Friday, 27 January 2017

Leak Detector for Underwater Camera

I've an underwater camera which doesn't have a vacuum system or leak detector and this scares the life out of me, especially when taking such expensive camera equipment underwater! So I had some bits lying around that could make a useful leak detector.

I've had a leak on a camera enclosure and strobe before and the only time you notice its flooded is when you go to use it and you see the water sloshing around inside! I'm not sure a detector is going to help that much but it's worth a try. 

Basically its an Arduino Pro Mini clone monitoring once a second, the water sensor via the ADC. When the sensor reports the presence of water, by returning a non-zero value, the audible alarm goes off and the on-board green LED starts to flash. The audible alert is quite loud and can be heard from outside the enclosure. Power is supplied by two 2032 cells in series i.e. 6v.

The code is written using a low power library which means it consumes virtually no power between the 1 second polling activity.  I can't read the power consumption on my meter as its so low, so worst case its 10mA, so with two 2032 cells I think that provides ~600mAh so 60 hours? As there is a switch on the battery container, you can switch it off when not in use. 

The Arduino Pro Mini clone is programmed from the standard Arduino tool with the FTDI USB programmer but you could get a different board with built in USB interface like a Leonardo/Trinket/Lilypad but the Pro Minis are so cheap.


Moisture Detector ( search ebay for "water detector arduino")
Passive Electronic Piezo Buzzer /Sounder 1-13V 
Pro Mini Atmega328P Board 5V 16M   ( look for cheapest 5v you can find)
6V Button Coin Cell Battery Holder Case Box With On-Off Switch CR2032



#include "LowPower.h"

int led = 13;
const int buzzerPin = 13;

// watch Board type, older ones are 5v,168 newer V2 ones are 5v,368

void setup() {
// initialize serial communication at 9600 bits per second:

  pinMode(led, OUTPUT); 
  digitalWrite(led, LOW); 
  pinMode(buzzerPin, OUTPUT);

void loop() {
// read the input on analog pin 0:

     LowPower.powerDown(SLEEP_1S, ADC_OFF, BOD_OFF);    //Sleep 1 second
     //  Water Sensor Alarm
     int sensorValue = analogRead(A0); 

     //Serial.println(sensorValue); // Put this back in for debugging but remember to remove the lowpower call otherwise it won't work.
    if (sensorValue > 6)
      digitalWrite(led, HIGH);   // turn the LED on (HIGH is the voltage level)
      while (1) //Sound Alarm
        digitalWrite(led, HIGH);
        tone ( buzzerPin , 523 , 200);
        digitalWrite(led, LOW);
    }else // else for test only
      digitalWrite(led, LOW);    // turn the LED off by making the voltage LOW


Here is the finished item:

The camera enclosure doesn't have a great deal of space inside so the next challenge is to fit it all in!

The CPU is really under utilised but going forward, it will start to take on more responsibilities, such as fire the strobes and monitoring pressure.  Watch this space.

Monday, 3 August 2015

Video Light for stills

I tried using one of my video lights as a light for my stills camera. I mounted it like so:

The articulated arm used to have a strobe on it but it flooded hence the need for some light!
Here are some examples of the picture, mostly in the 20-30m range, vis was pretty good at 5-10m, across multiple dives.

Crab was out in the open so no rocks to obscure the light. No particulates in the water either or very little, treated it like a strobe where you want as little of the beam in front of the lense.

White balance incorrectly set for the video light, suspect I had used it topside and not corrected to use the video light. I've dedicated a button on the Canon s95 to reset white balance, just do it before taking a series of shots.

Positioning of video light on arm all important.

Love these little fellas... nice and in focus with the light. These are old ropes on an LST tank landing ship.

Again, I think I was shooting ambient light, the vis was that good at 25m! But switching to using the video light you get the colour temperature differences..

Capstan on the Susan B. Anthony troop ship, stern. This was taken about 1m away and you can just see the light affecting the colour of the fried egg anemones. 


Cone of light visible but its in-focus and adds context..

Tuesday, 19 May 2015

New double video light rig

Trying something new for a video light right. When the lights are bolted onto the side of the Goodman handle I find the backscatter to be quite noticable, so this new rig is to push the lights further away from the camera and at an angle to reduce it... Its experimental at this stage... but here goes...


Loc-Line Coolant Hose Pack
Standard Pack of Coolant Hose containing 2 x 5.3/4" of 1/4" hose (10 Segments),
3 Nozzles (1/16, 1/8 & 1/4")
1/8" &1/4" Male NPT Connectors.

EPDM Rubber Lined P Clips Air Hose Pipe Clamps Cable Retention Tube Clip Mikalor
( 251259923794 )
Material: W1 Mild Steel Zinc Plated DIN 3016 (PHC-**)
Sold Per: 5 (Five) Clips Pack
Size: 26mm Pipe Retaining Clamp
£3.85        for the 4000lm torch

£1.35 for 10

( 111035553279 )

100g Polymorph Thermoplastic

To fix the loc-line to the Goodman handle I've cut the threaded section off the loc-line base. Its quite thin so a junior hacksaw will do it easily.

Then bolted the base to the handle with the bolts threaded through the base.

Once both loc-line bases are bolted securely to the handle I made the connectors for the torches. For these I used the nozzle ends provided in the pack, choosing the two largest ones. The finished product looks like this:

To make the hole for the bolt I used a soldering iron, it could be drilled but its quite a hard polished plastic so melting a hole was a lot easier.

Passing the bolt through...

I later added a washer either side but its not shown in these photos..

To make the hose clip connection more rigid, I used a small of piece of polymorph plastic to stabalise it. Take a small piece, say a lump the size a coin, melt it in boiling water and then place it between the hose clip and the nozzle end when you tighten the bolt up. It will form into the correct shape and solidify.

You can use the soldering iron to melt the polymorph plastic around the edges to make it neater.

To plug the whole thing together, lay the loc-line pieces in boiling hot water for a few minutes and then apply a little pressure.

The torches I've used in this example are the two LED torches I've previously reviewed on my blog, one with reflector removed for video and just a normal 4000lm torch with reflector.

I've added a chord to tie the torches to the handle, just in case the loc-line is not strong enough. Its 1/4" loc-line, out of water is seems a little flimsy but I'm hoping in water it will be ok. Its the same as the Ikelite kit.

I over tightened one of the nozzle heads and cracked it, so i'm going to reinforce them with some liquid metal or resin.. not done that yet.

First outing this week...

OK, the arms work really well, the connections stayed tight and secure for 6 dives. In the water they support the weight of the torches fine. Out of water they move a little.

The non-modified torch with its reflectors still in, was no good for video, hot spot ruined the video.

The video camera also developed a nasty fault of not turning on... so not to successful!

My 3mm gloves got ripped by the bolts connecting the arms to the goodman handle, so I think the bolts need to be reduced in length.

Friday, 9 May 2014

3x CREE XML U2 LED 4000Lm Dive Light Tear Down for a new video Light

Bought off ebay a triple XML torch for £30. It came from a UK seller and so arrived in 2 days but suspect it could be sourced more cheaply from China..

I'm following this thread as a lot of the hard work has already been done...

Here goes then:

 #1 Remove the chrome Bezel, I used some long nose pliers.. it will move but its very very tight...

This leaves the glass lense jammed in with an o-ring across the top. I used my nails to pluck the o-ring out.

The lense fell out with a little vigorous banging of the light head onto the floor. The LED module can then be pushed from the inside of the lighthead and it just falls out.

There is another smaller o-ring inside the light head, the glass sits on this and hence there is an o-ring above and below. You can easily miss this lower o-ring. I've left it in-situe as I'm not sure I can get it back in again.

Here is a top view of the LED module, the spring is in the centre and the hall effect sensor sits in a small groove in the metal casing. Interestingly, there is only one sensor while the slider knob, is pushed forward to make the light intensity increases. My only thought is that the magnet(s) in the slider has some sort of coding/positioning inside to indicate position on the length of the slide? Its seem to increase the light output based on the strength of the magnetic field, as you slide the magnet along it gets closer to the hall effect switch. Feels like 10 increments to full power?

There is a small knurled ring to keep the Slider magnet in place, this can be unscrewed enough to let the magnet come out. Might be tricky putting in back in.
The Light steps through 7-8 power steps as you push the slider forward.

To Disassemble the LED module, push a screw driver in the small slot opposite the hall effect sensor.

The reflector piece for the LED module can be detached by unscrewing it from within the LED module.

Once you have disconnected the reflector module, the LEDs are visible. These seem to be stuck to a washer liberally coated in heatsink compound.

Other observations:

I measured 2.26A on full power at the tail.

Alignment of the magnet slider and the hall effect switch is done by the LED module only fitting in two ways and its always aligned correctly, there are two notches on the module to align it.

The Magnet slider switch can be removed by unscrewing the knurled ring, pushing it back over the o-ring a little and pulling the slider out.

This is the underside of the switch, its got a small rubber insert which gives it the resistance when you push it back and forth.  After  a while I found the switch to stick, put some silicon grease in the slot, this seems to sort it.

Run times:

Initial test show:-

On TrustFire 2500mAh batteries from Dealxtreme, the light lasted ~36 mins...

On Xtar 2600mAh batteries from Amazon, the light lasted ~52 mins.

The lights don't gradually go out, they go off immediately.
The Trustfire are not 2500mAh, suspect they are a lot lower...
Need to double check all these time.. with another run...
I use my test rig here :-

Note the torch and sensor barely fit in the bucket.

Dived this weekend and noticed that the slider switch is impossible to feel in a pair of 7mm mitts, you just can't feel the knob, just have to rotate the light making the motion and hoping it engages and turns off.

It also started to flicker like mad after ~35mins full on use, it was so distracting (like a strobe) I had to turn it off. This makes me think it dies in different ways depending on the battery type (trustfires in this case xtars) . I'll do some more tests.

With the reflector in, its quite a floody light but because of its overall power it does throw a lot more that my other Chinese torch.

Pool shots of this triple XML light ( on left ) and the single version ( on right) with reflectors still in...

Triple light...

Single XML, much tighter beam below...

Single XML again..

Triple XML, less defined spot but much brighter..

All shots on manual 1/80 sec , f2.8, 1000 ISO, Canon S95 with Inon UWL-H100, FOV: 100.8° underwater.

Took it diving last week to 29m ( 95ft) and it was fine, vis was around 1-2m and it was dark, I actually prefered this light to my other Chinese single XML light.. The wider beam was actually more useful at this close range, wasn't much point have a long throw narrow beam.. Only issue for me is you need true >2600mA batteries to get an hour out of it ( two dives in my experience). Cheap batteries and it won't last the day..

In use at low viz... 28m..

Sunday, 20 April 2014

DIY Camera Floats / Buoyancy

I've got a Canon S95 in an Ikelite housing, an Epoque 150DS strobe on Fantasea flexarm and metal tray. Coupled with an Inon UWL-100 28M67 Wide Conversion Lens. Out of water it weighs 2.3Kg approx. In water its negatively buoyant and its hard work.

There are plenty of examples on the internet for DIY floats mostly using closed cell foam. I managed to find some on ebay ( used for surfboard cores). The trade name is DIVINYCELL H80 PVC FOAM. Its pretty robust stuff, you can't make a dent in it with your finger.

It came in 100mm * 100mm * 65mm blocks for £6.20

I tried the entire block on the camera in the local fresh water pool and the camera was way too buoyant. Decided to cut it in half and try again. This proved to be about right for me, slightly positively buoyant but not too much.

To cut the foam I used a new stanley knife blade, the foam does not like to be sawed, it needs to be cut with a blade, push the knife in and out, don't pull it through the foam.

To make the hole for the bungy for attaching to the strobe arm, I used a general purpose drill, be very careful because it can rip the foam to pieces. I did it on slow speed and gently. If you have a bench drill, use that.

You can rub the foam on sandpaper to remove any rough edges from the cutting action.

I also sprayed the foam black with Matt Black auto spray. Needs only two coats.

Took it on a trip to the Red Sea and found that is was a little too positively buoyant, if I let go underwater, the camera always gently floated up round my neck ( its attached to a chest D ring by a small retractable lanyard). It also floated on the surface quite nicely, so when dekitting to get in a rib, I wasn't worried about it plunging into the depths if it was dropped !

It went to 30m no problems i.e. didn't deform and change its buoyancy characteristics.

So, for Sea water rather than fresh water, it could be say 20-30% smaller..maybe next trip..

Wednesday, 9 April 2014

Video Light - Screenshots from Red Sea Videos

The following video captures are from the Sony Action Camera, using VLC in 1080p/25fps/120 degrees.

The full videos will be put on my Vimeo Channel.

All were taken using the DIY Video Light ( Double version) on a recent vacation to the Red Sea.

The first three shots are inside the Thistlegorm, lower decks in the holds, with just the Chinese torch in use not the Video lights... Not sure if you can make out the rows of open top trucks full of trailers and motobikes...

This is with the two light video light turned on. Spotted a nudibranch on the cab of a truck against the ships outer hull/rib. Distance to subject was less than a metre and very little silt in the water..

This is another hold, distance to truck cab is 2m, vis was good again, the hotspot in the top corner of the shot is the chinese torch not the video light. 

Hole in the top of a truck cab, 1-2m range.

Exiting the hold, foreground lit up with DIY video lights, quit silty at this point... you can see the particles in the water.

Truck radiator, cowling long gone. Middle deck at edge of hold entrance.

Daylight shots with the video lights on.

Swimming up to a small pinnacle, "The Alternatives" at Ras Mohammed, Egypt. Its a sequence of shots swimming up to the pinnacle, the lights are only starting to have any effect within the last meter.. The little shinney fish are glass fish.