Topics Topics Edit Profile Profile Help/Instructions Help Member List Member List  
Search Last 1|3|7 Days Search Search Tree View Tree View  

LED cube - basics of timing /frequencies

:: EPE Chat Zone ­:: ­Radio Bygones Message Board :: » EPE Forum Archives 2010 - » Archive through 14 July, 2013 » LED cube - basics of timing /frequencies « Previous Next »

  Thread Last Poster Posts Pages Last Post
  ClosedClosed: New threads not accepted on this page        

Author Message
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1394
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 08 July, 2013 - 04:09 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

I've spent many hours in the last days reading about POV and LED cubes.

Given an 8 x 8 x 8 cube and a refreshing frequency of 150, from all my reading I concluded the following:

T (to light them all) = 1/150 = 6667 usec

Time devoted to each LED in the sequence (whether on or off) = 6667 usec / 512 = 13 usec / LED

If I have to introduce some variation in the perceived luminosity of each one I should apply to every LED a PWM signal of suitable frequency during the asigned time slot of 13 usec.

I am not sure if I will build one cube or not (most probably no) but I am interested in the concept.

Please, I am not looking specifically for links but basically anyone telling me if what I say above is correct or not. Comments really appreciated.

Also, isn't a 3 layers cube as proposed, equal to a 2D frame of 64 x 8 LEDs on this regard?
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1395
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 08 July, 2013 - 07:48 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

In the post above, last line, please read: " isn't an 8 layers cube as proposed "
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

boris
Frequent Contributor
Username: boris

Post Number: 754
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 08 July, 2013 - 10:30 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

You don't have to light them one at a time. Depending on your chosen driving method, you could do line-at-a-time or layer-at-a-time or even update them all simultaneously with 100% duty.
Thank you for not breeding.
Top of pagePrevious messageNext messageBottom of page Link to this message

boris
Frequent Contributor
Username: boris

Post Number: 755
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Monday, 08 July, 2013 - 10:32 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Also, 150Hz refresh is much faster than necessary. TV/Movie is 25Hz and most people are comfortable with that.
Thank you for not breeding.
Top of pagePrevious messageNext messageBottom of page Link to this message

istedman
Frequent Contributor
Username: istedman

Post Number: 337
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Tuesday, 09 July, 2013 - 11:20 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hello Agustin,

I built a 5x5x5 LED cube a couple of years ago from a design on the internet. Like most designs I looked at, you control 1 layer at a time. The design I used, switched on 1-25 LEDs per layer for 1.8ms. Each individual LED was switchable but the software output 4x bytes, to control the LEDs, then enabled the constant current drivers for the entire layer.

So it took <10ms to light 1-125 LEDs as required, giving a refresh rate of 100Hz, much higher than needed.

The constant current drivers, OnSemi CAT4016, define the brightness. They are driven at 30mA each but for short periods of time. Do not use TTL logic to switch the individual LEDs, use constant current drivers. Logic 0 of a TTL device is 0 to 0.8V, this causes a large variation in LED brightness. Remember Iled = Vin-Vled-V_logic 0/Resistor, when V_logic0 varies by 0.8V the current changes a lot.

Be warned, LED cubes are time killers :-)

Ian
Top of pagePrevious messageNext messageBottom of page Link to this message

james
Frequent Contributor
Username: james

Post Number: 239
Registered: 02-2007

Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 10 July, 2013 - 02:35 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hi Agustin

Just set up a little breadboard experiment driving a single LED from a 555 in astable mode.

With low duty cycles of 5% to 50%, persistance of vision effect reduced at frequencies of below about 40Hz.
ie. LED flicker just became visible when the frequency dropped below about 40Hz.

This cutoff freq reduced significantly as duty cycle increased from 50% through to 100%.

At 2% duty cycle flicker is visible below about 70Hz.

Your figures for driving each of the 512 LEDs in its own time slot would mean a very low duty cycle :-

13/6667 = .2%

probably meaning that the light from each LED would be hardly visible. Not sure how the LEDs would handle being driven by a very high current for this short period of time to overcome this lack of brightness.
Much better idea to light the cube in sections, each section for a longer time period producing greater brightness.

Cheers

James
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1396
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 10 July, 2013 - 02:48 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Thanks to you all for replying.

My bad, I couldn't recognize from the beginning that it is one layer (8x8) at a time! Too evident...!

But, and this is still my stumbling block, what if I want to apply different whatever to each LED in each layer to have them with different luminosity? Have to see how could I.

In other words I could be needing up to 64 different PWM signals per layer when a certain layer is activated. (That led me to think driving each one individually).
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

istedman
Frequent Contributor
Username: istedman

Post Number: 338
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Wednesday, 10 July, 2013 - 11:35 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Agustin,

An interesting part is the Texas Instruments TCA6507. This has programmable LED intensity controls, fade on/off times all controlled via the I2C bus. You would need the 400 KHz I2C bus to control it for a cube.

If you don't mind SMT parts, I can give you the schematic and single sided PCB layout of my test board.

Experimenting with LEDs is fun!

Ian
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1400
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Thursday, 11 July, 2013 - 10:02 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hola Ian,

Thanks for the offer.

I was somewhat aware of chips like that. Point is that I am considering to have every LED controlled independently for different levels of luminosity. The 6507 would allow to control just two as I would like.

Discarding a hardware solution, last night I started to consider implementing in software what is called MIBAM. For that in the time slot asigned to one layer I should accomodate the succesive setting/resetting for 64 LEDs, repeated 7 times to comply with the BAM algorithm. Some 512 (fast) steps plus the unavoidable overhead.

In two hours, when having my espresso at the cafeteria here at the corner I will do the maths to see if it's feasible. My fastest 18F might be called for that.

I have to optimize (and I really mean THAT) the way to read the respective bit and ser/reset the LED in consonance in the fastest possible way.

It seems that I am looking for a nightmare as when I got my micro writing text on the scope.

BTW the outcome was good...nightmare or not!
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

externet
Frequent Contributor
Username: externet

Post Number: 183
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Friday, 12 July, 2013 - 06:08 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Hey Agustín, go all the way for complexity and real nightmare !

Use these: ----> http://www.alibaba.com/product-gs/694513426/high_brightness_4_pin_5mm_rgby.html
Abolish the deciBel !
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1401
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Friday, 12 July, 2013 - 07:28 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Oh no! Not yet Miguel... I was tempted to start with RGBs but decided to become insane step by step .

Even if I wrote no code yet I think that there could be a way. This weekend I am writing some minimal one to test a recent idea I had.

(Message edited by atferrari on 12 July, 2013)
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

externet
Frequent Contributor
Username: externet

Post Number: 184
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Saturday, 13 July, 2013 - 12:00 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

OK. will let you off the hook.
Then go fully the opposite way; no need to warm up the soldering iron nor fight with positioning leds. Go virtual, that is what compfusers are made for, right ?

Your 8x8x8 spots on 3D interacted gyrated screen; RGBYW; and interfaced to 5 band audio music. That will be a home run!

I will be your first customer for the program, from money saved in leds, solder, sweat, blisters, tears and nervo-calm.

Am willing to wait ! You can do it! -Will show great in 52" HD flatscreens !

Miguel
Abolish the deciBel !
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1402
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Saturday, 13 July, 2013 - 01:28 am:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Wow! That's encouragement, Miguel.

It is said that hardware comes first but, I feel, not this time.

Have to solve the esential part to have the 64 LEDs of each layer driven as I want and then the rest.

Sleeping on that tonight.
Agustín Tomás - Buenos Aires - Argentina
Top of pagePrevious messageNext messageBottom of page Link to this message

atferrari
Frequent Contributor
Username: atferrari

Post Number: 1403
Registered: 05-2005


Rating: N/A
Votes: 0 (Vote!)

Posted on Sunday, 14 July, 2013 - 02:49 pm:   Edit Post Delete Post Print Post    Move Post (Moderator/Admin Only)

Yesterday, one my worst days this year, found hard to concentrate on the algorithm. Late in the night, got something that solves the essential part by transposing data.

Have to test it in actual hardware.

Not bad after all.
Agustín Tomás - Buenos Aires - Argentina

Administration Administration Log Out Log Out   Previous Page Previous Page Next Page Next Page