Canon Film Camera Battery Tests

This web page compares metering information for various Canon cameras tested over 2 days. The purpose of these tests is two-fold.

1. The first reason is to see what affect battery voltage has on the meter's response and how linear that response is. This test was applied to Canon F-series cameras that use mercury cells which are no longer available domestically. Replacement cells of the same size usually have a higher voltage due to their composition or have some means of reducing the voltage to a more appropriate value.
2. The results from the first round showed serious variations in metering values. A second round of test we performed; this time it included newer A-series and T-series Canon cameras. The purpose here is to see if metering variations continue to appear in newer models.

The data presented below was obtained over 2 different days. Some of the tests were done indoors and are reproduceable. Others were done outdoors and thus are at the mercy of sunlight conditions.

Batteries and Test Group

2-March-2008

The mercury cell was chosen for use on older cameras because of its relatively flat voltage curve over time. As the battery drained, it maintained a fairly constant 1.35v. This made it ideal for camera use. The seven F-series cameras tested here do no have internal voltage regulators; they rely on the constant voltage source for accuracy. A replacement non-mercury battery needs to provide a similarly flat discharge curve to be useful (both in terms of voltage and over time).

For each camera, three (3) different batteries are used:

1. a 675 zinc-air hearing aid battery with a rubber O-ring to keep the battery centered in the camera housing. This battery has a voltage similar to the mercury cell but has a relatively short life span. Covering the air holes with tape when not in use can extend battery life significantly. The voltage from this battery tends to fluctuate a little so readings were taken at each test cycle.
2. a Varta V625 U alkaline battery. The voltage for this battery remained constant during the tests.
3. a PX625A alkaline battery (no-name). The voltage for this battery remained constant during the tests.

Alkaline batteries are known to have voltage decay which makes them unsuitable as constant voltage sources over time. However the tests I performed were very short metering tests which did not have any affect on the voltage of the battery. Each battery was measure before and after each test to make sure it had not decayed. I also spot checked the current draw on the FT-QL model and found that its current draw was between 0.01mA (low light) and 0.155 mA (bright light). Circuit resistance was measured on various camera bodies to double check these readings.

• bright light - typically 13-25Kohms (0.115 mA to 0.06 mA)
• no light - typically 130K-200Kohms (0.012 mA to to 0.0075 mA)

At these current levels, I think the test battery, regardless of its composition, will not sag enough under camera load to affect the metering results.

Cameras in Group1are TX, TLb, FT, TL, F-1, FTb, FTb-N. Generally speaking, these cameras have a more limited EV range than Group2 cameras.

9-March-2008

Canon A and T-series cameras do not have the same voltage problem as the F-series. These are powered by a variety of cells - 1.5v AA, 1.5v AAA, 6v PX28 (or equivalent), or 1.5v 357 (or equivalent). Two F-series cameras are also immune to the mercury cell problem - the New F1 uses a 6v battery and the EF has either internal voltage regulation is designed to be insensitive to voltage variations. Lastly, I tested a 20D digital camera. Since the lens on this camera is not as fast as the FD lens I'm using, meter values are different but the corresponding EV is still valid. The lens is a zoom lens set to cover the same area as the 50mm FD lens used in the rest of the tests.

The test data shows the voltage level of the batteries used in the second round of testing. For the cameras not affected by the mercury cell issue, these numbers are there purely for completeness and have no affect on metering performance.

Group2 cameras are New F-1, EF, A-1, AT-1, AL-1, AV-1, AE1-P, AL-1, T60, T70, T90 and 20D.

Test Methodology

2-March-2008

We got 6 inches of snow the day before and the sky is now clear, blue and sunny. It's also close to freezing outside. So for a bright subject, I stayed indoors and shoot a scene from of the bright snow from a window (test T1 in the table below). I only had about an hour to do this and I want to avoid light changes over time so everything had to be done rather quickly. For an average brightness scene. I metered a dark blue rug in the same lighting (test T2 in the table below). T1 and T2 are thus daylight tests.

In the evening, I made a second pass with each battery & camera combination. This time I aimed the camera at an incandescent lamp from a fixed location (test T3 in the table below).

8-March-2008

Unlike the first test, the second test was conducted after a full day of rain. No more snow! Once again, the sky was clear blue but with puffs of clouds blowing by in relatively stiff winds. Testing took more time as I had to make sure each test took place when the sun was affected by cloud cover. The results (test T4 in the table below) are applied only to Group2 cameras.

Another set of intermediate light level tests similar to test T2 was performed on Group2 cameras (test T5 in the table below). While similar in setup, T5 had less light than T2 due to weather conditions. Group2 cameras were also tested under conditions identical to T3. Finally, all cameras in Group1 and Group2 were given a low light test. Due to the different levels of meter sensitivity, Group2 was tested at EV1 (test T6) and Group1 was tested at EV3 (test T7).

Disclaimer

None of this scientific or rigorous so please don't complain about the lack of calibration, laboratory conditions, etc. The results apply only my test samples and your might well behave entirely differently. Nonetheles, these results are enough for me to get a feel for each camera's meter's characteristics.

Reference

For voltage measurements I used a Fluke 76. My initial reference was a Canon A-1. I just had a role of film processed from this camera yesterday and everything looked good. It had a CLA by Ken Oikawa in November '07. With Group2 testing, I added a Canon 20D digital cameras into the mix. The lens used is a 50mm f/1.4 FD BL chrome nose, no filter, f/1.4 to 16 (not used on the 20D for obvious reaons). All metering was done with the camera set to ASA100/DIN21. Every reading was taken with correct focus set and plenty of time was given for the meter to settle.

Though obvious, it's worth saying - Group1 cameras are old and their metering systems are not exactly factory fresh. Some of the exposures vary quite a bit from the reference A-1's meter. The F-1 is the original 1971 flavor. I tried to keep one parameter (usually the aperture) constant across each line (shutter speed or aperture) to make it easy to compare, but sometimes, that didn't work out. That's why I computed EV values. Some cameras were limited by a top shutter speed of only 1/500 instead of 1/1000 (TL, TLb an TX) which didn't help in keeping a constant aperture across the table.

Keep in mind that the term "reference" applied to the A-1 just meas it is used for comparison against other readings. It does not mean that the readings from the A-1 are "correct" or held to a higher standard than other readings.

Also note that the AE-1 and EF also had a CLA by Mr. Oikawa at the same time as the A-1.

Nomenclature

• If a meter reading is close to "exact" meaning the needles match fairly well, (or the needle is real close to the "dot"), then that's the value shown in the table.
• If a correct exposure seems to be more-or-less in-between two readings, I will write insert the 1/2 stop value even if it's not a setting on the camera. For example, if the exposure is between f/11 and f/16, I will write f/13 even though it is not a setting on most cameras. The same applies to shutter speeds. Cameras like the A-1 and T90 display 1/2 stop settings but most don't.
• If a correct exposure is just a little off a particular speed, I will add "-" or "+". In other words "125 +" means a little over 125 (towards 250) while "125 -" means a little less than 125 (towards 60). The base number is used to compute the corresponding EV.
• "over" means overexposure - the camera can not meter the scene.
• "under" means underexposure - the camera can not meter the scene.
• "unreliable" means the meter did not respond smoothly to changes in light level well enough to give an adequate sense of decent sensitivity to this light level.

Test Results

Tests Groups - Summary

Test Data

There are several ways to use this data; including:

• Reading across gives the variation in exposure with changes in voltage (Group1)
• Reading across in groups of 4 tests per camera gives an idea of the linearity of the meter with changes in lighting. When compared with the reference A-1 meter, this can be used to determine whether a linear compensation value can be used to correct the exposure.
• Reading down the columns, one can see the differences that slight voltages cause in the metering.
• Comparing results of cameras within a family and families within the overall sampling.

Exposure differences are calculated with respect to the Reference A-1 EV value column. A positive (negative) exposure difference means overexposure (underexposure); all relative to the A-1's meter.

Each camera is listed with it's EV range (according to spec) and sensor type. The line marked "comp" is the amount of compensation to apply to the camera (via the ASA dial) to make that camera and battery combination useful. This is explained later in the Conclusion section.

Group2 cameras are not tested for voltage variations. For completeness, here's the battery information:

• 6V Radion Shack Lithium 23-266 2CR 1/3N. (PX28 equivalent). 5.92v
• 1.5v Energizer 357 (AS76, MS76 equivalent) 1.591v
• 1.5v Lextron AA alkaline. 1.604v
• 1.5v Duravell AAA alkaline. 1.583v

A positive (negative) exposure difference means overexposure (underexposure); all relative to the A-1's meter.

Conclusions

This began as an exercise in determining how a change in meter supply voltage to an F-series Canon camera would affect the meter's linearity. The prevailing belief is that the increased voltage not only changes the exposure but is also non-linear thus the difference can not be compensated with a linear exposure adjustment (such as with the ASA dial). After I looked over the Group1 data, it became clear to me that the numbers were all over the map. Even the zinc-air hearing aid battery which is supposed to be very much like the old mercury battery both in voltage and decay curve showed a wide variation in meter reading when compared to the A-1. This led me to perform Group2 tests to see if the A-1 readings were consistent enough to be considered accurate.

The Group2 data shows decent corrolation (+/- 0.5 EV) between the A-1, EF, AL-1, AV-1, T-70 and T-90. In fact, the T-90 was identical to the A-1. The remaining cameras in this group fell within +/- 1 EV. The obvious exception is the AT-1, which is unlike the other A-series cameras in its use of a CdS sensor. Overall, I feel pretty good in treating the A-1 as a reasonable reference meter. Therefore, for these cameras, I conclude the following:

1. The A-1 seems to be a decent meter reference. Thus...
2. On the F-series cameras, a voltage closer to that of the original 1.35v mercury cell is prefered because...
3. A good number of these older cameras are likely to underexpose if a higher voltage is supplied to the meter. The 675 hearing aid battery is a decent replacement. However...
4. Because the meters are relatively inaccurate (relative to the A-1), and also non-linear regardless of the supplied voltage, there is no reason why a higher voltage battery can not be used with the exposure compensated via the ASA dial (see below).
5. As a group, the cameras in Group1 represent aging technology, that while usable, may have seen their best days pass them by. While I love old cameras, these simple test exposed some of the annoyances of older designs. For example, the battery cover on these cameras is simply the worse design possible. Visibility in the viewfinder can be very poor - in low light testing, it's sometimes impossible to see the meter's reading in the viewfinder. And the slow response time of the CdS cell is blatantly obvious. By contrast, the A-1, T70 and T90 were a joy to use with all the information clearly visible in the viewfinder. The T90's batteries last a very long time but of course, that's no replacement for a fully mechanical shutter. The EF is very nice, with full information in the viewfinder; but does suffer from legibility in low light.
6. On the whole the CdS meters used in the F-series (and the AT-1) have much greater variability than the Silicon Photocells (SPC). When they deviate, the camera with SPC sensors tend to overexpose at the higher EV levels and underexpose at lower EV levels (relative to the A-1 reference). The CdS cameras do not exhibit consistant patterns from model to model.
7. If I were to use any one of the bodies in Group1, I would compensate it individually since the group behavior is all over the map.
8. For those faithfully loyal to their F-series bodies, I suggest performing a series of tests like the ones I've done. You never know what you'll come out with.

Using the above test data, I can draw some conclusions on the amount of compensation that could be applied to each of the cameras listed above to yield reasonable results. The goal is to provide a +/- 0.5 EV accuracy (relative to the A-1 reference). Sometimes, this is not possible and the amount of compensation is biased to provide the best possible outcome given the test data. Other times, the meter is so inaccurate it's just not worth using. The line marked "comp" in each test camera's data set gives data as follows:

• "0" means no adjustment necessary.
• "+", or "-" EV adjustment. A positive (negative) value means the ASA dial needs to be adjusted to a higher (lower) than the actual film used. For example, for ASA100 film, a "+1" means the ASA dial should be set one stop higher (ASA200) while a "-1" means the ASA dial should be set 1 stop lower (ASA50). Remember that the numbers in the table are defined such that a positive error means overexposure, the camera has to be told the film is faster than it really is in order to reduce the amount of light exposed on the film. Compensation can also be done via the exposure compensation dial if available. I prefer the ASA dial.
• DNU - Do Not Use. Sometimes, no amount of adjustment is useful. The meter is simply not accurate enough. This is largely a matter of personal taste. YMMV.

30-April-2009

Note: The contents in these pages are provided without any guarantee, written or implied. Readers are free to use them at their own risk, for personal use only. No commercial use is allowed without prior written consent from the author.