Build a shutter tester for Focal Plane shutters - Cheap, Easy & it Works

[Niglyn]

Some ideas for push buttons. I have only listed UK sellers, so shop around.
Of course always far cheaper from the Chinese suppliers and Aliexpress.

For modules, with header pins, so no soldering required (but comparatively expensive)

Large Push Button Cap Board Module for Arduino Raspberry PI 5056517069424 | eBay

Find many great new & used options and get the best deals for Large Push Button Cap Board Module for Arduino Raspberry PI at the best online prices at eBay! Free delivery for many products.

www.ebay.co.uk

or for a set of 8 switches, great for tinkering.

Arduino Keypad 8 Push Button (E17) | eBay

Find many great new & used options and get the best deals for Arduino Keypad 8 Push Button (E17) at the best online prices at eBay! Free delivery for many products.

www.ebay.co.uk

Or for breadboard mounting, select 6x6x4.3mm

https://www.ebay.co.uk/itm/393131936303?hash=item5b887d4a2f:g:uLYAAOSwSUBgKGap&amdata=enc%3AAQAHAAAAwAOwiYHDRJPxlIXZGTpv1iwS6FAj5JOCtm1F87iGc9iLn6dezNwUi5blUcvdJXJOFCSdt6XFF0rDrUvLrqnAt%2BHTm7vxd9%2BRzzwuc5b07%2BBj93mMjfl9usjO49Ix5KRKGXBHhw3B08o6uy7dB2AZMHDCXytSxuYfNAhxMJ6MQDsxaZoO4BhOlO0InFN%2B5AN2JTsc3BsZVadhJaW54psQ%2FLNcKkK41fz9lYgCmh1IHwe2VaQr7QlcBCxPFIQzMVPM2w%3D%3D%7Ctkp%3ABk9SR-rQmtPnYQ

Also search on ebay for 'push-button switch' or 'Arduino button'

 

[snusmumriken]

Averages are coming....Averages are coming......(sung to the coca cola Christmas ad )

Latest wiring diagram below. Can you spot the changes?

Only change is the addition of two push button switches to pins D11 & D12.
These buttons are optional. One resets the average count to zero.
The other is for future ideas, possibly to toggle the LCD to show current or average readings.

In lieu of a push-button, you can just use a jumper wire between D11 or D12 and 0V, just be careful not to touch it to anything else!


View attachment 334401

Oh boo, I was just thinking through how to do that. I was also going to use a button to toggle the display between your original and a running average for repeat tests. Of course it is much easier if you do it, bless you .

Besides knowing about curtain speeds and bounce (what I think of as shutter health), the info I felt I most want to know is how bad each speed is likely to be. So over ten tests for each shutter speed, the display would show the running average speed as a fraction, and maximum deviation above and below in stops and thirds, eg “1/235, range -1/3 to +2/3 stops”

I was going to loop through all the speeds from 1 sec to 1/1000, and the screen would show the current nominal speed and which test it was on, eg “1/250, test 2 of 10” before displaying the running average and extremes as above.

I was dead sure I had a spare Nano in my box of stuff, but now I’ve dragged it out I find I have two Unos, a Leonardo, an MKR Zero and two Mini Pros, plus a dozen or so Hobbytronics Ardulog boards and a Teensy. Everything but! I’ll get the shutter tester up and running with a Uno and perhaps use a Mini Pro for the final build - but I’ve ordered a Nano Every too, which will be here on Monday. Also waiting for my son to get home from Uni so I can demand some Lego.

 

[ic-racer]

I added some buttons to my Wokwi sim. but the interrupts don't work as well as the real deal. Looking forward to your updates.

 

[Niglyn]

Hi, A uno will work fine. Nano is just a cut down version. For new projects, I'm going to move to the ESP32. Cheap as chips (pun intended) from China, cheaper than Nano & far more everything under the bonnet.

It would be possible to do a guided test, so screen says fire 10 at '1/30' then when these have been done, screen says 'fire 10 at 1/60' etc At the end, compile the results & spit them out. However, this would make the tester very inflexible for anything else.

Similarly, one could do 10 firings, then the tester stops & spits out the results, resets back to zero. But again, this makes the tester very inflexible.

As it works at the moment, a rolling average is kept from the last 10 firings (11th gets binned when new result is received). Latest result & and calculated rolling average is displayed at each firing. The average can be reset by the push of a button, so one is not limited to a certain number of firings. Min & max from the rolling average could also be displayed.

Fractions are a nightmare on the Arduino. I have cheated and vulgar fractions are only shown for times less that 1 second. Trying to display deviations in stops is a challenge I'm not going to take on )

 

[snusmumriken]

Just in case this causes anyone to destroy their gadget, my lasers have the 's' and '-' terminals the opposite way round to those shown in your post #36. Whichever side it is, 's' should be connected to +5V and '-' to GND.

 

[snusmumriken]

Fractions are a nightmare on the Arduino. I have cheated and vulgar fractions are only shown for times less that 1 second. Trying to display deviations in stops is a challenge I'm not going to take on

Ah OK, I'll take that on then. I made a spreadsheet to do it yesterday, to check my maths, and it should be simple enough to construct a string for printout/display.

The average can be reset by the push of a button, so one is not limited to a certain number of firings.

The only thing is that if you want to calculate deviation from the target speed, you need to know what the current target speed is. That's why I was going to loop round all the shutter speeds, doing a fixed number of tests on each.

For new projects, I'm going to move to the ESP32. Cheap as chips (pun intended) from China, cheaper than Nano & far more everything under the bonnet.

Curiosity aroused. WiFi and Bluetooth. Is this another photography-related project?

 

[OAPOli]

I'm my setup (single sensor), I check the measurement versus the usual shutter speed sequence, then show the closest one and the difference in EV. Easy enough to add +/- 1 EV if the speeds are off. And too bad for the cameras with 1/10,1/25 etc.. speeds.

 

[Flighter]

Blu-tack no more! I gave in and got a hot glue gun, also cut rebates in the wood for placing the lasers and sensors (at a 32mm spacing) and tidied up the wiring with the glue gun as well and then downloaded version 2.9.7.

 

[Niglyn]

Blu-tack no more! I gave in and got a hot glue gun, also cut rebates in the wood for placing the lasers and sensors (at a 32mm spacing) and tidied up the wiring with the glue gun as well and then downloaded version 2.9.7.
View attachment 334435 View attachment 334436

Yay, nice work & you will not regret owning a hot-glue gun (unless you leave it on for hours & it over-heats)
I guess you are now ready for code 2.9.9 )

 

[Niglyn]

The only thing is that if you want to calculate deviation from the target speed, you need to know what the current target speed is. That's why I was going to loop round all the shutter speeds, doing a fixed number of tests on each.

Standard deviation is now included, based on the spread of results, not variation from target speed.
The Nano memory is now completely full, I have had to take out some stuff for the averages & SD to fit.
The ESP32 offers more memory, faster speed and duel processor. This would allow the simple shutter tester to evolve, to set a target shutter speed, for example.

Results can also be output to a csv file & then fed into excel or LibreOffice, for analysis.

 

[Niglyn]

It's here, just in time for the end of April Fools day. Version 2.9.9 Change the .txt file extension to .ino

Now includes running average shutter & curtain speeds & standard deviation. As well as shutter bounce detect & flash sync.
Shutter blanking will also show & will be obvious by the unbalance readings or second laser reporting no change at all.

Alas, it comes at the cost of totally filling the Nanos little brain, so nothing else can be added, without having to lose something.
Library files must be added, as well as the built in Wire,

LiquidCyrstal_I2C
LibPrintf
RunningAverage

They are added the same way as LiquidCrystal of which there is a post on here with a link to the youtube. There are lots of getting started with & loading library videos on the youtube. If you get stuck, just ask here.

The LCD can only show the current values, more data is sent to your computer screen via 'serial monitor' function in the Arduino IDE computer program.

 

[Flighter]

Version 2.9.9 up and running!

Thanks again for this great project, not only can I now check how far off my shutter speeds are but I've also enjoyed getting a taste of Arduino.

Might turn out to be an expensive project as I think I need a few CLAs!

 

[snusmumriken]

OK, mine is up and running, though still waiting for the LCD display. I turned my lasers sideways because I noticed that the beam was elongated, and my shutter runs horizontally. Perhaps accuracy would be boosted by blasting the lasers through 1mm drill holes?

 

[snusmumriken]

I've only tested with version _2_9_6 so far, but I need some help with interpretation.

I was fairly sure my M6 needed a CLA, and it looks like it really does. Although the slow speeds (1/60 and below) are all bang on, all the higher speeds are way out. However, I don't fully understand these results. Can somebody walk me through them, please? (1/1000 at the top, running down to 1/125 at the bottom). Chiefly I don’t understand what is happening in the mechanism to make the laser 1 and laser 2 results so different.

I am surprised by the shutter bounce finding. Would I not expect to see evidence of it on negatives?

 

[Niglyn]

Chiefly I don’t understand what is happening in the mechanism to make the laser 1 and laser 2 results so different.

Hi,
I think you win the prize for best tester. the Lego looks great )

Each sensor should only be seen once and blocked once per shutter cycle. It is only the first seen/blocked time that is used for calculation. If the sensor is seen more than once, this is recorded as a bounce, but the times discarded.

'first curtain bounce' refers to the first sensor, the one on the right, so it is really not possible to have an actual shutter bounce there, so what is causing it, I do not know.

It is misleading that I have used the phrase 'curtain speed'. It should be 'travel time'.
It can be seen that curtain 2 has a longer travel time (16mS) than curtain 1 (7mS) , so curtain 2 is taking over twice the time to travel the same distance as curtain 1.

The left side seems to have the correct exposure for each of the speeds, but the right is always around 1/1000.
It is almost as if the second curtain is jumping out into the gate (blocking first sensor) and then waiting till the correct time to continue it's journey, thus giving the correct exposure on the right.
If the second curtain is jumping out & then back in again, until it's correct time to travel, this would cause the bounce warning.

However, until the cause of shutter 1 bounce has been determined, I would not rely on these results. Do you get shutter 1 bounce on lower speeds?

 

[snusmumriken]

Thanks, Nigel. Here is the readout for the lower speeds. They look pretty good.

 

[snusmumriken]

And weirdly the higher speeds look a lot better this morning. I have changed nothing, but I put the camera away overnight.

 

[Niglyn]

Perhaps accuracy would be boosted by blasting the lasers through 1mm drill holes?

I was thinking, mount the receivers inside a project box (with hot-glue - of course) with1mm holes for them to peep through.

Same thought for the lasers, but glue a piece of wood inside the box first, that has two holes drilled in it, at 32mm centres, the same diameter as the brass laser barrel. This would give a good mount for the lasers.

It would offer mechanical protection to the project. My sensors are always getting knocked out of alignment.
Arduino & LCD could also be built into one of the project boxes.

 

[snusmumriken]

I think you’d have to restrict both ends of the beam, so that the ‘exposure’ is crisp at the shutter blinds. But then, lining up two sets of holes with lasers and receivers would be quite a challenge. I’m mulling it over!

 

[Niglyn]

I think you’d have to restrict both ends of the beam,

A 'professional' tester uses just one light source & the sensors are in a little box with slits cut to allow the light through, so having an unrestricted laser should not decrease accuracy.

If I ever box mine up, the lasers and rx sensors would be reversed. with the box containing the rx sensors fitting against the film plane. It was just easier to line up, having the lasers shine on the curtain.

If I port the code to an ESP32 (the Nano memory is full) it would be easy to add an alignment routine, using Bulb on the camera, to ensure camera gate is aligned. Could also add an Oled screen and have far more data displayed.

However, not sure whether to do that or move on & build a light meter. The light meter maths maybe too complex & there are different equations for incident & reflected light.

Might just go back to brewing beer ) (no Arduino required)

 

[snusmumriken]

A 'professional' tester uses just one light source & the sensors are in a little box with slits cut to allow the light through, so having an unrestricted laser should not decrease accuracy.

If I ever box mine up, the lasers and rx sensors would be reversed. with the box containing the rx sensors fitting against the film plane. It was just easier to line up, having the lasers shine on the curtain.

If I port the code to an ESP32 (the Nano memory is full) it would be easy to add an alignment routine, using Bulb on the camera, to ensure camera gate is aligned. Could also add an Oled screen and have far more data displayed.

However, not sure whether to do that or move on & build a light meter. The light meter maths maybe too complex & there are different equations for incident & reflected light.

Might just go back to brewing beer ) (no Arduino required)

Exactly what I was thinking, about reversing the camera and sensors. I’d been following that other thread too!

FWIW, the Nano Every has 50% more memory and the same dimensions and pin-outs as the Nano.

 

[Niglyn]

Here it is.....drum roll.......
Version 3.0.1
The second pushbutton now toggles the LCD between the latest and average results.
First pushbutton resets the average. Averages are calculated from the last 10 samples.
Both buttons & LCD are optional, if you want a really cheap shutter tester (that works) )

Changed the wording to 'curtain travel time' to make it clearer. It is the time it takes for the shutter curtain to pass the two lasers. So longer the travel time, slower the curtain is travelling.

Changed the baud rate to serial monitor to 115200, so output is much quicker. (you will need to change the baud rate on the drop down box in the serial monitor)

The poor little Nano is now totally stuffed with code, so other than bug fixes, there is nothing else I can add. (so much for a simple shutter tester )

Please post here photos of your creation (Lego is winning so far )

 

[ic-racer]

I'd like to see this in continued evolution. Keep the older versions available, so people can build it as simple or complex as they like. It is one of the best shutter tester projects I have seen on the internet for focal plane shutters.

I'm actually building a 'pro' style tester from plans, but it uses the old school diffuse light and tiny openings for the sensors. Lasers are step ahead.

 

[Niglyn]

I'd like to see this in continued evolution. Keep the older versions available, so people can build it as simple or complex as they like. It is one of the best shutter tester projects I have seen on the internet for focal plane shutters.

I'm actually building a 'pro' style tester from plans, but it uses the old school diffuse light and tiny openings for the sensors. Lasers are step ahead.

Thanks for your kind comments.

I have been watching the other thread. No reason this one could not be put into a pretty box. I might do this as the sensors keep getting knocked. I would add another two sensors at 36mm for flash sync. The receiver modules for the KY-008 laser are really designed for 880nm light, they do not work that well with white light. The also have an amp & Schmitt trigger built in, so makes using them easy, without the need for additional components. I have not found anything else that is as good.

I might put this project to one side & work on a light meter. There are some good boards around with built in oled screen & lipo battery charger. Might also be fun to build new electronics for a defunct camera. I will keep an eye out on ebay.

 

[Flighter]

Cool, thanks for v3.0.1. By good luck my button switches arrived today so I was able to update my tester tonight. K1 and K4 used to give a bit of separation.