len - 1);
The scope is set to 100mV/div with 2.5ms/div.
Single ADV pulse, 2.5ms duration.
Multiple pulses, 2.5ms in duration at intervals of 100ms.
The smaller pulses located between the larger pulses are most likely due to other operations being run. The ramp up and ramp down shown just before and after a single pulse is most likely caused by the Bluetooth stack forcing the device to sleep when not advertising.
I've attached the Excel sheet for reference, here are the images of the graphs converted from voltage samples to actual current values. The scope's ADC measures 8 bit samples which are converted using the equation in the Excel document.
Appreciate the contribution whitephoenix92
1. Was WICED Sense powered by battery or USB when you took the Tx current measurements?
2. Were any additional measures taken to isolate the Silicon Labs USB to Serial converter?
Note that we have also found ground plane issues on both TAG3 and WICED Sense that would cause irregular current measurements.
There is a new board coming out in Q3 that not only adds the ability to use an external DC-DC (i.e. 50% reduction), it also resolves some of the ground plane and component isolation issues found on the existing eval/dev boards.
The early measurements we have observed on this board are inline with the power numbers represented in the current BCM20737S Bluetooth Low Energy SiP Module Technical Reference
I think the original plan was to use a battery but instead we ended up using a 3.3V supply on the battery holder terminals.
No we didn't take extra measures to isolate the FTDI, I think since we powered it from the battery terminals we avoided the FTDI? I'd have to take a look at the schematic again.
Mostly I was just curious to know the duration of the adv tx event, peak current wasn't so much a concern but it will be nice having the ground plane issues worked out.
Hi, w.r.t. my response to whitephoenix92, is it the right way to achieve optimal efficiency for the on-chip 1.2V LDO with an external 1.8V DC-DC supply rather than 3V? I am concerning majority of the time is just with light load.
Appreciated for any guidance on this subject.
Sorry for the late response, we would recommend the DC-DC, as mwf_mmfae said, our new TAG4 board, that is coming out soon, will have the ability to take more accurate power measurements with the external DC-DC.
This is a very good subject. With what you have in your excel sheet, majority of the time is at light load. So it is important to look at the efficiency at light load on the supply. However, most DC-DC efficiency is defined at its nominal loading. Another aspect is whether 3.3V supply is a good choice for the on-chip 1.2V LDO. Should it be reduced to 1.8V? I want to check with mwf_mmfae as well.
Thanks for your response. Could you also advise how the on-chip 1.2V LDO can be bypassed with an external 1.2V DC-DC?
Do you mean I can use the BCM92073X_LE_KIT to by pass the on-chip 1.2LDO? Greatly appreciated if you could provide some guidance on how to do so with the dev board.
So I can begin the works now while awaiting the BCM20737L.
Unfortunately, you cannot use the TAG3/BCM92073X_LE_KIT to by pass the on-chip 1.2V LDO.
There will be another board released later this quarter (TAG4) that you can however use for this function.
I checked the TAG3 board, and found it does keep the option to bypass the internal LDO and replace with the external one. Do we need BCM20737 firmware setting to disable the internal LDO for lower power comsumption when the external LDO is used ?
We do not support an external LDO on the TAG3 board.
You will need to wait for the TAG4 to be relased later this quarter.
Please confirm if it's needed with BCM20737 firmware setting to disable the internal LDO when the external LDO is used ?
No firmware changes need to be made if an external LDO is used.