Particle Photon consuming ~3ma in deep sleep, firmware version is 0.6.2


#1

I am using the below, simplified version of deep sleep example. Hardware setup is photon board with no peripheral connected and lipo battery connected to vin. My multi-meter shows ~2.95ma current when photon goes into deep sleep. Photon firmware version is 0.6.2.

include "Particle.h"
unsigned long firstAvailable = 0;
void setup() {
}

void loop() {
	if (firstAvailable == 0) {
		firstAvailable = millis();
	}
	if (millis() - firstAvailable > 30000) {
		delay(2);
	    System.sleep(SLEEP_MODE_DEEP, 30);
	}
	delay(1000);
}

#2

There are several threads about this already and one lesson to take from them was that a DMM is not a good tool to measure this as the measurement itself impacts the current draw.

Measuring µA currents isn’t an easy task.


#3

Thanks for response, but unfortunately none applies in my case. My DMM is very much capable of measuing uA currents. Also, although true that measurement impacts current (we all know Heisenberg’s uncertainty principle :slight_smile:) but as far I know current measurement introduces a tiny resistance in series and hence if anything, should lead to drop in current.

Anyhow, what I discovered (and then just now confirmed from an old post, ([Resolved] Photon Dead After Low VIN) that current drawn in active as well as deep sleep states is function of voltage at vin. And unfortunately, as voltage drops, current consumption increases (reported in the shared link also).

When I powered photon via vin = 5v, current in deep sleep was 70uA. And when I dropped the voltage at vin to ~3.6, the current in deep sleep increased to ~3mA.

ashish


#4

That current scaleing is not very linear. There is a big gap between 70uA and 3mA.

A good tool for measuring microamps is the u currentnGold tool.


#5

Thanks for suggesting the tool.

Is it true that current consumption is inversely (may in linear or non linear fashion) proportional to voltage at vin? Could you shed some light on it? Also, Is it generally the case for any step down regulator or is it specific to RT8008-33GB (of similar regulators)?


#6

I have a uCurrent Gold meter that I could use to verify your results. I’ll give it a try and report back what I see at different voltages.


#7

I have my own ideas but @bko could probably explain it better than me.


#8

Hi @ashishgarg

Don’t forget there is a 5V --> 3.3V switching regulator at the input to the Photon, so if you drop Vin to the 3.6V minimum, you are putting this device into a funny state. I think the switch regulator is entering a linear mode trying to produce 3.3V from 3.6V input and the regulator itself is consuming the extra current.

It would be easy to verify this on a modified Photon, but difficult on a stock one. You could back drive the 3.3V pin with a 3.3V source but the regulator will consume current in that mode too. Removing the regulator entirely would work, of course.

Also, your DMM may have a uA scale, but making good measurements in that range is hard, usually requiring a Kelvin connected shunt. If you paid less than about $2000 for the DMM, I would be suspicious of the numbers it produces in the under 10uA range. The uCurrent Gold is nice; it is really a low-burden amplifier attached to a Kelvin connected to a shunt.


#9

You are likely seeing the DCDC go into a permanent on state; this often happens with switchers when Vin is close to Vout. The datasheet is not so forthcoming about this from what I can see (eg example graphs of efficiency don’t go near to that state).

If you have an external 3.3v supply then use that and your ammeter to check? The regulator may not take much current in this mode, it’s a quick experiment

@bko HP/Agilent/Keysight 344xx meters are great down to sub microamp and cost way less than $2k. The main thing is you’re going to hit brownouts if the device wakes, but you can certainly put it in 1A mode, wait until the device enters sleep, then switch ranges.


#10

The HP/Agilent/Keysite 6-1/2 digit is $1400 new. The 7-1/2 digit is over $2800. I have mostly used a Keithley 6-1/2 digit that I like a lot.

My point was that a typical sub-$100 DMM isn’t really up to the task here.


#11

They do have some sweet meters/power supplies that I can’t currently justify buying right now but damn they nice. Love their vacuum fluorescent displays.


#12

I have an issue that I believe may be related to this thread - I’ll apologise in advance if it’s not. I currently have three photons running the exact same code. The code contains the following line:

System.sleep(SLEEP_MODE_DEEP,600);

On one photon that is running firmware version 0.6.0, I can consistently get 20+ days of battery life. On two other photons that are running firmware version 0.6.2, I routinely get a battery life of 6-10 hours.

The hardware is identical for all three photons. I am going to go back and flash the problematic photons with firmware version 0.6.0 and see if that fixes the issue. Just wanted to put it out there in case others are experiencing the same issue.


#13

Interesting. Please do verify and let us know.

If true @ScruffR knows the correct people to notify.


#14

@bjagpal, could you try one device with 0.7.0-rc.2?
If I’m not mistaken there was an issue with the radio module not powering down properly under some (rare) circumstances which shoule be addressed with 0.7.0-rc.1 following.

If this does not help, @rickkas7 might have some more input or defer to Brett.


#15

Running the code above on a Photon running 0.6.2 I consistently get 70 µA while in SLEEP_MODE_DEEP. I tried at 3.6V, 4.5V and 5.0V to VIN. I’m not sure why you’re getting 2.9 mA.


#16

So I tried loading the devices with the 0.7.0-rc.2 firmware and the results are the same. I have, however, realized that it is just one device (of three) that is experiencing reduced battery life. I’m going to look a bit closer at the PCB upon which this Photon is mounted; I’m beginning to suspect this is probably a hardware issue and nothing to do with the Photon module itself. Thanks.