P.O.V.  Blog



2009-06-23    --    Source code is now available.

Contact with any questions/comments.

pov.c
pov.h
characters15.h
characters8.h
font15_Chancery.h
font8_Plain.h

---

2009-03-15    --    More images. They're so easy to do now.

---

2009-03-14    --    Made a few more images for the new V3 board

   

Creating images for these boards used to be very tedious, but today I wrote a program to do the dirty work for me. All I have to do now is create a simple image in a paint program, then my program analyzes the image and spits out a definition file ready to compile into the POV program. It handles both the 8-LED and 15-LED versions automatically. Today's images were created using the new program. What would have taken me at least 30 minutes to do manually took the computer only .04 seconds total, and it doesn't make any mistakes during the process.

The first image is just a bunch of circuit symbols (didn't turn out as cool as I'd hoped), and the second one of course is a bunch of bugs.

---

2009-03-12    --    Nabbed a few images of the new V3 board in action

Today I worked out a bug that's been plaguing this version, so now I can move on to stabilizing the display under changing speeds and generating new images that use the full 15 LED's available on a side. I tried out a few today and grabbed some shots.

This first image is of a new font set I ported to the board called "URW Chancery L Italic Not-Rotated". It doesn't look as smooth as I'd hoped, but it didn't turn out terrible either.

Unfortunately the rear wheel on the bike is extremely unbalanced, so it wobbles greatly during testing. Perhaps it will look better on the road or on another bike.

There is still some kind of odd timing issue going on that can cause the display to "glitch out" occasionally. It just appears as a bright streak very briefly on the wheel. It's infrequent and definitely not catastrophic, but I certainly would like to figure it out. Overall I'm very pleased with the way the board works and looks at this point.

Whenever I go mountain biking I leave the boards attached to the wheel. It's too much trouble to keep taking them off and putting them back on. Lately the terrain has been extremely rugged (I'm surprised the bike hasn't broken), and the boards have held up just fine. I don't think there will be a problem with their longevity.

Here's a 15-LED version of an image originally made for the older 8-LED boards:



The new board pulls more than 4 times the current of the older models, so the little 9v rechargable batteries I've been using probably won't last very long when on the road. Probably will need to invest in some RC hobby batteries of the lithium variety. These have come down in price, and offer more than 6 times the capacity in a package that isn't really any heavier. They recharge much faster as well.

---

2008-10-16    --    Version 3 prototype is finally built

Finished the first prototype for version 3 of the POV board. This version is double sided (LED's on both sides), and has 15 LED's per side rather than just 8 LED's on one side. Because 30 LED's consume more power than the micro controller is able to supply, 30 transistors have been added to drive the LED's, which requires 30 additional resistors on the bases of the transistors. With all LED's lit, the board pulls over 3 watts of power. The board has other additional features such as a reset button and a larger voltage regulator, and may get more extras later.

The MCU has been upgraded to an ATmega644p, with 64k of flash (increased by another factor of 4 over the previous version). This MCU has more I/O pins, allowing 30 LED's plus two inputs: one for the rotation sensor and one for a "mode selection" jumper.

The LED's on each side are controlled independently of each other so that text can be displayed in the correct orientation on each side. If not handled this way, text is printed backwards on one side. I'll post shots of it in action when I'm finished debugging.

---

2008-07-27    --    More is better

Made a few more V2 boards and created more graphics:





Went on a ride with 3 POV-equipped bikes. Unfortunately there were few people out so we got little attention, but a few motorists paced us briefly to get a better look.

Running two boards on one wheel works great -- much better looking because the refresh rate is doubled (less flicker). Three or four boards would look even better.

---

2008-07-05    --    Version 2.0 rolls off the presses

OK, now finished debugging Version 2 of the POV thingy. I've upgraded to the ATmega168 MCU with 16K of flash, which is eight times more memory than the V1 board had using an ATtiny2313 (but it's still easy to run out of memory!). I've added a hall effect rotation sensor to track wheel movement. Added code to stabilize the image (in scale and translation) regardless of wheel speed. Added code to calculate speed in miles per hour and distance traveled in miles. The distance function should be extremely accurate, but the speed will be ballpark only because the internal clock on the MCU isn't very accurate. I can add an external crystal to make this spot-on, but it's not that big a deal to me. In an 8-mile test, my display was off by 6% from a normal bike computer. I can certainly live with that. Next step is to make a duplicate V2 board and run them both on the same wheel, positioned on the same side but opposite each other. This will cause the image to look much better, especially at slow speeds.

Here's a shot of the speed/distance display in action.

---

2008-06-14    --    Version 1.0 comes to life

I recently completed my P.O.V. (Persistance of Vision) project. A POV device flashes lights in a pattern very quickly. When the device is put into motion (perhaps by attaching it to a wheel or even just holding it and swinging your arm in a wide arc), the timing can be adjusted such that the light patterns form words or graphical images. I used the AVR ATtiny2313 micro controller and some old amber super-bright LED's for this project. I attached the tiny circuit board to the back wheel of my bicycle and turned the pedals to put the device in motion. Please see the pics below:

Some limited graphics capabilities are available:
















Here's a shot of the board mounted on the wheel:


And a closer look:



The ATtiny2313 really doesn't have enough memory for this. I can only get a few letters mapped/defined at one time, so the text that can be displayed is very limited. I'll have to switch to a bigger MCU to continue work on this project.

More to come...