CC2640R2F: Measuring BTLE current draw with a scope

PATRICK OHARA

Part Number: CC2640R2F
Other Parts Discussed in Thread: INA188, CC2650, CC2640

Hello all,

In an effort to measure current draw from my CC2640r2, I assembled the following circuit to get the voltage levels up to a point that they could be measured. I used a TI INA188 instrumentation amplifier for this purpose. Notice it uses a single gain resistor to set the gain. By using a single fixed resistor and trim pot I was able to dial in the amplifier gain to exactly 100 @ 1Khz. Also, notice that the inputs to amplifier are across a 1 ohm resistor. This allows me to measure voltage across the shunt resistor. The CC2640r2 LaunchPad is powered from an external bench power supply connected to the 5VDC pin on the board. The circuit shown below is inserted in the circuit return path as shown. Of course there are instruments that will do all of this for you, but they're expensive. This configuration allows measurement of the parameters needed to make a pretty accurate estimate of current draw and thus battery life.

This allowed me to make some measurements of current draw as a function of time with sufficiently high resolution on the scope to capture individual BTLE connection events. Following shows the device in advertising mode. Notice that the advertising period is set to 100mS. The RMS voltage is 0.8mA (remember I'm measuring through a 100X amplifier so 77.7mV measured is actually 0.8mV)

Following is a measurement of current while the CC2640r2 is connected via BTLE to a phone. Notice that since the connection events occur closer together ~30mS this condition draws more RMS current.

The other nice thing about using the scope to look at the current draw is that the individual connection events can be measured.

over 8 years ago

0 JXS over 8 years ago

TI__Guru** 105180 points

Hi Patrick,

That is some fine engineering!

What do the DC/DC recharge pulses look like in your scope?

Best wishes

0 PATRICK OHARA over 8 years ago in reply to JXS

Genius 3205 points

Hi JXS,

I am not able to fully decouple the CC2650r2 from the on board debugger. So the actual current draw is around 70mA, but my scope is DC coupled so that I'm only seeing the current draw from non-DC events like the radio. I think the DC/DC recharge pulses may be too small to see and buried in the DC signal. Please let me know if I'm wrong. I tried to disconnect all of the jumpers assuming that the CC2650 would be the only thing operating as was suggested in an alternate post, but unfortunately with all of the jumpers disconnected, the CC2650 doesn't function. I'd be happy to make some more measurements - suggestions?

Thanks,

Patrick

0 PATRICK OHARA over 8 years ago in reply to JXS

Genius 3205 points

OK, I may have found the DC-DC recharging that you suggested may be here. Unfortunately because I can't completely isolate the CC2640r2 from the on-board (CC2640r2 Launch Pad) debugger, I can't be sure that the signal that I measured is from any identifiable source - it could be the DC-DC converter but I'm not sure. Furthermore, these sequential scans are from paired vs unpaired conditions. I find it very interesting that 100mS current draw is larger while not paired with the phone. This makes me think that the cause is something unrelated to the CC2640r2.

Thoughts?

Notice the relative slow (~1kHz) broad current draw. This may be due to the DC-DC recharging event. A note, the scope is set to acquire in average mode while taking 128 averages, thus the fine detail, but in actuality the small broad current events go to nearly zero when paired as the device is in this case. This is a very nice technique to watch the current state change slowly while the device is paired and then unpaired.

Here with the device unpaired, the slow current events are more prominent.

Simply included to show a single BLE pulse as measured on the scope

Lastly, showing the performance of the amplifier and verifying its gain of 100 at 1kHz.

0 Fredrik K over 8 years ago in reply to PATRICK OHARA

TI__Guru** 106705 points

Hi Patrick,

Nice work!

See the very bottom of this page on how to apply an external power supply to the LaunchPad: dev.ti.com/.../

This will decouple all of the debugger part except for the level shifters. These only draw a few hundred nano amps though.

With this setup you can still debug and you can power the debugger part from the USB connector (only one external supply required).

Regards,
Fredrik

0 Nathan Conrad71 over 8 years ago

Prodigy 220 points

I've been curious about hooking up a current sense amplifier, but never got around to trying it. My solution ended up being to use a HP66309D to record the advertisement pulses, though it doesn't have a long enough record length to know how often the pulses happen. I posted an image of a capture at www.eevblog.com/.../

My board was a CC2640 (older version), plus some other stuff drawing current, so the idle current was somewhat higher than it should be.

0 PATRICK OHARA over 8 years ago in reply to Nathan Conrad71

Genius 3205 points

Hi Nathan,
A quick look at your screen shot shows that your measurement system HP66309D evidently doesn't have sufficient bandwidth to capture the detail found in the BTLE pulse. This combined with limited logging memory makes this instrument unsuitable for your application. Might I suggest the Tek MDO3014. It's the best darn scope I've ever had.
Patrick

0 PATRICK OHARA over 8 years ago in reply to Fredrik K

Genius 3205 points

Hello Fredrik,

Thank you for your kind words of encouragement. I took your advice and configured the CC2640r2 as you suggested - only the VSENSE jumper, GND to my circuit, and 3.0VDC from my lab power supply to the board connected. The results were similar to those shown previously but without the DC offset - likely due in large part to the LED. For completeness I've attached my results here.

With the device unpaired (disconnected) the CC2640 draws ~ 1.1mA RMS. Notice the small spikes. I'll show those later. This may be the DC-DC converter.