Collimating image from an LCD display

[DELTA12]

I have an LCD display I'd like to use in a collimated HUD display, however it seems that I can't get the light (image) from it to collimate properly. I'm trying to use a lens with the focal distance of 4.5 cm. When I place the display 4.5 cm away from the lens' center I expected to achieve the collimation however the only thing I can see in the lens is a blurry mess.

Questions is: Maybe I'm putting the lens in the wrong side of the display - in order to get a collimated image from it do I actually have to pass already collimated light through the display (replace the display backlight with a collimated light source) or is it somehow possible to collimate the image coming out from it (it's shown this way on all schematics of collimated displays/collimators I've been able to find)?

Optionally, are there other means of collimating an image from an LCD display?

 

[CWatters]

Perhaps I misunderstand but ... If you put an object at the focal point won't the image be at infinity?

 

[DELTA12]

Perhaps I misunderstand but ... If you put an object at the focal point won't the image be at infinity?

Yes, that's how HUDs and collimator sights work if I understand it correctly.

 

[DELTA12]

The projection unit in a typical HUD is an optical collimator setup: a convex lens or concave mirror with a Cathode Ray Tube, light emitting diode, or liquid crystal display at its focus. This setup (a design that has been around since the invention of the reflector sight in 1900) produces an image where the light is collimated, i.e. the focal point is perceived to be at infinity. (https://en.wikipedia.org/wiki/Head-up_display)

 

[mfb]

How precise are the 4.5 cm of the lens and your 4.5 cm distance? Can you vary the distance a bit and see if it gets better?

 

[Aaron Crowl]

Can you get the screen in sharp focus at one point but the rest is blurry? Is it all blurry all the time? I ask the first question because your lens may have a limited aperture. That will mean that it can only collimate a small portion of the image and the rest will be blurry.

 

[mfb]

Chromatic aberration might be another issue: the focal length will be different for different colors.

 

[DELTA12]

I can get a focused image, although it's not the goal I'm aiming for, instead I need to achieve collimated image from the display (focused at infinity).

I've tried the method I described above (replacing the display backlight with a collimated light source) with good results although the collimation wasn't perfect due to the quality of my light source it still was noticeable. Queston still stands: is it possible to collimate the divering image from the display without replacing the backlight?
For additional information I'm using a lens from oculus rift dk2 with diametet of 4cm and focal distance of 4.5cm. I also have ordered a 10 diopter 5.5cm diamater lens, perhaps that'll give different results

 

[Drakkith]

If you look through the lens at the display and move the lens forward and back, is there any distance at which the display becomes focused?
When you say you can get a focused image, are you looking at that image directly through the lens, or are you projecting it onto a surface and seeing the image that way?

 

[DELTA12]

Yes, if I move the lens back and forth I get a focused image at some point when looking through the lens

 

[mfb]

Then where is the problem? That is the right distance.

 

[DELTA12]

Then where is the problem? That is the right distance.

Well is that a collimated image? Because as far as I know, doing that simply results in the light rays to focus at some point on the other side of the lens, but I need an image focused at infinity

 

[DELTA12]

One of the major features of an HUD is that the produced image is collimated and thus always in eye focus independent of the current eye accomodation. That's what I'm aiming for.

 

[mfb]

If you look through the lens and see a sharp image with the eyes adapted to infinity (not necessarily, but that is the typical application), your HUD works. That is exactly the use case.

 

[DELTA12]

If you look through the lens and see a sharp image with the eyes adapted to infinity (not necessarily, but that is the typical application), your HUD works. That is exactly the use case.

Okay, much thanks for the info!

 

[TomGo]

Reheating old subject, but i`d do with a word of advice.

I`m about to start my own fight with pretty much same problem, and for exactly same use.

As of now, on theory basis (parts being shipped and that's all i can do at the moment), i can see two forks.

starting from end point, that id windscreen:
image reflected off either windscreen (possibly a 3% reflective film on inner surface to avoid ghosting)
or
dedicated shaped perspex with reflective coating between driver and windscreen

next:
mirror fitted to dasboard. shape TBC

yes, this means the image will be reflected twice. first on the dashboard mirror (face up and not visible by driver) and then again on windscreen/perspex

source:
modified 3.5 inch raspberry pi display, hdmi.

fork:

a)
to use built in backlight and try to collimate by one or few frensel lenses in front of screen (easy and.. less likely to work, but i assume this is what author of this thread tried)

b)
to strip the backlight off the LCD panel, use non- collimated point light source, lens at focal length from the light source, and LCD panel between these twoanyone tried ?

(parts to arrive in about a week)

 

[berkeman]

image reflected off either windscreen

What windscreen? Are you trying to install a DIY HUD in your car? Have you looked at commercial versions?

 

[TomGo]

Volvo xc70, UK (right hand drive),
Yes, I did look at commercial options, none is satisfactory from my point of view (ask and I'll give list), possibly closest to ideal, but still lacking would be (not sure of name) Hudly unit
But it's sold only in Taiwan.

Yet, idea of fitting screen/projector comes just from that.

 

[TomGo]

https://www.growly.io/hudly-cool-display-for-cars/amp/

By the look of photos, and few reviews claiming this projects perceived image few metres in front of car, light is not necessi collimated, but projected in fan like, spreading fashion.
Glass "screen" is slightly convex, and that's the part that serves as collimator.
This way small screen can be seen as much bigger

I want to mimic this.

 

[berkeman]

https://www.growly.io/hudly-cool-display-for-cars/amp/

By the look of photos, and few reviews claiming this projects perceived image few metres in front of car, light is not necessi collimated, but projected in fan like, spreading fashion.
Glass "screen" is slightly convex, and that's the part that serves as collimator.
This way small screen can be seen as much bigger

I want to mimic this.

Thanks, that was interesting. At least they are positioning the small HUD out of the direct view of the road by the driver. It seems that any HUD that would be in the direct line of sight of the roadway for the driver would need NTHSA approval (or whatever safety agency applies where you live). It would be way too easy for a marginal design to hinder the driver's ability to see or notice problems in the roadway that were developing. What size display are you shooting for, and how far out of your tactical visual driving window are you going to position it?

 

[TomGo]

True answer would be... I do not know until parts arrive.

Can't say I'm getting much, at this stage, but..
A 3.5" hdmi LCD screen off e-flea (they claim it to be 1920*1280, I'll be very surprised if it is at this price)
Several Fresnel lenses, various sizes, with focals 120-300mm,
Some stricte mechanical parts, like 5mm threaded rods with nuts, I need to figure out some frames to hold screen, lenses etc. Rods wi hold these together allowing adjustments.
And that goes on the workbench.
Make it work first, then try to fit it to car, adjust positions etc

 

[TomGo]

Ok..
So I got some parts.

And I run into some issues.
It must be common knowledge, for many around here, but.. it isn't for me.

So I got a 3.5" raspberry hdmi LCD display.
And a bright LED torch.

I stripped the display, removed original backlight together with wedge light guide, then I removed white dispersing film, still from rear, and then again, underneath was another film, looking like kind of Fresnel lens, which I removed as well.

In theory, this should leave me with an LCD panel that has horizontal filter on the back, LCD matrix in the middle, and vertical filter on front.

Then I connected random signal (from my satellite box), power, and put torch behind it.

It's even darker than on its original backlight.

What am I doing wrong?

 

[TomGo]

I'm getting suspicion that I'm actually doing it right..

Need to make an adjustable test rig.
Light holder (led stripped off that torch, maybe with its lens, maybe without), lens holder, screen holdr, another lens holder (may be needed), reflector/combiner.

Need to calculate curvature of that combiner, too, and blow mould (as opposed to vacuum mould) piece of perspex, so I end up with some 5-7 inch image on combiner, from a 3.5 inch screen.

And some light shield.

 

[Tom.G]

It's even darker than on its original backlight.

What am I doing wrong?

You are removing way too much of the illumination optics. Keep everything assembled except the light source, which you then replace with a high intensity source.

Getting that appropriately designed and implemented will likely keep you busy for awhile.

Cheers,
Tom

 

[TomGo]

Actually.. I had it in the "bench" yesterday again

I didn't remove enough.
There was one more layer of adispersing film.

Now I got it working
Next step, make it shorter.
I collimate light just before it hits the LCD, currently I have a 180mm Fresnel lens, I need something around 35-45mm to be able to make unit short. Otherwise it will take too much space on windscreen and will be absolutely illegal.
Or I'll just use few 180mm lenses.
Need to prove it practically, but it looks like I don't need to worry much about light aberration.
It passes LCD as last step before hitting combiner, there's no lens to aberrate it between these two.

 

[TomGo]

Here's what I'm trying to do.

 

[PaulCarew]

Here's what I'm trying to do.

Looks like an interesting project.
For the last year, I have been working on a different type of 'collimating HUD'. My project is a HUD is for attachment to Binoculars. This brings with it a whole set size/weight and portability issues.
I have been using a combination of Convex and concave lenses to collimate between the oled display (IE not a TFT LCD) and the combining screen.

Still a work in progress

 

[berkeman]

My project is a HUD is for attachment to Binoculars

Welcome to PF.

There are already binoculars with internal displays, so why would you try to add an external HUD for data display? As a retrofit maybe?

https://ak.picdn.net/shutterstock/videos/5750420/thumb/1.jpg

 

[PaulCarew]

Hello @berkeman , thanks for the welcome.
Your find is interesting too, however, my project is based around adding capability to ordinary binoculars.
The use case is actually for tracking sailboats in a sailboat race. Each of the racers will be carrying a small gps tracker that relays their position amongst other info. This is relayed and processed and then basic facts relayed to the binocular HUD/AR device. The device includes a gyro and compass for binocular orientation.

I so far have all of the pieces operational from the trackers to the relaying and processing of the data.
Working on a physical binocular attachment at the moment that includes the lens system. This bit is the *most* challenging :-)

 

[anorlunda]

my project is based around adding capability to ordinary binoculars.
The use case is actually for tracking sailboats in a sailboat race.

I tried researching your question. I didn't find a heads-up-display for binoculars, but I did find a comment on another forum. There's no more info there than the following.

Spyglass is an essential offline GPS app for outdoors and off-road navigation. Packed with tools it serves as binoculars, heads-up display, hi-tech compass with offline maps, gyrocompass, GPS receiver, waypoint tracker, speedometer, altimeter, Sun, Moon and star finder, gyro horizon, rangefinder, coordinate converter, sextant, inclinometer, angular calculator and camera.

 

[PaulCarew]

Hello Anorlunda,
I took a look at the 'Spyglass' Android app. Looks interesting in it's own right. Not quite what I'm looking at, but interesting never the less.

 

[berkeman]

The use case is actually for tracking sailboats in a sailboat race. Each of the racers will be carrying a small gps tracker that relays their position amongst other info. This is relayed and processed and then basic facts relayed to the binocular HUD/AR device. The device includes a gyro and compass for binocular orientation.

Why not just use hull numbers or names (I'm not sure how sailboats are labeled -- whatever is equivalent to tail numbers for aircraft)? How do you plan to handle sailboats that are side-by-side and so in the same line of sight from your location?

Working on a physical binocular attachment at the moment that includes the lens system. This bit is the *most* challenging :-)

Are you able to open up the binoculars to add your optics inside, or are you trying to do this all external to the binocs? If all external, maybe use something like Google Glass or similar HUD glasses, rather than trying to add something to the front entrance of the binocs?