The Mercury Battery Replacement FUD (fear, uncertainty and doubt)

The basis of this FUD

 The batteries are described in the SRT owners manual in this way.  
"The Minolta SR-T 101 uses a 1.35v, button-shape mercury battery for photo-graphic applications (Mallory PX-625, PX-13, Eveready EXP-625, EPX-13 or equivalent). "
These mercury batteries your SRT was designed to use are now very hard to locate, as production has been discontinued in most of the world for very good environmental reasons.  There are a number of possible replacements, but none of them result in exactly the same meter readings as the original design, at least without some effort at recalibration.  This is because the meter is a very simple circuit which depends upon a specific stable voltage source.

There is a tremendous amount of data on the web regarding ideas for replacement power sources.  I think most of these solutions are viable for the SRT application, assuming you have the capability to recalibrate your camera's mechanical and electrical systems.  Ok, I'll repeat it, I don't think there is a really ideal solution, without doing some recalibration.  It might not matter too much, because it is entirely possible that that 25 year old camera of yours needs recalibration, even if you still use mercury batteries.   In any case, once you get the hang of it, and build yourself a tester, recalibration is not very hard to do.

Possible Replacements

Here is a table detailing some of the possible mercury battery replacement options.  This is ordered with my preference at the top and least desirable solution at the bottom.  I highly recommend that whatever solution you choose, that you consider calibrating your meter for maximum performance with that power source.  See my meter tune-up page for more information.

possibility availability
voltage characteristics
relative lifetime
voltage variation
voltage changes with current variation
resistance to voltage variation with temperature
plus
minus
Schottky diode adapter with 357 type silver oxide button cell
357 batteries are available almost anywhere.  You can build your own adapter or purchase the closely related MR9 (which I have not tested).
Briefly 1.4
Nominally 1.35
Quick drop off at end of life
more stable than mercury
some
good
May be cheaper  than 76 type photo batteries, may not require recalibration.
Expensive, unless you make your own adapters, which is not difficult to do.
357 cell with O-ring to fit
357 batteries are available almost anywhere
Briefly 1.6
Nominally 1.55
Quick drop off at end of life
more stable than mercury
no
good
Simple, easy and cheap.   May be cheaper than 76 type photo batteries.
I've had some problems with o-rings interfering with good electrical contact.  Requires recalibration since the voltage is so far off.
imported mercury replacements
hard to come by, apparently illegal to sell in many places
1.4 or 1.35 volts, the energizer technical page specifies both.  The import I recently tested read 1.46 volts under load!
very stable
no
poor low temperature performance in some situations
original technology
Apparently illegal to sell, mercury is very toxic,  it is unclear whether voltage is 1.35 or 1.4 volts.  Perhaps both versions exist.
Schottky diode adapter with 76 type silver oxide button cell
76 batteries are available almost anywhere.  You can build your own adapter or purchase the closely related MR9 (which I have not tested).
Briefly 1.4
Nominally 1.35
Slow decline towards end of life
Possibly less stable than 357 variation and mercury. I've read somewhere that the charts on energizer.com are no longer accurate and they now have stable voltages over their lifetime.
some
good

Expensive, unless you make your own adapters, which is not difficult to do.
76 type cell with O-ring to fit
76 batteries are available almost anywhere
Briefly 1.6
Nominally 1.55
Slow decline towards end of life
Possibly less stable than 357 variation and mercury. I read somewhere that the charts on energizer.com are no longer accurate and they now have stable voltages over their lifetime.
no
good
simple, easy and cheap
I've had some problems with o-rings interfering with good electrical contact.  Requires recalibration since the voltage is so far off.
zinc-air cell example Energizer (AC675) or Wien MRB625
available in photo specialty stores only
Briefly 1.35
Nominal 1.3
Quick drop off at end of life
Voltage varies depending upon recent usage for 675 cells.  I haven't tested Wien cells.
Voltage varies depending upon recent usage for 675 cells.  I haven't tested Wien cells.
good, but at high temperature the lifetime is shorter

Batteries are only good for a few months.    I haven't tested Wien cells.
alkaline 625 available almost anywhere
1.5
Slowly declines over lifetime of battery
voltage varies greatly over life of this battery - basically unusable for this application
no
good
same form factor as original
Variation in voltage over life of battery makes this a less desirable solution than others
modify camera electronics with voltage regulator
how good at electronics design are you?
whatever you choose
most stable
should not
depends upon circuit

Requires camera modifications.  lowers lifetime of selected battery quite a bit, do to increased current draw

Battery tester

Battery Testing

The new Everready combined 303/357 has a poor voltage/discharge curve, and many other manufacturers don't report such data at all.  In order to find a battery that would perform satisfactory, I decided to build a device that would  discharge a battery in a controlled manner and capture the resulting voltage curve.  With this device and a voltmeter, I will be able to determine the following information for each battery that I test.  I will not be testing under varying temperate conditions.
Once I have tested the battery and have complied the results, I will post them here.  If you have a battery that you want me to test, I will do it, with the only restriction being that any results that I gather will be posted on this web site.
The tester itself is a fairly simple device.  Those of you who have seen my universal timer/counter, will recognize that the base micro-controller design is the same.  I have removed the interface to the photo transistors and replaced it with a battery test interface.  It uses a LM334 current source to provide the controlled drain on the battery.   This controls the amount of current that the battery will have to provide over the course of the test.  Once the battery is unable to provide the 2 milliamp current, it is assumed that the battery is close to being drained.  An AVR micro controller monitors voltage and current through an ADC0832 dual 8 bit analog to digital converter.  The micro controller logs the voltage data to non-volatile memory.  A transistor is also included in the circuit to shut off the current in case of power failure.   The combination of these two  features provides the ability for the test to automatically resume where it left off in case of power failure.  It also contains an RS232 interface with user controls to start and stop the test and a way to dump results.  A LED is programmed to report via morse code, voltage and current information, in case someone needs to get battery status without connecting a serial port.

 
battery graph


CVS Zinc-Air
Eveready 303-357
Rayovac 303-357
Generic Mercury
Duracell 76 S
Measured milliamp hours 486 145
136
252
173
Expected lifetime in days with continous 200 uAmp load
96 days***
30 days
28 days
53 days
36 days
Brand new battery measured no load voltage 1.388 1.62
1.598
1.447
1.591
Brand new battery measured voltage with 2 milliamp load 1.25 volts 1.60
1.561
1.427
1.569
Measured voltage drop over useful lifetime* .059 volts .27 volts
.059 volts
.1 volts
.098 volts
Calculated internal resistance 50 ohms 9 ohms
18.5 ohms
10 ohms
11 ohms
Calculated average voltage with 50 uAmp load 1.35 1.35
1.56
1.30
1.53
Calculated average voltage with 200 uAmp load 1.34 1.35
1.55
1.30
1.52
Calculated average voltage with 400 uAmp load 1.33 1.35
1.55
1.3
1.52
Calculated broken in battery voltage with 200 uAmp load** 1.37 1.49
1.59
1.34
1.54
Calculated old battery voltage with 200 uAmp load** 1.31 1.21
1.53
1.24
1.48

*    Does not include new battery break in and after voltage curve starts dropping off
**    Fairly new battery and failing battery are arbitrarily based on when new battery break in curve and old battery failing curve convergeto/from fairly slow/steady voltage fall off
***    Does not account for zinc-air batteries limited lifetime due to exposure to air

Mercury Battery Note

I have recently tested a generic (manufacturer unknown) mercury battery obtained a few years ago.  Turns out that performance is good, but an ordinary hearing aid type zinc-air battery has a somewhat flatter discharge curve.  Voltage output under a 2 milliamp load varied from 1.34 volts when broken in to 1.24 volts when near end of life.   Note that internal resistance is very low, so that output voltage should not vary much over different loads.

Note that a number of the other kinds of batteries still contain a tiny amount of mercury, thought the RoHS effort in Europe may eventually lead towards elimination of all mercury in batteries.

Zinc Air Battery Note

My test of a CVS branded 675 hearing aid battery resulted in impressive performance under stable load.  Very flat discharge curve and very long life are evident.  Output voltage is very similar to that of a mercury battery.

At www.energizer.com, I read that zinc-air batteries vary output voltage depending on current.  This is do to a fairly high internal resistance.  In fact, basic testing has shown that, with an ordinary 675, zinc air cells, voltage output isn't a simple curve.  Voltage varies depending upon current and in addition there is a history effect (historesis).    If they have been used recently, voltage will tend to be lower than if they have been in circuit that was switched off.  Since I turn my meters off, when not in use, I don't think I'll make use of these cells.

I found a web page about collector watches that lists problems with this type, including a comment that they may emit water which can increase risk of corrosion.  You may need to drill holes in your battery compartment to get the needed air to the battery.

Also keep in mind that zinc-air batteries tend to have a fairly short storage lifetime, once the seal is removed.

Silver Oxide Battery Notes

There are many varieties of Silver Oxide batteries and it is apparent that they cannot be lumped into a single catagory.  Basic types include the 76S photo battery, 303(SR44SW) and 357(SR44) watch batteries, and the new combined 303-357 batteries.  The 303 types are designed for continuous low drain applications, such as watches.  The 76S photo battery are designed to handle situations where higher peak loads can occur temporatily.  The 357 batteries are somewhere in the middle.

303/SR44SW Batteries

 Sodium hydroxide (NaOH) is typically used as the electolyte.  

357/76S/SR44W Batteries

Potassium hydroxide (KOH) is typically used as the electrolyte.  This offers less internal resistance and allows the battery to be more efficient under heavier drains.  This means that KOH batteries will last longer under higher loading.  The lower internal resistance will also provide a more stable voltage output under varying loads.  Batteries with KOH are harder to seal than NaOH.  As a result NaOH batteries usually exhibit more salt resistance.  Batteries using KOH will also work better at lower temperatures.

Eveready 303/357 Battery Test

A year or two ago, Eveready combined the previously separate 303 and 357 into a single product called 357/303.  Since, in the past, I considered them interchangable, that initially didn't bother me.  However, Eveready also put a new data sheet on their website that reveals a must poorer discharge/voltage curve.  The second half of the batteries life, looks especially bad.  I sent them an email and asked them what happened, but did not get a useful answer.   My test results pretty much matched the results shown on energizer.com.  In my opinion, there are better batteries available for camera meters sensitive to the voltage of power supplies and the Energizer 303/357 should not be used.

Rayovac 303/357 Battery Test

This is a brand new battery purchased from a nearby Sears Essentials Store (electronic department).  I have also seen them in Walmarts Jewerly department (near watch bands).  The battery is labelled 357 on the back, and rayovac's web site "Material Data Safety Sheet" indicates that all there silver oxide batteries are the potassium hydroxide type.   The discharge curve for this battery is outstanding.   Only a few minutes are needed for "break-in".  This is so brief that my battery tester doesn't even recognize it, as can be seen from the chart.  The voltage drop, when it is exausted, is by far, the steepest I have tested.  Since you don't need to deal with the issues with exposing zinc-air batteries to the environment, and there is virtually no break in required, this is my recommended battery for the moment.

Duracell 76s Battery Test

Very similar to Rayovac 303/357.  Longer life, with flatter voltage/discharge curve.

The Schottky Diode Adapter Notes

Due to varying output voltage over different load conditions, it couldn't hurt to recalibrate when using an schottky diode adapter.

Note that all schottky diode adapter information on this page is based on prototypes built with a type 1N6263 schottky diode purchased from Active Electronics.

You can make your own schottky diode adapter or buy a similar device called the "MR9".  One problem with purchasing the MR9 is the cost of nearly 30 dollars each. I found an excellent  web site describing how to make your own shottky diode type adapter. The link can be found on my "For more information" page. The only problem with the design at this site is that it requires usage of an adhesive conductive material from 3M. This material is not expensive per foot, but as far as I can tell, it can't be found in small quantities. I found an inexpensive source of adhesive backed copper foil at a nearby Michael's Arts and Crafts store that is able to serve the purpose. It is labeled Foil Tape, and manufactured by Glass Tile and Metal Works (Lafranc & Bougeois). Get the "Classic Copper" version. I believe that it is made of real copper. This stuff is not wide enough to fill the entire bottom of the adapter, but it is not necessary. I heard that stores carrying extensive supplies of doll house components might have a similar copper foil that is used for lighting doll houses. I made one other change when I built my second unit. I replaced plastic as insulation with relatively thin paper card stock. It will not melt from the heat of the nearby soldering iron needed to solder the final connection of the tape to the diode.

So What Am I Using?

Stay tuned as I will likely alter my battery selection as I find a few "good" items in the process of doing my battery tests.

I am currently using the home made schottky diode adapters combined with 357 silver oxide batteries in place of mercury batteries.  In addition, all of my SRT's have been calibrated to match a relatively new Pilot 2 meter over a wide range of lighting conditions.  See my meter tune-up page for instructions on how I did this.

I am also using one of these adapters in a Canonet.  Even with no calibration done, I have had very good luck. Well, the battery check function doesn't work, but that is only a minor inconvenience.

More details on meter performance with the different power sources.

To better understand meter performance, I measured CDS cell resistance under several different light intensities.   Then I added an arbitrary 4k to this base resistance to account for the resistors in the meter circuit.  The result is shown here as the two lines.  It is clear that the CDS cells in the meter meter react to light in a logarithmic function.  For more information, see my pages on SRT operation.  Also it is important to note that the adjustment resistor has a much larger effect at higher light levels.
resistance-light
I measured voltage of several power sources with each of these resistance's.  I then calculated current through the circuit and constructed the chart shown below.
batteries and voltage
 This chart shows calculated current at different light intensities with a variety of batteries.  Note that the horizontal bars that show the nominal current expected at various F-stops.  The amount of error a given battery has at a given level of light can be determined by finding the intersection of the blue bar positioned on the theoretical 1.35 volt mercury spot and the line for the battery you are interested in.   For example. at 256 lux the silver oxide battery (adjusted under low light) is (in theory) almost 1 stop off.  At 2048 Lux it is off the chart.  Note I have only tested one battery in each category, which leaves a whole lot of room for sample error.

Descriptions of test batteries

Important note:
Adjusting the meter by using the adjustable resistor only is problematic. Although the meter can be made to match at some given setting, other settings will  diverge from matching the good meter. If you are going to use this approach, it would be better to adjust to match an good meter in brighter light rather than dimmer.  See my meter tune-up page to find out how to make the meter accurate in broad range of lighting conditions.



Back to Mike's Hobby Home Page