housing-iso_l housing-exploded-iso_l

Disclaimer

I am not a professional underwater videographer/photographer, nor am I a mechanical engineer or have any other official credentials.  I’m not responsible if you choose to try this yourself and end up with a waterlogged hunk of plastic and metal, but it’s really not that hard either if you choose to try this, I’d encourage you to do so.  Just make sure you test things in small steps first before risking your expensive goodies.  Also don’t let the fact that you’re using a camera distract you from your responsibilities under water.  Keep checking your gauges, and watch your buddies… even the ones not in front of the camera!  A camera is a lot to think about beneath the waves (you’ll be surprised), so if you’re not ready for it, don’t take it!

During the summer of 2004 I’ve finally got around to gearing up and getting involved with a local SCUBA club.  This is not an inexpensive proposition,
but this summer should see the end of the major budget hits for a while.  I went on my first dive trip with the EUC (Etobicoke Underwater Club) to Tobermory for the August long weekend.  A couple divers brought their digital cameras with waterproof housings and were able to get some great shots while sitting on the bottom.  I’d brought my video cam and got lots of video on the surface of activities on the boats, but that wasn’t really why I was there…  I want submerged video.  Ikelite makes a housing for my Canon ZR-70, but the cheapest I could find it for was $640 US.  I paid $440 US for the camera itself!  Now I understand these nice cases give you pretty much full control of your camera using zoom and start/stop recording as well as full use of the digital still camera capabilities of this particular model, but I reasoned about this, and I don’t feel the need for any control over the camera besides pointing it at interesting things.  My reasons are as follows.

1)  The digital still quality of this camera is poor by my standards, and wasn’t really on my list of requirements when I bought it.  – Don’t need it.

2)  I can’t remember where I heard this, but it made too much sense to ignore.  Most underwater photography (including videography) is in relatively close quarters  5-15ft, so you want to use a wide angle lens if you’re able.  One of the reasons I selected the ZR-70 was that it came with a wide angle lens, however the zoom function is extremely limited with the wide angle lens in place.  – Scratch this too.

3)  Now what about that all important start/stop recording button? Well obviously this would be the first control I’d want implemented, but a few thoughts on that.  I’m not going to have use of the viewfinder, so I have no feedback about the camera’s state of recording.  I’ve had cases where I get myself messed up, and end up recording when I don’t want, and then as soon as I want to capture a moment, I bring the camera up, press the record button, and end up pausing the camera instead…  that’s when I DO have use of the viewfinder.  So why increase the risk of that happening?  Finally, any additional controls on a SCUBA rig increase the workload of the diver no matter if they plan to use them or not.  Any device you bring with you on a dive is part of your rig.  A camera may not be life-support equipment, but it can be a definite distraction.  Why make something any more complicated than point and shoot?  Anyone with a current PC has the capability of editing video now anyways.  – It’s settled…  No start/stop recording button.

My Plan

In researching the problem and available solutions, I came across this fellow’s site.  The simplicity of his Mark III design appealed to me.  It’s truly elegant in its simplicity, and his thinking seems to be along the same lines as mine.  He now sells a production version of his housing here. If you’re at all nervous about building your own housing, then you should contact Paul and see if the housing will fit your camera.  From what I can see, he’s a craftsman and I don’t think you’ll be disappointed.

The design I’m using is based on Paul’s, though I plan to add a few refinements (which you’re free to incorporate if you read this Paul).  The housing body is built from clear acrylic plastic tube and sheet.  The body is made up of a 5″dia. disk 1/2″ thick,  5″OD tube with a 1/4″ thick wall approximately 7″ long that are welded together. The open end of the body is closed off by another 5″diameter disk that 3/4″ thick.  This plug has a step machined in it for an o-ring. The whole thing will be held together with a couple draw-latches.

drawlatch_l

drawlatchopen_l

I have
selected 1864A13 from McMaster-Carr for $9.23ea.  I picked these guys for three reasons.

1)  304 stainless steel construction

2)  Ability to adjust their length (flexibility is nice when you’re dealing with hand construction)

3)  I really like their low-profile smooth lines.  Too many of the latches I looked at had upturned ends and such that could snag on lines and either open if the latch isn’t pinned, or just cause nuisance hang-ups with any lines that might be in the water.  Why not streamline wherever possible? One drawback is that the cleat that comes with this draw latch mounts in the same orientation as the latch (Unlike Paul’s Mark III who’s cleats mount on the flat surface of the plug).  This means that in order to mount it to the plug, I have to cross-drill the holes and then have shear force on the bolts.  I’m afraid of the acrylic cracking if I do that, so I’ll have to work on another solution.

Any means of fastening should be good.  One could even use thumb screws.  Anything that puts a bit of axial pressure on the o-ring.

As I mentioned, the construction will be acrylic.  I obtained some 1/2″ sheet, 3/4″ sheet and 24″ of 5″dia tube.  This is enough material for three housings by this design, and it only cost me $160 CAN.  I called in a couple favors at a local machine shop and they fabricated the three pieces I needed.  Thinking about it now, I believe I can make the disks on my 7×12 lathe, but I still can’t easily true up the ends of the tube.

All three pieces are made of clear acrylic, though the two ends have protective paper/plastic still on them.

camandplastic_l

Parts from the Machine Shop

plug_l

Plug End with step for O-ring

endcap_l

Lens End

For this design, there are two points of failure.  The obvious leaking that most likely will occur in the first few feet of descent, and crushing of the housing.  Leaking obviously endangers the camera, but is relatively minor compared to a catastrophic implosion that would destroy the camera with hundreds of bits of acrylic.  Obviously both these scenarios must be tested before the camera will be placed inside.

A Bit On The Seal

The trick is to maintain a seal at the surface.  In theory anyways.  Any mechanism that holds the end cap (plug) on at the surface and seals the housing will become redundant as soon as you start to submerge past about 15 feet.  At this point, you’ll have roughly one-half atmosphere of pressure on the housing.  Both ends of the canister will have about 7.5PSI pressing them together.  For my 5″ diameter cylinder, this works out to over 300lbs of force!  That means for most diving (25+ feet) you’d be hard-pressed to remove the plug.  Further, not only could you undo the latches at this point without actually risking your camera (no I won’t prove it to you with a camera in the case…just trust me), but also the seal will only get tighter and tighter as you go to greater depths.

With this simple design there are two styles of seal that I’m going to try.  First is the basic o-ring that Paul used for his Mark III, is a quad o-ring which is approximately 4x the price, but gives you some redundancy by doubling the number of seals.  From Able O-Rings in Toronto, the regular O-ring is $1.10, and the quad is $6.39.  I haven’t decided on the frequency of o-ring replacement yet, but regardless of which I end up using, I think the cost of replacing a quad o-ring before each trip is well justified.

The o-rings of course will be smeared with silicone grease to make the job of the o-rings a little easier.

Testing

After further research into camera control glands (I don’t plan to use them on this housing, but we’ll see…  plans change), I found Tom’s site which is not only encouraging for home-brewed dive accessories, but had a couple interesting tidbits for someone like myself.  At this point, the important thing is that he turned me on to an inexpensive pressure testing apparatus that I had not previously heard of.  The device is called a “pressure paint pot” and is a large canister that is meant to be pressurized to dispense paint to a spray gun.  He pointed me to one at Harbor Freight for about $80 US.  Unfortunately we don’t have Harbor Freight in Canada, though I’ve visited a number of their outlets on my travels (even buying my air compressor from them in Memphis).  Up here the next best thing is Princess Auto.  They have a virtually identical model for about $160 CAN. A little more than I wanted to spend, but still doable (and I’m still well under the price of any other housing out there).  I have yet to pick it up, but rated at 80PSI as Tom pointed out, this will allow me to test to roughly 160fws.  This will come in handy when I decide to try to improvise control glands and want to make several controlled tests.

Construction

I have the three main pieces of the body shown above, but never having done any serious work with acrylic before, I need to work on my technique before I try to cement the fixed end onto the cylinder.

Other parts that have to be fabricated are:

1)  Latch Mounts – I don’t want to drill and tap holes into the tube itself so I have to fabricate a couple of blocks to mount the latches to.

2)  Retaining Bar – The cleats that came with the draw latches mount so that the bolts will have a shear force on them that makes me nervous of cracking
the acrylic.  To deal with this, I’m going to fabricate an aluminum bar that will span across the back of the plug, but won’t be fixed to it.

3)  Camera Tray – The camera will mount to the tray using its 1/4″-20 tripod mounting hole and then slide into the housing.  The tray will be guided into the housing so the camera lens will be aligned properly.  The tray also serves to open a space below the camera where lead can be placed to adjust the boyancy.

4)  Lanyard tie – You really need to tie equipment to something, and I prefer not to tie a lanyard to the latches.

4)  Protection Ring – A ring of acrylic tube that will be cemented on the front of the housing concentric with the lens of the camera.  It simply protects the acrylic surface in front of the camera lens.  The plastic is easily scratched, so the ring will try to protect it.  In the future it may even receive a colour correcting filter.

5)  Handles – I don’t have a plan in place for these yet. It’s still percolating.  The whole package might end up looking something like this…


housingwhandles_l

Latch Mounts

Using a bandsaw I took a scrap chunk of smoked (brownish tinted) 1/4″ sheet and ripped it into a 1″ wide strip.

Using a flat surface, I sanded one edge of each of the 1/4″ strips smooth.  first with 120 grit waterproof sandpaper (you need to use water when sanding acrylic), then moving up to 220, 360, 600, 1000 and finally 1500 grit.  These steps are probably overkill, but it meant that I needed fewer strokes at each step before I was able to move on.  I can’t argue with the results, after the 1500, the surface feels as smooth as the polished surfaces.  A quick touch with a buffing wheel, and it’d come up beautifully.

Note: Waterproof sandpaper is available from any automotive store that has body paint and other detailing supplies.  The waterproof sandpaper is quite resilient if you keep it wet. You can even rinse it under a tap to get the waste shavings out when cleaning up and reuse the paper.

sanding_l

Sanding the Edges

The strip then had its end trimmed off square and a length 3″ long was cut off to work with.

I wrapped 120 grit paper around a piece of left-over 5″ tube and clamped a straight guide (can be wood, metal or plastic) parallel to the axis and on the outside of the tube.  This setup was used it to shape the profile so that the mounts would sit firmly against the side of the housing without too much gap.  The shaping of that inside curve took the most time out of any other step in this project.

Once shaped, the strip was cut into two 1 1/2″ blocks and the mounting hole locations for the draw latches were marked, center-punched, drilled and tapped.  Take it easy with the tapping as the plastic can tent to grab.  Back the tap out regularly to clear chips, and  use water as a cutting fluid.

latchmount_l

The Latch Mount

latchmountend_l

The Latch Mount – End View

<Construction Details to be Completed>

Testing

I’m sorry, I moved kinda quickly on the final assembly.  It really did go to fast to keep up with, and I’ve been lazy about documenting it, but I promise I’ll finish that soon.

To test the housing, the harshest test for this design (without control glands) is also the easiest to perform.  Leaking is most likely to occur at the surface or within the first few feet of decent.  So I drew a cold bath, set a couple weights from my belt (6lbs) into the housing so it would sink.  Then a regular o-ring was selected and greased lightly with silicone.  The housing was clamped shut with the draw latches and lowered into the bath.  I left it sit for an hour and returned to find a completely dry interior.  I then repeated the experiment with the quad o-ring which should only improve the safety margin.  Again the results were perfect.

On to proper weighting.  To get things moving quickly for the maiden voyage, I went and bought a two-pound belt weight and used an old luggage strap to tie it onto the housing.  Then with the camera installed and sealed inside the housing, I put everything into the bath to see if it would sink or float (I expected it to sink).  Then through repeated trials I gradually drilled into the lead weight using a drill press until the whole thing just barely floated.

You’ll have to decide for yourself what your preference is when it comes to positive or negative buoyancy.  The trade offs are pretty straight forward.  Negative and the camera will sink, and stay put on the bottom where you dropped it.  Positive and it’ll float to the surface, but be subject to currents and winds making it harder to find later.

Maiden Voyage

Well the day finally came to test the camera and its housing beneath the waves.

The site of this auspicious event was the Innerkip Quarry in Ontario.  There happened to be a bunch of people from the club going and it’s closer to my place than most destinations.  Some would think this is a terrible place to test a video camera, but it actually turned out to be perfect as it highlighted a few points I’d failed to consider (I’ll get to those later).

For those who don’t know Innerkip it’s just an old quarry with the usual assortment of hardware sunk on the bottom for divers to find…  and find they must… visibility ranges from nil, all the way to 12 feet or so depending on the amount of student traffic.  We figured on our first dive we passed within 10 feet of two airplanes not having seen a thing.

I was feeling brave after the bathtub test and was eager to get on with things, so I nixed the thought of diving with an empty canister. So the camera was loaded in.  The housing was prepared as during the final testing with one minor change.  I had forgotten to grab a bit of foam to wedge behind the camera to keep it pressed forward in the housing.  I looked around, but the best I could find in the car was one of my just removed socks.  So in it went.  I powered on the camera, started recording and closed up the housing careful to keep lint and dog-hairs out of the seal.

We got in the water and set out on our first dive of the day. As we went it quickly became apparent to me that something had changed and for a bit I couldn’t figure it out.  The housing now was sinking slowly instead of rising when I let it go.  I scratched my head for much longer than I should have on that, then remembered the (clearly visible) sock.  It was enough to throw out the balance.

The dive was made with a couple minor stops on the surface while we tried to sort out interesting destinations.  We finally found one of the school busses after about 35min of bottom time, and then moved on after everyone (besides me) had swum the length inside the bus.  I guess I started the camera too soon on shore, ‘cuz after about 45min I ran out of tape.  Of course this was just as we came across the second interesting thing, an upside-down car.  When we returned to shore and had relaxed a bit, I reviewed the tape.  Very quickly I could see something was wrong. Almost none of the shots were in focus.  Right there I realized that I was going to have to use a separate check list for the camera configuration. I’d left the camera zoomed about half-way through it’s optical range, and with the wide-angle attachment on the camera, it won’t focus on anything if you zoom any further.  Darn. I had to scrap that tape.  Besides that, the housing kept the camera perfectly dry and proved to be not too bad for diver comfort.

Dive Two

Correcting the zoom setting, I again prepared the camera for another dive (recharged and rewound).  On the second attempt, we saw a lot more and I got a few interesting moments on tape.  At first I was pleased with the set up of the weight exactly as it was because the housing leveled itself nicely and all I had to do when not swimming was nudge it around in the water.  From the outside this felt great!  But when I reviewed the film I could see that the housing was pitching and rolling around a fair amount. Moving the center of gravity up some would improve this as would turning on the image stabilization (doh!) on the camera.  Also having real handles out away from the body would dampen the operators movements somewhat as well.
Other observations that came out of filming in this muck was that the auto-focus tended to be counter-productive as in the murk, the camera tended to hunt around trying to find solid edges.  This made some otherwise fine scenes with divers at the limits of visibility useless.  That’s fine though, all in all, the housing proved itself though I never exceeded 25 feet, I’m confident that it’ll do well in the future.  On the second dive I turned on the camera immediately before getting in the water, and the tape lasted beyond the duration of the dive this time.

The biggest surprise to me was the shear amount of work a video camera is under water.  I’m not referring to physically carrying the camera, that’s not bad at all, and it actually kept me from using my arms too much, so I’m sure my buoyancy control will improve faster.  No, I found that the task loading for the operator went waaaay up.  It was no longer “fun” diving, it became real work just trying to find interesting things to point the camera at, keep yourself relatively steady (without using your hands), keep your range right, etc., on top of the usual monitoring of your life support, buoyancy and buddies.  No wonder people often feel that diving with a photographer is diving alone.  This doesn’t discourage me from my project, but I’ll be forced to be more selective about the dives I take it on, perhaps even leaving the camera behind on most new dive sites.  Time will tell.

Some Notes: Things to Think About

- Turn off Auto-focus + manually set for ~10ft

- Image stabilization on/off ? judgment-call based on weight placement and other requirements

- ballast weight placement – low-mid body…  low CG makes camera self-level, but camera can oscillate if allowed to float freely… mid-body cam won’t self-level as quickly, but also won’t oscillate

- Tape loading door…  mine is obstructed by any tripod/platform that screws into the available mounting hole…  camera must be unscrewed before tapes can be changed

- battery life

- ballast adjustments should be carried out using complete kit…  any changes will affect buoyancy.  (remember the sock?)