Power Arduino with 1V

Hello,

sorry for my english!

I've searched for a while but could not find an answer for question. Before I start my tests, I was wondering if someone can help.

I know some of you may report that is not possible since datasheet states 1.8V of minimum Vcc, but I will explain the problem and the infos I got.

I need a uC for temperature control, but I only have AA battery (1.5V - fully charged and 1.0V - Almost dead). Since my system uses lot of current for the heater, it is preferable to stack all them in parallel. In worst case I will get 1V.

I can build a simple regulated charge pump to reach 3V or 5V and power Atmega328p, but many IC available need voltage higher than 1.5V. If I could only drive Atmega328p for a few moment, just enough to charge pump to 1.8V and 3.3V/5V. Later I can change the clock prescale and switch for 8MHz internal clock.

Looking deep into the datasheet, I found that the uC works, but it runs a rotine to check if clock is ok. If it is, it enables all the functions. Will digital pins work for that tiny period of time?

Is there someway of doing that?

Reference the section of datasheet where it says that… but that would happen before any user code runs anyway.

There are boost converters that work off single cell AA. Very little effort turned up the AAT1217 which claims to work on as little as half a volt… it is hardly unique. Be sure to study datasheet carefully - they often have sections where they reveal substantially lower performance under the conditions you care about… but yeah, just use an appropriate boost converters IC.

I did a project with PAM2803s to drive 3W leds off a stack of parallel alkaline batteries. Worked a treat, except for red, where a bit of enginuity was called for

DrAzzy: Reference the section of datasheet where it says that... but that would happen before any user code runs anyway.

Main purpose of the delay is to keep the AVR in reset until it is supplied with minimum V CC . The delay will not monitor the actual voltage and it will be required to select a delay longer than the V CC rise time. If this is not possible, an internal or external Brown-Out Detection circuit should be used. A BOD circuit will ensure sufficient V CC before it releases the reset, and the time-out delay can be disabled. Disabling the time-out delay without utilizing a Brown-Out Detection circuit is not recommended.

ATmega48A/PA/88A/PA/168A/PA/328/P [DATASHEET] pg.29

Not directly stated, but in many different parts it explain why it stop for voltages lower than 1.8V. The main idea is to disable time-out and BOD. The clock source is compromised, but I don't care, since it works for a little time.

DrAzzy: There are boost converters that work off single cell AA. Very little effort turned up the AAT1217 which claims to work on as little as half a volt...

As expected from a good design indeed, but where is the fun? =)

I will do some tests to see what happens.

That's not monitoring anything, it's just saying that during power on, you can set a delay to give power time to stabilize.

The reason you shouldn't use a voltage below 1.8V is because it's not designed for it. Atmel (now Microchip) put that as the recommended minimum voltage for a reason. If it was capable of running off of a single cell, they would have advertised that fact in the datasheet quite prominently.

I need a uC for temperature control, but I only have AA battery (1.5V - fully charged and 1.0V - Almost dead). Since my system uses lot of current for the heater, it is preferable to stack all them in parallel. In worst case I will get 1V.

I think you don't understand electricity well. Heat comes from power, and power is voltage * current. If you put 2 batteries in series to get twice as much voltage, you only need half as much current to get the same amount of power.

Your heaters are running off the batteries too? What are they?

Hi Jiggy-Ninja,

I thank you for your time and help. Sometimes is difficult to keep the post short and objective. I will try to clarify some of your presented points.

That's not monitoring anything, it's just saying that during power on, you can set a delay to give power time to stabilize.

I'm not saying it is monitoring. Datasheet is...

ATmega48A/PA/88A/PA/168A/PA/328/P has an On-chip Brown-out Detection (BOD) circuit for monitoring the V CC level during operation by comparing it to a fixed trigger level.

11.5 Brown-out Detection - Pg.49

The voltage level has the clear objective to keep it working nicely, it is presented in the datasheet. To be more specific, the clock source. If the voltage is too low, the clock source is bad and the instructions as well. As I mentioned, what I want to do is hack the chip. Disable initial checks, boost the voltage at a minimum voltage. After that, in code, active BOD, time delay, increase clock frequency and etc...

If you put 2 batteries in series to get twice as much voltage,...

I'm really sorry, but I don't know where I wrote that I need more voltage. If you read carefully my post, I wrote:

Since my system uses lot of current for the heater, it is preferable to stack all them in parallel.

If you look for any AA batery datasheet, you will see that it has a maximum short current capability. Putting them in series will not increase that capability. However, in parallel they are able to drive more current. If it is not clear, lets make an assumption: Think of a battery as an equivalent Thevenin. It has an internal resistance and a voltage source. If you put in series, you will reduce its current capability due internal resistance voltage drop for a constant resistive load. if you put in parallel. What happens?

My daughter (baby) is suffering from colics. The heater is for her belly and to keep the temperature constant. Thus I built a heater with Ni-Cr wire. I took care to keep the resistance lower as possible R=0.5ohm. With that design I am able to put 3,4,5... AA rechargeable batteries in parallel to run all over the night.

I think communication is complicated. I'm not a native English speaker, which difficult it. Please, be more patient...

I think communication is complicated. I’m not a native English speaker, which difficult it. Please, be more patient…

That’s fine.

My daughter (baby) is suffering from colics. The heater is for her belly and to keep the temperature constant. Thus I built a heater with Ni-Cr wire. I took care to keep the resistance lower as possible R=0.5ohm. With that design I am able to put 3,4,5… AA rechargeable batteries in parallel to run all over the night.

This is exactly the kind of misunderstanding that I was talking about. You don’t need to keep the resistance as low as possible, and you don’t need to put all of the batteries in parallel.

Here’s an illustration.

Suppose that you have two batteries in parallel providing 1V to a 1 ohm resistor. Through Ohm’s law (V=IR) and Watt’s law (P=VI), that works out to 1 amp of current and 1 watt of heating power. The 1 amp of current is divided evenly between the two batteries, so each supplies 0.5 amps.

Now put the two batteries in series, to provide 2V to the heater. With the previous resistor, that will give 2 amps of current and 4 watts of power, which is too much.

The solution is to change the properties of the heater and make it have a higher resistance. 4 ohms of resistance will take 0.5 amps from the batteries, which works out to 1 watt of heating. The batteries are also giving 0.5 amps of current each, just like the first set that were in parallel.

Whether they are in series or in parallel, the power draw is distributed evenly among the batteries. A parallel arrangement is capable of sourcing more current, but it also needs to give more current to the heater for the same amount of heating, which cancels out any benefit you think you’re getting.

Put 3 batteries in series and make a heater with higher resistance to compensate. 3 AA/AAA cells will power any of the 3.3V Nano or Pro mini variants comfortably.

Don’t forget safety! Have sanity checks so that you can detect is the temperature sensor is broken or malfunctioning so you can turn off the heater. Waking up from colic is better than her getting a stomach burn. Also make sure that even if it’s completely on, it’s not hot enough to injure her.

Hi Jiggy-ninja,

I know the maths and I want to keep the resistance because of the way batteries work. Not because P=VI.
Series resistance in battery will always cut the maximum short circuit current. Since I will have a control system, I want my heat to reach temperature real fast, lot of current, which is blocked from series batteries.

I may be mistaken, but set up a simple model (mathematical, I mean), modeling the batteries as a voltage source in series with a resistor, heater as resistor, etc, run the numbers for serial vs parallel - I don't think you actually benefit from the parallel configuration, except if running on a user selectable number of batteries with relatively small change in output is a necessary design requirement.

There's also the modern option of a lithium ion battery. Those could allow you to reach temperature within seconds - and with the wrong design, promptly and dramatically exceed it. Actually on second thought, maybe it's best to stick with AA's...

No need at all... Check the AA datasheet: http://www.jmargolin.com/furnace/Ultra-Power_AA_MX1500.pdf

Look for internal resistance at 20% of load (~0.3ohms). Imagine that I need 2W. For 3 series batteries. I = 2W/3V = 0.6A. R = 5ohm Now put that current in series resistance. V = 0.6 x (3x 0.3) = 0.5V

Doing the math again with voltage drop: P = (3 - 0.5)^2 / (5 ohms) = 1.25W

Lost of 0.75W. For a portable system (who ever had designed one) knows that it is a huge waste.

To solve this issue, I may need a higher voltage, more series batteries and operate with PWM to lower the voltage.

Anyway... I was looking the help to hack the uP.

thalesmaia:
No need at all… Check the AA datasheet:
http://www.jmargolin.com/furnace/Ultra-Power_AA_MX1500.pdf

Look for internal resistance at 20% of load (~0.3ohms).
Imagine that I need 2W.
For 3 series batteries. I = 2W/3V = 0.6A. R = 5ohm
Now put that current in series resistance. V = 0.6 x (3x 0.3) = 0.5V

Doing the math again with voltage drop:
P = (3 - 0.5)^2 / (5 ohms) = 1.25W

Lost of 0.75W. For a portable system (who ever had designed one) knows that it is a huge waste.

To solve this issue, I may need a higher voltage, more series batteries and operate with PWM to lower the voltage.

Do the math again with the batteries in parallel. You’ve either got 0.9 ohms internal resistance with a load of 5 ohms (series configuration), or 0.1 ohms in series with a 0.56 ohm load (parallel) to get the same amount of power. Both situations will waste exactly the same amount of power in the batteries’ internal resistance, so there is no efficiency benefit to parallel or serial.

Anyway… I was looking the help to hack the uP.

And you’ve been told that it won’t work. MOSFETs need a certain minimum amount of voltage to properly conduct. Below that voltage, and there’s no way it’ll work. Even the recommended minimum 1.8V requires a significant compromise to the maximum speed you can run at, so it’s not like there’s some hidden extra operating range that Atmel forgot to include.

If you’re lucky, you might be able to pull off 1.6V or 1.5V, but insisting on 1V is crazy and pointless. I’ve already demonstrated to you that parallel batteries provides no benefit over series.

I’ll repeat it again: You are correct that parallel batteries can provide more current than series, but because of the lower voltage the load requires more current to get the same amount of power. These two things exactly cancel each other out.