When [Caleb Kraft] was in full production for Hackaday he pumped a pile of awesome videos. But not every project worked out. He’s been a fan of the Fail of the Week posts, and sent in his own recollection of a project gone wrong. Above you can see his phosphorescent CD player. He prototyped the project in May of last year but technical issues and looming deadlines meant it never saw the light of day. We’ll fill you in on his fail after the jump.
Editor’s note: We need more tips about your own failure! There are a handful of submissions left in our reserves, but to keep this topic as a weekly column we need help tracking down more failed projects. Please document your past failures and send us a link to the write-up. If you don’t have a blog to post it on you may do what [Caleb] has done and email us directly. Remember to include any images and links to video which you may have.
The gist of the project is that [Caleb] wanted to use an LED to charge up some glo-in-the-dark paint. The easiest mechanical setup he could think of was to use an optical drive, with the disc itself as a moving display, and the lens sled as a way to move the light source. [Caleb’s] own words tell of the successes and failure he encountered:
Hey Mike and gang, let me start out by saying that I really love the new direction of the site. You guys are kicking some serious butt in my opinion. One of the new things I really love are the Fails. I thought I would share a fail from my time there at hackaday, a project that didn’t get very far, but I think has some really cool potential.
As many already know, I’m a little obsessed with glowy things. I was really inspired by the Ghost Matrix and immediately started thinking of things I could do along similar lines.
What I came up with, code named Enceladus for no apparent reason, was going to be a Compact Disk sized glowing visualization.
In the beginning, things were looking good. I grabbed the clear protective CD off the top of a stack, scuffed it up, and sprayed it down with glow in the dark spray paint. It cracked a little, but I didn’t care since this was going to be a proof of concept. I ripped the carriage out of a CD player and replaced the laser with an LED. At first I used a white one, then I ran to the store and bought some UV LEDs
I thought it looked good enough to continue. So I did! I fixed the LED spacing issues so it wouldn’t rub. Then I added controls for moving the LED as well as controlling PWM on the spindle motor.
At this point I found that the poor motor couldn’t handle low speed at all. It was fast or nothing. I had intended to rip it out and put a stepper motor in there for much better control, but got distracted by my almost coast to coast hackerspace tour that summer.
I’m pretty sure I tossed the hardware a long time ago, but this is a project I wouldn’t mind revisiting in a more refined manner in the future.
thanks,
Caleb Kraft
www.calebkraft.com
www.thecontrollerproject.com
The failure here is actually quite similar to the one that we featured last week. The issue is that the CD needs to move slowly enough for the UV LED to “charge” up the phosphors. It also needs to be accurate enough to achieve some type of meaningful resolution. We’d like to know your thoughts on this project. Is a stepper motor replacement the best way to go, or could this be driven reliably with the right brushless motor controller? Even if the motor control was sorted out do you think the single light source will be able to charge the phosphors fast enough for a message to be read before the glow material fades back to normal?
Fail of the Week is a Hackaday column which runs every Wednesday. Help keep the fun rolling by writing about your past failures and sending us a link to the story — or sending in links to fail write ups you find in your Internet travels.
Hey! I made it on hackaday!
You should use a laser diode next time. Something around 405nm should work
I used a cheap >5mw “royal blue” laser and a hacked together read head unit (into something resembling a galvanometer) from a CD drive. It worked quite well but even at low speeds having a single point of light is not quite bright enough. A brighter laser might work, but this one was already bright enough that it made your eyes ache after watching the device operate for 10 minutes or so.
I found that a particular shade of green acrylic plastic is quite effective at filtering most of the UV while letting some of the green through, although with the size of my hacked galvanometer I have yet to work on a way of miniaturising it to the point where I can fit a filter over the device .
A video of it operating in a slow “night rider” mode is below- It’s hard to see but it is drawing dots, capable of about 20 pixels high.
http://www.youtube.com/watch?v=BmfjW9e0HZU
Here’s a video with words being printed;
http://www.youtube.com/watch?v=h5FgM11-6nA
I drove it using a series of belts and a stepper motor from an old portable cassette player. The whole thing is mounted in a dead hard drive, and the screen is the drives disk on it’s original bearings. Horizontal resolution was great but the sticking point was really the laser brightness/image persistence.
Multiple UV LEDs do work well – I advise using a 5mm rounded package as anything else (square package, SMD) doesn’t focus the light enough to make nice looking pixels.
Very cool project. I’d love to see a finished product. Have you ever considered a focusing lens between your LED “write head” and the write surface? For a engineering project back when I was in college I used a live light rig that you’d see at a music concert to write to a projection screen coated with a luminescent spray (people passing by could draw pictures or messages on an old stylus style touchscreen laptop and then see their work projected up on the screen mounted on the side of the engineering building). In the light rig I replaced the whole lighting system just to use the motor system that was capable of precision control (like those laser light systems that make shapes with high powered lasers that are moved at extremely fast rates to achieve a bleeding effect so the shape seems consistant). With the lighting I used it barely activated the luminescent coating at extremely slow speeds and the light had too much spread making the effect very rough. All I had to do to fix this was a matter of getting a focusing lens (I think it was from an old projection tv). This tightened up the light scatter and activated the luminescent spray so well that I was able to increase the write speed by a factor of five. I also noticed that persistance improved greatly as well. The only hard part was altering the lens slightly to get the correct focus shape and intensity. I know there are very small form factor focusing lenses that could fit in your project. Maybe you would have to shift the head back a few mm.
There may be a suitable lens inside the CD player optic already. Just be very careful removing them, they scratch very easily.
It’s almost as if people forgot lenses exist, which which you can focus light.
Neat project, regardless of success… :D
I would increase the led output (high wattage or maybe a blu-ray laser) in order to charge the phosphor faster, this would help improve the resolution.
With the rotation of the CD stabilizing the motor speed, wouldn’t the usual POV setup (opto-flag and a timer) be adequate?
I’m the failer from last week.
Despite all my experience to the contrary, I still think that driving these motors at low speed and reasonable torque should be possible. It would definitely need some feedback, either inductance between coils or separate HALL sensors. It also wouldn’t be very efficient, but perhaps efficient enough to not overheat the motor.
That brushless motor driver project that I linked to in the write up above claims it can work without hall effect sensors. It apparently takes readings from the coils not being driven in order to judge the position of the rotor.
Brushless motor controllers for RC models mostly do exactly this. The problem is that to get a decent reading from the ‘unused’ coil as a magnet goes past, the thing has to be turning at a certain minimum speed. Sensorless BLDC controllers run open-loop like a stepper driver up to the minimum speed and then start commutating properly. Their low-speed performance is not good, as a result.
That’s interesting. So is there some type of “jump-start” algorithm to get the motor up to the minimum speed for the readings from the coils to work using a sensorless BLDC?
Why can’t you strobe the UV LED fast enough to make it work, over multiple rotations of the disc? If the disc is spinning at 150rpm, it just means that a given point is under the LED more frequently, but for a shorter period of time. Am I misunderstanding how the phosphors charge up?
You should try that. Sounds reasonable.
That was my thought as well.
I also thought of this, it was the first thing that came to my mind after reading the problem. I’m really interested to see if this is possible.
I was wondering something similar. Add a single mark and an IR emitter/detector pair to signify location. Basically, POV. If you pulse the LED quickly, you should be able to bump the current output to it to allow for more intensity.
Sure you could, but that isn’t what I wanted. I wanted to do things like sine waves and square waves along the disk. I wanted to be able to bring it to a stop, then reverse, etc.
You could theoretically flash your LED and write to the disk like a dot matrix display, but it would be very limited. A bank of LEDs would be much better.
Yeah, i was thinkinking something like a row of 0402 UV leds on the thinest copperclad you can find instead of your single LED. Then it would be pretty much like any other POV thing, except of course for the phosphor and the sled being able to move (easy sine-wave text scrolling like in the oldskool demoscene)
Ah, i see.. I should read the comments before i reply to them XD
You might have hit the nail on the head! Or you might not, I dunno. Does the phosphor do well with shorter brighter pulses, or longer dimmer ones? Or even longer ones with the same brightness, just slow the spindle down!
ISTR the record player version of this was much slower, phosphor paint’s not really a fast-working kind of thing. In the record player, the phosphor had faded quite a bit by a quarter-turn round, much more by a half-turn.
Caleb could try perhaps using more phosphor. Glonation.com sell “glow in the dark powder” online. Quite a few other companies seem to sell it, in different colours, though standard green is by far the most efficient at retaining a glow. Perhaps make up your own paint using the powder at a higher concentration. Would be expensive in quantity but you only need to buy a small sample of the stuff.
Glonation.com sell all sorts, with stuff from less than $6. It had good information on mixing your own paints, too. The paint is made from the powder in a solvent, there are several would work. Stuff like acrylics and resins too. I think the point is, all paints start off with the powder, it’s the “pure” form of the glow.
They seem knowledgable and helpful. And I totally don’t work for them! Bit of an unsolicited testimonial tho I’m not even a customer of theirs. They could be Russian Mafia for all I know, send ’em your credit card number and you’ll end up with someone’s organs in the post. I’ve no idea. It’s just a web site!
They do lots of colours. Glow-retention time on some of them is pretty miserable, 15 mins vs hours for the green. But that wouldn’t matter for this project.
Other thing, maybe more powerful light source. As others mentioned here, UV LEDs aren’t really very UV in the sub-$500 price range. You could try blue or violet, maybe white. Or maybe a UV tube, the little ones they use in currency checkers. Then some system of mirrors or maybe fibres, to get the light to the “disk head”.
I have some geared motors rescued from an old laser scanner here, slow rotation.
Might work for this..
Yes, bluray would also work as would 0603 UV SMD LEDs.
Did something with phosphorescent materials, a bicycle wheel mod.
This works well but its worth mentioning that infrared light quenches the phosphorescence so you really need to add a bandpass filter.
Some UV LEDs have quite pronounced infrared emissions, some don’t.
I believe this will be specific to the paint chemistry – I have used “plasti-kote” glow in the dark spray and (although it would have been cool) was unable to get any noticeable quenching.
I didn’t know that, how fast is the quenching? If it’s faster than charging maybe Caleb could do a “white board”, charge the disc with a bank of UV LED’s and then discharge sections with an IR LED to make the patterns, it would make things mechanically simpler anyway.
Now I want to try this :)
If the exposure time is the issue, then simply slow down the frame rate. Expose the disc multiple times over multiple rotations.
Caleb, if you have a green laser pointer try this:
Charge up your glow in the dark disc (or whatever) then shine your green laser pointer on the charged up part for a few seconds.
:)
Hi, the quenching occurs in a tenth of a second.
I could actually perceive individual flashes from a TV remote affecting the phosphorescent (zing sulphide) coating on the exit sign material.
Then using a “white board” method would work, charge the disc with a battery of UV LEDs and quench it with the IR LED.
I would have thought that spinning faster like a similar whirling display, but perhaps the decay is too fast? Huh.
I made a similar project, but I used a turntable instead, and 7 LEDs.
http://www.youtube.com/watch?v=8PFDqtOrgy4
I think the turntable spins too fast, but the motor is direct drive, and is integrated into the original PCB, so I’m just using that to drive it. I don’t think it would be very easy to replace the motor with a stepper, unfortunately.
I did the same thing – the way I slowed down the rotation was to just put a voltage regulator on the motor. The unit I had ran at 12 volts so a 3v regulator caused it to crawl round, which gave quite a nice effect. A trim pot on the regulator lets me tweak the speed.
Put the disk on a giant gear, with the motor driving a smaller one? If it doesn’t need to be precise, using rubber contacts instead of gearing might reduce noise.
Galvo from an old hard disk driven with a sinewave (PWM or regular) would work well, a single LED could then be used and the existing head cable as the power connections.
You can get these down to 0402 in UV 2mW now, used extensively as case mods and for their size they are extremely bright.
Use IR LED as well as a quencher if needed but this can simply be a single line of lights mounted on the underside of the head arm.
Someone should make a variant of this with a broken Microdrive, one SMD UV LED, infrared quencher and a gel filter and then wear it to Comic-Con etc.