So i'm building a camera controller - Ideas please!

[Whiternoise]

After taking a few time lapses in Poland, i realised a few things.

1. Clicking a remote control every 8 seconds for 3 hours gets boring quickly
2. I should probably buy something that will do this for me!

So, on goes the internet and up comes Google. Turns out there are a few 3-rd party intervalometers available. Trouble is they're not all that customisable and they cost an arm and a leg.

[URL=http://www.pclix.com/pages/pclix_main.html]Pclix[/URL is the one that seems most.. usable.. cost: $140 for the device + more for an IR cable for the D90.

Then i wondered, hey, the remote just sends out a pulse of IR - nothing else - that should be replicable with a fairly simple circuit right? (at this point i was thinking, well an Arduino is £20 and IR LEDs cost pennies).

It turns out that a few other people had similar ideas and the protocols for various infra red remotes are known (and have been posted online). As a result there are quite a few guides on how to make your own remotes (certainly cheaper than the Nikon ML-3). So, knowing that there are a couple of people who have implemented intervalometers too - i.e. automated time lapses. A couple of people sell kits or finished products and what surprised me was how much they wanted. A simple device with 10 preset timings cost something in the region of (shock horror) 56EUR. Staggering, given that the components cost under £5 and the box is probably worth less.

http://www.bigmike.it/ircontrol/ - IR Signal needed

http://www.glacialwanderer.com/hobbyrobotics/?p=167 - The basis for this controller (does effectively the same thing, but you can do some other things too), the downside? He mentions that a kit would cost $300+?!?!

http://www.gizmoforyou.com/shop/inde...&cPath=3_58_59 - A design customised and based on the above, note the unit cost... (but, it is professionally made, fully waterproof and can do multiple cameras)



So, where is this going?

In the end i decided to build it from scratch using a well documented processor. The Atmel AVR series is programmable in C which is very cool, so i'm using one of those.

Well, i used a handy little application called Proteus to flesh out the hardware (using an Atmel microcontroller). The reason i used it was because it has the ability to run micro-controller firmware. Whilst i'm waiting for the actual components to arrive, i've been developing the software on my computer and making it do what i want.

So far i've got a nifty little set up like this:

Features:

LCD screen (16x2 bog standard screen, backlit of course)
ATMega168 Microcontroller - programmed in C
IR LED for communicating with the camera
Analog inputs for external sensors (i'll get to this)
Various buttons for input, control, etc



Modes:

Manual mode:

Press the shoot button and it'll take a picture (simple!)

Interval mode:

Bit more interesting now. Set a user defined interval from 1 second to 9 hours, 99 minutes and 99 seconds, the camera will then send a "shoot" signal to the camera every (how ever long you want) until you tell it to stop. In theory i thought i'd add the option of putting in tens of hours, but i thought, well these days triggering every hour won't make much difference so the more photos you take, the better quality the video will be afterwards so no point adding stupidly long delays.

Trigger mode




Here's where it gets interesting. I learnt that some other people had crafted devices that let you plug in say, a microphone, and then the device will take a picture when you clap your hands. Why is this good? Well, it makes high speed photography really easy.

You plug in a mic, set the trigger condition to something high (ie. when the noise level spikes) and then the device sends the signal after a predefined delay. Why a delay? If you want to take a picture of say, something smashing, the sound will go high as soon as the thing impacts the ground. If you set a 30ms delay, you'll get a shot just after the collision. A good example was taking a picture of a cricketer. If you set the delay to zero, you'll get a picture of the ball hitting the bat, if you set the delay to just after, you'll get a picture of the ball flying away. You could also get it to take a picture of a gun firing, etc.

The great thing about the trigger is that it simply converts an input voltage into a 10-bit value (that's 0-1023 by the way). This means you could plug in any sensor feasible and you can customise the trigger threshold to suit the situation and triggering device. So, light meter, microphone, pressure sensor, humidity meter (dunno why but..!), etc will all work.

The beautiful thing about the hardware is that it's compatible with any camera with an IR port, provided the "shoot" signal is known. As far as i know the signals for Nikon, Canon, Pentax and Olympus are online, so i can add it into the software later, only the shoot function is hardware dependent. As i've only got a D90, i can only test to see how it works with Nikon.

Cost wise, so far i've got it down to £15 (not using a printed circuit board). The LCD cost me £6 from eBay so by far the most expensive component. The IC is £1.50, switches come in at around £4 including some nice ones and the rest of the components are dirt cheap (leds, resistors, caps, crystal oscillator all under £2) and of course the battery which will be a basic PP3 9V jobby. The enclosure should be relatively cheap too. I'm going to work on the PCB issue, but it'd likely add at least £5-10 on (from BatchPCB).

But, now i need your help, i'd like to know what you as photographers would like if you had a device that could control your camera. Would you like some other functionality, etc?

Ideas people!!

[kalniel]

I like the scripting functionality of CHDK for my Canon, but I'm not expecting that kind of flexibility from something like this As well as a custom interval I think the next simplest thing to add would be a custom number of shots - with infinite until stopped being like your current setup.

Combine that with being able to start the sequence after a custom delay and you've got a great remote series controller.

The next most useful feature would involve a bit more brand specific information, namely, adjusting a parameter like shutter speed to enable bracketing shots.

[TheAnimus]

I'd definitely say go down the route of a PIC or Atmel, but why use IR?

The cable interface just seams to make more sense? If you've got a time lapse and there is sunlight etc often the infrared doesn't work.


But what might be cool is to have a trip trigger? Use a ultrasounic pair, if anything comes in close, it could fire?

[Whiternoise]

Yeah, CHDK is pretty cool, but a bit beyond the scope of this as you say - unless someone's reverse engineered Nikon's USB protocol...

Yep, custom number of shots would be the next thing to add, with zero being shoot until stopped. Adding an initial delay would be trivial in any case.

It's going to be Atmel, i've already done the code for it, plus a bit of optimising and i've made it reasonably modular (so it's easy to add more modes if need be). To be honest i never really considered PIC, in hindsight perhaps i should have. Initially i was going to use the Arduino bootloader because the language is a bit higher-level, but i think you trade off a bit more flexibility. With AVR-LibC, it's highly portable (you can run the same code from any AVR chip, just change the makefile accordingly) and there are some great libraries to do things like highly accurate delays (which is what i need for the shoot signal) - and all the examples to do with DIY remotes are Atmel based for some reason.

As for why IR? Nikon use a proprietary cable format and as far as i know they expect users to buy their wireless remote these days (unless you've got a D300+ i think). In any case it's a 10-pin cable that you have to buy, so it either means buying cheap Chinese rubbish from eBay and hacking the cables off or buying expensive Nikon ones and hacking the cables off. Either way it costs a lot more than an LED (40p currently). I've never had any problems shooting in the sun with my remote so far, but we'll see.

Trip trigger would be pretty easy to do, simple way would be to use a light dependent resistor (i.e. photoresistor) and point a laser pen at it. Then you trigger when the sensor goes low (when the beam is blocked) and you have a basic trip wire.

If it generates a changing resistance to stimulus, you can trigger from it. You'd need to calibrate it with a resistor so the readings you get are sensible, but that's just trial and error.

EDIT: Sorely tempted to get one of these to use: http://cgi.ebay.co.uk/128X64-OLED-LC...4.c0.m14.l1262

Bit of an overkill though!

[TheAnimus]

EDIT: Sorely tempted to get one of these to use: http://cgi.ebay.co.uk/128X64-OLED-LC...4.c0.m14.l1262

Bit of an overkill though!

That does look quite sweet. But talking to the built in controller i'd recomend using an assembly language rather than C just for performances sake.

But we all know a simple HD44780 (iirc) would be easier and cheaper, with better battery life.

I would still say go for ultrasound.

Photodetection is problematic, you would have to use the ADC as a comparator in all honesty, balancing it with a variable resistor would be a lot of effort and just not work properly. Software calibration would be the only reliable way to slide with the daylight without making the circuit too complex.

So ultrasound, nice and simple, would require no complex calibration, just a simple filter for say 44khz on the echo. That way you wouldn't have to do any FFT in software. I really would advocate this technique over using a photodiode or LDR..... Really would.

You could do a beam, have a reflector or something which it breaks, and have a simple band pass or something to detect only your frequency, but again these things in my experiance are plagued by sunlight. The other option is a PIR, but all to often the off the shelf modules require large heat blobs.

So ultrasound again

[Bobster]

id just tether my camera to my laptop and tell it with the canon software to take a shot every whatever..

[TheAnimus]

liekwise with the pentax software, but the ability to have action trigger is good.

[BUFF]

I would just use the built in intervalometer on my KM 7D ...

aren't there fairly cheap intervalometer remotes on ebay for the D90?

[Whiternoise]

D90 doesn't have an inbuilt and the ones that you can buy are either overpriced or badly made - there are surprisingly few good quality ones. Oh, and they use the cable not the IR. From some google researching, they've been known to damage the camera's input port so i've steered clear from them so far. This is both cheaper and does more (the high speed triggering is pretty useful i think) - and it's customisable too.

As for a laptop, yeah, i would just use Camera Control Pro, but i don't always want to take my laptop with me, CCP is expensive and i think that sums up why i'm not going down that route. The major benefit of using a laptop is that you're not tied to the SD card for memory... Plus, you can't do high speed photography with most laptop programs iirc

@Animus, yeah it's an industry standard HD44780 controlled display, specifically a JHD162A. There are plenty of good C libraries for it so i'm saving time and using them, they can do just about everything i want within reason (nice being able to use sprintf to send formatted functions out!).

Surely OLED would use less power than an LED backlit LCD screen?

I'm liking the ultrasound idea, with an LDR, the company that manufactured someone's custom design basically sells a range of pre-calibrated modules with different triggers (i.e. mic, ldr, etc) so you'd just need to work out once what the right resistor to use is and then marry it up with the device to make sure you sample a good range of lux's.

http://www.rapidonline.com/Electroni...nsducers/36077 something like that? I'm assuming i'd need two though? One for sending out the pulse and one that looks to see if one comes back?

Could use a dedicated module i guess, would save on the pins: http://www.rapidonline.com/Electroni...er-SRF08/60296

And yeah definitely easier to go with the HD44780, but i'm thinking maybe going for a 16x4 rather than x2 on the next hardware revision. Also thinking about adding in an RF trigger somehow...

[Flibb]

There are alternative bits of software for tethered shooting, some do interval


http://thephotogeek.com/choosing-nik.../#DIYPhotobits

just fiddling with
http://www.bernd-peretzke.de/index.p...tware&Itemid=3
which does liveview and interval, and is free.


But really like you hardware solution

[TheAnimus]

Surely OLED would use less power than an LED backlit LCD screen?

Ah sorry I thought it was a graphic one? Those generally draw a fair bit more juice, then again, if its only on for the 15 seconds after someone pushed a button, who cares!

http://www.rapidonline.com/Electroni...nsducers/36077 something like that? I'm assuming i'd need two though? One for sending out the pulse and one that looks to see if one comes back?

Thats the ticket, if your good with your casing, it can still be waterproofed!

Pins.... and modules....... No, generated the 40khz yourself using one pin. Then have an op-amp that is set to around 40khz for the resonance, i'm still at work, and look like i will be till close of NY, so doubt i'll be bothered to dig out any old schematics when i get home...... But its not too hard, and will only take 1 pin for rx, 1 pin for recive.

What you then do is clear a timer when you send, then count before you see a response from the op-amp. Its really straight forward, and i'd guess would allow you to make it fit into a case neater.

If you wanted you could probably get a package with 4 op-amps on (use double stage for the ultrasound Rx might be easier) and then use the rest for a microphone input, the problem is filtering out background noise is a nightmare of maths, so you end up just calibrating it via an ADC so that once it hits a threshold it goes. This is often as useless as an LDR.

[Whiternoise]

Right

Got a nice package of bits and pieces from Rapid this morning, just waiting on a couple more pieces and the all important programmer and i can get started.

Need to do some serious code optimising though, i realised that i could make the battery last about 20 times longer with a bit of creative power management!

[Whiternoise]

Right, here is the prototyped board.

Apologies for the hideous monstrosity on the left, i'm going to replace it with a MAX619 (http://datasheets.maxim-ic.com/en/ds/MAX619.pdf) in the real thing, only needs four small caps - none of this inductor/massive electrolytic nonsense. The MAX only needs 2x 10uF low EMS electrolytics and a couple of ceramics, far more space efficient! The chip there is an LT1307, 5V step up which will be connected to 2 AA's, i ordered it from Farnell before i realised that they stocked the other chip and that it would serve my purpose a lot better. The top strip is the 5V output (currently not connected) and the strip below is the 3V "raw" source from the battery.

The middle section is the LCD header, a quick fix for the moment. The trimmer is for the LCD contrast (set and forget, hopefully). The lozenge just below that is the 4MHz crystal for the microcontroller (when actually made it'll be much closer to the micro).

And of course on the right is the ATMega168V, connected up to the LCD and the LDR just below it ("hot swappable" with mic/ultrasound/thermistor/potentiometer). The button there is a reset button. There wasn't enough room to put any of the rest of the buttons on there, so for the moment testing is limited to grounding the relevant pins by hand with some jumper wire (the effect is basically the same). The pins have internal pull-up resistors enabled which is why there's no need for external ones. In the final "build" i'm going to be using a 3.5mm jack for the trigger input and a 10-20cm cable to go with the LED

I need to solder up the LCD too, so when that's done and the damn programmer has arrived i'll be able to fire up my code and test it. I've checked the power supply and that works as expected, although i'm considering tweaking the resistor that controls the output voltage - it's currently pushing out around 4.75 which is ok, but nearer to 5 would be better i think.

The bottom left of the micro will be where most of the buttons are (the pins with the red wire going over the top of the chip). Also not yet placed are the status/ADC LEDs.

The plan is to use MOSFETs to switch the LEDs on. I think a 2N7000 will do the trick, resistor in front of the pin to lower the output voltage (VCC) to around 3V and then the LEDs are connected straight up to the 3V rail (from the batteries) with series resistors to regulate the forward voltage and current. This is necessary for the IR LED in particular as it needs around 100mA to operate.



I'm going to write the project up (when it's finished) and enter it into the Instructables Digital Days contest, so fingers crossed!

I've added support in the code for power management so everything except the onboard timers are off for around 99% of the time when in time lapse mode. I've also added in code to support Canon and hopefully i'll extend it to Pentax and Olympus when i get round to it.

PS: Ashamed as i am to be abusing Farnell's free delivery whenever i notice a component i've forgotten, they're awesome!

[TheAnimus]

Yeh farnell know the score thou, every student/hobbiest that actually then goes on to do something will just use them forever because they serve you so well.

Its looking really good.

I'd say also mabye add a small tantilum bead cap as a decoupling for the MCU? If your worried about space you could also use some for the MAX step up?

If your are going to have a awsome little all in one box, how about using a rotary encoder for the input device? I always used them because i love the infinate twiddly properties they offer, and they are so easy to code agains?

If your using really high power IR leds, you might find better battery life from one of the maxim IR drivers? That way you could use a 2 to 3v logic rail? Then you'd only be using a charge pump for the LED, that was tuned for it? It would let you make it smaller, and if it was a kit for newbs, it might be easier for them (as they wouldn't have to test the output of that device really).

[Whiternoise]

Yeah i've got the FETs on order as well as the maxim charge pump, it'll look better when it's actually soldered and the LCD header is wasting a lot of rows on the breadboard so it'll be pretty small when it's done.

It's tempting to get a dedicated driver, but since it'll only be pulsed at a 50% duty cycle for 0.1s it should be fine. I'm going to play with the resistor that sets the current to work out how much is needed for reliable shooting (why pump it out to 100mA when 50 might be fine? ).

I need the 5V rail though, for the LCD!

On the plus side, the programmer arrived today so i'm going to test out a blinky light program to make sure it works and then solder the display

[TheAnimus]

Yeah i've got the FETs on order as well as the maxim charge pump, it'll look better when it's actually soldered and the LCD header is wasting a lot of rows on the breadboard so it'll be pretty small when it's done.

yeh and I take it your going to solder it on strip board? You can really reduce the size normally just because you can make breaks!

It's tempting to get a dedicated driver, but since it'll only be pulsed at a 50% duty cycle for 0.1s it should be fine. I'm going to play with the resistor that sets the current to work out how much is needed for reliable shooting (why pump it out to 100mA when 50 might be fine? ).

I need the 5V rail though, for the LCD!

Yeh if the LCD needs 5V (shame ) then there is little point having a driver for the LED, because you need it for the whole thing.

On the plus side, the programmer arrived today so i'm going to test out a blinky light program to make sure it works and then solder the display

Awsome the fun can start!

I remeber when doing my final year project, I was up at about 4 AM almost daydreaming in MASM, before i noticed the loverly smell. I still think they should put some notch or something to prevent tired students putting the MCU in upside down.