ich weiss net ob das das richtige ist weil ich net englisch kann. sonst schaust du einfach auf:
http://www.gbalight.com
ok:
Progress Update
2-3-02 Finally, a new LED has been tested! This LED (although requiring 5.6v) is quite a step up from the rs 276-320. If you've checked the bulletin board here lately, you've seen the preliminary info that I've been posting about this new LED.
I've got some good news and some bad news following the conclusion of these final tests. Please read on...
The LED
This latest LED that I've been working with in this latest experiment is known as "the world's brightest LED" per the supplier. The first drawback is obviously the voltage requirement of 5.6v which is quite a bit more than the current voltage that we're used to working with (~3.2v). This means that no, you will not be able to just use this LED in place of the rs 276-320 because the 3v will hardly run this LED. You need at least 5.5v, and that is most easily accomplished by using 2 3v batteries.
LED Pros
Extremely bright, completely drowns out the rs276-320
Very decent battery life, with 12 hours on a set of batteries it's still nearly as bright as the 276-320 on a new set of batteries!
Would be very useful in many other applications(besides lighting the GBA -see info page later on)
LED Cons
20 degree viewing angle degree rating which means that unless you can mod a device to position it at least 4 inches above the screen, you will need to use a diffusing lens.
Requires 5.6v. Most commonly derived from 2 3v batteries(lithium or button cell type), but can also be supplied by multiple 1.5v batteries.
The batteries Note: at 1:am EST time, the LED is still lit, that's 15.5 hours so far!
The batteries that I tested these LEDs with were the 3v BIOS type button cell batteries. Two were used and they each provided ~3v. Unfortunately, these batteries are on the expensive side at $1.96 ea.(@ walmart) However, to use this LED, you would not need to use this type of battery, you could use an alternative such as 4 1.5v triple or double As. That fact should be obvious, but I want to make it clear, because if I end up offering these for sale, they will include 2 of the cr2016 cells.
Pictures
These pictures were taken under controlled conditions and each picture is labeled with details.
Set 1
In this first set, the new LED is running on a brand new set of cr2016 batteries, and the rs276-320 is being powered by the link port on a brand new set of batteries.
Set 2
This set shows the new LED after 12 hours of constant operation on the same set of batteries. Although significantly dimmer, the new LED is still (after 12 hours) as bright as the rs276-320 running on a brand new set of batteries. Note: I said "as bright as", in the pictures you can plainly see that the rs276-320 diffuses the light more and the new LED is more of a "spot".
Conclusion
What can I say? this is one bright LED and will run on one set of batteries for quite some time! This LED IMHO would be the best LED for GBA external lighting as long as you can meet it's 2 main needs:
1. Provide the proper voltage of at least 5.5v -the cr2016's supply 5.6v
2. House it into a structure/device that will hold it at least 4.5 inches above the screen
I've been experimenting with these LEDs for about a week and a half now. I've built 2 different supports/structures that hold the LED and work pretty well. I've not perfected a support/device, but if and when I do, I will post complete instructions on how anyone can do it with a few simple materials (that I bet you have in your house right now!).
Now about availability/where to get these...
I currently have a source for these LEDs, but will not be able to sell any unless there is enough interest. Even if you do not want to use these as a GBA lighting device, there is good news...They are already in the form of a product
Confused? Click here for the revelation!
--------------------------------------------------------------------------------
1-20-02 Additions to this page will be added if anything new comes up.
The last real technical advance was the 2 LED mod.
7-28-01 I was a little hesitant to post this story, but at the advice of a few readers, I have decided to do so. I believe that by posting even this failed attempt, it will inspire people to keep pushing this idea further and may eventually lead to someone else learning from this attempt and possibly taking an aspect of it and implementing it into a future design.
The goal
The goal of this experiment was to try to find a way to supply the LED/s with at least spec. voltage (3.6v). Since the voltage coming out of the link port is only 3.1v in it's lowest state (while playing a GBA game), the LED/s are not as bright as they potentially could be. You'd be surprised at how much of a difference an extra .5v can make.
The method
I wanted to find a solution that would fit into the LS's case, that was readily available, and that was inexpensive. After researching different calculator, camera and computer BIOS batteries, I decided that possibly a certain type of these batteries would be the best way to go. I purchased some 3v calculator batteries and a couple 12v garage door opener batteries. After purchasing the batteries I checked their voltages and measured the voltage across the LED that I planned to use in the mod. Everything seemed fine, the LED was getting about 3.8v from the two 3v batteries I decided to use. Had I tested a little more extensively before I got down to business, I would have saved myself a lot of work
The Solution(or the thought to have been solution that is!)
I decided to use the LS's dimmer (potentiometer), switch and plastic housing as a basis for the mod. Since I wanted to keep everything packed into the tiny case, and have it fit securely, I decided to leave the potentiometer mounted to the board and to cut away parts of the tiny copper "runs" (the "wires" of a printed circuit board). I did this so that I could leave the potentiometer mounted to the board without having it short out the circuit once it was wired to the terminals, switch and LED leads. It sounds drastic, but I was very careful. I ended up using single strand copper phone wire to connect the potentiometer to the switch and to the LED leads and battery terminals. I created tiny battery "terminals" from 2 pieces of phone wire that would hold the 2 3v batteries in place. Also note that since the batteries were fitted in between the link port extension and the tiny on/off switch that they had to have tiny pieces of black tape at each end so that they didn't short out. I had planned a more permanent solution than black tape though if the mod would have worked out
The Problem
Once everything was back together it seemed to work fine, the link port still worked as an extension, and the light turned off and on with the switch and was able to be dimmed. Plus, everything fit into a nice little self sufficient package
Then the problems started, upon testing the voltage across the LED I saw that for some reason the current was steadily dropping... not extremely fast, but definitely at a steady rate. And of course since the voltage continued to drop, the LED became dimmer
After trying many different tests and eventually just connecting the LED directly to the batteries and measuring the voltage, I reached the conclusion that something must be wrong with the batteries. The only thing I could think of was that the batteries were not providing enough current to the LEDs. When Andy returned from vacation I e-mailed him explaining the problem and he verified that yes, it was the batteries since those type of batteries are designed to provide a "trickle" current, and were simply not strong enough for this application. It's really too bad though since as you can see from the pictures it would have been a really nice solution.
Oh well, back to the drawing board!
--------------------------------------------------------------------------------
7-16-01 Many new conclusions and revelations have been reached in the past week or so. The first and most important of these conclusions is concerning the mythical "perfect" LED that I've been searching for.
I've basically come to the conclusion that there will not be a "perfect" LED, without a "perfect" source of power for the LED. Let me explain; I've tested at least 6 different LEDs and only 2 of these LEDs are viable candidates in the search for the perfect light solution: The ever popular radioshack 276-320 and the just tested Nichia flat NSPWF50BS.
The Nichia NSPWF50BS shows GREAT promise and has an unbeliveable wide, white coverage area. These characteristics are obviously an advantage over the 276-320's blue "halo" effect, but the 276-320 does not require a reflector that the flat Nichia NSPWF50BS definitely requires. However, when supplied with the close to spec voltage of 3.54v the Nichia provides much, much more uniform coverage than the 276-320, and offers a much purer white light as opposed to the 276-320's bluish hue.
Have a look at these controlled test pictures:
A. B. C.
Descriptions:
A. Comparison shot showing the physical characteristics of both LEDs.
B. A comparison shot of the 276-320 and the Nichia NSPWF50BS in total darkness in a highly controlled testing environment, each placed 6 inches from the target grid. Notice the blue "halo" that the 276-320 is now infamous for. Also notice the unbelieveable white and uniform light emitted by the NSPWF50BS.
C. A comparison shot of the 276-320 and the Nichia NSPWF50BS in total darkness in a highly controlled testing environment, each placed 3 inches from the target grid. 3 inches is approximately the height that the LED is positioned in above the GBA's screen when the Pelican LS is in it's nearly vertical state. Notice the blue "halo" of the 276-320 is less evident in this picture due to the fact that the LED is much closer to the target and also due to the fact that through the camera's "eye" a lot of the blue is cancelled out due to the brightness. It is still highly evident though that the 276-320 casts a blue hue while the NSPWF50BS exhibits a much purer white light.
Revelations and Conclusions:
The Pelican LS only puts out about 3.1v when playing a GBA game. This is, unfortunately, not enough voltage to utilize the potential of the Nichia NSPWF50BS. The Nichia LED would also require the use of a reflector(note: NO reflector was used in the controlled testing) if installed into the pelican shield. The original reflector, although not tested for this purpose would probably work fine.
Here is a list of pros and cons for each LED:
Radioshack 276-320
+ Bright enough while using only 3.1v
+ Readily available
- Blue "halo" effect
- Although not needed, a diffusing lens helps this LED out a lot
Nichia NSPWF50BS
+ Whiter, more pure light than the 276-320
+ Much wider coverage area that does not require the use of a diffusing lens
- Does require a reflector to control the light
- Requires at least 3.4-5v to be bright enough to use in this application
- Not as readily available or as convenient to find as the 276-320
Although in the above comparison the Nichia has one more con than the 276-320, I strongly believe that the Nichia might just be the way to go. The GBA's screen really needs an even white coverage in order to be effectively illuminated. The problem now is coming up with either A) a redesigned circuit in the LS which would include a dc-dc power boosting chip (Andy is working on a schematic and prototype for this right now), or B) using an external source of power rather than trying to draw power from the link port.
Anyway you look at it, we are that much closer to finding the perfect light!
--------------------------------------------------------------------------------
7-7-01 While I was away I finally received the light diffusing LED covers from Huw!
Thanks again!
Ok, these lenses are designed to fit over the LED and diffuse or disperse the light in order to create a wider viewing angle. Although the lenses do help cut down on the glare and the spotlighting effect, unfortunately there is a price to pay.
But first more good news about the lenses: They significantly reduce or more or less eliminate the blue "halo" which these radioshack LEDs seem to emit. They also do away with the "rainbow" effect that I'm sure many of you have seen with the lighting devices that are out there (including the modded LS).
The main problem with the lenses is that in diffusing the light, quite a bit of light normally focused on the screen dissipates too much and covers less of the screen (most notably the bottom corners). I even tried the lenses in conjunction with a lower viewing angle degree rated LED (a 13 and a 20), and the results were pretty much the same. Here is a comparison shot of SMA with and without the lenses:
Conclusion: So after all this waiting on the lenses, are they worth it? Well, I really don't think it's as simple as that. I think that you must make up your mind for yourself and decide if you would rather eliminate some of the lighting "artifacts" of the LEDs (i.e. blue halo effect, rainbow/oil effect) or if you'd rather have a brighter screen while sacrificing a more pure light. Personally, I would rather have a brighter screen and just deal with the rainbow effect (since it never really bothered me that much in the first place), but I'm sure there are people who would rather eliminate the rainbow and "halo" effects instead of having a brighter screen.
Where to get them? Ok, so you want to try them out? Huw sent me a link to a site that sells these lenses in the states and Canada, you can find it by clicking here. I don't have any info regarding the price of the lenses, shipping or such. From the looks of the site you must request a price quote or request samples in order to get the lenses. It isn't a simple "click and buy" site so be warned. This is probably a good thing though since the most you'd need would be 1 or 2 so maybe just try filling out their sample request form? Good Luck!
--------------------------------------------------------------------------------
Further Testing and Thoughts: I've just finished running some tests on the radioshack 276-320, the 5600mcd 20 degree, the radioshack 3700mcd 13 degree, and the glowguard's LED. I was testing in hopes of finding a way to diffuse the light from these LEDs while at the same time keeping the light as bright as possible and was also trying to determine overall usefullness of the LEDs for this application.
In the testing environment: The testing environment consisted of a fixed base in a fixed position(6" from target) that each LED could be placed in and tested for brightness/diffusion aimed at a piece of white paper. This has been the most controlled test that I've conducted so far.
The radioshack 276-320 exhibited a blue halo or "ring" effect that was very evident against the white paper, but for whatever reason(most likely the high viewing angle degree rating) it still seems to have the most going for it. I've read a few other reviews of the radioshack 276-320 LED and the reviewers seem to be of the opinion that it is really more like a 60 or so degree rating contrary to the "100" listed on the back of the package. I can't say that I'd agree with that assumption or not because I simply don't have the kind of equipment(or even know what equipment is used for that matter) to do such specific testing.
The glow guard's LED appeared to have a degree of about 30, not much more of a spread than the 5600mcd 20 degree ones that I tested a few days ago. This is obviously the reason that the glow guard puts a bright "spot" on the screen, the viewing angle degree rating is rather low compared to the radioshack LED which has a rating of 100.
The radioshack 3700mcd 13 degree LED (received them today) is a very poor LED for this application, it is the bluest "white" LED I've seen yet, not to mention that it is very much a "spot" light with it's pathetic 13 degree rating.
The 5600mcd 20 degree LED is also more or less a "spotlight". You can see more info further down the page from the initial testing of these LEDs.
Light diffusion and more conclusions
Ideally, If the LED could be diffused without the means of additional products, materials, or other covers, it would be great. But it doesn't seem like that is going to happen anytime soon. I was testing in particular one option of diffusing the light of the LED's; using #0000 steel wool to frost the lens in hopes of bettering the diffusion. I must say that I had originally thought of using sandpaper, and that Steve e-mailed me suggesting sanding the LEDs as well. I frosted the lens of one of the 5600mcd 20 degree LEDs first. I powered it on in the testing setup and it looked very hopeful; the "spot" light was gone, the light spread out more evenly and it had a very nice effect. However, once I soldered the LED into the light shield and put it to a "real life" test, the results were terrible
First off, yes, the light was diffused and there was no "spotlight" effect, but at the same time the light being emitted was much dimmer and was only tolerable while playable in pure darkness.
Needless to say, I didn't bother wrecking any more LEDs(one reason that I've been putting this test off is because the LEDs are kind of expensive).
So it looks like it's "back to the drawing board" for now, and back to waiting for the diffusing lens covers from Huw. Note: There will most likely be a delay in site updating and testing though since I will be out of town next week. I'm not even sure if I'll have net access at that time to update the site and answer e-mail. I really hope to find a way though, because I like keeping the site fresh and updated at frequent intervals.
If anyone is interested, I also wired 3 of the 5600mcd LEDs together in a triangular pattern to see just how much light these things would put out when in a small group. The effect was an amazing amount of light! Check out the pictures here if you want
--------------------------------------------------------------------------------
New LEDs Tested: Yesterday I received another LED that I had very high hopes for.
I say "had" because the results were rather disappointing. The LEDs were rated at 3.6v, 20mA, 5600mcd and had a degree rating of 20. For the sake of comparison the current 276-320 radioshack LED's ratings are 3.6v, 20mA, 1100mcd and a degree rating of 100. There are 2 issues that keep these new LEDs from measuring up to the current LED used in the mod:
They tend to emit a yellowish green hue opposed to the radioshack 276-320's bluish white hue.
The viewing angle of 20 degrees is just not efficient in illuminating the entire screen.
I have my fingers crossed while waiting on the lens cover samples that Huw is sending me.
I really hope that they will possibly diffuse the 5600mcd LED's light enough to get over the hurdle of the limiting 20 degree viewing angle. By the way I tested 4 of these LEDs so it was not a bad one that exhibited these results.
Here are some comparison pictures, do not pay attention to the glare, look at the color of light cast onto the screen and the amount of the screen that is lit. The degree measurement is the viewing angle and the mcd measurement is the brightness rating:
