• Welcome to Photrio!
    Registration is fast and free. Join today to unlock search, see fewer ads, and access all forum features.
    Click here to sign up

My Dollar Densitometer

Street photo Nashville

A
Street photo Nashville

  • 2
  • 0
  • 54
Rome

H
Rome

  • 2
  • 2
  • 67

Recent Classifieds

Forum statistics

Threads
202,547
Messages
2,842,175
Members
101,375
Latest member
JoannaG
Recent bookmarks
0

Alan Townsend

Member
Joined
Dec 7, 2025
Messages
206
Location
Peoria, IL, USA
Format
Multi Format
I use a homemade densitometer that cost only a dollar. Here's how;

1769569361143.jpeg

This densitometer is shown in my darkroom on the baseboard of my 35mm enlarger. It consists of a 3mm CdS cell, a good quality digital multimeter set to ohms, and two wires soldered to the CdS cell on one end, and jacks that plug into the DMM on the other. I bought a bag 20 small 10 Mohm CdS cells on Amazon for $5. I built a mounting board using some thin mdf that includes a strain relief, and used jacks I had already. I did not include the DMM in the cost, since everybody already has one. For highest accuracy, a good quality DMM is required.

It's not well known, but the lowly CdS cell is a very linear converter of light to current, It's completely passive, robust, and accurate provided a good quality DMM is used to measure the resistance of the cell. Then, we need to convert the resistance to conductance and make a few calculations to get us to density, but this is all very simple if you don't mind some number crunching. The enlarger lamp is a convenient way to measure density of materials placed directly on the CdS unit, or of negatives projected there from a negative. I have the enlarger all the way up to where it would make a 3 x 4 1/2 print, and measure my negative densities there. The sensor is about the right size for sheet film directly or 35mm at this size.

To measure a density, we turn on the enlarger, then measure the resistance without the material , and write that down A. Then we place our material to measure on the sensor and measure the resistance, and write that down B. This sensor is a 10M ohm one, so to calculate a conductance in arbitrary units, we calculate 10,000,000/resistance in ohms AR and write that down. Then we calculate the same for B in the same way and write that down BR. The amount of light measured is proportional to these conductances, so we are making a ratio AR/BR. This is the ratio of the amount of light measured. With ratios, the units cancel, so don't. With film, the ratio of two light transmissions is always the same their densities. The common LOG of this ratio is the density. It took a while to explain, but is easy to do. Put 10000000 into the calculator memory, then divide this by the two resistances, find the ratio, and take the LOG. This also works for finding density ratios, only we measure two points on a negative to compare them. If we pick the brightest and then the dimmest point, we have the density ratio of that negative.

The reason I used 10,000,000 is because that makes the calculation small whole numbers, and much easier to do. My meter is an old Radio Shack model that has an RS232 interface for reading on the PC directly. I used to have a simple program for doing these calculations on DOS, but it doesn't work on Windows without changes. When I was in high school, I bought a Science and Mechanics light meter that came with several probes, a small meter with a needle, and a computation circular calculator. They included directions on making a densitometer/sensitomer so I did that. It turns out the probes were CdS cells, the meter was a conductance meter, and the calculator converted everything to convenient units, like density for exampe, EV's and camera setting.

My Minolta SRT101 and my Gossen Luna Pro meter both use the same sensors in the same way, although they also have bridge circuits that adjust logarithmic current meters to work properly. These passive devices have no electronics to fail. The current meters are moving coil type, so can fail mechanically. DMM's also have bridge circuits for each range that were adjusted at the factory. My dollar densitometer is a little better than 1% accurate over a range of 10,000 to 1, which is a density of 4.00. Needs no calibration since we are using ratios. I really enjoy using it even it is a pain sometimes when in a hurry.

Comments and suggestions are welcome.
 
I use a homemade densitometer that cost only a dollar. Here's how;

View attachment 416718
This densitometer is shown in my darkroom on the baseboard of my 35mm enlarger. It consists of a 3mm CdS cell, a good quality digital multimeter set to ohms, and two wires soldered to the CdS cell on one end, and jacks that plug into the DMM on the other. I bought a bag 20 small 10 Mohm CdS cells on Amazon for $5. I built a mounting board using some thin mdf that includes a strain relief, and used jacks I had already. I did not include the DMM in the cost, since everybody already has one. For highest accuracy, a good quality DMM is required.

It's not well known, but the lowly CdS cell is a very linear converter of light to current, It's completely passive, robust, and accurate provided a good quality DMM is used to measure the resistance of the cell. Then, we need to convert the resistance to conductance and make a few calculations to get us to density, but this is all very simple if you don't mind some number crunching. The enlarger lamp is a convenient way to measure density of materials placed directly on the CdS unit, or of negatives projected there from a negative. I have the enlarger all the way up to where it would make a 3 x 4 1/2 print, and measure my negative densities there. The sensor is about the right size for sheet film directly or 35mm at this size.

To measure a density, we turn on the enlarger, then measure the resistance without the material , and write that down A. Then we place our material to measure on the sensor and measure the resistance, and write that down B. This sensor is a 10M ohm one, so to calculate a conductance in arbitrary units, we calculate 10,000,000/resistance in ohms AR and write that down. Then we calculate the same for B in the same way and write that down BR. The amount of light measured is proportional to these conductances, so we are making a ratio AR/BR. This is the ratio of the amount of light measured. With ratios, the units cancel, so don't. With film, the ratio of two light transmissions is always the same their densities. The common LOG of this ratio is the density. It took a while to explain, but is easy to do. Put 10000000 into the calculator memory, then divide this by the two resistances, find the ratio, and take the LOG. This also works for finding density ratios, only we measure two points on a negative to compare them. If we pick the brightest and then the dimmest point, we have the density ratio of that negative.

The reason I used 10,000,000 is because that makes the calculation small whole numbers, and much easier to do. My meter is an old Radio Shack model that has an RS232 interface for reading on the PC directly. I used to have a simple program for doing these calculations on DOS, but it doesn't work on Windows without changes. When I was in high school, I bought a Science and Mechanics light meter that came with several probes, a small meter with a needle, and a computation circular calculator. They included directions on making a densitometer/sensitomer so I did that. It turns out the probes were CdS cells, the meter was a conductance meter, and the calculator converted everything to convenient units, like density for exampe, EV's and camera setting.

My Minolta SRT101 and my Gossen Luna Pro meter both use the same sensors in the same way, although they also have bridge circuits that adjust logarithmic current meters to work properly. These passive devices have no electronics to fail. The current meters are moving coil type, so can fail mechanically. DMM's also have bridge circuits for each range that were adjusted at the factory. My dollar densitometer is a little better than 1% accurate over a range of 10,000 to 1, which is a density of 4.00. Needs no calibration since we are using ratios. I really enjoy using it even it is a pain sometimes when in a hurry.

Comments and suggestions are welcome.

pretty inventive. I give you that, but it won't put any densitometer manufacturers out of business.
 
Cool, I had a really fancy one given to me by a lab friend who was closing. I gave it away. I think it ended up being scrapped. I just make prints and fiddle around. I usually make several different ring around exposures. Dry the prints and come back later.

My brain needs time to figure out what I like at least overnight.

I watch @Andrew O'Neill working in his lab, he clearly knows the way to use this equipment, I'm just slow.
 
it won't put any densitometer manufacturers out of business.
Nope. That already happened long ago.

My dollar densitometer is a little better than 1% accurate over a range of 10,000 to 1, which is a density of 4.00.
While the electrical part may have this accuracy (repeatability), things might get (very) sketchy starting around 1:100 due to issues with flare etc. The general effect is that the density measurement ends up lower than anticipated.

The general principle of course works.
 
To measure a density, we turn on the enlarger, then measure the resistance without the material , and write that down A. Then we place our material to measure on the sensor and measure the resistance, and write that down B. This sensor is a 10M ohm one, so to calculate a conductance in arbitrary units, we calculate 10,000,000/resistance in ohms AR and write that down. Then we calculate the same for B in the same way and write that down BR. The amount of light measured is proportional to these conductances, so we are making a ratio AR/BR. This is the ratio of the amount of light measured.
No !
For a CdS photoconductive cell, the conductance S (inverse of resistance) is NOT proportional to illumination L. Rather, there is a relation S proportional to L^gamma, where gamma is generally in the range 0.35--0.8. See example below; it is a little difficult to find such a graph, since CdS cells nowadays are used more to detect presence versus absence of light rather than as a light-measuring device.

How come, will you ask, CdS cells are (were) used in cameras and hand-held light meters? The design of the electrical circuit of these light meters takes into account the properties of the particular model of CdS cell, via the shape of a cam (needle-follower) or the ad hoc graduations on a galvanometer, or other tricks.

This also works for finding density ratios, only we measure two points on a negative to compare them. If we pick the brightest and then the dimmest point, we have the density ratio of that negative.
You probably mean transmission ratio, or opacity ratio. One seldom needs to take the ratio of densities; differences of densities, yes.

You can make a 10-dollar densitometer using a photodiode. That has a linear relation of current versus illumination, as long as it is operated in photovoltaic mode (see below). Photovoltaic mode ensures the diode has zero volt across its terminals; this is accomplished with just an operational amplifier wired as a transconductance amplifier (google is your friend): just the amplifier, the diode, and one resistor. Logarithmic conversion will cost a little more.


CdS1.png
 
Thanks for the response, and Yes.
For a CdS photoconductive cell, the conductance S (inverse of resistance) is NOT proportional to illumination L.
It is within some limits over a limited range of values. With my setup and appropriate lamps, I begin to go out of linearity outside of 1000 to 1 range, and the error is about 1% at 10000:1. These ranges require specific amount of light. With my enlarger as shown at full aperture, I can only cover a density of 0 to 3.00. To measure higher, I need more light, so use a superbrite white LED for that. My ortho litho film goes up to 4.5 density range, so reading that has about a 3% error. In photography, a 10% error is acceptable usually. At higher ratios, the CdS cell starts going way over with more nonlinearity. I verified the linearity using an acurate illumination meter that I bought. Since these are more common, they are much less expensive the photo meters or densitometer.
Rather, there is a relation S proportional to L^gamma, where gamma is generally in the range 0.35--0.8. See example below; it is a little difficult to find such a graph, since CdS cells nowadays are used more to detect presence versus absence of light rather than as a light-measuring device.
Which is unfortunate, since they work rather well, especially with computer modelling applied to the data
How come, will you ask, CdS cells are (were) used in cameras and hand-held light meters? The design of the electrical circuit of these light meters takes into account the properties of the particular model of CdS cell, via the shape of a cam (needle-follower) or the ad hoc graduations on a galvanometer, or other tricks.
Yes, they used much better CdS cells than mine, and all kinds of tricks. My Gossen Luna Pro covers an amazing range of light intensities, not sure but I believe about 20 fstops, about a million to one. That much takes lots of tricks like bridge circuits and specially designed meter coils and possibly cams, not sure about that one.
You probably mean transmission ratio, or opacity ratio. One seldom needs to take the ratio of densities; differences of densities, yes.
Density ratios are required for so called alt printing. Actually, all printing is alt today.
You can make a 10-dollar densitometer using a photodiode. That has a linear relation of current versus illumination, as long as it is operated in photovoltaic mode (see below). Photovoltaic mode ensures the diode has zero volt across its terminals; this is accomplished with just an operational amplifier wired as a transconductance amplifier (google is your friend): just the amplifier, the diode, and one resistor. Logarithmic conversion will cost a little more.
And the lamp will need feedback control and the system will need frequent calibration. Op amps have linearity limitations as well. Photodiodes are limear, but a million to one? Not without seperate gain stages with discrete ranges.

I hate to mention I am an electrical engineer who has worked mostly with embedded control systems rather than optical metering, but I do like to understand things. Thanks for the feedback.
 
A project that an Electrical Engineer would not be able to ignore. . . I must start making . . .

Thank you
 
You tested your theory by reading the densities on a commercial step wedge?

How about a null meter. Raise and lower the enlarger head (no lens) with the negative material over the sensor to null your meter. Calculate the density based on the intensity of light as produced by the enlarger height.
 
Are you checking against a calibrated step wedge? That's how I do it when using my various densitometers to read "someone else's" negatives.

For my own, I don't care but when I read other people's negatives I want my readings to be sensibly compared to a calibrated standard.

So my usual workflow is... create and develop a sensitometry step wedge in a batch of film I'm processing. When it's dry, I'll pick whatever densitometer I feel like using that day and read it as-displayed zeroed on B+F (or subtract out B+F after reading). I don't care that the densitometer can't calibrate because I'm just checking my processing and graphing for my own pleasure.

When I do it for someone else, I'll use the visual densitometer and compare their negative step to the closest calibrated step wedge step and interpolate the readings, that way I'm never more than 0.15 away from a calibrated reading.
 
You tested your theory by reading the densities on a commercial step wedge?
+1

Or, use the aperture of the enlarger lens. Start at, say, f/4. note ohm-meter value. Close to f/5.6: current should be halved, i.e. "resistance" twice as much. Close to f/8; resistance again twice larger... Etc...

Please report results.
 
Last edited:
Are you checking against a calibrated step wedge? That's how I do it when using my various densitometers to read "someone else's" negatives.
No, I have no need for that.
For my own, I don't care but when I read other people's negatives I want my readings to be sensibly compared to a calibrated standard.
Good
So my usual workflow is... create and develop a sensitometry step wedge in a batch of film I'm processing. When it's dry, I'll pick whatever densitometer I feel like using that day and read it as-displayed zeroed on B+F (or subtract out B+F after reading). I don't care that the densitometer can't calibrate because I'm just checking my processing and graphing for my own pleasure.
With my setup, in reading two points on a negative for density range, the base plus fog cancels out, so does not need to be read. I am not plotting characteristic curves. I enlarge negatives for alt processes. I read the density range and highlight value of a negative, then apply that to get a desired enlarged negative with a certain density range.
When I do it for someone else, I'll use the visual densitometer and compare their negative step to the closest calibrated step wedge step and interpolate the readings, that way I'm never more than 0.15 away from a calibrated reading.
I do have a few known density filters that I use to estimate density. Thanks for the response.
 
+1

Or, use the aperture of the enlarger lens. Start at, say, f/4. note ohm-meter value. Close to f/5.6: current should be halved, i.e. "resistance" twice as much. Close to f/8; resistance again twice larger... Etc...

Please report results.
I did this 25 years ago with my first, and 3 years ago when I built this one. This very easy kind of test is automatic with me. My enlarger only have 5 stops, so that's only a 32 to 1 range, but yes the conductance, not the resistance, is linear over that short range. I also have a few calibrate ND filters, like 1.00, 2.00, 3.00 which extend that range for testing. I have another light meter I use to measure all of these. There is no doubling of anything with twice the light. It's ratios of two conductances that are used. The aperture on my enlarging lens is not more than a few percent accurate also.
 
You tested your theory by reading the densities on a commercial step wedge?

How about a null meter. Raise and lower the enlarger head (no lens) with the negative material over the sensor to null your meter. Calculate the density based on the intensity of light as produced by the enlarger height.
I have three calibrated ND filters 1,2,and 3 densities that all read properly. That's all I need. There no nulling of my meter, since I use it differentially. Also no calibration needed for same reading.
 
This is very interesting. My math skills are good, but my electronics knowledge is poor. Here are some questions for you.

Are you measuring the resistance of the cds cell by itself, without any voltage applied? I thought cds cells needed a battery and you measured current variation due to light.

Why do you bother with the 10,000,000 factor? Why not just divide the two resistance values, and take the log of that? I don't see how the 10,000,000 factor does anything useful, and just makes things more confusing.

Your original description talks about conductances. I find this confusing, since it's not really clear what you're measuring with the multimeter, and how.

My take on this is that I could place a cds cell and clip the probes of my multimeter to each lead of the cell. I would then measure the resistance with different film densities over the cell. I would then divide these two resistances, and then take the log of the result. This would give me the density difference between the two pieces of film. Is that the gist of it, or is there more to it than that.
 
This is very interesting. My math skills are good, but my electronics knowledge is poor. Here are some questions for you.

Are you measuring the resistance of the cds cell by itself, without any voltage applied?
The DMM has an Ohm function, which is the units for resistance. We need conductance, which is the reciprocal of resistance. So we calculate conductance=1/resistance the units would be Siemens, but that has no meaning on context of density or transmission.

I thought cds cells needed a battery and you measured current variation due to light.

The DMM has an internal voltage reference it uses to measure resistance. It is likely measuring current internally to determine that resistance. Using the DMM current function is likely not sensitive enough to measure the CdS current, and doing that would require a constant voltage reference. The old cameras and meters used mercury cells that put out a nearly constant voltage, but those don't exist any more. So we need to calculate the conductance using the resistance measured by the DMM.
Why do you bother with the 10,000,000 factor?
My CdS cell has a maximum impedance of 10 M Ohms, that's 10,000,0000 Ohms. If I divide 1 by ten million, I get 0.0000001 Siemens. Takes pushing the zero button alot. I can put ten million in the memory once, and use it several times. Siemens is a non useful unit. Nano Siemens would be more useful, but extremely more confusing. My 10M Ohms would be 100 nano Siemens. Multiplying by 10M makes the smallest possible reading a 1. For measuring a density of 3.00, my CdS cell is at about 2M. I write that down as 5 T, which is conductance in the Townsend unit which is 10,000,000 Siemens. This is known as an "arbitrary unit" made up purely for convenience.

Why not just divide the two resistance values, and take the log of that?
That would give a useless number. With a CdS cell, the resistance is NOT linear with illumination. On a log-log scale, it is nearly linear over some range. The conductance is linear on linear-linear scale.

I don't see how the 10,000,000 factor does anything useful, and just makes things more confusing.

It saves from the tedium of pressing the zero button a lot more.
Your original description talks about conductances. I find this confusing, since it's not really clear what you're measuring with the multimeter, and how.
We measure impedance in Ohms, then convert to an arbitrary unit of conductance.
My take on this is that I could place a cds cell and clip the probes of my multimeter to each lead of the cell. I would then measure the resistance with different film densities over the cell. I would then divide these two resistances, and then take the log of the result. This would give me the density difference between the two pieces of film. Is that the gist of it, or is there more to it than that.
No. Measure resistance and convert to conductance, then find the log of the ratios. This is based on a log simplification rule whos name I don't recall.

LOG(A) - LOG(B) = LOG(A/B) So we find the differences of two densities using the ratios of two illumination intensities. Confusing, I know, but the math works because we are using ratios that cancel the units. LOG(A) is the density of the air between the enlarger lens and the CdS cell, which is zero and LOG(B) is the density we are measuring. OR they are both densities we are measuing. In both cases, we get a correct result.
 
One thing I forgot to mention is the color sensitivity of CdS cells. They have very low sensitivity to blue, but good green and red sensitivity. So, reading xray film, the lite blue tint shows up as much more than it would with a silicon sensor. For most black and white work, this shouldn't be a big problem, just need to know and adjust for that.
 
The DMM has an Ohm function, which is the units for resistance. We need conductance, which is the reciprocal of resistance. So we calculate conductance=1/resistance the units would be Siemens, but that has no meaning on context of density or transmission.

I thought cds cells needed a battery and you measured current variation due to light.

The DMM has an internal voltage reference it uses to measure resistance. It is likely measuring current internally to determine that resistance. Using the DMM current function is likely not sensitive enough to measure the CdS current, and doing that would require a constant voltage reference. The old cameras and meters used mercury cells that put out a nearly constant voltage, but those don't exist any more. So we need to calculate the conductance using the resistance measured by the DMM.

My CdS cell has a maximum impedance of 10 M Ohms, that's 10,000,0000 Ohms. If I divide 1 by ten million, I get 0.0000001 Siemens. Takes pushing the zero button alot. I can put ten million in the memory once, and use it several times. Siemens is a non useful unit. Nano Siemens would be more useful, but extremely more confusing. My 10M Ohms would be 100 nano Siemens. Multiplying by 10M makes the smallest possible reading a 1. For measuring a density of 3.00, my CdS cell is at about 2M. I write that down as 5 T, which is conductance in the Townsend unit which is 10,000,000 Siemens. This is known as an "arbitrary unit" made up purely for convenience.

Why not just divide the two resistance values, and take the log of that?
That would give a useless number. With a CdS cell, the resistance is NOT linear with illumination. On a log-log scale, it is nearly linear over some range. The conductance is linear on linear-linear scale.

I don't see how the 10,000,000 factor does anything useful, and just makes things more confusing.

It saves from the tedium of pressing the zero button a lot more.

We measure impedance in Ohms, then convert to an arbitrary unit of conductance.

No. Measure resistance and convert to conductance, then find the log of the ratios. This is based on a log simplification rule whos name I don't recall.

LOG(A) - LOG(B) = LOG(A/B) So we find the differences of two densities using the ratios of two illumination intensities. Confusing, I know, but the math works because we are using ratios that cancel the units. LOG(A) is the density of the air between the enlarger lens and the CdS cell, which is zero and LOG(B) is the density we are measuring. OR they are both densities we are measuing. In both cases, we get a correct result.

Alan, thanks for the feedback.

I'm still thinking you're making the math too complicated. If A is the conductance of the first measurement, then it's 1/(first resistance measurement). call this 1/R1
If B is the conductance of the second measurement, then it's 1/(second resistance measurement). call this 1/R2

Then A/B = (1/R1)/(1/R2) = R2/R1 It's just the reciprocal of the resistance measurements, and just take the log of that to get the density difference. The denser negative will have a higher resistance value, so just put the higher resistance over the lower resistance, and you're done.
 

This article is about my Popular Mechanics A3 Light Meter that I bought when I was a kid and was the basis for my first building of a dollar densitometer about 25 years ago. I used mine mostly in the darkroom, but also built a densitometer for it as well. It was too clunky to use as a light meter, although I did use in on a Crown Graphic groundless a few times. If only that meter movement had a locking function... as it was, it was a two person job using it that way. 🙂
 
Alan, thanks for the feedback.

I'm still thinking you're making the math too complicated. If A is the conductance of the first measurement, then it's 1/(first resistance measurement). call this 1/R1
If B is the conductance of the second measurement, then it's 1/(second resistance measurement). call this 1/R2

Then A/B = (1/R1)/(1/R2) = R2/R1 It's just the reciprocal of the resistance measurements, and just take the log of that to get the density difference. The denser negative will have a higher resistance value, so just put the higher resistance over the lower resistance, and you're done.
You are correct about that. Thanks for the tip. 🙂
 
Photrio.com contains affiliate links to products. We may receive a commission for purchases made through these links.
To read our full affiliate disclosure statement please click Here.

PHOTRIO PARTNERS EQUALLY FUNDING OUR COMMUNITY:



Ilford ADOX Freestyle Photographic Stearman Press Weldon Color Lab Blue Moon Camera & Machine
Top Bottom