When should one declare primary batteries depleted?

[TheWelly888]

Dave mentioned that one $100 multimeter could work from a PP3 battery with as little as 3.9V output and shown that it could still measure resistance accurately and gave it a thumbs up.

I questioned this and realised that the discharged primary battery output voltage level is debatable.

I work on maintaining and repairing medical equipment at my local hospital. An example is a temporary heart pacemaker where the electrodes are inserted to the patient's heart and the pacemaker fitted externally until the patient can get a permanent one or the heart condition goes away. It is powered from a PP3 battery ( we only use Duracells because they don't leak! ) and the pacemaker indicates that the battery is going when the voltage is approx 7.25V and completely flat approx 6.5V ( I cannot recall the exact ranges off hand ) So I can understand Dave giving the thumbs down to multimeters which declares the battery low at 7.5V when something as critical as a pacemaker is happy with 7.25V!!!

Clearly a rechargable battery is different because their discharged voltages are clearly specified and equipment working off them must be able to work down to that and switch off below.

Any ideas when one should declare primary batteries flat?

[EEVblog]

Primary alkaline batteries are deemed to be completely flat at around 0.8V per cell. For a 9V battery that is 4.8V for the battery. So an ideal "well designed" 9V product should continue to work down to and declare low battery at 4.8V.
Voltage drops off extremely rapidly after this point in a greatly varying way, so in practice it's debatable whether or not the UEi would actually have time to declare low battery before it simply switches off.
My thumbs up was for the fact that the meter uses essentially every last drop of the battery capacity while still being operational, and I think that's rather exceptional.

Any product designed to use rechargeables will have a user selectable cut-off point to cater for the two battery technologies. Multimeters with hundreds of hours life are of course not in this category.

A meter that has a cut-off of say 7.2V is wasting around half of the capacity of the battery.

Dave.

[jimmc]

For a graphical confirmation of Dave's reply see the Duracell datasheet:
http://www1.duracell.com/oem/Pdf/new/MN1604_US_CT.pdf
Graph top right shows battery voltage falls over time with constant 2mA load...
Life 340hrs to 4.8v, 230hrs to 7.2v

Jim

[TheWelly888]

Thanks Dave and Jim, I'm now enlightened! 0.8V per cell - I'd better remember that! 

Out of curiousity I looked for the GP PP3 data and the datasheet is not as comprehensive as Duracell's but here it is: http://www.goldenpower.com/pdf/English_Spec/Greenergy_Batteries/Greenergy_Alkaline_Batteries/GL6F22A.pdf I'm sticking with Duracell! 

[Excavatoree]

Thanks, everyone.

Now I know that my wireless headphones that stop working when the batteries get to 1.3 volts per cell are CRAP!!!!   It will operate down to 1.1, but when the voltage gets to 1.3, the receiver drifts, and I have to constantly adjust the tuning capacitor to receive the signal.  It gets more and more frequent until I give up and change the batteries in frustration.

[Kiriakos-GR]

Measuring an battery with just the multimeter , can be an illusive measurement .

Thats why I had to seek for one more proper solution , with Load !!

https://www.eevblog.com/forum/index.php?topic=448.msg6650#msg6650

[logictom]

Is there a standard way of utilising a battery till it is flat? Like a voltage booster or similar rather than designing the circuit for such large voltage variation?

[saturation]

Is there a standard way of utilising a battery till it is flat? Like a voltage booster or similar rather than designing the circuit for such large voltage variation?

Yes, DC-DC converter similar to this:

http://www.maxim-ic.com/datasheet/index.mvp/id/1881

[ThunderSqueak]

When I read the title of this page it reminded me of the "Joule Thief" for some reason 

http://en.wikipedia.org/wiki/Joule_thief
http://www.emanator.demon.co.uk/bigclive/joule.htm

[Zero999]

Now I know that my wireless headphones that stop working when the batteries get to 1.3 volts per cell are CRAP!!!!   It will operate down to 1.1, but when the voltage gets to 1.3, the receiver drifts, and I have to constantly adjust the tuning capacitor to receive the signal.  It gets more and more frequent until I give up and change the batteries in frustration.

Have you measured the battery voltage whilst the batteries are in the device?

Yes, it's ideal to work down to 0.8V per cell but it's not always possible. Suppose the device uses two AA cells and a PIC microcontroller which is only guaranteed to work down to 2V, it might work down to 1.8V but the chances are it won't.

When I design, I aim for 0.8V per cell and will settle for 1V per cell but I wouldn't consider 1.1V per cell.

[EEVblog]

Yes, it's ideal to work down to 0.8V per cell but it's not always possible. Suppose the device uses two AA cells and a PIC microcontroller which is only guaranteed to work down to 2V, it might work down to 1.8V but the chances are it won't.

That's when you use a DC-DC converter. You can power virtually anything from a single cell if you really want to.
Depends on your design spec and what the important criteria are.

Dave.

[Zero999]

Yes, it depends on the application, a DC-DC converter might not be cost effective, it could take up too much space, be noisy or the loss of 10% of the power might reduce the battery life more than only discharging to 1V.

[Excavatoree]

...  It will operate down to 1.1, but when the voltage gets to 1.3, the receiver drifts, and I have to constantly adjust the tuning capacitor to receive the signal.  It gets more and more frequent until I give up and change the batteries in frustration.

Have you measured the battery voltage whilst the batteries are in the device?

Yes, it's ideal to work down to 0.8V per cell but it's not always possible. Suppose the device uses two AA cells and a PIC microcontroller which is only guaranteed to work down to 2V, it might work down to 1.8V but the chances are it won't.

When I design, I aim for 0.8V per cell and will settle for 1V per cell but I wouldn't consider 1.1V per cell.

Right now, with the device off, the cells are at 1.30 volts (Fluke 87-3, standard mode)  WIth the device on, the cells are at 1.26.  It's operating fine now, but when it starts to drift, I'll measure the voltage, and then continue to measure it as it gets worse, up to the point when it becomes too irritating to use.  The numbers in my earlier post were from memory - I check the voltage of the cells when I replace them.

I'm not an RF guy, so I don't know what trade-offs had to be made in terms of cost, size (fit into headphones), etc. etc.

[ModernRonin]

I keep wondering why multi-meter designers use one 9V battery instead of 3 AAs. The AAs have 4-5 times the amp-hours, and don't lose by much in weight. (http://www.solarbotics.net/library/pieces/parts_elect_pass_batcomp.html) If you use lithium AAs, they win on weight as well.

Gotta be a cost thing. 9V must be cheaper than AAs + DC-DC converter. (Though with a good meter costing ~$100, does a $2 DC-DC converter chip really make that much of a difference in the overall price?) Or maybe it's the cost of the mechanical enclosure for 3 AAs...

Edit: LadyAda's "MintyBoost" is a great example of how much more you can get out of AAs - http://www.ladyada.net/make/mintyboost/process.html . And that's only 2 of them.

[Zero999]

EMC might be another reason: an SMPS will require good shielding and filtering to stop radiating energy which could potentially interfere with a sensitive circuit being measured.

[saturation]

Fundamentally, many use old designs and chipsets, even if new often a variation on the great grand daddy of accurate single chip DVM, the Intersil 7106 and driven by a ~9v.  It drive LCDs with ~ 6V with onboard LCD drivers.  Many LCD panels use less power driven at higher voltages.

Many DVMs of modern vintage do run off AA batteries, or Li or even 1-2 Li watch batteries.

I keep wondering why multi-meter designers use one 9V battery instead of 3 AAs. The AAs have 4-5 times the amp-hours, and don't lose by much in weight. (http://www.solarbotics.net/library/pieces/parts_elect_pass_batcomp.html) If you use lithium AAs, they win on weight as well.

Gotta be a cost thing. 9V must be cheaper than AAs + DC-DC converter. (Though with a good meter costing ~$100, does a $2 DC-DC converter chip really make that much of a difference in the overall price?) Or maybe it's the cost of the mechanical enclosure for 3 AAs...

Edit: LadyAda's "MintyBoost" is a great example of how much more you can get out of AAs - http://www.ladyada.net/make/mintyboost/process.html . And that's only 2 of them.

[kb3pxr]

The issue you have to watch with discharging cells down too far is the potential for leakage in primaries. In the PP3 form factor you should be fine with plastic cased batteries, I don't know if they do this for any primaries other than maybe lithium. For the AA cell multimeters you have to really be careful. Over discharge could cause excessive build up of pressure in Alkaline or excessive zinc can consumption in carbon zinc which can result in leaking electrolyte which can damage your device. I have a lithium PP3 in my multimeter to prevent such issues.