Connect with us

Batteries for voltage reference

Discussion in 'Electronic Design' started by Hal Murray, Oct 22, 2007.

Scroll to continue with content
  1. Hal Murray

    Hal Murray Guest

    I remember tales from many many years ago about using fresh batteries
    as a simple, cheap, and readily available voltage reference.

    Has anybody tested that recently?

    Assume I'm using standard AA alkaline batteries. How much do they
    vary over temperature, age, and from brand to brand? How much
    do they vary over manufacturing time within the same brand?

    That might be a fun science project.

    Are there other inexpensive references that are easily available?
  2. D from BC

    D from BC Guest

    On Sun, 21 Oct 2007 20:16:14 -0500,
    Search for voltage reference on Digikey or Mouser or other.

    I bet some are so stable your measuring equipment won't be sensitive
    enough to measure a change.

    D from BC
  3. There has been extensive discussion on this in the past, Google will
    show you.
    But this idea is just so antiquated now it's not even worth
    Yes, bandgap voltage reference IC's are available that do just this.
    They can cost anywhere from a few 10's of cents to a few 10's of
    dollars depending on the specs.

    Digikey have thousands of items listed:;keywords=voltage reference

    0.1% ones cost only a few dollars for example.

  4. whit3rd

    whit3rd Guest

    If you use a battery technology that doesn't include a gaseous
    byproduct, it works OK. Mercury batteries (and best, the Weston
    style standard cells) are fine. They're also 'WAY more expensive
    than a good IC reference. TL431 (bandgap) or uA723 (zener) are
    low tempco and low drift references, under $1, and many trimmed
    IC with buffered references are available in the $3-and-up range.

    Carbon-zinc and even Mn alkaline are NOT recommended; they
    have barometric sensitivity and can 'burp' a little gas past the
    seal and do a ramp/step kind of AC output when loaded.
  5. Boris Mohar

    Boris Mohar Guest

    On Sun, 21 Oct 2007 20:16:14 -0500,


    Boris Mohar

    Got Knock? - see:
    Viatrack Printed Circuit Designs (among other things)

    void _-void-_ in the obvious place
  6. Eeyore

    Eeyore Guest

    Google 'standard reference cell' or somesuch. They used to be the norm.


    More modern miniature standard cells were perhaps 5cm x 5cm x 2cm,
    made from glass tubing in an H shape, IIRC.

    I don`t think any version was ever RoHS-compliant.

    Mercury batteries have a fairly stable voltage during their life- a
    characteristic that was used in relatively instruments such as light

    But if you only need that level of stability and tempco an IC
    reference is better, cheaper and won`t poison anyone.

    Best regards,
    Spehro Pefhany
  8. Barry Lennox

    Barry Lennox Guest

    On Sun, 21 Oct 2007 20:16:14 -0500,
    Batteries are a PITA. But, before cheap IC refs were readily
    available, there was a number of "Electronic" reference cells
    available, designed to replace Weston Cells and the like. Typically
    they used carefully selected Zener diodes of about 5.2 volts,
    somewhere about there Zeners/Avalanche diodes have a zero TC.

    Cropico was one such manufacturer in the UK. I see they are still in
    business, but no longer seem to have references.
  9. Ecnerwal

    Ecnerwal Guest

    If you don't mind being years out of calibration (usually) and ROHS
    issues, old standard cells are not so hard to come by. The one I have is
    built into an old millivolt potentiometer instrument (there's some
    vintage electronics for you - standard cell and a really fiddly
    galvanometer). However, given the choice of buying such a thing or
    buying a modern solid state reference, buy the reference - it takes up
    less space, among other things. It may also be temperature-compensated,
    while I seem to recall that standard cells require a specific
    temperature to be correct.
  10. Tam/WB2TT

    Tam/WB2TT Guest

    Back in the dark ages when they sold VTVM kits, one of the manufacturers
    directed the builder to calibrate the meter by assuming the supplied D cell
    was 1.xx Volts.

  11. Phil Hobbs

    Phil Hobbs Guest

    The biggest problem with standard cells for lab use is their horrible
    temperature coefficient. Batteries are a great deal quieter than
    voltage references, which sometimes matters more than long term
    stability...but as someone else suggested, a LED can be useful when you
    need a quiet reference.

    A red LED with its cathode grounded, running at constant current and
    buffered by an NPN emitter follower makes nice quiet 1V reference with
    lowish tempco, due to the TC of the LED and the Vbe more or less
    cancelling. Give it some reasonable current, and this can easily be 15
    dB quieter than a bandgap. (Bandgaps use junction forward voltages too,
    but have to multiply up a smaller voltage, leading to more noise.)


    Phil Hobbs
  12. Hal Murray

    Hal Murray Guest

    Thanks. That's probably what I'm remembering. Mine was an Eico.
    (I think.)

    I measured all the never-used batteries I could find.

    1.622 Mar 2014 Safeway AA
    1.619 Mar 2124 Duracell AA
    1.580 Mar 2007 Safeway AA
    1.575 Mar 2008 Duracell AA

    1.575 Jan 2002 Duracell D
    1.436 Jan 2002 Duracell 9V (8.62/6)

    The dates are the expiration date.
  13. Jim Yanik

    Jim Yanik Guest

    (Hal Murray) wrote in

    yes,and what sort of resolution could you get on the VTVM scale?
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day