I recently acquired an Olympus OM-1n, and like a lot of people, I want to get the light meter working with a silver oxide battery. In searching for a solution, I came across the following thread looking for a way to modify its light meter that doesn’t involve adding a series diode to the circuit.
I thought that the question Chris Douglas asked deserved an answer, but I could not find one the internet.
The camera I bought had a broken rewind knob, and I ended up buying a “parts only” OM-1 to fix it. I now had a parts only camera with a working light meter, so I decided to figure out how the resistor values of the light meter circuit could be modified to work with silver oxide batteries. It certainly seems like this should be possible.
Below is a schematic of the light meter circuit as determined from the service manual and by taking apart the circuit in my donor camera. The values of the resistors were measured using a cheapo digital multimeter. The currents are indicated for the purpose of network analysis.
I was somewhat surprised that the meter block has a constant resistance regardless of the exposure setting on the camera. Adjusting the aperture and shutter speed rotates a shaft that disappears into the meter block. What exactly this shaft adjusts, I do not know. Perhaps the spring tension in the moving coil. Setting the ASA appears to move either the shaft or the meter block with respect to the shaft, so all the exposure parameters are handled mechanically. There are actually two photoresistors (one on each side of the viewfinder), each with three terminals. However, they are connected in parallel electrically, and can thus be modeled as a single block. The terminals are connected to a red wire, a pink wire, and a green wire. The red wire connects to the battery, and the pink and green wires connect to the resistor network.
Prior to measuring the resistor values (which required disassembly of the circuit board), I had presumed each of the two halves of the circuit would have similar values, and thus the circuit could be simplified to the following:
The idea is to alter the values of R5 and R6 so that the meter voltage Vm has the same values across the range of photo resistor resistances Rvar with a 1.55 volt battery as it did with a 1.35 volt battery and the original values of R5 and R6. Of course, the resistance of Rvar affects the meter voltage Vm (that’s how the meter works), so this isn’t a simple matter of changing a voltage divider ratio.
My original hope was that simply connecting another resistor in parallel with R5 would lower Vm sufficiently without throwing off the meter too much. Plugging the above equations into an Excel spreadsheet showed me it was not going to be that simple. I could only get the correct value of Vm at a single exposure level. I later had a brainstorm and determined that if the load RL the light meter circuit presented to the photoresistor resistor Rvar was kept the same as in the original circuit, the voltage compensation would be kept consistent across all values of Rvar. This was the key to solving the problem.
To characterize the photoresistors, I pointed the camera at a lighted area which I had previously measured using the meter in my Canon A1. I then measured the resistances of the photoresistors (in parallel) at each aperture setting (1.4 through 16) with an Olympus 50 mm lens mounted to the camera. As it turns out, the pink and green outputs of the photoresistors have significantly different response curves. Using a least-squares curve fitting technique (okay, I eyeballed it) to match each curve to the equation 1/(k+E^γ), the RV1 (Red to Pink resistance) has a k=1.00E-06 and RV2 has a k=1.00E-05. Gamma is 0.5 in each case. The actual measured values were:
EV lux (E) f RV1 RV2
12 10250 1.4 11500 1280
11 5125 2.0 13000 1420
10 2563 2.8 18200 1890
9 1281 4.0 28100 2940
8 641 5.6 42100 4200
7 320 8.0 61200 5760
6 160 11 81200 7640
5 80 16 101500 8810
In view of this 10:1 difference, I felt I had to model the entire circuit rather than rely on the simplified version. I don’t want to turn this into a tutorial on electrical networks, so I’ll just post the transfer function for Vm:
That function is a bit much to deal with, so I decided to select values for (R1, R4) and (R2, R3) separately using the simple circuit, which has a much simpler transfer function. I created an Excel spreadsheet using the equations for the simple circuit across a number of values for Rvar, and adjusted the values of the resistors iteratively until the load resistance RL and Vm for the new circuit was equal to that of the original circuit. Typically, I would converge on a solution by adjusting R1/R2 upward until Vm at Rvar=0 matched, adjust R3/R4 downward until RL matched, then rinse and repeat. When I was finished, this is what the sheet looked like.
I then plugged the new values for R1-R4 into the transfer function for the full circuit, and low and behold, the numbers worked. Images of the results are shown below. Note the ratio of Vm is within 4% over EV ranging from 0 to 16.
I’m pretty confident that this method will work, but the question remains, is it worth doing? Truth be told, I’ll probably just buy the diode based adapter rather than try and modify the circuit board. But I like knowing how it could be done.
I’ve probably glossed over a lot of stuff in this post, so I welcome any questions or criticisms. If someone has access to SPICE or some other simulation program, or can shed some light on how the meter block works, that could also be helpful. Even better would be some input on how to design the meter circuit from first principles.
Olympus OM-1n exposure meter conversion
Yup, another discussion about mercury batteries, but different than I could find in the archives. I don't want to re-start the same old discussion on Wein cells and hearing aid batteries. The best solution to the mercury battery problem is to adjust the meter circuit to use silver oxide...
I thought that the question Chris Douglas asked deserved an answer, but I could not find one the internet.
The camera I bought had a broken rewind knob, and I ended up buying a “parts only” OM-1 to fix it. I now had a parts only camera with a working light meter, so I decided to figure out how the resistor values of the light meter circuit could be modified to work with silver oxide batteries. It certainly seems like this should be possible.
Below is a schematic of the light meter circuit as determined from the service manual and by taking apart the circuit in my donor camera. The values of the resistors were measured using a cheapo digital multimeter. The currents are indicated for the purpose of network analysis.
I was somewhat surprised that the meter block has a constant resistance regardless of the exposure setting on the camera. Adjusting the aperture and shutter speed rotates a shaft that disappears into the meter block. What exactly this shaft adjusts, I do not know. Perhaps the spring tension in the moving coil. Setting the ASA appears to move either the shaft or the meter block with respect to the shaft, so all the exposure parameters are handled mechanically. There are actually two photoresistors (one on each side of the viewfinder), each with three terminals. However, they are connected in parallel electrically, and can thus be modeled as a single block. The terminals are connected to a red wire, a pink wire, and a green wire. The red wire connects to the battery, and the pink and green wires connect to the resistor network.
Prior to measuring the resistor values (which required disassembly of the circuit board), I had presumed each of the two halves of the circuit would have similar values, and thus the circuit could be simplified to the following:
The idea is to alter the values of R5 and R6 so that the meter voltage Vm has the same values across the range of photo resistor resistances Rvar with a 1.55 volt battery as it did with a 1.35 volt battery and the original values of R5 and R6. Of course, the resistance of Rvar affects the meter voltage Vm (that’s how the meter works), so this isn’t a simple matter of changing a voltage divider ratio.
My original hope was that simply connecting another resistor in parallel with R5 would lower Vm sufficiently without throwing off the meter too much. Plugging the above equations into an Excel spreadsheet showed me it was not going to be that simple. I could only get the correct value of Vm at a single exposure level. I later had a brainstorm and determined that if the load RL the light meter circuit presented to the photoresistor resistor Rvar was kept the same as in the original circuit, the voltage compensation would be kept consistent across all values of Rvar. This was the key to solving the problem.
To characterize the photoresistors, I pointed the camera at a lighted area which I had previously measured using the meter in my Canon A1. I then measured the resistances of the photoresistors (in parallel) at each aperture setting (1.4 through 16) with an Olympus 50 mm lens mounted to the camera. As it turns out, the pink and green outputs of the photoresistors have significantly different response curves. Using a least-squares curve fitting technique (okay, I eyeballed it) to match each curve to the equation 1/(k+E^γ), the RV1 (Red to Pink resistance) has a k=1.00E-06 and RV2 has a k=1.00E-05. Gamma is 0.5 in each case. The actual measured values were:
EV lux (E) f RV1 RV2
12 10250 1.4 11500 1280
11 5125 2.0 13000 1420
10 2563 2.8 18200 1890
9 1281 4.0 28100 2940
8 641 5.6 42100 4200
7 320 8.0 61200 5760
6 160 11 81200 7640
5 80 16 101500 8810
In view of this 10:1 difference, I felt I had to model the entire circuit rather than rely on the simplified version. I don’t want to turn this into a tutorial on electrical networks, so I’ll just post the transfer function for Vm:
That function is a bit much to deal with, so I decided to select values for (R1, R4) and (R2, R3) separately using the simple circuit, which has a much simpler transfer function. I created an Excel spreadsheet using the equations for the simple circuit across a number of values for Rvar, and adjusted the values of the resistors iteratively until the load resistance RL and Vm for the new circuit was equal to that of the original circuit. Typically, I would converge on a solution by adjusting R1/R2 upward until Vm at Rvar=0 matched, adjust R3/R4 downward until RL matched, then rinse and repeat. When I was finished, this is what the sheet looked like.
I then plugged the new values for R1-R4 into the transfer function for the full circuit, and low and behold, the numbers worked. Images of the results are shown below. Note the ratio of Vm is within 4% over EV ranging from 0 to 16.
I’m pretty confident that this method will work, but the question remains, is it worth doing? Truth be told, I’ll probably just buy the diode based adapter rather than try and modify the circuit board. But I like knowing how it could be done.
I’ve probably glossed over a lot of stuff in this post, so I welcome any questions or criticisms. If someone has access to SPICE or some other simulation program, or can shed some light on how the meter block works, that could also be helpful. Even better would be some input on how to design the meter circuit from first principles.