PDA

View Full Version : HDD Clock ????



jsaddiction
03-28-2012, 09:57 AM
I saw this online and thought that it would be a cool thing to add to my desk at work.
http://www.youtube.com/watch?v=K1asNB0te0o

I have already sorced the HDD and made the appropriate modifications but now i am trying to decide what route to go with for a controller for the LEDs and time keeping functions. I would like to incorperate a LCD screen and only use the Platters when i want to show it off.. Maybe have an external switch or something to turn the HDD display on. I was thinking of using the chip out of a chronodot for time keeping with a super cap to eliviate power losses. Some buttons or a 5 way joystick would be nice user interface to change display settings and time. I have never programed with the Propeller but it lookes like it wouldn't be hard to do since the OBEX will have half the code for this thing.

Has anyone done this or know if the Prop would be up to the task? The disk i have is 7200 RPM which would be 120 FPS each frame would have segments of the display divided by say 6 degrees giving you an RGB value for 60 different positions 120 times a second. CAN IT BE DONE

Cluso99
03-28-2012, 10:55 AM
the prop is definately up to the task. I am sure you will find lots of help. The projects done with these povs are great.

Ahle2
03-28-2012, 08:51 PM
Of course the Propeller is overkill for a simple HDD clock.
What could you do with all those unused cogs then?!


http://www.youtube.com/watch?v=6CX2E9Xh1hM

Phil Pilgrim (PhiPi)
03-28-2012, 09:23 PM
I'd be reluctant to unbalance the platter by cutting a slot out of it, then revving it to 5400 or 7200 RPM. Holy flying shards of aluminum, Batman!

-Phil

jsaddiction
03-29-2012, 02:37 AM
I have never used a prop but i am now looking forward to expand my knowledge. I am looking to design a piece of electronic artwork for my desk. I will be making a printed board and will attempt to enclose it all in one piece including power supplies. Actually Phil I have already opened up 3 broken drives one of which was a server drive designed to spin at 15k RPM. I was quick to tear this thing apart so i cut the slot out before figuring out how to keep the thing spinning. Apparently most newer drives will shut down the spindle drive circuit if no data is read from the read arm. I got it to spin up to its designed speed and no flying shards and to be honest with you it really wasn't that out of balance. I removed one of the screws on the opposite side of the slot and the vibration was almost eliminated. I had almost the same amount of vibration as an untouched drive. i was digging around with my scope trying to find a logic level to control the spindle motor and found that it was controlled over a serial line and then i messed up and shorted a couple of pins with the tip of the probe... oops. no more spin... tore open another drive only to find out that it was controlled the same way. I found one last drive and took a look at the board and found that the drive circuit was not just a chip but several components and stages. i disconnected the read arm and bingo... still spins. no modification needed!!!.. I cut a slot out of the top platter with a dremel and filed down the sharp edges i left a little material (1/8th" or so) at the outer most and inner most edge. I removed that one screw opposite the slot and whamo zero vibration and very quiet in comparison to my first attempt. This drive is 7400 rpm and very stable. my next step is to attach a small magnet like the one on the read arm parking assembly to the bottom platter and re-balance. My goal is to mount a hall sensor to the base of the frame and use this as a zero degree flash.


I am interested in any input you guys have. As said before i love a challenge so learning the prop with this project will be an experience. Please fill my brain with ideas and tips maybe even some insight into how this could be or should be programmed. Thanks for all your help.

Phil the precision RPM gauge with the BS2pe Mobo is working very well Thank you

Cluso99
03-29-2012, 04:49 AM
I would mount the magnet on the slot side as this will tend to balance out the slot. Interesting that they are detecting data from the read heads - manufacturers are becoming too smart for the hackers :(
Now have to get the prop to drive the motor too!

average joe
03-29-2012, 08:48 AM
I think this has some way cool potential. Haven't dug into the tech of this yet, but would love to play with this once my project is finished. @Phil, I tend to agree with your statement of unbalancing the platter. I worry that it could work fine for quite some time, until enough stress fractures have built up to create what Phil suggested. Once I read about the magnet, I tend to think it's a bit safer if balanced properly. I am also wondering about not cutting a complete slot. I have a few junk hard drives so I will have to dig into one and check it out sometime.

Cluso99
03-29-2012, 09:06 AM
How about just putting a pcb onto the platter with the prop and a battery at the centre and use it like the conventional POV displays?

Alternately, mount the leds on the underside of the platter with a tiny hole above each led and mask around the led to stop the light from getting to other led holes. (mount the leds with epoxy and paint this black). Os some similar concept.

Warning - I take no responsibility for parts flying off - so make sure you safely enclose the device - perhaps under a thick piece of perspex/accrylic.

average joe
03-29-2012, 09:11 AM
How about just putting a pcb onto the platter with the prop and a battery at the centre and use it like the conventional POV displays?

Alternately, mount the leds on the underside of the platter with a tiny hole above each led and mask around the led to stop the light from getting to other led holes. (mount the leds with epoxy and paint this black). Os some similar concept.

Warning - I take no responsibility for parts flying off - so make sure you safely enclose the device - perhaps under a thick piece of perspex/accrylic.

All things I was thinking about too. Not sure what would be the most safe. I want one so bad now, but just like a jacobs ladder, this one might have to wait till i get a man cave.

jsaddiction
03-29-2012, 10:07 AM
Sorry about the confusion with the placement of the magnet. You are absolutely right. The magnet should be placed on the same side as the cut. My latest drive does not have a cut all the way. It has some meat left at the spindle and some left at the outermost edge. I think using this technique, balancing the drive as best as possible, filing the edges of the cut (produce less turbulence) and then covering the whole thing with lexan should provide enough safety. I think as long as its not dropped we should be OK.

i am loving all the ideas. Some pretty cool thoughts. I like the propeller pov display using a propeller... kinda makes since . I think since i haven't seen too much with this type of hdd clock when this is all ironed out i will post some design notes and some how to for others.

The tech for this thing seems to follow a certain pattern for the most part. As far as i can tell we need to measure the time between passes of the hall sensor and divide this time by a certain number of possible positions of the clocks arms. For instance. If we measure a 7200RPM drive we would get around 120hz pulse from the hall sensor or 8.333 ms between pulses. If we decided on a 6 degree arch then that would give us 60 possible places for the hands to exist. This shows we need each "position" to be
138us per hand position. If this is put into an algorithm each rotation and the display is drawn the next rotation then we have a display that will not be sensitiveitive to RPM.

What do you guys think?

Cluso99
03-29-2012, 10:52 AM
I am interested to know if the different hard drives were from different manufacturers and if the platters were identical. i.e. were the mounting holes in identical positions?

jsaddiction
03-30-2012, 01:06 AM
The hhds are from different manufacturers and the platters are interchangeable. Take a look at a drive or find a picture online. Hard to explain but the platters are held in with spacers and a small plate on top. The hole in the middle of the platter seams to be standard but the thickness is not. May take several hdds to make an odd platter configuration.

I installed a hall effect sensor pulled from a cpu fan and it works beautiful. 120Hz pulses = 7400 rpm. I made a transistor circuit to use the pulse to strobe an led. This helped me align the top platter so pulse happens when the slot on the top platter is at the 12 o'clock position.

All i need now is a prop. Any suggestions on a good development board that will help with the development and the wallet?

Cluso99
03-30-2012, 02:10 AM
For your pcb, a lot depends on how much room you have. I have a 1" sq board without the FT232 chip. It has a 3v3 reg from 5v in and the transistor reset cct so you can use a cheaper CP2102 board for the propplug (if you dont have one). All I/O pins are brought out to the edges. Another solution might be the Prop protoboard with or without usb. You can trim this pcb down if necessary and it has a breadboard section around the prop. There are lots of other boards too.

jsaddiction
03-30-2012, 02:22 AM
I think the direction i would like to go is develop the circuit on an external board. either breadboard or some other solution. using that i would then design a more permanent printed pcb made to fit where the read arm was located and include a usb connection for further program debugging and development. I will then use the development board for other projects and keep the finished HDD clock

jsaddiction
03-30-2012, 07:04 PM
would it be a good idea to use an LED driver ic to control PWM and attain a smoother gradiant of colors or would it take too long to serial clock that data in.. It looks like 3 bit per channel is common so 9 bits of data? in 138usec

Cluso99
03-30-2012, 07:41 PM
How many leds do you want to use? The prop has a total of 32 I/O pins and has a pair of counters capable of PWM in each cog, making 16 PWMs available. But PWM can also be done in software, so a cog could control many I/O pins in a software PWM mode. Remember, we do not use interrupts so each cog is effectively autonomous.

jsaddiction
03-30-2012, 08:27 PM
Thats true. I guess i was thinking more about constant current being a good way to control the leds but a resistor should work fine. I will have a fairly reliable 12v source. These leds are what I was thinking http://www.superbrightleds.com/cgi-bin/store/index.cgi?action=DispPage&category=BARS&Page2Disp=%2Fspecs%2FFLS-RGB.htm
i guess i could replace the smd resistors and reconfigure the led strip for 5 volts either way