MDBT42Q on coin cell

Posted on
  • Hey everyone,
    I just wanted to clarify something, will the MDBT42Q bare module run perfectly wired to just a CR1225 coin cell? With no voltage regulator, just the battery to the module? I'm assuming it is but I just wanted to clear things up. I don't have a module to try it on, sorry.


    P.S. Will i be able to program the MDBT42Q with mbed?

  • Sun 2018.11.25

    Let's check the data sheets.

    The load:

    2.5 - 16v voltage input, 20uA power draw when advertising

    Note the caution: Do NOT get the polarity wrong
    Also see note beneath heading 'Bare Module' about battery selection

    The supply:­f

    From the graph, ~600 hrs @ 60uA

    Number of hours in a month ~720

    While just advertising, under ideal conditions, at room temperature, expect no more than 3 months of continuous use with no additional hardware attached and no LED's illuminated. Mileage may vary. I'd be more worried about what the other interface hardware load requirements were, then select the power supply.

    Might be able to play with (following link) to extend that some:­setConnectionInterval

  • I'm just using two seven-segments and 5 LEDs, I think a coin cell would provide about 5-7 days of continuous power(LEDs will be used about 5 times an hour each use being 1 second long).

  • As LED's are more than ten time uP usage, we'll just use LED draw
    7+7+5 = 19 seg's - we'll use 20 for easy math
    Approx ~20ma ea draw is 20 x 20 = 400mA total draw

    5 x 1sec = 5sec/hr x 24 = 120sec/dy

    Capacity 50mAh

    Total draw is 8 times capacity, so 60min / 8 ~ 7min total

    I get 3.5 days+

    A CR2032 might be a better choice @235mAh nearly five times the capacity­f

    Use a PWM signal at 50% duty cycle maybe?

  • @Marty_McFly you're asking about the bare MDBT42 module, not the breakout board?

    If so just be aware that @Robin's quote of "2.5 - 16v voltage input" isn't right. Putting much over 3.6v on a bare module will destroy it, however the module can function down to 2v or so.

    But you can connect non rechargeable lithium cells directly since they're between 2 and 3v, and @Robin's suggestions on actual battery sound good.

    In terms of the LEDs, if you're looking at seven segment displays, you can 'scan' them out, so you're only leaving the LEDs on for a fraction of the time. If you do that you should be able to get the peak power draw down significantly (in fact you'd have to, because you'll struggle to draw more than 50mA or so from a Lithium cell reliably).

  • the module can function down to 2v or so.

    What do you mean by that? The MDBT42 docs page says it accepts 1.7-3.6v range.

    you can 'scan' them out

    Do you mean multiplexing and using persistence of vision?

  • Mon 2018.11.26

    Agreed, one needs to read carefully, reading beneath heading 'Bare Module'

    'All you need to get the MDBT42Q working is to apply power between the VDD and GND pins'

    could get one in trouble, if one applies the wrong range as indicated. I mis-interpreted '2.5 - 16v voltage input' so might others.

    Perhaps a 'Bare board features' beneath 'Breakout board features' summary in order to get it right the first time ?

    Or a sub-heading beneath 'Features' ?

    Also, overall outside dimensions for both would be nice. (which one is listed beneath the features heading? bare?)

    FYI A word of planning Caution: Do you @Marty_McFly have a good plan on the use of the bare module? It's beyond my ability (16mm x 10mm) to easily solder it, and the breakout, although larger is also a bit of a soldering challenge for me, even with magnification.

  • @Marty_McFly,

    'scan' them out

    means exactly what you are saying.

    @Gordon suggested to let the scan run free to reduce the flicker - see Driving LED matrix with shift registers such as 74HC595 and Retro Bubble Displays driven with 74HC595. I used timeouts how long I would turn on a segment (and how much to wait after turning it of and turn on the next segment). Free running gives the eye the most snapshot like image of all segments. On the other hand - if you can afford from brightness / dimness - you could think of timing the next free run to save battery. Without, one segment would be always on and be the determining factor for power consummation.

  • @Robin I'll update the docs to make it very clear.

    What do you mean by that? The MDBT42 docs page says it accepts 1.7-3.6v range.

    By '2v or so' I meant exactly that. If you look at the CR2032 discharge curve the battery is basically dead by the time you get to 2v anyway.


    As @allObjects says, instead of wiring each segment to a pin, you just wire them in a 'grid' - so probably you use 7 IOs to drive your 7 segments and 5 LEDs, and then another 3 IOs to select whether you want to light up the 1st or 2nd 7 segment, or your LEDs.

    It gives you a lot of control over the brightness and power consumption, but also saves you a bunch of IOs and makes wiring easier.

  • Tue 2018.11.27

    ' I'll update the docs to make it very clear'

    Thank you Gordon.

    'What do you mean by that? The MDBT42 docs page says it accepts 1.7-3.6v range'

    'By '2v or so' I meant exactly that'

    From the manufacturer datasheet heading 'Information' >> MDBT42Q Datasheet p.20


    Paraphrased: Power on-chip power-on set circuitry may not function above 60ms power on rise time.

    Range 1.7 - 3.6 with 3.0 nominal

    Me thinks that @Gordon in indicating by '2v or so' that as the chip functions best at the nominal 3.0 voltage, and as when using a coin cell as a supply source, that when the battery discharge threshold is significantly below nominal, that the power on rise time becomes much worse (longer) such that the chip is unable to set itself up correctly as power is applied; and that the coin cells ability to maintain a sustainable working voltage within range is severely compromised. So effectively the battery voltage droop is the issue before the lower limit is reached. Having a stable regulated supply, might allow the chip to function down to that 1.7v limit, but the manufacturer is suggesting a 3.0v optimal value.

  • Just as a NB I've had a bare MDBT board connected to a 4 digit 7-segment LED (usually off) that's lasted about 2 weeks on a dodgy CR2032. Dodgy because it was about 3 years old when I found it and put it in the device. There's also a .22µF ceramic bypass cap which may be helping with the rampup time.
    Even a new 2032 has about a 30Ω internal resistance.
    If the LEDs are always on 3.5 days sounds about right - I mux them and also have a button that turns them on for 10 seconds.

  • Thanks! Actually it's really helpful to have a decent size capacitor across the battery too. A smaller bypass cap is good, but an additional biggish (~10uF) capacitor across the battery could really help.

    I'm not sure if it's just the internal resistance @rmd6502 mentions, but you get significantly less life out of a battery if you try and draw more than a milliamp or two. It's not really advertised very much but if you look at­5.pdf and the 'Load vs. Capacity' graph at the end, you'll see that with a 1k load (~3mA) you get 150mAh, but with 10k it's nearer 200.

    It's not on the graph, but you can see that after 3mA the battery life falls off a cliff. Having a capacitor and then 'pulsing' the LEDs by multiplexing them/scanning them out means that hopefully you'll be drawing most of the peaks of current out of the capacitor, not the battery - giving you much better life.

  • Ah, yeah 10µ FTW - time to rerev the board...

  • Post a reply
    • Bold
    • Italics
    • Link
    • Image
    • List
    • Quote
    • code
    • Preview

MDBT42Q on coin cell

Posted by Avatar for barry_b_benson @barry_b_benson