Using "Shutdown"/"Standby" mode on voltage regulators to conserve power on battery operated devices?

Mahonroy

Hey guys,
I am designing a battery operated device using the propeller, where the propeller will be sleeping for a few minutes, wake up to read sensor information, transmit data via a transceiver, then go back to sleep for another few minutes.

Is it feasible to have an external timer chip interface with the "Shutdown"/"Standby" mode on a voltage regulator? In this case, the propeller would do all of its necessary operations then using one of its GPIO's, bring the Shutdown to ground. This would cause the voltage regulator to cut power to the propeller/sensors/transceiver and would run in a mode that would have very little quiescent current drain (as low as 1uA depending on the voltage regulator). The external timer would eventually bring the Shutdown back up, and everything would power back on, at which point it would gather sensor readings, transmit data, then go back to sleep and repeat.

Is this a legitimate way of doing this? I am using 2x CR123 batteries (3.0V each) and want to get the maximum life out of them.

I had the option of using them in parallel for the 3.0V and no regulator, but they quickly drop down to 2.5V to 2.8V which is too close to brown-out for the propeller. It would be better to use a step-up/boost regulator for this.
The better option I came up with was to run the two batteries in series for 6.0V, then use a buck regulator to drop it down to 3.3V. From what I have read, this should be equivalent but slightly more efficient?

Without using the "Shutdown"/"Standby" feature of a regulator, how are you supposed to get a small quiescent current drain?

Thanks guys and any help/advice is greatly appreciated!

kwinn

Two batteries in series with the buck regulator is the way I would go. Avoids the inherent problems of batteries in parallel. You might save having an external timer if you used a diode and small supercap to allow the propeller to be powered long enough to do the timing.

jmg

Mahonroy wrote: »

Hey guys,
I am designing a battery operated device using the propeller, where the propeller will be sleeping for a few minutes, wake up to read sensor information, transmit data via a transceiver, then go back to sleep for another few minutes.

Is it feasible to have an external timer chip interface with the "Shutdown"/"Standby" mode on a voltage regulator? ...as low as 1uA depending on the voltage regulator.

Is this a legitimate way of doing this?

Sure, if you have a system that spends a lot of time waiting, then an external time chip can help. The current of that timer chip of course matters, but many RTCs can get under 1uA, or close to 1uA.

For a total system design, you need to consider ENERGY.

Total power removal, will require your prop to reboot, which means EEPROM read and boot energy.

You might want to look at stopping the clock to P2, or making it very low.

Take some measurements.
Data tables suggest 600nA Icc static, but graphs indicate 3-4uA (Brown out circuitry disabled ) ?

jmg

There is more low-power info in this thread
http://forums.parallax.com/discussion/129731/prop-limbo-how-low-power-voltage-can-it-go/p2

and the device I mentioned PCF85063A is now available in SO8 and TSSOP8

Be interesting to see what the Prop Icc is, with BOD off, and running on an external 1024Hz 'sloth clock'.

You could also look at the EFM8SB1 as a low power timer, that includes an ADC so could manage some battery monitoring too.
The SB1 could change the ExtCLK over a wider range than the PCF85063, from a very frugal 256 Hz to 24.5MHz (no PLL needed)

Tracy Allen

How accurate is your timing requirement and have you already included an RTC? My favorite RTC chip (ISL1202x) has a repeating interrupt mode that can be made to go low for 250 ms periodically at say 1 minute or 10 minute intervals. That could be inverted to a shutdown input on a regulator or to a p-mosfet power switch. An RTC would probably need its own 3.3V regulator, something like the MCP1702 with 1µA quiescent current. Oh, the circuit is complicated by the 3.3V i2c interface requirements to the powered-down Prop.

If the timing requirement is not critical, there are many approaches you could take. I'm recalling a system I made back in the 1980s that used the Radio Shack model 100, which had a "powerdown" command that brought it down to a few milliamps. That caused an SCR (silicon controlled rectifier) to drop out to cut the power off completely. But a timer using a PUT (programmable unijunction transistor) after a few minutes re-fired the SCR and turned on the T100 for another round. The battery backed clock in the T100 kept running even during power down to keep things more or less on time.

dMajo

I will do it with an RTC. If you search, you can find several RTC that have an alarm output that can be programmed to trigger at specific date/time or by masking parts of date/time every x second/minute/hour/day/month. Some keeps the alarm state on pin until cleared (eg through their spi interface). Current consumption when running the RTC on battery (looking also for alarm condition) is <1uA. The VBat pin usually accept or a battery or a supercap.

You can connect the RTC's VBat to your battery, the RTC's Vcc to the regulator output and the RTC's alarm to the regulator's output enable and, perhaps, to an propeller pin (just to monitor its state). The latest through a jumper that can be open when not under development. I don't know how much current can drain the unpowered prop through its pin.

When the alarm fires it will turn on the power supply through the regulator output. After the prop boots and do its job can clear the alarm in the RTC thus shutting down the regulator again. If you have a repeated alarm eg every 2min30secs you even not need to set the new alarm time and mask on the RTC, only a spi write to clear the existing alarm.
This setup needs a parallel switch/jumper (by-pass) to the output enable to allow the turn-on of the regulator when you need to program the RTC's alarm parameters (if done with propeller, usually the very first time or during debug/development).

As jmg has said you need to consider in terms of energy, if it costs more to boot the prop from its eprom or to freeze/hold it keeping the regulator running.
I think that the same alarm output can be used to enable the propeller clock. I don't know if through a single gate (and/or) in the crystal path (between crystal and XO) will somehow work, certainly can be done if using an external oscillator.

jmg

dMajo wrote: »

As jmg has said you need to consider in terms of energy, if it costs more to boot the prop from its eprom or to freeze/hold it keeping the regulator running.
I think that the same alarm output can be used to enable the propeller clock. I don't know if through a single gate (and/or) in the crystal path (between crystal and XO) will somehow work, certainly can be done if using an external oscillator.

It would be hard to disable a crystal Osc to sub uA levels. The feedback resistor will consume current, even if you drive hard one way, then you have the startup time.
External oscillator is probably best, and I would test first with the RTC supplying that via CLKOUT. (less parts)
PCF85063A(etc) can go down to 1024Hz, which allows the Prop to WAIT, and do its own timing.
Once Awake it can use RCFAST to finish quickly, then go back to Wait_1024

Looking more, the EFM8SB1 can generate 32768 out, but not a divided version, whilst PCF85063A(etc) can generate 1024 or even 1Hz. Prop wait is 1 SysCLK granular, so you could run at 1Hz
Allow some seconds after wait, for the few opcodes to enable RC-FAST...