2 AA batteries with step up converter to Vcc pin VS 3 AA batteries to RAW pin



  • Hello all,
    I'm still in the learning phase prior to making an purchases. I was planning to use mainly arduino pro mini 3.3V. I'm trying to keep things cheap, so I had one question:

    What is the reason the examples here seem to always mention using 2 AA batteries with a step up converter (which I assume would connect to Vcc) instead of 3 or more AA batteries connected to the RAW pin. This seems like it would eliminate the need to purchase the step up converter for every sensor. What am I missing?

    Thanks


  • Hero Member

    @hooraysimpsons
    The step up will suck as much power out of those batteries as possible.


  • Contest Winner

    However, a step up converter also has losses. Has anyone done the maths? I guess there is a breaking point in the total power consumption where overall battery life actually improves without a step up converter due to the leakage it introduces?


  • Hero Member

    Don't buy the 3.3V boost / step-up listed in the "store" if you intend to run it on batteries. It has a quiescent power draw of 1mA or so (a modern boost will have 5uA o so by comparison!!), which will drain even your 2xAA batteries in well under two months for sure no matter how lean you run the rest of the circuit. Unfortunately there is no better boost that I've found for sale. You'll have to either wait for the official mysensors hardware lineup which will start with a boost circuit. It should be ready any day now? Right? :-) @axillent

    Or you can run most of the sensors on just 2xAA (in series) directly, no boost or anything in-between. Sure the voltage will soon drop below 3V but it will most likely take months and most sensors as well as the nrf24 and arduino pro mini will cope with that. You wont be able to suck every last drop out of the batteries but it shouldn't be too bad. Far better than the 1mA quiescent.

    If you had a boost, and limited space, you could also power your sensor node with just one AA. Like the official battery powered sensor node that is forthcoming.

    If you're handy with SMD and can source a few capacitors and inductors you could always build some yourself. Texas Instruments has free samples available if you want to try. Still trying to get mine to work though :-/ I can't figure out what I did wrong here. I thought about starting thread about it but I'm holding out for the official modules.

    Btw. Battery Calculator: http://oregonembedded.com/batterycalc.htm


  • Hero Member

    isn't this duino tolerant up to 6V in input ?



  • @bjornhallberg: thanks for the warning! I was about to buy the step-up in order to drive my sensor but I think I will wait.

    One problem though, I'm planning on building quite a lot of battery powered temp sensors, and as far as I understand, the DS18B20 temp sensor require at least 3V, anything below that will potentially give false temperature readings. Has anyone got any ideas on how to best drive a DS18B20 based sensor using batteries?

    If I use 3 AA's, then I would have enough voltage to drive the DS18B20, but then I would have to step down the voltage for the nRF24L01+ transciever, right? SInce it accepts only up to 3.6V?


  • Contest Winner

    Some (if not all) arduinos have a regulated output (VCC). I know that at least the Nano have an onboard 3.3V regulator that can be used for powering the radio. However, if you want to run the radio with increased power, the onboard regulator might not cut it (it provides too little current).



  • I was planning on using a stand alone ATmega 328 with the Arduino bootloader and program it with the sensor code by temporarily place it in an Arduino Uno and then solder it directly to a board with the temp sensor and transceiver, in order to keep the power usage to a minimum.

    But maybe this is a bad idea and it would be better to use an Arduino Pro Mini 3.3V so that I get the regulated output? How much current does a Pro Mini draw in comparison to a stand alone ATMega 328?


  • Hero Member

    @bjornhallberg said:

    Don't buy the 3.3V boost / step-up listed in the "store" if you intend to run it on batteries. It has a quiescent power draw of 1mA or so (a modern boost will have 5uA o so by comparison!!), which will drain even your 2xAA batteries in well under two months for sure no matter how lean you run the rest of the circuit. Unfortunately there is no better boost that I've found for sale. You'll have to either wait for the official mysensors hardware lineup which will start with a boost circuit. It should be ready any day now? Right? :-) @axillent

    Or you can run most of the sensors on just 2xAA (in series) directly, no boost or anything in-between. Sure the voltage will soon drop below 3V but it will most likely take months and most sensors as well as the nrf24 and arduino pro mini will cope with that. You wont be able to suck every last drop out of the batteries but it shouldn't be too bad. Far better than the 1mA quiescent.

    If you had a boost, and limited space, you could also power your sensor node with just one AA. Like the official battery powered sensor node that is forthcoming.

    If you're handy with SMD and can source a few capacitors and inductors you could always build some yourself. Texas Instruments has free samples available if you want to try. Still trying to get mine to work though :-/ I can't figure out what I did wrong here. I thought about starting thread about it but I'm holding out for the official modules.

    Btw. Battery Calculator: http://oregonembedded.com/batterycalc.htm

    In testing, using a sensor with 2 DS18B20's, I had problems almost immediately (at about 4 days). As the batteries dropped below 2.3vdc I started to see skewed temp. results.

    Now in my tests I was using rechargeable EBL 2800mAh NiMH batteries, so that my have been part of the problem. With only 1 mod (removal of the Arduino LED, I'm getting about ~4 months. My sketches are not stock, and I'm only sending updates when changes exceed .5 delta.


  • Hardware Contributor

    I have to disagree with some information regarding the 3.3V step-up. I use it like this and I'm very satisfied. Don't now if it's a good value for money though, but I don't know what else to buy?



  • Does anyone have the spec sheet for the Chinese step up regulator? I bought a few of these to test
    http://www.pololu.com/product/2561


  • Hero Member

    @ServiceXp Are those NiMH batteries low self discharge (LSD) or not? I looked them up on Amazon and saw someone had tested them and found they were some sort of middle ground. Better than most NiMH but not LSD per se. Now I was about to rant about how bad NiMH batteries are but after actually looking up discharge curves of Alkaline and at least LSD NiMH I'm not so sure. If prices have come down a bit (have they?) they could work as well as Alkaline. Maybe. I haven't really found a test yet that uses as slow discharge as we are using here (understandable perhaps as it would take a great deal of time to complete). When doing something like a constant discharge of several hundred mA, the 1.5V starting point of the Alkaline will soon drop below the 1.2V starting point for NiMH, but in our case with extremely slow discharge, Alkaline should still hold their edge for a good portion of their lifespan.

    @cygnus Yeah, the DS18B20 doesn't seem as forgiving as some other sensors. So running off of AA directly may have its downsides, but at least you haven't bought anything you may regret.

    @m26872 70uA sounds phenomenal. I'm going by this old thread though:
    http://forum.mysensors.org/topic/171/efficiency-of-voltage-boosters

    @Rek A bit fuzzy on the quiescent power but should be better than the Ebay ones. Would be interesting to see some results. Still, I'm holding out for the official modules.


  • Hardware Contributor

    @bjornhallberg
    I admit that 70uA sounds a bit to good to be true. The figures in that thread says min 500uA for "china 3.3V without led) I think I'll have to do some more measuring to confirm my results. So far I've only used my Fluke 87 in uA-range with burden voltage "display value times 100uV", which hasn't been calibrated since due date in 2012.


  • Hero Member

    @bjornhallberg I would agree; while these EBL's are probably the best NiMH batteries I've ever used, they probably are still not quite as good as good Alkaline's.


  • Hero Member

    The reference for mini pro says:

    Input Voltage	3.35 -12 V (3.3V model) or 5 - 12 V (5V model)

  • Hardware Contributor

    @bjornhallberg As I suspected this was wrong! I now measure total sleep mode current draw ~1mA (with one DHT at Vbat 3.1V. My project thread/post is now updated. Sorry.



  • @epierre said:

    The reference for mini pro says:

    Input Voltage	3.35 -12 V (3.3V model) or 5 - 12 V (5V model)
    

    Does anyone know what is the power loss of the on board voltage regulator (to compare to the loss from the step up converter)?

    My thought process was that maybe using 3 AAs (or perhaps 4) and letting the on board regulator step the voltage down would result in a longer time until battery changes being needed. I thought maybe even as the batteries discharge and their output decreased, it would still be overall be higher than 2 AAs and therefore require less current to achieve the regulated 3.3V and therefore a slower discharge rate.

    If power losses from step up converter are comparable to power losses from on board regulator having more batteries seems like it would obviously increase the run time, increase the size of the sensor (although wouldn't require space for the step up converter), decrease intial cost, possible increase long term cost. Right?


  • Hero Member

    @hooraysimpsons The "official" Pro Mini is supposed to have a Micrel MIC5205 allegedly.
    http://www.micrel.com/_PDF/mic5205.pdf
    Marked "KB33". Same as on my cheap clone. On paper it has phenomenal quiescent current draw.

    mic5205.png

    In practice, people are not as impressed:
    https://www.sparkfun.com/products/11114#comment-504d47b1ce395fb72d000001
    Don't know what to make of this. But it still seems miles better than the useless Ebay boost? Our own build page claims you can save 220uA by removing or disabling the regulator. Is that figure so high only because the regulator is getting back-fed otherwise as you bypass it?

    Bottom line, are we 100% sure the on-board regulator is so crappy that it in fact has to be circumvented? I mean, for inputs when using 3+ AA batteries? I mean, personally I'd rather not use so many batteries and bulk up the sensors, plus the extra cost of batteries would soon pay for an external boost I guess.


  • Hardware Contributor

    I have now set up a test node with 4 AA to the raw-pin (mic5205) and with the pwr-led removed. External loads are one DS18B20 and a 5+1Mohm battery monitoring circuit (and the radio of course). With battery voltage 5.82V (not fresh batteries) I get 115-120uA in sleepmode. Not bad. Power <25% of my 2AA china-stepup nodes. On the downside are a little worse (higher) low-limit voltage and little less available 3.3V-power to sensors. And I don't think the different final voltage level would matter since it is near the end of "the curve".
    Considering Vin = Vout +1V = 4.3V according to the datasheet above, I'm sure 3 AAs not will be using enough of their working range. But 4 AA looks like a good choice to me, if space isn't an issue.
    It would be interesting to see the economy with 4 AAA batteries. Calculations anyone?


  • Hero Member

    @m26872 Interesting findings! A not at all unreasonable setup either. As to AAA ... I don't know. They're less than half of an AA typically in terms of capacity usually. At the same per unit price. Still you'd probably get up to a year on 4xAAA. Or close enough. For just a few sensors I could probably live with this.

    http://oregonembedded.com/batterycalc.htm


Log in to reply
 

Looks like your connection to MySensors Forum was lost, please wait while we try to reconnect.