bob's not quite sane hmd project.

[bobv5]

Its based on a trimerson headset, which I'm not really happy with. I have a feeling that I am somewhat overengineering it, but it needs to last until I can buy something better for a decent price. This is likely to be at least 5 years.

I don't think anything I put here will be particulary intresting to many people who read this now, but one wanted to see it, and it may well be usefull to people reading this in the future. (Much of the vr stuff I read is 5++ years old)

The goal of this project is a hmd with at least 60 degree fov, and at least 1024x768 resolution. From my (rather dodgy) calculations, 60 degrees is enough to fill the main part of my vision (The oposite of peripheral vision. I dont know the words for that.) And 1024 is high enough to read gagues and stuff. 90 degree fov is my ultimate goal, as this is (I believe) the highest that "normal" software can support, without needing special techniques, like fisheye rendering or somesuch thing.

I plan to fill my peripheral vison with a clone of philips ambilight. Search for "boblight" if your intrested. I didn't make boblight, we just have the same name.

I am currently working on a mechanism that holds the optics and the displays. The idea is that it will be adaptable to hold any optics that I might decide to use, and adjustable to fit my head, rather than trying to adapt my face to fit the hmd.

I will post a quick render of my current design, I would be happy to hear any ideas to improve it, and I am aware of the fact that some parts of it are impossible. Its not finished yet! (Right eye side shown only) Construction will be possible with hand tools, but it needs to be quite accurate. (about 0.2mm)

[Okta]

Dude, what teh hell is that?

[ERP]

I assume that's just a display mount with some sort of screw adjustment?
How big are the displays and what the intended configuration.
i.e. How are the displays mounted relative to the eyes?

[PalmerTech]

Are you going to try to make it 3D capable?

Not to be a downer, but I think the least of your worries is the mounting system. Finding panels with that kind of resolution in the sub-$2000 range is going to be really, really hard. Well, not the panels so much as the driver boards. I have some ripped out of projectors, but those driver boards are large and heavy, and making a custom one is a big project.

If you are going to experiment with lots of optics and displays, I would suggest you make a benchtop mount, something you can mount on a work table, to hold the parts in place. If/when you find a suitable LCD/optics combo, you can translate that to a head mount.

[bobv5]

The headset is already 3d capable. It has already done stereo 3d. (its pretty awesome)

The idea is to include any optics I may aquire in the future. I am aware that the optics I want are very expensive at the moment. I assume they will be cheaper in the future.

Okta,,,
That is my best attempt at making a vr set. You only get to shout at me if you can make somethibng better.

[Okta]

Okta,,,
That is my best attempt at making a vr set. You only get to shout at me if you can make somethibng better.

Not having a go, im just not sure what im looking at in the picture. Could you add some more angles and a description?

[bobv5]

Fair enough.

"Dude, what teh hell is that?"

Is not exactly encouraging.

Never mind.

This optic mounting system is intended to give easy ipd adjustment, and also to be able to adjust the distance between the optics and the eyes, and the angle to the eyes. All I'm really trying to do at the moment is make my headset usable. But rather that modifying a headset every few years, I want a headset I can put the new optics into. I will put labels on the next render....



ERP----
"i.e. How are the displays mounted relative to the eyes?"
I don't know. That is the whole point. I will be able to move them if they are in the wrong place. Perhaps I have misunderstood your question?

[PalmerTech]

I assume they will be cheaper in the future.

Not to rain on your parade, but they will not. Optics are one of the few kinds of technology that simply does not progress much, if at all. Our capabilities have not improved very much since the late 80s/early 90s.

You see, optics need to be made in large, large batches to be cost effective at all. And not only are the optics you seek "expensive"... They simply do not exist. As in, they have never been made. There are physical limits to what you can do without severe distortion, and getting 60 degrees of FOV or higher out of the 0.44" Kopin LCDs is somewhere around impossible.

Even if it could be done well, the optics will not simply drop in price over time. There would have to be someone making an enormous amount of lenses made specially to magnify small LCDs... And if that ever happened, it would probably be because someone is making something awesome.

I can see the theoretical advantages of your plan, but I really don't think you will get much out of it. Especially considering all the extra weight of this system on your face as opposed to a simple fixed mount.

[bobv5]

When I said "optics" I wasn't being accurate enough. What I really meant was the screen part of the optic/display system.


"There are physical limits to what you can do without severe distortion, and getting 60 degrees of FOV or higher out of the 0.44" Kopin LCDs is somewhere around impossible."

Getting 1024x768 from a 640x480 display is equally impossible. The higher res kopin displays are also physically larger, not that I plan on using them anyway. The 1024x768 x 60 degrees is my goal in a few years time, not tommorow, not next month.

I am aware that the price of the glass will not really change, but the price of the displays will fall hugely, and in fact already has or I wouldn't have any displays at all. I mentioned (in oktas thread) that I want to try 22x loupes as optics. The reason I wish to try this is that someone else used 15x loupes of the same type on larger displays and was happy with the results. Either this person has very resiliant eyes, or the results are acceptable.


"Especially considering all the extra weight of this system on your face as opposed to a simple fixed mount."

A fixed mount would be better in many ways, if it is fixed in the right place. However I don't really know how to measure my eyes. And even if I did I don't know how to translate that to display position. Even if my mount is more complex than needed the total mass will only be slightly more. It will be balanced by whatever stuff I have mounted at the back of the hmd, and if it's not I will stick a counter weight on it. The total mass of the hmd is not a major concern for me.

Unfortunatly the whole project is on hold at the moment until I build a new pc. But I will eventually complete it, and if it's rubbish, other people will know how NOT to build a hmd!

Mr Palmertech, I saw somewhere that you also thought of using an ambilight clone in a hmd. Did you ever try this? What was it like? I did a few experiments, and it seemed that it would be helpful, but the colours were distorted due to my drive circuits being inadequate and it was horrible to use because the tape that stuck it in place seemed to prefer my hair to where it was supposed to be.

[Chiefwinston]

Hi bobv5, I'm a design engineer. I'd like to see more of your design if possible. What software for design are you using. I'm designing products in Solidworks. I could easily help you if you need it. Your proceeding with modeling is really the proper approach for a device like this. You can design products like this and know they'll work in a PC without spending a dime. I think I see where your going. If your going to modify an exist head mounted display. Modeling that unit will help considerable in design. I like your approach.

cheers everyone

[bobv5]

Hi Chiefwinston,
I will have to dig the files out but I will try to get them to you in the next few days. I was using google sketchup at the time. It's free to download but I expect there is a way to view in solidworks. I'm not opposed to using different software, but I wouldn't like to spend too long learning to use it when I could use that time on other things. The design hasn't really changed since I started this topic, was using rubish computer's for a while untill I built my new one.

Unless you can spot anything wrong with it I'm happy with the top bar/ ipd adjustment thing. It's not easy to explain, but the bit below that, which is intended to slide the optics closer or further from the eye I am not too happy with, but I think it will work. If I can find a way to hold the physics violating joint that will be needed between two of the parts. Was considering moulding plastic round them to hold it as glue is not going to last long enough. It's especially hard to explain as it is only half finished in the picture I posted. The lowest red piece is actually two pieces, it is intended to rotate about the center of that bit, to adjust if it is pointing towards the chin or the forehead. The initial plan was to have teeth on either side so it could click into position, but this is likely to be too hard to mould at home. The other plan was too have some sort of friction washer between them, I don't really like this sort of hinge, but it gives stepless adjustment is going to be easier to make. I supose that way would actually be better if I could build it in such a way that it didn't spring back a little as these joint's often do.

I am planning on using the ratchet adjustment harness from a builder's helmet to make it more stable on the head. I am unsure as to if I will fit the harness into the original head mount, or build a new mount from the builder's helmet.

An unrelated question-- at one time I was looking for a program that would allow me to build a 3d model, specify the materials, then test how strong it would be. I couldn't find anything. Any sugestions?

[PalmerTech]

An unrelated question-- at one time I was looking for a program that would allow me to build a 3d model, specify the materials, then test how strong it would be. I couldn't find anything. Any sugestions?

SolidWorks.

[Chiefwinston]

Thanks Bobv5, There are many 3d Engineering programs out there. SolidWorks has a very cheap and in some cases free student version. Pro-E would be another, Catia another, there are a few others. Solidworks I'll forwarn you can have a steep learning curve. But I've seen enthusiastic guys pick it up pretty quick. By modeling you'll open the doors to parametrics, fininte element analysis, and stereo lithography. Use a program that supports .stp, .step, .igs, .iges formats. Solidworks does have the structural analysis aspects you inquired about. But don't get too hung up with material strengths at this point. I don't see to much that would require that type of study in an HMD design. If you can get the geometry in place I can tell you if something needs to be beefed up or not for strength. But getting the geometry in place would be a great start. Even if the geometry is in the approximate position. I've worked with people that are not engineers but have great ideas many times over the years. These miniture displays are advancing and improving rapidly now. I can see why there's interest in these DIY HMD's.

cheers everyone

[Tone]

Have a look at an actual implementation of the mechanism being discussed...

http://www.vrtifacts.com/hmds/its-all-i ... -the-eyes/

[cadcoke5]

Bobv5, said "An unrelated question-- at one time I was looking for a program that would allow me to build a 3d model, specify the materials, then test how strong it would be. I couldn't find anything. Any sugestions?"

The process of using a CAD model to test its strength in a given situation is not as simple as it might sound. There are a lot of situations where the process can give you totally bogus information. You are probably best off contracting with another company that specializes in this. Otherwise you may very well end up spending even more time and money resolving problems you created for yourself.

Real-world testing is often cheaper and a whole lot more reliable.

Some of the rapid prototyping methods are fairly close to end product, in terms of strength. The Fused Deposition method often generates parts with 80% of the compression strength of an injection molded part. Prices are also more affordable every day. See ShapeWays.com for one mail-order service that offers good pricing.

Even if you don't do FEA testing, you might still want a modeler. SolidWorks, Inventor, and Pro/Engineer are the three of the lower-priced parametric modelers, but still plan on starting at around $5,000 USD. Things like optics analysis are add-on packages.

There is one more, Alibre, which has a "Personal Edition" for $100. But, it really only works for prismatic shapes, and not curvey ones (unless it is a simple extrude) and has very limited import/export abilities. They do have a $1,000 pro version, and a surfacing addon (Don't know the price)., but in general, it is not nearly as full featured as the ones in the $5,000 range. Still, depending on your needs, it might work. Evaluating a CAD program can take A LOT of effort. You should contact someone who knows your work, and also knows a lot about CAD.


Joe Dunfee

[Chiefwinston]

bobv5, which ever way you decide to proceed is fine. With Cad without cad it doesn't really matter. I believe you can make a nice DIY HMD. The landscape for small displays is rapidly changing. With some of the new tech having usefulness in your project. When I look at your current diagram the only thing that throws me off a little is the scale of your diagram relative to a human head. In any event please keep use informed of your progress.

cheers everyone

[bobv5]

"Solidworks does have the structural analysis aspects you inquired about. But don't get too hung up with material strengths at this point. I don't see to much that would require that type of study in an HMD design."

The structural analysis was not for the hmd. I can't even remember what it was for, just that I couldn't find it. Not knowing what to look for was probably a factor. Havn't had chance to find the model yet, real life stuff getting in the way.

Tone, that's brilliant, thanks for that. I had allready looked at your vr4 teardown vid, and had based my design from what I saw, but a closer view is always good.

Cadcoke, I agree that what you say would be the best way to go, if this were my work. This is a one man project, with a far lower budget than most of the other home build hmd's on here. The numbers you are talking about, I could just buy a nice hmd and be done with it!

[cadcoke5]

Yesterday I attended a meeting of the Maker club in Allentown PA. A guy was showing off his MakerBot rapid prototyping machine. I was surprised at how strong the output was. A version that is relatively easy to assemble (compared to some other home-brew machined) is the Makerbot Thing-O-Matic 3D printer. The kits cost around $1,300 USD. http://www.makerbot.com" onclick="window.open(this.href);return false;

Even if the kit doesn't interest you, you might see if you can find someone local who has one. They may be interested in helping you with your design by printing some parts for you. The guy who did the demo didn’t' have much practical use for it himself. Rather, he was just interested in the technology. So, someone approaching him with another interesting technology like VR would give him something practical to use the machine for. I suppose you would have to check the Makerbot forum to find someone.

Joe Dunfee

[bobv5]

Now that is a good idea. Why didn't I think of that?
Cheers.

[bobv5]

Started work on this again, now that I have some half decent optics to use. Will be using a simplfied version of the mounting system in the first post, I have modified it so that it is only a little more complicated than a fixed mount, yet still includes most of the adjustments on the original. Looking for a way of doing the actual "head mount" part. Was thinking of using a builders helmet. The type with the ratchet adjust ment at the back. Any thoughts?

[cadcoke5]

I can't really speak from experience... well actually I can, but only as a puppeteer, and a person who has built a couple of mascot costumes. I have played a little with them, but never wanted to spend the money to get the good helmets that had all the adjustments.

I know there are some helmets that have a simple round strap, but others have an additional part that reaches further down the back of the head. It obviously grips the head more securely. I would think that since you are attaching some weight to this helmet, and moving it around a lot, that this additional feature would be desirable.

I just did a quick internet search, and was quickly overwhelmed by the variety of choices. I did learn that the additional part I was thinking about is called a "Napestrap". There are also versions that allow for add-on parts, like chin straps, and ear guards. The ear guards are especially interesting, since you will probably also want to add sound.

The major hardware chains have at least a few choices. I would be surprised if there weren't stores specifically geared towards construction equipment. This way you can try some on and gauge how easy they are to adapt.

Joe Dunfee

[Tone]

Headbands - http://www.oaktreegroupllc.com/html/products.html

2a-front.jpg