Hardrive NipKov Disk Hard Drive Clock?

[spinnaker]

  I am currently working on a relatively simple HDD clock that does nothing but illuminate cut out numbers.  Once I get that down I would like to move on to something a bit more challenging.

I was wondering if anyone could tell me how this guy did the following project.    I am pretty sure it is based off of the Nipkov disk.   But what I can't figure out is how he did the lighting from the LEDs.

In my project there will be 8 "compartments".  Each illuminated at the appropriate time to display the selected number in the selected location.   Fairly straight forward.

I assume he would also need individual compartments but what I can't figure out is how many he would need to something like this.   Can anyone guess what this looks like behind the disk? 

[dmills]

If you look closely at the stationary disk you can see LEDs sticking thru the surface, you only need one for each ring of pixels. 

I would guess that the other platter(s) have been removed to make space and a thin PCB with the LEDS glued on, then add some wireless power stuff and maybe a bluetooth chip and either FPGA or small processor, job done.

The disk is clearly not spinning all that fast, so I don't think he is using the original motor drive at original speed (Probably mechanical issues with the balance).

A nice starting point for these is an old video head drum assembly, you can use the former head wiring to provide a fairly decent serdes link to the outside world.

[spinnaker]

If you look closely at the stationary disk you can see LEDs sticking thru the surface, you only need one for each ring of pixels. 

I would guess that the other platter(s) have been removed to make space and a thin PCB with the LEDS glued on, then add some wireless power stuff and maybe a bluetooth chip and either FPGA or small processor, job done.

Not sure what you mean by one for each "ring of pixels".   

Attached is how my current project is going to work.  Are you saying that instead of those compartments that I have, he has rings of compartments?  I will try and take a closer look at the video but was not seeing it. I'll take another look at the video.

The disk is clearly not spinning all that fast, so I don't think he is using the original motor drive at original speed (Probably mechanical issues with the balance)

A brushless motor controller solves the speed issue.   I have one that you can take the motor to a fairly slow speed.

A nice starting point for these is an old video head drum assembly, you can use the former head wiring to provide a fairly decent serdes link to the outside world.

 

Wow I am lost there.   And I have one torn apart right no, that I just salvaged for parts, w sitting in my basement.

[spinnaker]

Here is another way to do it.    No holes just a strobe.   Not exactly sure what is going on here.    It is another Nipkov disk.

[helius]

Yes, the OP video is a Nipkow disk. It states that the resolution is 120x16, which corresponds to 16 holes drilled in the disk. You can see the holes at around 0:59 if you freeze frame. They are drilled on 8 radial sectors: so there are two holes on each sector. This means that the space behind the disk must be segmented by baffles into 8 radial compartments, with each further segmented into an inner and outer half, with an RGB LED in each one. Each LED pulses 15 "sector" pixels before switching to the next "track".
The system also needs a way to synchronize with the disk's rotation, which could be a small magnet glued to the disk or an optical sensor.

[dmills]

I was seeing the holes as being SMT LEDS actually sticking thru the disk with with the electronics mounted on and rotating with the platter, but a real disk with holes is possibly simpler at the cost of resolution and increased flicker.

Note that the holes do not need to be in a spiral, and there may be something to be said for setting up a pattern that gives you an interlaced output.

Regards, Dan.

[spinnaker]

Yes, the OP video is a Nipkow disk. It states that the resolution is 120x16, which corresponds to 16 holes drilled in the disk. You can see the holes at around 0:59 if you freeze frame. They are drilled on 8 radial sectors: so there are two holes on each sector. This means that the space behind the disk must be segmented by baffles into 8 radial compartments, with each further segmented into an inner and outer half, with an RGB LED in each one. Each LED pulses 15 "sector" pixels before switching to the next "track".
The system also needs a way to synchronize with the disk's rotation, which could be a small magnet glued to the disk or an optical sensor.

So you  think it is a radial grid?    Sort of like my design (Image in reply #2)  except the compartments go all the way around the circle  and there are multiple rows?  How large would each compartment be?
 

[spinnaker]

I was seeing the holes as being SMT LEDS actually sticking thru the disk with with the electronics mounted on and rotating with the platter

So more of a conventional propeller clock?   I don't think so.   As mentioned above if you look closely in the one frame, you can see the  Niplov spiral.

, but a real disk with holes is possibly simpler at the cost of resolution and increased flicker.

Note that the holes do not need to be in a spiral, and there may be something to be said for setting up a pattern that gives you an interlaced output.

Regards, Dan.

What would the LEDs behind the disk look like.   That is what I am trying to figure out.

[helius]

The holes are two spirals. You need a spiral pattern so that the LEDs can be spread far enough apart—and you need two of them because if you divided up the 360° of the panel into 16 compartments only 22.5° across, the LED wouldn't give even illumination. So instead there are 8 compartments 45° across and there are two LEDs in each of them, one for the inner part and one for the outer part.
A classic Nipkow disk, like in the Baird television, has a single "compartment", because it is supposed to scan a rectangular raster. Only one hole passes over the raster at a time, and the light that illuminates the whole rectangle evenly is modulated. If you add more rasters around the diameter of the disk, you can produce a much larger image, but there will always be a hole in the disk's center. This requirement for multiple rasters has an impact on the disk's resolution, though, because instead of only one hole passing over the light, every hole in the disk is always lit. So you need a separate light source for every hole, which is very different from Nipkow's conception.

[spinnaker]

Sorry but I wanted to drag this old thread off of the shelf.   I have been away for a while as I usually put the soldering iron down at the start of the summer.   

This thing is still bothering me as to how it is done.  How does he get the pixels so close together in columns   I understand the 16 compartments but wouldn't there need to be far more to get the columns so close together?

[helius]

Well, it's a raster scan (the Baird television/Nipkow disk, and the Farnsworth CRT are both the same in this respect). So it's a method of chopping up an analog video signal into lines. Within each line, the resolution isn't inherently limited: it's as good as the analog signal can reproduce given the signal path. The number of lines is a ratio of the line rate (the horizontal refresh for a CRT) and the picture frame rate (the rotational speed of the disk; the vertical refresh for a CRT). One difference is that mechanical TV does not require any blanking intervals as CRTs do (because there is no capacitance or inductance to overcome in deflection circuitry).
In this clock, there are 16 rasters being scanned simultaneously: 16 parallel analog signals. And each one is RGB (3-bit color, which I think means 1 bit for each primary and not 3). So the MCU needs to output 48 bits for each "pixel clock". All the rasters are doing the same thing, but the holes in the disk go over them sequentially, so each raster is "delayed/advanced" by one line relative to its adjacent raster.

[spinnaker]

": 16 parallel analog signals. And each one is RGB (3-bit color, which I think means 1 bit for each primary and not 3)."

But how are these LEDs arranged?   that us what I am trying to figure out.   I understand the Nipkov disk and the arrangement of holes.   I guess that would give you your vertical scan or columns.    I can clearly envision . How you would get a column of dots to display,   What I can't figure out is how you get so many columns   

Is it something like in my attachment?   Where each"compartment" would be lite by LEDs?   If so, how does he seem to get so many rows of dots?   I would think the most he could get was 16 rows and that is spread all around the disk.

P.S.  I understand how he is getting 16 rows because basically he has 2 nipkow disks intertwined.    I would be happy just to figure out 8 rows i.e. 8 holes in the disk.. 

And sorry for the delayed response.   I have notification turned on now.  Thought I had that on by default.

[helius]

I already described how the compartments are arranged, read the thread.
Each compartment has a RGB LED, either a normal one (with a 3 channel LED driver) or a Neopixel addressable type.

[spinnaker]

I am more of an image type of person.   I am trying to figure our what you were describing.   Is it pretty much as I attached?


I am still trying to figure out how he gets so many pixels on the horizontal (or rings.)    And maybe you mentioned this but I am not getting it either.    Lets forget about the colors for now and think just one color to keep it simple.  Is it that a given hole can actually be illuminated at various places over a given compartment or sector as long as the illumination does not encroach    on a neighboring hole? 

Now that I reread your post above it seems like that is what you are saying.   

Also I think I am getting the two spiral thing now.  Is it two  compartments are being lite at the same time?  One on one side of the disk, the other on  the other side?
I would assume if I was willing to work with only half of the disk I will be fine with one spiral?





Thanks for all the help and being patient.   For some reason I am having trouble making this one sink in.