Jump to content


Photo
- - - - -

DIY housing for E-M5

diy housing e-m5 cnc

  • Please log in to reply
59 replies to this topic

#1 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 08 January 2014 - 03:27 AM

Hello everyone

 

I've been a recreational diver for the past six years. I live in Finland where there is not much to see while diving so I only do it on longer vacations once every one or two years. I have been building my own underwater housing for my cameras for ten years now. I'm not much of a photographer, instead my interest in building housings comes more from an engineering viewpoint.

 

The success with the DIY housings (4 by now) has been mixed during the past years but by now I have only destroyed one camera partly. On the other hand not one of the housings has been completely leak proof. Anyway, the housings have been evolving to more and more professional level by the years and now I thought it is time to share the current project with all of you as I have found only a very limited number of other DIY housing projects in internet. Maybe it's because everyone else values their time and gear too much to spend hundreds of hours of time in order to make a housing which implodes in 30 meters...

 

As a reference below is a picture of my previous DIY housing which is for Olympus E-PL2 with 20mm Panasonic pancake lens. The body of the housing is machined from a solid block of aluminum and the port is turned also of aluminum and attached with screws to the body. Front lens is acrylic and the back cover is 10 mm thick polycarbonate. This housing was made two years ago for a 3 month trip to Central America. There were problems with the back cover as the polycarbonate developed cracks around some threads. This may have been caused by the ethanol I used to clean oil from the polycarbonate. The 10mm thickness also was not enough for deep dives as the back cover bended visibly. Therefore I used the housing mainly for snorkling and for that it functioned pretty well.

 

old_small.jpg

 

Currently I am designing a housing for my E-M5 which I'm planning to use with Samyang 7.5 mm fisheye lens. Me and my girlfriend are leaving to the Philippines at the end of this month so I am starting to be pretty low on time to finish the project. The structure of t he housing is basically the same as that of the E-PL2 housing, however, it will have a dome port, hopefully better ergonomics and I dare to hope that this housing will finally be leak free. Below are two images from the CAD model in its current state. I guess I will have to finalize the CAD model during the next weekend in order to start machining the parts next week. I am planning also to print a 3D model of the housing to test the ergonomics before I start machining it. We'll see if I can make that happen. Oh, in case you are wondering how I get to use such nice tools as 3D printers and CNC machines, the answer is that I work at the machine design department of the largest university in Finland.

 

front.jpg

 

back2.jpg

 

There are still a couple of problems to solve with the housing design. First would be how to attach the dome to the port. The dome is separated from an Ikelite 5503.15 dome port which was leaking from the glue seams between the plastic parts. I was planning to use an o-ring between the dome and the aluminum port but then there was the problem on how to press the dome against the o-ring. Also, I was worried that the dome might develop cracks when the pressure presses it against the aluminum. Therefore, I decided that it might be the best option to just glue the dome to the port with epoxy. This will make it leakproof and the glue also should distribute the surface pressure between the dome and the port evenly. The downside is that when the dome is glue, it can't be removed.

 

Another, still unsolved problem is the trigger button. As you can see in the image of the CAD model, it is still unfinished. If it is located above the trigger button of the E-M5 it will be quite far from the left side of the housing and my (short) index finger will not reach it easily. Therefore the button should be moved to the left somehow. The commercial housings have nice lever systems for the trigger but I'm afraid they will be difficult to manufacture in a DIY project. I would be glad if you could give me some ideas on how to solve this.

 

I will add more pictures and info later about the current state of the project and of course add pics when I start machining the parts. Now I will unfortunately have to start doing the work I get paid for.

 



#2 troporobo

troporobo

    Sting Ray

  • Member
  • PipPipPip
  • 289 posts

Posted 08 January 2014 - 03:38 PM

That is some awesome work!  I really admire your ambition to build DIY housings.  My own DIY skills are nowhere near that level (although I do build tube powered audio gear and high efficiency speakers, so I know all about expensive labours of love).  

 

When you have dates for your Philippines trip please get in touch, if we can meet up I will gladly provide drinks on the beach at Anilao so we can compare notes!



#3 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 10 January 2014 - 08:23 AM

Thanks troporobo. I've also sometimes considered building an amplifier and/or speakers but never got into it. Oh, I did build a subwoofer when I was in the Finnish equivalent of junior high school but I don't count that as a proper speaker. Were leaving at the end of January, however, sadly we will fly straight from Manila to Coron so no drinks in Anilao this time.

 

I have now reserved a 3D printing session for Monday so that means I will definitely have to finish the CAD model by then. Busy weekend. One thing for which I would like get some help from you is the optical sync port on the upper left side of  the housing. I will probably not have time to make my own strobe for this trip and I will not pay 500+ euros for a strobe so my pics will be done only with natural lightning. However, I will make an optical sync port for future use. It seems that the optical sync cables used in the commercial housings have just a cylindrical connector, with a small bump in its side, which is pushed into a cylindrical hole. Is this correct? So there are no claws or anything to hold the connector in its place? Could some of you measure the diameter of the connector and the hole that it is fitted into so I could make my housing compatible with the standard cable in case I one day end up buying a standard strobe. I could not find the info anywhere for the dimensions of the cable.

 

Also I will not make buttons on the housing for every button in the camera. I thought I will manage with the trigger, fn1&fn2, video, play, ok and power buttons with the thumb and index finger wheels. Any opinions from E-M5 users if these are enough? I guess the wheels compensate for the missing arrow buttons in all situations?



#4 troporobo

troporobo

    Sting Ray

  • Member
  • PipPipPip
  • 289 posts

Posted 10 January 2014 - 05:08 PM

You are correct about the optical cable connector.  Here is a quick photo of an Inon cable for a Nauticam housing.  

 

I estimate the outside diameter of the ridge on this connector to be about 7mm, and the inside diameter of the hole also seems to be 7mm.  Unfortunately I don't have a vernier caliper at home to measure more accurately. 

 

Alternatively you could use just a bare end of an optical cable, that would require only a very small hole.  I estimate the outside diameter of my Inon cable to be 2mm.  I do not know if this is standard across other manufacturers.

 

Maybe you could try contacting Wetpixel member Edward Lai of Nauticam (see the OM-D EM-1 thread) to see if he can tell you more accurate dimensions

 

As for the buttons, I would not want such restricted access.  You can map some adjustments to a subset of buttons but you need to be able to get to either the menu or the super control panel and navigate for other settings.  You might get away with setting everything up before loading the camera into the housing, and mapping white balance to an Fn button, but then you could not (for example) change ISO, metering, or focusing.  I use the super control panel a lot, but that requires the OK button and the four way buttons.

 

 

 

11879927566_bcaf6ab541_b.jpg



#5 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 11 January 2014 - 04:33 AM

You are correct about the optical cable connector.  Here is a quick photo of an Inon cable for a Nauticam housing.  

 

I estimate the outside diameter of the ridge on this connector to be about 7mm, and the inside diameter of the hole also seems to be 7mm.  Unfortunately I don't have a vernier caliper at home to measure more accurately. 

 

Alternatively you could use just a bare end of an optical cable, that would require only a very small hole.  I estimate the outside diameter of my Inon cable to be 2mm.  I do not know if this is standard across other manufacturers.

 

Maybe you could try contacting Wetpixel member Edward Lai of Nauticam (see the OM-D EM-1 thread) to see if he can tell you more accurate dimensions

 

As for the buttons, I would not want such restricted access.  You can map some adjustments to a subset of buttons but you need to be able to get to either the menu or the super control panel and navigate for other settings.  You might get away with setting everything up before loading the camera into the housing, and mapping white balance to an Fn button, but then you could not (for example) change ISO, metering, or focusing.  I use the super control panel a lot, but that requires the OK button and the four way buttons.

 

Thank you for the pic and the dimensions. I have actually never worked with optical cables. So you think it would be possible to use just the optical cable without a connector? Direct contact with water won't ruin the fibers or something like that? I think this might be the easiest option and I guess I could then use any cheap toslink cable without the connectors. Then I would just need to somehow ensure that the cable stays in the port...

 

The super control panel is the one opened with the button next to the viewfinder? This button could be a little difficult to access in the housing since it needs to be pressed from the side. It could be done with a rotating button in the back panel. However, all the options in the super control panel seem to be available also in  the menu which you get by pressing the ok button (at least on P-mode which I will probably use). The options can be selected with the with the wheels and the ok button. Also, I will have ISO adjustment assigned to the fn1 button for easier use. White balance will be adjusted with the front wheel and exposure time with the back wheel.

 

The Samyang 7,5 mm lens is manual focus so AF/MF selection is not needed. I wont' even have access to manual focus with this simple port. I have understood that if i focus the lens to about 30 cm, pretty much everything will be in focus (probably not very very close to the dome but I will just have to live with it). Metering will be difficult to adjust but to be honest I rarely use it even on land.


Edited by T.J.L., 11 January 2014 - 05:56 AM.


#6 troporobo

troporobo

    Sting Ray

  • Member
  • PipPipPip
  • 289 posts

Posted 11 January 2014 - 04:59 PM

It is indeed possible to use a bare fiber optic cable at the housing end, in fact, the Nauticam housing has tiny holes for just this purpose.  I don't know how you could manage the attachment at the flash end though.

 

The super control panel has to be activated through the menu first (Custom menu D -> Display -> Live SCP -> On)  Then it is accessed by simply pressing the OK button not the button next to the viewfinder.  If you have not tried this yet, I recommend it.  However I believe that the method you outline will also work for the shooting you want to do, so it is really a matter of preference

 

Also, if you have not yet seen the excellent configuration advice provided by coroander at the following link, you will want to review it ASAP before finalizing your design:

 

http://wetpixel.com/forums/index.php?showtopic=48625 



#7 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 12 January 2014 - 12:38 AM

Thanks again. I've just bought the camera a few weeks ago and used it only a couple of times so there is much  to learn. The SCP seems easier to use than the live control. The SCP is also usable with only the dials but it would be faster with the arrow buttons. I'll think about making also them for the housing. It's not that much more work, however, they make four more possible routes for leaks.



#8 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 14 January 2014 - 11:59 AM

Now I have a 3D printed plastic model of the housing's two main components i.e. the body and the port. The process was not as magically easy and fast as I had expected and the result is not exactly 1:1 copy of the CAD model. Below is a quick cell phone pic from the end of the printing process. I'll add better images when I transfer them from my camera.

 

3D.jpg

 

The parts are made with relatively cheap (~3000 euros) fused deposition model 3D printers i.e. machines that extrude heated plastic. There were initially some problems with the settings of the programs and in the case of the port plastic did not stick properly onto the plate where it was extruded on. This caused the port to be significantly warped. However, the body turned out pretty good. The printing of the body took 19½ hours so the rapid prototyping was not so rapid. The part is not exactly dimensionally accurate, instead the part has sort of expanded 0.5 mm to every direction. For example 3 mm holes are in reality 2 mm in the printed part and the inside of the body is 1 mm too narrow etc. Anyway, the camera fits inside the body pretty tightly and the housing feels pretty good in hand. I will machine the back cover and the camera tray maybe tomorrow or during the weekend and then I will really get to test the ergonomy of the housing.

 

For the flash sync port I decided to make a hole which has a M10 (10 mm major diameter) internal thread. I can install into this thread a part with as large a hole as I want for the optical cable. I can use a 7 mm hole for the commercial connectors or I can make a 2 mm hole for just the fiber. I'll decide that in the future.



#9 TomekP

TomekP

    Clownfish

  • Member
  • Pip
  • 27 posts

Posted 16 January 2014 - 01:00 AM

Generally nice idea, I always was impressed by the people who made their own lamps, rebreathers and now camera housing, which looks sometimes as good or even better than this stuff which we have to buy in the diving shop for much more money :) 

 

Is there will be only 1 o-ring or 2?

What is the maximum depth for that housing?

Did You ever try to immerse some parts build in that 3D printers to deep water? I think this printers are commonly used to produce models, not the material or elements with some reasonable strength. And it is possible that there will be some small holes which will be compressed at can generate cracks. So please let us know about results of pressure test which You should do few times before You will put camera inside and go for diving. 



#10 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 16 January 2014 - 02:46 PM


Is there will be only 1 o-ring or 2?

What is the maximum depth for that housing?

Did You ever try to immerse some parts build in that 3D printers to deep water? I think this printers are commonly used to produce models, not the material or elements with some reasonable strength. And it is possible that there will be some small holes which will be compressed at can generate cracks. So please let us know about results of pressure test which You should do few times before You will put camera inside and go for diving. 

 

Each button will have one O-ring as will the port and the back cover.

 

I have no idea about the maximum depth. That would require experimental tests or finite element simulations. I'm pretty sure the housing would operate down to 50 meters or more, but probably the deepest it will ever be is 30 meters. I can't see the aluminum body breaking under pressure and the 15 mm back cover will surely be thick enough for high pressures. I am a little worried about the dome since it has small cracks in the corners but the epoxy glue and the port will support it.

 

The 3D printed parts are just for testing ergonomics etc. The final parts will be machined from solid blocks of aluminum.

 

Tomorrow I'll upload some pics of the 3D printed parts. Now, after a 13 hour work day, I'll go get some well deserved sleep.



#11 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 17 January 2014 - 10:11 AM

Here come the pics of the 3D printing process.

 

Even though the walls are only 4 mm thick, the printer (or the software controlling it) makes them hollow to save material and speed up the printing process.

printing.jpg

 

After 19 hours and 28 minutes, the printing is done.

ready.jpg

 

And this is how the body and the port look like when the port is attached. The port/dome looks huge compared to the body but that is just how it is. If the body works properly, I may later make a smaller dome port. The front surface of the the body is very rough because that is where the supports were printed during the printing process. The rest of the body is printed on top of the supports because the machine can't print on air, there has to be something on which the plastic is extruded. Anyway in this case the surface finish is of no concern. The inside of the body is more important

ready2.jpg

 

And here is the inside. The port is mounted on the body with four screws. The screws are inside the port O-ring so they do not have to be sealed in any way. In the final part the holes for the latches do not penetrate through the walls like in the printed part so they do not have to be sealed either.

inside.jpg

 

The main reason for printing the prototype was that I could test how it fits my hand. As you can see I can fit my fingers between the body and the port which is of course important. Here the problem with the trigger button is visible. Even though the button on the housing is already moved a lot to the left compared to where the button is on the camera, the natural position for the index finger is even more to the left and forward. I will move the trigger button still a little forward and left in the final design.

hand.jpg

 

And this is the environment where I was printing. It is next door to the laboratory where I work and now I can basically use it whenever I want for free. Nice.

addlab.jpg



#12 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 17 January 2014 - 10:23 AM

And a couple of more images of the more or less finished CAD model.

 

Front side.

front_CAD.jpg

 

Back side. I'm pretty sure I will manage without the arrow buttons when I have the dials. It'll save me some work.

back_CAD.jpg

 

Here is the CAD model without the body, port and back cover so that all the small parts are visible. The green part is a mockup of a PCB where I will maybe later make a DIY trigger flash. The only parts I have bought are  the latches and the stainless screws, all the other stuff I will have to manufacture. I will soon start to panic, since in two weeks we will already be in Hong Kong. So tomorrow and Sunday will be spent machining and I hope that I will get a lot done. Sadly, it always seems to take more time to machine something than what you initially thought...

 

parts_CAD.jpg



#13 troporobo

troporobo

    Sting Ray

  • Member
  • PipPipPip
  • 289 posts

Posted 17 January 2014 - 04:46 PM

Looking good! I must need a new project of my own because I'm really enjoying your progress reports. Thanks a lot for posting.

By the way, are you sure about using only 4 screws to attach the port? It looks like it leaves a lot of unsupported circumference to possibly flex. Since the screws are inside the o-ring (great idea) why not add a few more?

#14 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 18 January 2014 - 12:46 AM



By the way, are you sure about using only 4 screws to attach the port? It looks like it leaves a lot of unsupported circumference to possibly flex. Since the screws are inside the o-ring (great idea) why not add a few more?

 

You just might be right. My previous housing had 5,5 mm thick walls and I always thought that the body could withstand anything. This one has 4 mm walls and it might be possible that it could bend a little under pressure. I don't think that it will be a problem, but the pressure force 30 m deep is pretty large, it equals about 300 kg of mass on a 10 cm by 10 cm surface.

 

I added two more screws for the port. There is no space for screws on top and bottom, but I moved the screws in the "corners" a little closer to each others and added screws to the sides. Although I guess the commercial housings do not have anywhere this rigid mountings for the ports. Actually, I don't know about that since I have never held one in my hands :lol:.

inside2.jpg



#15 rrodex99

rrodex99

    Starfish

  • Member
  • Pip
  • 6 posts

Posted 18 January 2014 - 01:49 PM

Fascinating project. I hope all goes well. You discuss the pressure capability and possible deflection. What looks right often is right, but there is a free Finite element analysis program Z88 Aurora which will give you the answer in seconds. If you have produced the model for the 3D printer I think you could feed it straight in and check the stresses and deflections. Not at all difficult.

One question, you do not show any springs on the buttons, does the camera alone release the button or have you just not shown them?

 

Rod



#16 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 19 January 2014 - 12:16 AM



Fascinating project. I hope all goes well. You discuss the pressure capability and possible deflection. What looks right often is right, but there is a free Finite element analysis program Z88 Aurora which will give you the answer in seconds. If you have produced the model for the 3D printer I think you could feed it straight in and check the stresses and deflections. Not at all difficult.

One question, you do not show any springs on the buttons, does the camera alone release the button or have you just not shown them?

 

Rod

 

Actually even the CAD program I have been using, Solid Edge ST3, does FEM analysis. I had just totally forgotten about it. I tried to make a quick analysis with it but did not get a sensible results so I imported the model to Comsol Multiphysics. I did an analysis with only the body of the housing and a 300 kPa (30 m depth) load on the outer surfaces. The inner surface of the port was selected as the fixed part of the body. The resulting displacement is shown below. The maximum displacement is about 0.2 mm in the edge behind the buttons. On the other hand, this part of the housing is partly supported by the pressure force on the back cover, which is now not simulated. The displacement on the bottom of  the housing is probably real and there is not much I can do for it besides increasing the thickness of the walls. I think a 0.15 mm displacement there is just fine.

 

displacement.jpg

 

The CAD model does not show springs since I have not had time to model them. There are springs under each button and there are stainless steel washers between the springs and the o-rings. In addition to those, there will be pieces of plastic at the ends of each shaft of a button so they do not scratch the camera buttons. Otherwise, the CAD model should be complete.

 

Yesterday I got half of the back cover machined, not as much as I had expected. So now back to work.



#17 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 19 January 2014 - 12:41 AM

Well, I modeled the back cover also. 15 mm polycarbonate is thick enough (surprise...). Maximum deflection 0.1 mm which has little significance.

 

FEM_back.jpg

 

I also added a 3600 N load on the contact surface of the body and the back cover to simulate the pressure on the back cover. It removes most of the deflection in the upper part of the body. The bottom deflection (max. 0.16 mm) remains.

 

FEM_body2.jpg



#18 TomekP

TomekP

    Clownfish

  • Member
  • Pip
  • 27 posts

Posted 19 January 2014 - 03:37 PM

I think You are a bit crazy .... but in positive way of thinking  :) You are using very advanced  tools to make a toy :) :)

Good luck :)

 

P.S. Are You sure that at 30m there is 300kPa??



#19 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 19 January 2014 - 04:26 PM

It's 2 am in Finland and my machining weekend just ended. Did not get as much done as I had hoped for but the housing is starting to take its shape. Here are the results this far.

 

 

More pics to come but now is not the time to write long posts...

 

 



I think You are a bit crazy .... but in positive way of thinking  :) You are using very advanced  tools to make a toy :) :)

Good luck :)

 

P.S. Are You sure that at 30m there is 300kPa??

 

Yes I have also been wondering why am I even doing this, especially now that my schedule is tight and I know there will be long days, evenings and probably nights of machining ahead. Probably it is a symptom of some mental disease. I think I also just want to prove to myself that I am able to do this.

 

The pressure in 30 meter depth is about 400 kPa and inside the housing the pressure is 100 kPa so 300 kPa pressure difference.

Attached Images

  • machined1.jpg


#20 T.J.L.

T.J.L.

    Triggerfish

  • Member
  • Pip
  • 30 posts
  • Gender:Male
  • Location:Helsinki, Finland

Posted 20 January 2014 - 12:17 AM

Here's a step by step story of machining the inside of the body. I always enjoy machining pics, I guess there must also be others with this kind on personality disorder.

 

Here is where the machining starts; a rough 3½ kg billet of aluminum. Of course before starting, the milling paths have to be made (took me like 5 hours for this part) and the tools have to be set up. The milling process itself takes less than an hour.

machine1_alku.jpg

 

First, the billet is machined to as precise as possible a rectangle with the desired dimensions. Now when the orientation of this rectangle is changed and it is clamped, it will always be accurately in the desired orientation (as opposed to the initially not-so-rectanglish billet).

machine1_alku1.jpg machine1_alku2.jpg

 

Then, the back cover o-ring groove is machined with a 2.5 mm end mill.

machine1_alku3.jpg

 

Next the main job i.e. digging the hole for the camera with a 12 mm flat end mill. The middle part is for mounting the part during future machining steps. This was as pretty messy part of machining since the pit was full of cutting fluid and I had to empty the chips from there with pressurised air. There were chips and fluid everywhere within 3 m radius of the machine.

machine1_alku4.jpg

 

Then finishing the sides and chamfers with a 14 mm ball end mill and some pockets to the bottom with 8 mm flat end. The tools had to be long so they vibrated a little. Therefore, I did not get exactly the surface finish I wanted. I guess I will sand also the inside of the housing if I have time. That was pretty much it for now.

machine1_alku5.jpg

 

The worst part is always the cleaning. This time it took about an hour. It seems that half a liter of solid aluminum turns to quite a large volume of chips. At times like these I really love our industrial vacuum cleaner which can handle the chips and the cutting fluid.

machine1_loppu.jpg







Also tagged with one or more of these keywords: diy, housing, e-m5, cnc