Making a more colorful oscilloscope // Modified CRT Television, DIY XY scope with video input
ฝัง
- เผยแพร่เมื่อ 14 พ.ค. 2024
- I've seen tons of CRT TVs turned into basic XY scopes in the past, but after making this one the other week I spent a bit of time experimenting with video signals as well and wanted to make a video about what I learned since I hadn't seen this sort of thing covered before. You can do some pretty cool stuff with this thing!
I ended up going a sort-of-educational route with this video, but it also assumes a lot of knowledge and I don't go crazy in depth into anything. The unique thing here was learning how the video signal was being interpreted after the modification, so I mostly tried just going into that.
More info & schematic: www.gleix.net/crtscope
All music in this video is my own and comes from a series of improvisational live performances.
www.gleix.net
/ gleix
0:00 - Intro
1:02 - Basic info/demo
2:16 - Adding color
3:16 - Waveform coloring explained, basic + horizontal lines
4:55 - Vertical lines
5:28 - Dialing in some looks
9:10 - Arrangement
10:18 - Wrap-up
#diyelectronics #oscilloscopemusic #audiovisualart
The main thing this did was to shatter what I thought I knew about Color CRT TV :P
I am also curious about the aliasing artifacts. It seems to be unable to make full curved lines even when the input color is white. I mean, curved lines seems to always be segmented into lines. Is this because of the Shadow mask on the screen?
If you look into what a composite video signal is comprised of & how an image is built on a CRT TV, you'll notice that the full signal consists of some additional data that isn't just the image itself! The waveform appears segmented into lines because the raster image itself is actually segmented into scanlines.
Things like the blanking intervals, sync signals, color bursts, etc... which are outside of the visible image we see on screen, manifest themselves in this modified display as empty gaps as there's no image data there to interpret. You can think of each dotted line here as the active image portion of a single line in the raster scan (there ought to be 525 in this instance, I think), and each gap as the section which contains the horizontal sync and color burst data, which shows up at the beginning of each line in the scan.
The larger visible gap between the start and end of the waveform, like you see at sections like 2:16, represents the section of the signal that contains things like the vertical blanking interval, which would only appear at the very beginning of the raster scan.
Sending a colour gradient to the TV was genius, absolutely incredible!
Sometimes I get sad about old technology going away but I am also thankful im born at the right time to experiment with old technology before it's gone, while also having access to information and tools of the current era
Yeah, our access to used equipment for affordable prices has never been better. Equipment costing thousands in the past is accessible for $50 today.
Dude, you just bring "oscilloscope"' to a completely new level! And it's pretty nice, too! 👍
how does this only have 922 views wtf? a video this cool and with really good production like this should have at least 100k views.
this is so sick I can't wait to try this sometime, thank you
Commenting here because the RGB registration/purity is so nice that beautiful colors can be implemented.
Constructive comments:
Screen burn in the center could be avoided by adding a circuit to detect lack of deflection signals. These are used in commercial video game displays such as in ‘Asteroids’ and other great games. The beam is cut off after a fraction of a second. The delay covers the occasional legitimate times when the deflection voltages are momentarily at zero.
The blanking is showing. There may be no simple way around that if using composite video input. Better to disable the blanking and use separate analog non-blanked RGB inputs direct to the video amplifiers on the CRT socket board. Since blanking is part of most video recording, a different approach may be required such as multichannel continuous signal method akin to a high bandwidth analog audio stream.
Be sure to use a closely substituted and non-saturating inductance connected to the H sweep circuit for the horizontal deflection coil when leaving the deflection coil and CRT convergence and purity magnets on the CRT neck unmolested. The inductor is required because the H yoke winding for which it is subbed is part of a tuned circuit required for correct operation of the high voltage power supply. It’s a reliability and safety issue. Leaving the original yoke etc unmolested is the best way to get good color purity as displayed in the video. Mouser and others have chokes for substitution. Again, the substitute choke is a safety and reliability issue.
The LM358 is very low power. I used 20-30W tube amplifiers when I did this because those are hard to destroy. Not so with solid state amplifiers facing the transients of inductive kicks.
The deflection yoke is highly reactive and far from a 4 or 8 Ohm load, so choose and oversize the amplifiers prudently and use those with fast rectifiers to clamp transients to the power supply rails. Or else.
To produce accurate waveforms, the H and V deflection power amplifiers must have current feedback from the yoke current so that delta Vin = delta Iout. It's the yoke current that positions the beam, not the voltage. The amplifier must slew faster than the yoke for this to have meaning. DC coupling is a plus.
The speed at which the yoke can traverse the beam with a step voltage is inversely dependent on its inductance. Therefore, the highest voltage output stage should be used to give the highest frequency. The yoke current to hold the beam at the edge of a decent sized screen is 2 to 4 amperes, usually. A 30Khz (60KHz -3dB) screen-wide deflection response may require a peak amplifier output voltage of 100-300V from the amp, so don't count on a full audio-range accurate scope without a box of amps next to it. Don't be discouraged though because the treble in audio never goes full screen in a music waveform. In graphics, YMMV.
I mainly alert experimenters to the sluggishness of electromagnetic deflection and the challenges to overcome according to the unsuitability of the stock TV set yoke, which has to be kept to assure proper convergence for color registration. In special and military vector displays, the yoke is quite different, with very low inductance (fewer turns) of larger wire for low resistance.
Here are some links to monochrome vector display manuals and one RGB unit showing or explaining current feedback power amps for deflection, and ‘spot killers’ or beam blankers. The documented working examples are scalable. These products are discontinued and not supported. Just like your TV set which I hope will get a new life.
bunkerofdoom.com/lit/G05-801.pdf
bunkerofdoom.com/lit/Texscan_DU-127_XY_display.pdf
color vector display: www.jmargolin.com/xy/xymon.pdf
Cool! Where can i learn more about this off circuit, id like it implement it.
@@EosKarlsson page 14 of this manual: bunkerofdoom.com/lit/G05-801.pdf
Also search for x-y displays, vector display, vector monitor, CRT spot killer, vector video game, to find manuals for more displays having variations on the "spot Killer" circuit. These other links MAY help so look at them all too. It's a niche circuit so there will be digging to be done, but all professional arcade game vector monitors from the olden days used the circuit. Vector display Manufacturers were Wells-Gardner, Electrohome, and others.
www.jmargolin.com/xy/xymon.htm (best, experienced tech)
www.retrorgb.com/vector-monitor-xy-kit.html
hackaday.io/project/2871-build-an-arcade-xy-vector-monitor
good luck!
Not only you can visualize each horizontal scanline, but also the h-blanking interval within each gap of the dotted lines
Sooo cool! Great job
I've been happy playing with my B&W TV for years.... but now I've GOT TO get a colour!
This is amazing!
This is awesome! It verifies a cool I dea I have with simple laser pointer projected pho- color TV, with rotating mirror's and pwm flashing the lasers
Omg thank you. This video is good.
This is flippin AAAWWEEESSSSOMMMMME!!!!!!
I have no intention of hijacking the OP's video or comments. If it's not appropriate, feel free to delete. It's just the result of drinking too much coffee.
So, Others have commented about drawing circles (presumably using two equal sine waves 90 degrees out of phase) which turned out to be non-circular when displayed. OK, the goal is to have an accurate display. The overlooked technique is that the deflection amplifier must use current feedback, not voltage feedback. In electromagnetic deflection (using a yoke) the amplifier's output Current must linearly follow the amplifier's input voltage. This is the opposite of the idea where, for example, a hi-fi amplifier's output Voltage must linearly follow the amplifier's input voltage.
Amplifiers use feedback to make them linear. To convert a voltage amplifier (common audio amp, etc) to a current amplifier, one removes the voltage feedback loop (primarily a high value resistor connected between at the amp output to deliver a negative feedback voltage to an early voltage amplifying stage such as a differential stage, and then one places a low-value non-inductive resistor between the 'cold' (circuit ground) side of the yoke winding and its return (ground). At the junction between the cold end of the yoke and the new resistor, a voltage will appear that is a linear copy of the yoke current. Therefore, the yoke current waveform can now be used as the negative feedback voltage, by feeding it to the same early stage in the amplifier.
I can't be more specific because there are thousands of audio amplifier designs and there are no such step by step instructions on how to do this, or how to control and select colors, simply because there are thousands of different models of TV set out there and each will absolutely require careful experiment, measurement, and some calculations. The links I provided are a good start for deflection amp feedback ideas.
About the vertical yoke, it is slow and low-frequency because it is designed to run at 60Hz and has a lot of inductance. Personally I still say leave the original yoke unmolested, substitute the suitable inductor for the H winding, and use an amplifier capable of higher voltage to drive the V yoke at the higher frequencies.
The first topic to understand is the inductive reactance of the yoke coil you want to drive (whether H or V sections), and its effect on the current through the yoke with a given sine wave voltage that varies in frequency, because the goal is to get equal deflection over a range of audio frequencies. At some point it becomes impractical but still very fine effects can be had out to a few Khz. Some electronics books have the formulas, and there are online calculators and charts to visualize the concepts.
A very gross example (and do not rely on the resitances or figures given here): A hypothetical vertical yoke winding has 1 Ohm resistance. With DC (0 Hz), it may take 3 Volts from the amplifier to push 3 Amps through the winding and fully deflect the spot to the top of the screen (and negative 3V to do the same to the bottom of the screen). At 60Hz, it may take 20V peak (+ and -20V) to fully deflect the spot, but of course being AC now it will be a line stretching from top to bottom of the screen and the same 3 Amps, though those will be the (+) and (-) peak amps. At 200Hz, it may take 200V peak to get 3 Amp peaks (same full V deflection), and so forth increasing as frequency is increased.
So from the above paragraph, the yoke coil in question has 1 Ohm resistance (some call 'dc resistance' in conversation). It can be considered that this resistance is effectively in series with the yoke's inductance. The indictance has a specific reactance (X) at any given frequency (hence "XL" inductive reactance @ 60Hz). If it takes 3A to deflect the spot fully, and we know that the DC resistance of the yoke is 1 Ohm, and it takes 20V to deflect the spot at 60Hz, then we can tell or measure that the yoke will have 17 peak volts across it (20V - 3V = 17V). What is the inductive reactance of the yoke coil at 60hz? It is 17V / 3A = 5.67 Ohms.
Ok, and what is the inductance?
- A chart online gives about 15 millihenrys (15mH).
or do the math:
- The formula for inductive reactance is:
XL = 2 * pi * F * L.
5.67 = 2 * 3.1416 * 60 * L (XL=inductive reactance and L=inductance).
5.67 = 376.992 * L
5.67 / 376.992 = L
5.67 / 376.992 = 0.015040 Henries. = 15mH.
either way is close enough for fun.
--------------------
- The H and V windings have very different inductances. Why? because it was most cheap and techncially simple to make the TV that way. So do this for the H winding and the V winding.
- Most all yokes actually have two H windings and two V windings. This is because it makes a more uniform magnetic field in the CRT. The windings are often in series. In some cases, it is easier to drive a yoke having high inductance by putting the windings in parallel. Inductors in parallel act mostly like resistors in parallel. The value is about half, depending on their magnetic coupling which I don't want to get into because yokes are different in this respect. Grossly though, the XL (inductive reactance) is cut in half. (as well as the inductance and the resistance).
- There is a term 'impedance' we all know from speakers and radio equipment. The impedance is the sum of the resistance and reactance. Impedance is expressed as "Z".
-------------------
I hope this is not overbearing and will encourage people to experiment with this fun kind of project!
If there are mistakes (other than silly grammar, spelling, etc), someone please point them out with the correct explanation.
Excellent TH-cam, cheers!
I like to learn the oscillating ways of the Gleix
most Glexelent!
That’s real neat. One “issue” is that you’re bound to the H-sync and V-sync periods of the television still, giving you large and small black areas respectively. I’d want to control the brightness of the electron gun with dedicated circuitry, instead of using the video input. I’d also put some sort of VCA on them to dim the beam at lower X/Y amplitudes, as not to burn a dot in the middle. Ideally a constant brightness per unit length would be used, but I think it would have to be approximated without a microcontroller, or including the brightness signal alongside the X and Y signals as a pre-processed time varying value.
One pleasant surprise about putting this video out is that people much more knowledgeable than I have shared their thoughts on features that can solve the main issues that this basic implementation has, with the biggest being that there's no way to avoid a hot beam in the center. A microcontroller based solution that can dim the output of the electron gun when the audio signal is low is a fantastic idea that might be the easiest to implement; especially on CRTs with an accessible brightness & sub-brightness trim pot. Solving the non-uniform deflection amount of the X and Y axis when fed the same signal is the other priority, and would require some research I've not done :)
I've actually run into the sort-of inverse of the hot-beam problem after attempting the modification on a few newer CRTs; some newer ones have deflection protection built-in, so that when the amplitude of the deflection signal it's expecting isn't received the gun refuses to fire (or fires very dimly), or refuses to deflect properly. Locating and bypassing (or modifying) this built-in protection seems like much more trouble than it's worth, and doesn't solve any hot-beam issues though.
I want to implement this.. where can I find more info on this?
The most challenging thing associated with this is NOT burning out the center of the CRT.
LOL
There is a Brazilian channel where the owner made one of those with double channel! his name is Rodrigo Negrelli
Nice video. Is there a tutorial linked to show you how to add the colour to the oscilloscope and do this step by step?
I'd also really appreciate a link to some sort of tutorial. This one seems quite good but doesn't incorporate new new circuit and pots - th-cam.com/video/UozU04smXlQ/w-d-xo.html
Thanks! I'm just using the stock composite video input on the TV, so nothing special there. If you were to build a scope out of a TV that has a composite input jack, like I did here, you could do the same thing - the only trick is making the graphics to send to the TV. At the moment I'm not planning on any tutorials on how to make or animate graphics, that's a whole other story :p In this video I was creating the colors & visual effects in Touchdesigner, but you could use anything you're comfortable with, like After Effects and whatnot. What I really wanted to demonstrate here was how it works so that if you were to make your own scope, you could then make your own graphics and know where to start and how it works :)
@@gleix_ Thank you for your comment! I was also wondering, if you could make a CRT Oscilloscope in your style/way including adding the stereo left/right audio into a mono tv unit? Your video is awesome and it might be too much to ask for but I really enjoyed your videos as of recently ever since lockdown. Thank you so much for making the unique and quality content that you do!
@@MyConfessionToMadness Do you mean adding the ability to send stereo audio directly to a TV with a single mono speaker? That should be as easy as summing both signals together / sending them both to the one speaker. If you were using the TV as a scope with separated audio channels for X & Y though you'd need to figure out a way to isolate those channels so the X & Y on the yoke aren't also getting the same signal. Appreciate the kind words!
sorry to ask but, how do you keep sending video to he yoke after tu've cut the 4 cables connected to the crt pcb? do you atually not cut them an just send video through them at the same time you send the audio signal to each magnet?
How would it look if you were to put it through an oscilloscope art making software such as osci-render or OsciStudio?
I really dont get where you can send video if the x an y are cut off and hooked to an audio source
Hey great do you get fast enough vertical deflection on your CRT? Also which CRT screen did you use in this video?
This was made with a Toshiba MV13L3 - I got it for free due to it having vertical collapse. I recapped it, resoldering just about every joint, and replaced the IC that handles vertical deflection but none of that fixed the issue unfortunately. All this to say, it may be a unique issue to this specific one I've got since it already had some wonky issues with deflection, but also in general the horizontal and vertical deflection don't typically act identically in CRTs from my experience - trying to make a circle will typically end up giving you a weird oval, or some other strange shape depending on the gain of your audio. Maybe something to do with the fact that the full scan is 4x3, not a perfect square.
@@gleix_ hey thanks it does loo very good. I just wonder a part the rescaling of the XY that could be far from perfect, what about the frequency, can you go as high on vertical as on horizontal? Or you have to substitute the coils?
Hello! Im trying to make your wave forms using your audio directly to the XY of my modded tv, but I can't repro the same form, Its more like a circle on my screen, What do you think I am doing wrong?
Make circuit of oscilloscope that run on video input composite of TV with out open any tv or any other device
I so wish I could make this as its would help with my disability im sure
are u using rga cable to play video?
Just curious how you set it up to send color to it
Saw the note get it now
What is the rating of the potentiometers you used?
I've got a schematic @ gleix.net/crtscope - I believe I used 100k, but any pot that lets you attenuate audio before going through the amp would work.
Can anyone tell me how he routed the video signal?!? How tf does this work 😅
Video still comes in via the stock composite video input :) It's just processed differently after the mod.
@@gleix_
Sorry for the noob questions but this is fascinating!
so you don’t have to mess with anything else, just re wire the coils - then plug a video source into Theo standard video input?
Could you hypothetically run a vcr whilst feeding an audio signal to make the waveforms the colours of each movie scene?
Also, what parameter controls the colours produced on a colour screen crt? Varying Voltage of L&R channels respectively controls the pattern, gain of incoming signal determines the size/amplitude of the waveform on screen, so what changes the colour? What happens to the electrons fired by the cathode that produce colour on the screen?
Thanks for your help and your video!
now play Jerobeam Fenderson's Oscilloscope Music into it