Smartphone app to measure shutter speed

[Denis L]

I wanted to correct a misunderstanding I had yesterday about "a bit more advanced testing" but can't edit the original post. I thought the uneven exposure was due to non-parallel shutter curtains while it appears it's due to the shutter slit width varying throughout the shutter's movement across the film gate. The testing method should still work to identify these issues, although it should be added that you have to be aware of how the shutter curtains are moving. If they are moving horizontally, you need to cover the film gate vertically (so left and middle part and then middle and right part for example).

 

[reddesert]

I see it's an old thread but I just ordered my first film SLR on eBay and was wondering how to test it for light leaks, shutter times and light meter and auto-exposure modes on zero budget and without ruining film, which is now getting more and more expensive to buy, develop and digitize. So if you have any tips on zero-budget film SLR testing, please let me know!

My two cents, since I contributed to this before, are that if this is your first film SLR and it's a relatively common or conventional camera, don't overthink it. Most 35mm SLRs have shutters that either work or they have some obvious flaw or don't fire at all; or at worst a mechanical slow speed timer is sticky. Hold the camera up to a light without film and fire the shutter at a variety of speeds to see that the shutter is opening all the way, and to check for pinholes in the shutter curtain (not common on newer cameras). Listen to the delay on the slow speeds to make sure the time delay changes and sounds about right - you can use the the free audio version of the Shutter Speed app to time the speeds from about 1/30 to 1 second. (It will not be useful above the sync speed of a focal plane shutter and the audio is not so useful faster than about 1/60 on a leaf shutter.)

If you want to be more accurate then you may need to buy or make a PhotoPlug (they are not too expensive), but you probably don't need to. I would spend more time worrying about (a) the light seals - check for dried up or missing foam near the back hinge and the latch side, and (b) understanding exposure and metering in manual mode, shutter speed, f-stop, etc.

 

[Niglyn]

Hello Everyone,

Making an optical shutter tester is not too difficult.

I use an Arduino Nano £5, a laser transmitter & receiver £6, LCD £6 and some wires £4. All parts easily available on Ebay or even cheaper via Aliexpress.

Photos below & more details on the 'Zenit EM complete CLA ongoing thread (pic heavy)'

I thought about starting a new thread detailing how to build, not sure which section to put it under.

If enough people want them and do not feel capable of programming the Arduino, I'm happy to order the parts in bulk, program the Arduino & post out, for cost price + a beer )

Testing my cameras, the Canon T70, which at the time was reputed to have the most accurate shutter, comes out almost perfect. The Zenits are terrible, with great variation of exposure from one side of the frame to the other & shutter speeds way off the actual setting.

View from above, showing Arduino (bottom left),
LCD (top left)
Two laser transmitters (left)
and two laser receivers (right).

I mounted the laser tx & rx on a simple jig consisting of a piece of plywood and two wooden blocks,
using hot-glue to mount the tx & rx, which allowed time to align them whilst the glue cooled.
A much better jig could be made, with the tx & rx on adjustable arms & all put into a nice project box.

Note:- two lasers shown in the photo, only one is needed.

Two lasers shining at the back curtain ready for testing, the red lights just above the frame are reflections.
Note:- only one laser is needed.

LCD, showing three readings of Zenit set at 1/125.
Tests made at right, middle and left of shutter.
Huge variation of exposure across the frame can be seen.

 

[Alan Edward Klein]

What doe that mean?

You said: "LCD, showing three readings of Zenit set at 1/125.
Tests made at right, middle and left of shutter.
Huge variation of exposure across the frame can be seen."

 

[Niglyn]

What doe that mean?

You said: "LCD, showing three readings of Zenit set at 1/125.
Tests made at right, middle and left of shutter.
Huge variation of exposure across the frame can be seen."

Hi,

Hope the below is a little more clear )

I set the Zenit shutter speed to 1/125

Ignore the second laser in the photo, pretend there is just one.

I positioned the camera so that the laser was aligned to the right hand side of the shutter curtain (right hand laser as shown in the photo).
Then I fired the shutter. The exposure was measured at 1/112.

I then moved the camera to the right, so that the laser was now in the centre of the curtain and again cocked and fired the shutter.
This gave a reading of 1/182

Finally I moved the camera to the right, so that the laser was now to the left on the shutter curtain and again cocked and fired the shutter.
This gave a reading of 1/207

Taking the highest and lowest reading, gives a 59% difference in exposure from one side of the frame to the other, which is over 1/2 stop

The Zenit camera curtains move horizontally from right to left, so it can be seen that either the first curtain is slowing down towards the end of it's travel, or the second curtain is speeding up
and catching up to the first curtain.

The latter is what is happing as described in one of the posts above, where there is a 2mm part of the film unexposed on each photo. this is caused by the second curtain actually catching up to the first curtain before the first curtain has completed it's journey.

 

[Alan Edward Klein]

So how does someone determine the actual shutter speed if you get three readings?

 

[gone]

Phone apps do work, they have light meter apps that work great. But a shutter speed sensor? I don't see how. You can buy or make a simple electronic shutter tester that uses a regular light source and works on Audacity on your computer.

 

[Niglyn]

So how does someone determine the actual shutter speed if you get three readings?

The camera shutter tester is showing that the shutter accuracy on the Zenit is terrible.

The actual shutter speed is 112/ at the beginning
1/182 in the middle and
1/207 at the end.

So you are getting all three exposure values and everything in-between as the shutter curtains travel across the frame.

It is not as simple as recording the sound of a shutter, or taking one reading from a random point during the shutter operation. The two shutter curtains need to be balanced
to give an even exposure across the frame & to ensure the two curtains have not met before they have travelled the entire frame distance, causing an unexposed portion of the frame, as as in the case a poster was experiencing with their camera.

In the case of this Zenit, I have tried adjusting the shutter mech but I am unable to get satisfactory calibration, so it will require a full strip down & cla.

I have another Zenit here, on which I had similar poor accuracy and a very draggy first curtain (as I found, caused by a very gummed up first curtain assist & mirror lift mech). I have given it a full cl this week and now it is sitting here waiting for the a.

 

[Niglyn]

Phone apps do work, they have light meter apps that work great. But a shutter speed sensor? I don't see how. You can buy or make a simple electronic shutter tester that uses a regular light source and works on Audacity on your computer.

Yep, you can make a shutter tester using a phototransistor wired into the soundcard of your computer. You are using Audacity like an oscilloscope and measuring the time the shutter has opened.
The phototransistor interface causes a voltage change on your soundcard input, as if there was no audio & then you gave it large into a microphone.

The phone app works in much the same way as the PC/Audacity method, by building a little phototransistor circuit which plugs into the phone's mike jack. The app on the phone measures the silence/loud period of the audio input.

The Arduino method, again, detects a changing light source, I used a laser as it is cheap, easy to use and gives a focused beam.
There is no poking around with the computer or phone inputs and can be used as a stand alone device.

You could of course use an oscilloscope, light source, phototransistor & a few other components to do the same thing.

I am not the first to use an Arduino to measure & adjust shutters, there are others around 'tinterweb. I just had the bits lying around & decided to make one, posting on here in case anybody would also like one but maybe is not as knowledgeable with electronics or microprocessor programming.

 

[Niglyn]

I think there maybe a misunderstanding of how a focal plane shutter works.
It does not instantly open and close. There are in fact two shutter curtains,
the first opens, gradually exposing the film, as it travels from one side to the other.
The second curtain now follows the first, gradually blocking the light as it travels.
This speed will be the flash sync speed.

For faster shutter speeds, the second curtain starts moving across the film, to block the light, before the
first curtain has fully travelled across, so only part of the film is exposed at any given time.
This is why you cannot fire a flash at higher speeds.

So a sync speed of 1/30, the first curtain will fully open before the second curtain starts to close.
The whole film frame is exposed to light, so a flash can fire.

For 1/60, the second curtain will start to close when the first curtain has travelled half way across the frame.
For 1/120, the second curtain will start to close when the first curtain has travelled a quarter

The shutter mech only actually moves at one speed, even if the camera is set to 1/30 or 1/1000.
(this is proved by my two laser shutter tester that can actually measure shutter curtain travel speed)
It is the time before the second curtain starts to travel, that controls exposure.

Here is a good video showing a vertical focal plane shutter in slow motion
(164) Inside a Camera at 10,000fps - The Slow Mo Guys - YouTube


Here is a good video explaining how a focal plane shutter works and although not discussed directly,
it can be seen why both first & second curtain must be balanced for correct exposure.
(164) Shutter Speed - Focal Plane Shutters - Episode 5.1 - YouTube

 

[Alan Edward Klein]

Phone apps do work, they have light meter apps that work great. But a shutter speed sensor? I don't see how. You can buy or make a simple electronic shutter tester that uses a regular light source and works on Audacity on your computer.

I use my cellphone to record the sound and then measure the results in Audacity. It;s accurate from 1/125 and slower.

Here's a test of 1/15th of a second on a large format lens with built-in shutter. .067 seconds= 1/15. I actually got .064 seconds measured.

 

[Alan Edward Klein]

I think there maybe a misunderstanding of how a focal plane shutter works.
It does not instantly open and close. There are in fact two shutter curtains,
the first opens, gradually exposing the film, as it travels from one side to the other.
The second curtain now follows the first, gradually blocking the light as it travels.
This speed will be the flash sync speed.

For faster shutter speeds, the second curtain starts moving across the film, to block the light, before the
first curtain has fully travelled across, so only part of the film is exposed at any given time.
This is why you cannot fire a flash at higher speeds.

So a sync speed of 1/30, the first curtain will fully open before the second curtain starts to close.
The whole film frame is exposed to light, so a flash can fire.

For 1/60, the second curtain will start to close when the first curtain has travelled half way across the frame.
For 1/120, the second curtain will start to close when the first curtain has travelled a quarter

The shutter mech only actually moves at one speed, even if the camera is set to 1/30 or 1/1000.
(this is proved by my two laser shutter tester that can actually measure shutter curtain travel speed)
It is the time before the second curtain starts to travel, that controls exposure.

Here is a good video showing a vertical focal plane shutter in slow motion
(164) Inside a Camera at 10,000fps - The Slow Mo Guys - YouTube


Here is a good video explaining how a focal plane shutter works and although not discussed directly,
it can be seen why both first & second curtain must be balanced for correct exposure.
(164) Shutter Speed - Focal Plane Shutters - Episode 5.1 - YouTube

If I understand you correctly, then any point across the film is exposed for the length of shutter time selected. So if the curtains are adjusted correctly, and both cross the full equally in speed, then you should be able to select any point to place the light to measure the shutter speed.

 

[Niglyn]

If I understand you correctly, then any point across the film is exposed for the length of shutter time selected. So if the curtains are adjusted correctly, and both cross the full equally in speed, then you should be able to select any point to place the light to measure the shutter speed.

Correct.

Therefore, the invers of this is true. If one selects different points in the shutter travel to measure and the results vary, it shows that the curtains are not travelling at the same speed and the exposure is varying.
Testing in one place or making an audio recording will not show this type of fault on a focal plane shutter.

I don't think these old cloth horizontally travelling shutters will ever be that accurate. There is no indexing between the two curtains, they each just do their own thing.
Add to this 30+ years of wear, dust, crud and aging lubrication.

 

[Alan Edward Klein]

Correct.

Therefore, the invers of this is true. If one selects different points in the shutter travel to measure and the results vary, it shows that the curtains are not travelling at the same speed and the exposure is varying.
Testing in one place or making an audio recording will not show this type of fault on a focal plane shutter.

I don't think these old cloth horizontally travelling shutters will ever be that accurate. There is no indexing between the two curtains, they each just do their own thing.
Add to this 30+ years of wear, dust, crud and aging lubrication.

Well the example I showed above was for a leaf shutter buolt into the lens. Whether that's more accurate then testing a focal plane, I don;t know.

I did check electronically controlled focal plane shutters on a Nikormat FT3 which were shown to be very accurate using the same sound procedure. Of course, you could be right about the curtains. But the test was accurately shown on all speeds from 1 second to 1/125.

 

[Niglyn]

large format lens with built-in shutter

Hi,

Thanks for posting, interesting to see.

Do you have a trace for a much slower shutter speed? I would expect to see two distinct areas of noise, firstly where the leaves spring open, a silent pause for the exposure and then more noise when the leaves spring back closed

The shutter in your lens will be a leaf shutter they and work in a differently to a focal plane shutter.

(164) Shutter Speed - The Leaf Shutter - Episode 5.3 - YouTube

 

[Alan Edward Klein]

Here's 1/4 second on 90mm lens with leaf shutter. Measured spend is .229 sec. .250 would be 1/4 if accurate. The lighter blue is where I measured. Obviously, depending on where one sets those limits defines the measure time.

 

[Alan Edward Klein]

Here's 1 sec.

 

[Alan Edward Klein]

1/2 second.

 

[Alan Edward Klein]

For comparison, heres a focal plane test on an electronic shutter for Nikon N6006. Very accurate at all speeds. This one shows 1/4 second.

 

[Alan Edward Klein]

On my last post, I;m assuming that the light blue I set is the actual full open and close. It measures .248 seconds which is pretty close to ,250. I could be wrong and its on another blip. What do you think?

 

[Niglyn]

Thanks for posting these Alan, they are interesting to see.

The leaf shutters can clearly be seen opening and closing, with the less noisy part between, being the exposure.
A leaf shutter only has one 'curtain' or set of leaves, so timing these audibly seems to work quite well.

Focal plane shutters have two curtains, which move slower than the exposure time.
For speeds below the sync speed, as in your example, audible timing will probably work quite well.

One will hear the mirror flip up & the first curtain open,
a silence during exposure,
then noise again as the second curtain closes and the mirror returns.

Speeds above sync speed, both curtains are moving, with a slit between them (slit width varying with shutter setting)
so there is no silent exposure time to measure.

This slow-mo shows a shutter at 1/1000 Second :-
(165) Slow motion camera shutter - Canon 5D Mark II 2,000 fps - YouTube

Shutter cycle starts at -60mS
First curtain starts to open at 0mS
Second curtain follows first and closes at 6mS
Mirror drops and stops bouncing at 95mS.

So the whole cycle took 155mS
The actual 1/1000S or 1mS exposure took 6mS to complete.


Vertical multi-blade shutters like on your FT3 for example, are filled with precision Japanese engineering and should be reasonably accurate, even after years of service.

I don't think there is a chance of getting an old Soviet era cloth curtain shutter anywhere near this accuracy. Even the genuine Zenit E service manual gives little optimism and uses the phase 'acceptable limits' with without being specific.

After a full strip down, removing old lubricants and CLA, the best I can get my Zenit E to are,
at 1/500 setting, 1/532 on left and 1/419 on right.
at 1/30 setting 1/44 on left and 1/33 on right.

 

[DWThomas]

As a long time compulsive tinkerer in many different areas, I've been following this thread with interest. One question I have in this matter of the apparent variation of speed at different positions in the frame: have you done any tests of repeatability at the same position? To me it seems likely that some varieties of old mechanical shutters, especially focal planes, thanks to both wear and dirt, could potentially vary a lot between actuations. It would be cool to be able to check two or three locations during a single pass (which I suspect is probably where your two sensors were headed). I suppose capturing the data from multiple sources might be more of a challenge.

I have had urges to get into the Arduino/Raspberry Pi stuff, but having once made a living playing with embedded systems, I'm afraid I'll disappear into my basement workshop and never be seen again.

 

[Niglyn]

As a long time compulsive tinkerer in many different areas, I've been following this thread with interest. One question I have in this matter of the apparent variation of speed at different positions in the frame: have you done any tests of repeatability at the same position? To me it seems likely that some varieties of old mechanical shutters, especially focal planes, thanks to both wear and dirt, could potentially vary a lot between actuations. It would be cool to be able to check two or three locations during a single pass (which I suspect is probably where your two sensors were headed). I suppose capturing the data from multiple sources might be more of a challenge.

I have had urges to get into the Arduino/Raspberry Pi stuff, but having once made a living playing with embedded systems, I'm afraid I'll disappear into my basement workshop and never be seen again.

Hi DW,

It is good to know others have been following the thread, please shout out & say hi.

Yes, I have taken multiple shutter readings at one location and it is surprising how inaccurate Zenits are )
So I take multiple samples at each point & take an average.

I have also tried a more modern Canon T70 with a vertical multiple blade shutter & this gives far more consistent results, variation of only 0.1mS in any one place, but even this camera shows a difference in exposure from one side of the frame to the other, below are a couple of outputs as shown on the pc screen, which gives far more info than the LCD

-----laser at bottom of T70 shutter-------------------
Laser1 Start: 105821328
Laser1 Stop : 105825204
shutter Speed Microseconds : 3876
shutter Speed Milliseconds : 3.876
shutter Speed Seconds : 0.004
shutter Speed fraction : 1/258.00
shutter Speed fraction : 1/258
---------------------------------------------------------------

----laser at top of T70 shutter---------------------
Laser1 Start: 141819928
Laser1 Stop : 141824052
shutter Speed Microseconds : 4124
shutter Speed Milliseconds : 4.124
shutter Speed Seconds : 0.004
shutter Speed fraction : 1/242.48
shutter Speed fraction : 1/242
------------------------------------------------------------

Using two lasers was a good idea (and probably still is) as it halved the number of tests required and was able to calculate the travel time of each curtain. However, as I was getting wildly different results between first & second curtain, decided to use just one laser, in case timing errors were being introduced by the code.

Arduinos are very addictive )

I purchased a kit to make a trigger timer, the kit was quite good, but analogue based, using 555 timers & pots to change the values. Whilst it worked fine, it took ages to set up each time, having to fine tune each dial.

So along comes Arduino and with just a Nano & LCD, was able to make a digital readout, which showed what the delays were, making setting up again a breeze. I did offer the idea & code to the manufacturer for free, but after initial communication, the offer was not taken up. Happy to supply details if anybody has this device & wants to add the digital display.

Next was to make my own timer-trigger device using an Arduino. This turned into a monster (did I say Arduino's are addictive? )
It has inputs for sound, light, lasers, all breaking or making.
Not only does it fire the flash after a user set delay, it will also open the camera shutter if the shutter latency is lower than the delay time (warns you if it is not)
It can control three high current outputs, in my case solenoids, which open/close period can be controlled, multiple times in one exposure.
Water drop photos for example, rather than just having the blob coming back out of the water, having a second water drip to hit the first as it rises, gives the exploding crown effect.

Sorry, getting well off topic here

Second water drop coming down, hitting first (on its way up) creates the crown

Arduino digital readout added to an analogue camera timer product.

 

[DWThomas]

Wow! I'm impressed -- love the water drop. Circa 1959, along with another tinkerer friend, we captured a .22 bullet going through a blob of Jello, a bag of flour, and a light bulb. We used an aluminum foil strip with the rifle mounted in a bench rest and a high voltage discharge from a capacitor, plus some crude electronics -- and a relay! The trigger had to be ahead of the target enought to compensate the relay's delay. The end point of the bullet was a steel plate at 45º deflecting into a box of sand. The most amazing part is that my parents allowed us to do it in the basement! We won a science fair award with it. Unfortunately I think the results have pretty much disappeared in the course of a move or three and 63 years. (Ah the good ol' daze!)

Thanks for your additional input on the data. And keep up the good work -- I may yet find an Arduino on my work table.