Updates to my mechanical set

Forum for discussion of narrow-bandwidth mechanical television

Moderators: Steve Anderson, Dave Moll, Andrew Davie

Updates to my mechanical set

Postby smeezekitty » Sat Mar 24, 2018 2:27 pm

I decided to update my old mechanical television that I first built when I was 12. Here is the other thread I made: viewtopic.php?f=1&t=2413

I made a new disc with the holes drilled more precisely with a smaller bit (1mm instead of ~2.5). It's not perfect, but it is certainly much better.
I scrapped the nail bearing and used a video head from an old broken VHS camcorder and also replaced the pulley on the disc side with a 3D printed one (I have a 3D printer now :D)

I also built a sync circuit based on an Atmega328P microcontroller. I have been running it at 15 FPS to reduce the flicker compared to that of 12.5 FPS. But I would like to get it even higher.

video_head.jpg
The video head with 3d printed parts (blue)

The cradle for the video head was 3d printed. As was a plate to make it flat so the disc can sit against it. I didn't disassemble the head and free the bearing and just used it as is and only removed the connections and the BLDC motor. The video head conveniently has two threaded holes which I use to attach the disc. It might be ideal to have 4 but I haven't had any issues so far.


sync_board.jpg
The sync board

The sync board isn't very pretty but it is functional.


sync_sep.jpg
Sync separation scope
sync_sep.jpg (179.1 KiB) Viewed 479 times

Sync separation looks good!


reed_switch.jpg
Reed switch
reed_switch.jpg (74.93 KiB) Viewed 479 times

Unlike most designed I see here, I used a reed switch next to the the disc and a tiny rare earth magnet glued to the disc to provide a frame sync pule (no line sync).
The reed switch provides a solid, sharp signal to the microcontroller. I had much better results out the reed switch than I have with infrared emitters and phototransistors in the past.

Unfortunately I don't have any pictures of it fully assembled right now with the motor, pully and rubber band (err...belt).

And natürlich results:

smiley.jpg

NBTV.jpg


Now the next biggest problem I can see is that it lacks DC restore so pure white (er...red) displays fade out while black ones glow. There is also vertical banding.

Anybody have any suggestions what I should do to improve this?
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby Harry Dalek » Sat Mar 24, 2018 8:44 pm

Great work at 12 ! i would been lucky to work out lighting a light globe at that age .\
I like the reed switch idea closest i have used is hall effect sensor ic and magnet yes most go down the optical road less mass to deal with .
Disk is pretty good for a home made too.
The electromagnetic spectrum has no theoretical limit at either end. If all the mass/energy in the Universe is considered a 'limit', then that would be the only real theoretical limit to the maximum frequency attainable.
User avatar
Harry Dalek
"Fester! Don't do that to 'Thing'"
 
Posts: 3696
Joined: Fri Sep 26, 2008 4:58 pm
Location: Australia

Re: Updates to my mechanical set

Postby Klaas Robers » Sun Mar 25, 2018 2:55 am

For DC-restore use the simple one transistor BUZ11 circuit from the NBTV Handbook. Although that circuit looks te be too simple, it works well and does DC-restore by a simple small diode. More over that circuit has inhaerent gamma correction, which gives a better rendering of grey scales. The pictures that you showed are only black and white, but you will soon go to natural pictures with shades of grey.

For the banding it is the disc that gives this. However I have seen it worse. Making an almost perfect disc is very difficult. I advice you to be proud on what you have done. Only a computer drilled or punched disc is slightly better.

I also can advice you to find yellow or amber LEDs for the light box. Our eyes are more sensitive to the more yellow light than to the red light.
User avatar
Klaas Robers
Frankenstein was my uncle.
 
Posts: 1367
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: Updates to my mechanical set

Postby smeezekitty » Sun Mar 25, 2018 5:04 am

Harry Dalek wrote:Great work at 12 ! i would been lucky to work out lighting a light globe at that age .\

To be fair, my father helped me cut the wood and stuff (my parents didn't like me working with the saw :lol: ) but I was a nerd for as long as I can remember :roll:

I like the reed switch idea closest i have used is hall effect sensor ic and magnet yes most go down the optical road less mass to deal with .
Disk is pretty good for a home made too.

The magnet is so tiny (3mm x 1.5mm), the mass isn't really an issue. It's easy to get a solid trigger on the digital input of the micro using a reed switch. Surprisingly, there isn't any noticeable bounce on the scope.


For DC-restore use the simple one transistor BUZ11 circuit from the NBTV Handbook. Although that circuit looks te be too simple, it works well and does DC-restore by a simple small diode. More over that circuit has inhaerent gamma correction, which gives a better rendering of grey scales. The pictures that you showed are only black and white, but you will soon go to natural pictures with shades of grey.

Unfortunately I don't have access to the handbook. I never really got around to join the NBTVA since I've been so on and off (and a broke college student :lol: )
And you're right, gray scale reproduction is a problem. It isn't even close to linear.

For the banding it is the disc that gives this. However I have seen it worse. Making an almost perfect disc is very difficult. I advice you to be proud on what you have done. Only a computer drilled or punched disc is slightly better.

Thanks! I know that the banding in the pictures is caused by disc defects (and also a rolling shutter effect from the camera). There is a different kind of banding that might be caused by lack of DC restore that is bugging me (dark objects get vertical light bands on them)

I also can advice you to find yellow or amber LEDs for the light box. Our eyes are more sensitive to the more yellow light than to the red light.

I'm just using the LEDs that I've had in it for a while. I need to rebuild the LED array to be brighter and probably a different color.
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby Klaas Robers » Sun Mar 25, 2018 8:42 pm

I will attach the Hand Book chapter about the single transistor LED-driver. The Handbook gives more interesting chapters on NBTV, however Vic Brown, who is the person normally "selling" these, suffers some medical problems at this moment.
Attachments
06 Single transistor LED driver.pdf
(45.61 KiB) Downloaded 26 times
User avatar
Klaas Robers
Frankenstein was my uncle.
 
Posts: 1367
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: Updates to my mechanical set

Postby smeezekitty » Mon Mar 26, 2018 1:36 pm

Awesome thanks! I'll try it as soon as I get ahold of a BUZ11
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby smeezekitty » Sun Apr 01, 2018 1:19 pm

While I am awaiting parts, I came up with a workaround for the lack of DC restore: I added a new feature to my NBTV generation software to make give the NBTV signal a carrier wave which is AM modulated. It makes the picture a little noiser and reduces the maximum brightness a bit but it fixes the DC restore problem by having even pure white have an AC component.

am_mod_nbtv.PNG
(5.21 KiB) Not downloaded yet


In case anybody is interested, I am attaching a schematic of a the sync seperation circuit. It is "super-diode" circuit based on an opamp and it has very few components.
syncsep.png
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby smeezekitty » Sun Apr 15, 2018 9:12 am

I built the LED driver that from Klaas and replaced the 6 LED array with one with 12 diodes (3 x strings of 4).

I made a couple of small modifications; I increased the LED resistors from 22 ohms to 150 ohms since I didn't have an 22 ohm resistors (hopefully that doesn't impact circuit performance too much).
Since my set doesn't have a +12V supply, I skipped the 15K resistor and tied the high side of the brightness pot to the +5V supply on the sync board.

Everything seems to work. There is a noticeable brightness increase, much better gray scale reproduction and DC restore action. The picture even seems to be sharper.

led_array.jpg

32_line_px.jpg
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby Andrew Davie » Sun Apr 15, 2018 5:55 pm

Trying to recall how this was all calculated, but given 4 LEDs in series, then about 2.2V/LED that would make 8.8V drop across the string. You don't have a 12V supply you say, but don't mention the voltage you do have. But going with 12V for my calculations, the resistor is "adjusting" 12V-8.8V = 3.2V and with a 150 Ohm resistor, 3.2/150 --> 0.021 (i.e, 20mA give or take) which sounds just about right for LED current. I don't see how 22 ohms would have worked at all. Hmmm. If the voltage was much lower it wouldn't be able to light a string of 4 LEDs. So, what IS your voltage?
User avatar
Andrew Davie
Frankenstein was my uncle.
 
Posts: 1407
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Re: Updates to my mechanical set

Postby smeezekitty » Mon Apr 16, 2018 2:57 am

Andrew Davie wrote:Trying to recall how this was all calculated, but given 4 LEDs in series, then about 2.2V/LED that would make 8.8V drop across the string. You don't have a 12V supply you say, but don't mention the voltage you do have. But going with 12V for my calculations, the resistor is "adjusting" 12V-8.8V = 3.2V and with a 150 Ohm resistor, 3.2/150 --> 0.021 (i.e, 20mA give or take) which sounds just about right for LED current. I don't see how 22 ohms would have worked at all. Hmmm. If the voltage was much lower it wouldn't be able to light a string of 4 LEDs. So, what IS your voltage?

It's +~19V. The circuit schematic in question calls for 20V
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby Klaas Robers » Mon Apr 16, 2018 4:22 am

I think 150 ohm will do.
- A current of 20 mA through 150 ohm is 3 volt.
- 4 LEDs in series is indeed 8.2 volt, say 9 volt.
- Then LEDs and resistor = 12 volt.
- from your 19 volt remains 7 volt for over the transistor. That seems to be enough.
The resistors are only to distribute the current of the transistor evenly over the 3 strings.

Your observation of a better grey scale rendering is the effect of the right gamma of the circuit. Gamma is a non linearity in the light output that compensates for the non linear sensitivity of our eyes.
User avatar
Klaas Robers
Frankenstein was my uncle.
 
Posts: 1367
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Re: Updates to my mechanical set

Postby smeezekitty » Mon Apr 16, 2018 5:14 am

Klaas Robers wrote:I think 150 ohm will do.
- A current of 20 mA through 150 ohm is 3 volt.
- 4 LEDs in series is indeed 8.2 volt, say 9 volt.
- Then LEDs and resistor = 12 volt.
- from your 19 volt remains 7 volt for over the transistor. That seems to be enough.
The resistors are only to distribute the current of the transistor evenly over the 3 strings.

I turned the brightness pot until I got 60mA of LED current so (3 x 20mA) and I measured 10.3V across the array and ~8.8 across the transistor.
They're red LEDs so 1.8V / diode seems about right.
Your observation of a better grey scale rendering is the effect of the right gamma of the circuit. Gamma is a non linearity in the light output that compensates for the non linear sensitivity of our eyes.

I figured it was something like that. Thanks again!
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby Klaas Robers » Thu Apr 19, 2018 6:09 pm

The non-linearity of our eyes is because we are used to look at objects and photos under different lighting conditions. In bright light the difference in brightness between white and mid-grey is much more than in dim light. But you want to have the same subjective difference. That is why our eyes are more sensitive to a certain grey-step in the dark area than in the bright area. You will see the same subjective grey-step from brightness 3 to 4 as from 60 to 80.

So when digitizing video signals you need small stepsizes in the low area, 3 and 4, say steps of 0.1, while for 60 to 80 the stepsize could be steps of 2. The gamma correction in the BUZ11 produces the increasing stepsizes. The camera does the inverse of course. In TV this is already common practice from the very beginning. However I am not sure that Baird was aware of this effect already in 1929. But he had no problems wit digitallization.

But added noise introduced by a radio link follows the same rules. The same amount of brightness noise added to low brightness areas of your picture is much more visible than the same amount of brightness noise near white. The gamma correction egalises this discrepancy.
User avatar
Klaas Robers
Frankenstein was my uncle.
 
Posts: 1367
Joined: Wed Jan 24, 2007 8:42 pm
Location: Valkenswaard, the Netherlands

Color!

Postby smeezekitty » Thu May 10, 2018 4:04 pm

I found a stash of green LEDs so I decided to add them to the existing LED array. I constructed a second driver circuit and now I have red/green color.
It looks a lot better than I expected it would. The down side is now I am using both channels for video and have no sound. I'm trying to figure out a way around that.

color1.jpg


color2.jpg

Believe it or not there are NO blue LEDs. The blue is caused by the auto white balance on the camera correcting for the lack of blue. It's amazing how well it reproduced all three colors though. But even to the naked eye, blue is mapped to green and things you expect to be blue actually look blue because of brain tricks.

color3.jpg

Skin tones are rather normal with a red/green color system

MTV_Fire.wmv
(1.4 MiB) Downloaded 5 times
smeezekitty
Misunderstood Genius
 
Posts: 89
Joined: Fri Jan 29, 2010 11:42 am
Location: USA

Re: Updates to my mechanical set

Postby Andrew Davie » Thu May 10, 2018 4:29 pm

Do I see top cat?
User avatar
Andrew Davie
Frankenstein was my uncle.
 
Posts: 1407
Joined: Wed Jan 24, 2007 4:42 pm
Location: Hobart, Tasmania, Australia

Next

Return to Mechanical NBTV

Who is online

Users browsing this forum: No registered users and 3 guests