2xAAA battery => 3.3v/250mA step-up converter with logic-driven power control [superseded]

Question

EDIT: The requirements in the question below appear to be practically unsolvable, but happily after a bit of digging I actually found an alternative equivalent RS232 device which runs off a 3.7V Li-Po instead of 3.0V AAAs.
Moral of this story: hold off purchasing for as long as you can - it can prove extremely beneficial :)
I've asked a new question for my new set of requirements.

---

I need to connect a Wi-Fi/RS232 module to a low-powered monitoring device which needs to periodically fire up, "ping" a server, then shut down again. For maximum portability, I'd like to power the Wi-Fi module via the pair of AAA batteries built into the device itself.

This decision was considered viable based on the fact that the device in question is very low-powered and takes just over a month to drain its batteries under normal use, and the fact that the Wi-Fi module will only power on for a few moments at a time and will generally use relatively small amounts of power. I'm very confident that I'll be able to get at least a few days' worth of battery life with the Wi-Fi module attached, and it won't be a problem to put new rechargeables in the unit every so often.

Now, the Wi-Fi module I've selected draws 3.3v at a peak of 250mA, so I need a step-up converter to run it off two AAAs. However:

1. The converter needs to fit inside a compartment 50x54mm in size, approximately 4-5mm thick. For what it's worth, this area is insulated, so insulation thankfully doesn't need to be taken into account.
2. The unnecessity of insulation aside, the step-up converter needs to produce virtually zero EMI/RFI due to the electromagnetically sensitive medical environment the device will be used in, which won't tolerate even low levels of continuous EMI/RFI radiation particularly well (although short Wi-Fi bursts will be fine).
3. I've never dealt with step-up converters before, and I'm unsure what effects the converter's electrical behavior/characteristics might have on the other electronics (a CPU, etc) in my RS232 device. What might I need to consider in this regard?
4. The Wi-Fi module draws 8mA when idle, which would kill a pair of AAAs fairly quickly. I can control the serial device's DTR line like a GPIO, so it should be possible to set DTR high or low to electrically switch the step-up converter on or off. I have absolutely no idea how to go about this electronically however, although I can say that the RS232 device uses TTL signal levelling, instead of standard RS232 voltage levels.

I must admit that SMD component-based solutions wouldn't work out for me very well; I have passable non-SMD soldering skills, but I've never worked with SMD components, and if possible I would like to build this device myself, preferably successfully.

If an inexpensive premade power converter module exists that meets the above requirements, that would be a very attractive option.

Alternatively, since the adaptor/converter won't be being moved after it is fitted, directly soldering non-SMD components' legs to each other would also be an acceptable solution.

Answer 1

1. Size constraints: The challenge would be the 4-5mm height, rather than the other two dimensions: A boost converter will typically require an inductor, likely to be the tallest part.

   For instance, this prebuilt 1 V --> 3.3 V boost converter module is 12x22mm, but the inductor on it is 6 mm tall:

   

   This example does not offer an enable pin, though - My cursory search shows several 5 Volt and 12 Volt boost converter modules with an enable pin, but not for 3.3 Volt output.

2. EMI constraints: This is a tougher problem. A typical boost converter works by switching currents on and off using an oscillator, hence even with the filter capacitors on the board, there will be some EMI at the oscillation frequency and its harmonics. These modules certainly are not medical-grade electronics. "Virtually zero" is virtually impossible.

   An option is to use a 5 Volt boost converter like this one, followed by a low drop-out, low-ripple linear regulator to produce the desired 3.3 Volts. The 3.3 Volt rail then will have negligible noise.

   

   Direct EMI radiation, especially from the boost converter's inductor, will still be an issue. To address this, shielding the entire assembly with a metal box or mesh would help. Only your own testing will reveal whether this serves the purpose sufficiently.

3. Microcontrollers typically work perfectly fine even with fairly noisy (50 to 100 mV ripple) power rails from boost converters. Also, a hybrid incorporating an LDO, as suggested above, addresses this issue completely.

4. The "enable" pin in the 5 Volt boost converter suggested above could be toggled by the DTR pin - but then how would you be supplying power to the serial device when the boost is off?

   For similar applications, I have used an unregulated supply (just capacitors) from the AAA cells into a minimal microcontroller such as the ATtiny85, and used the microcontroller's output pins to toggle the boost controller supplying the rest of my device. Consider whether this works for your design.

Answer 2

Could you use a 'two step' approach?

Use a cheap and small microcontroller with (non-wifi) wireless radio, and then a 'repeater'-like device that forwards the 'ping' to Wifi?

You could then use a JeeNode Micro ATMega based radio module and a PC, laptop or even something like a Raspberry Pi with a radio receiver and a Wifi module.

The JeeNode micro is really low power and can run of a coin cell or two or three AAA/AA batteries. If the 'ping' is small and only very infrequent it can run for months on a coin cell.

All these use standard off the shelf controllers (ATMega / ATTiny) and radios (Hope RF) and can be cheaply built on breadboard/perfboard if needed.

Answer 3

After learning how it's virtually impossible to obtain practically zero RFI/EMI from a switching design, I actually did a bit more research and turned up an alternative (equivalent) target device which runs off 3.7v instead of 3.0v, so all I need to do is remove 0.7v now.

I've asked a new question for my new set of requirements.