Hi list,
After a lot of discussion on this list of mercury battery
replacements for the OM1N etc I had previously submitted
some measurements and comments to the list as I was
in the process of testing and modifying my OM1N.
Here are some follow up comments with a lot more
measurements. This time I have included some comparisons
of diode modified and unmodified meter measurements when
operating off a silver oxide cell.
I have also added some other modification alternatives one of
which is cheaper than using a schottky diode and the part(s)
are more readily available.
Regards,
Tim Hughes
hi100@xxxxxxx
My original posting includes other information not repeated
here so you may also want to (re)read that.
Basic requirement for adaption is to drop the silver-oxide battery
voltage from 1,55V to 1.35V. (i.e. by 0.2V)
Here is a summary of my findings. Most people will probably
want to skip over the testing details and measurements included
at the end of this post and just read the initial bulleted comments.
Some more electrical discussion is also included at the end for
electrical technocrats.
EV values quoted below assume 100ASA film as is commonly
done. Most of my calculations assume no more than about EV16.5 ,
if you shoot film under EV's much greater than this (bright sun at noon
in the snow! ) errors will be greater.
* The good news: In an emergency substituting a '357 ,1.55V silver oxide
battery for a '625 mercury cell will give rise to negligible errors at "very
low
light" levels.(EV3)
* The bad news: In an emergency substituting a '357 ,1.55V silver oxide
battery for a '625 mercury cell will give rise to exposure errors of up to -3
stop at "high" light levels.(EV16)
* Light levels in between give rise to errors in between the above extremes.
Light levels below approximately EV8 (when using an F1.4 max aperture
lens) give rise to errors under 1 stop.
* The errors are NOT dependent on Fstop set, ASA set or shutter speed
set.
* The errors depend somewhat on the lens maximum aperture. (Fast is
worse) and if a ND filter or tele-extender is used (reduces error). For
example using an F4 lens should give errors under 1stop up to about
EV11.
* Simple diode adapters ( criscam MR9 ? ) **if properly designed **
should yield errors of about 0.6stop or less (depending on diode used)
with their worst errors being at high light levels. Such simple adapters
may have greater errors over a wide temperature range or if they use
an inappropriate diode.
* If modifying your own Om1 with a diode, a germanium diode (surplus
germanium transistor like a 2N1305) gave the best performance of the
diodes I tested.
* If you store your camera in a really dark cupboard and forget to switch
it off the battery will still last for years.(The current drain is very low
in
complete darkness/ cool conditions)
***Comments on Possible Alternative Battery Adapter Circuits***
*The readily available high current schottky diodes of the 1N5817/8/9
series reduce maximum errors to about -1.2stops if used in a silver oxide
to Mercury cell adaption, but the results may be bad at high temperature
and may depend a lot on the luck of getting a low leakage diode. Two of
these diodes in series might be better with much smaller errors (0.6stop
estimate ) **if** the samples I tested are typical, but the high temperature
performance may still be a problem.
*The low power HP schottky diodes 1N5711, 2800 series etc. mentioned
in my previous post can be used but will overcompensate somewhat.
Connecting 4 of these diodes in parallel helps to reduce the error to about
+0.8stop max. At high temperature the error would be reduced.
* The best choice of a series schottky diode would be a larger geometry
device than the HP products but not as large as a power device like the
1N58xx.
Low current devices tend to be aimed at radio frequency applications so
there
may not be much choice between power devices and RF small signal devices.
* A low power schottky 1N5711 etc. in parrallel with a resistor of about
1kOhm can be used (max estimated error +0.8stop) instead of the 4 diodes
in parallel. Trimming the resistor value under bright light conditions (to
yield say 1.30V at max light) should reduce errors to about 0.4stops
* Using a series resistor alone and adjusting it to produce a voltage of 1.3V
under bright conditions should produce max errors of 0.8 stop or less but
whereas the error with a diode occurs only at the highest light level the
error with a resistor occurs over a range of EV values. Over a wide
temperature range this may also yield bigger errors depending on the
temperature compensation (if any ) circuit used in the camera. (See
below for a discussion of why this simple adapter should work). The advantage
of this fix is that the part is readily available anywhere e.g.Radio Shack
etc.
* A germanium junction diode (not a point contact diode) may be better than
a schottky device but they are difficult to get. (Only one vendor, and not
vailable retail) A better, more readily available alternative that can be
purchased from electronic surplus stores would be to use the collector to
base junction of a germanium transistor (i.e.used as a diode.) This should
give rise to worst case errors of about 0.5 stops or less. The leakage
characteristics should be better than schottky diodes over temperature.
I tested ten different recycled devices from my junk drawer and found them
all to have low leakage and good interchangeability despite their vintage
(30years).
* One list member suggested using a shunt regulator diode ( e.g. LM385-1.2).
(Note true"zenner" diodes also mentioned on the list are not easy to get at
low voltages and have very poor tolerance). The regulated voltage of the LM385
and similar devices is 1.23V causing about 1 stop error at high EV's. This
would
require using a series resistor of about 300 Ohms from the battery to the
meter
circuit and the shunt regulator connected from the meter circuit to battery
common.. (That is, connected across the meter circuit on the meter side
of the on/off switch.) . This would draw about 0.66mA which would give a
battery life of about 250hours, independent of the light level. You would
need to be very picky about switching off when storing your camera as the
battery will then go dead in 11 days if left on. From an exposure accuracy
point of view this is bit worse than a good series diode but with the added
disadvantage of a short battery life.
* What solution am I going to use? :Since very low voltage amplifiers are not
readily available I plan to use a discrete 3 transistor series regulator
which
will use 10uA current giving me an estimated battery life of about 18000hrs
(2.5yrs) if left on by accident. I still need to see if there is space for
this in
my OM1. This should perform as well as a mercury battery. Although not
complicated, this is probably beyond the average home repairer's electronic
skill level to construct.
TESTS on an OM1N:
===============
OM1N current consumption is less than 1uA under completely dark and cool
conditions. (meter switched on, with eyepiece taped over and lens cap on.)
This implies if the camera is stored in a really dark cupboard it makes
almost
no difference if the meter is left on or not as the battery should still last
more than
2.5 years. This may not be as true at high temperature.
OM1N Current Consumption with a mercury cell (1.35V)
=========================================
Ambient temperature of test: 18 deg C
Note: meter current consumption is dependent only on light
intensity (not on aperture,speed setting or ASA settings).
Current Consumption does depend on lens maximum aperture
since metering is at full aperture.
Sealed eyepiece AND lens cap <1uA
EV2 (F2,1sec,100ASA) 16 uA
EV7 (F2.8,1/15sec,100ASA) 86 uA
EV16 (F16,1/250sec,100ASA) 471 uA
EXPOSURE ERRORS VERSUS BATTERY VOLATGE:
Test settings: ASA 100 (except as noted)
OM1N using 50mm,F1.4 lens
Light source incandescent with diffuser
The camera meter was balanced at a battery voltage of 1.35V
The "battery" voltage then varied and the meter rebalanced
using speed and/or aperture rings.
Approximate errors in stops were then estimated from the balance
change with voltage. This calibrates the sensitivity to battery
voltage error at a given light level.
Note: Reference Mercury battery voltage : 1.35V
Nominal Silver Oxide Cell voltage : 1.55V
=====================================
Voltage match setting current error
1.6V F2/1sec 15uA 0stop
1.55V F2/1sec 14.7uA * 0stop
1.4V F2/1sec 13.1uA 0stop
1.35V F2/1sec 12.7uA * REF
1.25V F2/1sec 11.7uA 0stop
======================================
1.6V F5.6, 0.5sec 71uA -0.8stop
1.55V F5.6, 0.5sec 69uA* -0.8stop
1.45V F5.6, 0.6sec 64uA -0.2stop
1.35V F5.6, 0.8sec 59uA* REF
1.25V F5.6, 1.0sec 54.4uA +0.2stop
1.6V F5.6, 1/40sec 167uA -1.2stop
1.55V F5.6, 1/35sec 162uA* -1.0stop
1.4V F5.6, 1/22sec 150uA -0.6stop
1.35V F5.6, 1/17sec 140uA* REF
1.25V F5.6, 1/12sec 130uA +0.7stop
======================================
1.60V ASA 50,F18, 1/1000sec 486uA -3.2stop
1.55V ASA 50,F14, 1/1000sec 480uA* -2.8stop
1.45V F12, 1/1000sec 448uA -0.6stop
1.35V F8.6, 1/1000sec 418uA* REF
1.25V F5.6, 1/1000sec 387uA +1.0stop
=======================================
Adapter using :
4 diodes (1N5711) in parallel to create higher current diode
Input 1.55V silver oxide cell
current OutVoltage error
EV2 12.7uA 1.35V 0 stop
EV6 57uA 1.31V 0 stop
EV7 135uA 1.29V +0.4stop
EV16 400uA 1.26V +0.8stop
========================================
Using a Germanium Diode (transistor C-E junction)
Estimated Performancefor Ge "Diode" :
EV16 0.42mA 1.35V +0 stops
EV8 0.15mA 1.38V -0.5 stops
EV6.5 0.1mA 1.40V -0.2 stops
EV6 0.05mA 1.42V -0.2 stops
EV2 0.015mA 1.45V -0 stops
========================================
Estimated performance using series resistor of about 500Ohm
(actual resistor value must be set depending on the
particular camera)
Input 1.55V silver oxide cell
current OutVoltage error
EV16 400uA 1.35V +0 stops
EV8 200uA 1.45V -0.6 stops
EV7 135uA 1.47V -0.8 stops
EV6 63uA 1.52V -0.8 stops
EV2 15uA 1.54V +0 stops
====================================
Estimated performance when using Schottky Power diode
EV16 1.425 0.44mA -1.2 stops
EV8 1.46V 0.16mA -0.8 stops
EV6.5 1.47V 0.1mA -0.6 stops
EV6 1.48V 0.05mA -0.2 stops
EV2 1.46V 0.015mA -0 stops
====================================
Because the battery voltage sensitivity error is lower at low
Light levels correcting the error at maximum light levels
tends to reduce errors fairly well over the whole range.
This is helpful when using diodes too, as the leakage currents
are less significant than at low light levels.
At high light levels having too low a simulated battery voltage
introduces slightly less error than if the error in voltage were
on the high side.This helps reduce the errors from the low
power schottky diodes which drop too much voltage
at maximum light levels.
==========================================
Tests on 5 Schottky, 1Amp power diodes (1N5818)
At 0.52mA voltage drop: 0.127-0.132V
At 0.20mA voltage drop 0.101-0.109V
At 0.05mA voltage drop 0.065-0.070V
==========================================
Test on 4 different Germanium Transistors
Using C-B junction (2N1305) (i.e. used as a diode)
At 0.52mA voltage drop 0.198-0.205V
At 0.20mA voltage drop 0.163-0.172V
At 0.15mA ~0.155 estimated
At 0.10mA 0.15V estimated
At 0.05mA voltage drop 0.127-0.130V
==========================================
Test on 3 different 1N5711's low power schottky diodes
At 0.52mA voltage drop 0.308-0.325V
At 0.20mA voltage drop 0.295-0.320V
At 0.05mA voltage drop 0.255-0.270V
=========================================
Test on 4 parallel connected 1N5711 schottky diodes
At 0.52mA voltage drop 0.274V
At 0.20mA voltage drop 0.261V
At 0.05mA voltage drop 0.224V
< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >
|