Increase Sharpness Using Optimal CRT Brightness Calibration

The place for all discussion on gaming hardware
Post Reply
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Increase Sharpness Using Optimal CRT Brightness Calibration

Post by LukeEvansSimon »

TLDR: This guide explains how to improve sharpness through optimal brightness calibration. Optimal brightness calibration (without using any hardware modifications) maximizes cathode bias voltage, minimizes G1 bias voltage, and maximizes G2 bias voltage. The limiting factors will be the maximum voltage your CRT's RGB drive amps can output and the maximum screen voltage your flyback can output.

CRTs provide 3 to 4 different mechanisms for calibrating brightness, and apparently many people are not aware of how they differ. All brightness settings are used to manipulate the CRT's triode. If you do not understand basic triode theory, then you won't understand the ideal way to tune brightness. This post outlines the 4 mechanisms for tuning CRT brightness, uses basic triode theory to explain how they are different, and describes how to tune them in order to have proper black levels and the smallest possible cathode ray spot size (for thick blank scanlines and a sharp image). Note that the ideal way to calibrate a CRT is using a colorimeter, but this article will describe how to get a good calibration without a colorimeter, as long as you have access to 240p Test Suite and a very dark room. This post will refer to a single test pattern in 240p Test Suite called "Color Bars" (screenshot below):
Image

Definition of "Brightness"
The term "brightness" refers to the lowest possible light emission levels of a display. This is sometimes called the black levels of the display. When brightness is too low, the darker colors in the 240p Test Suite color bars: black, dark grey, dark blues, dark reds, dark greens, etc, will all "crush" or "clip" into black. When brightness is too high, the darker colors in the color spectrum do not get dark enough. Black looks grey, and dark grey looks like light grey, and similarly for dark green looking more like light green, etc. You can think of "brightness" as controlling the separation between the darker colors in the color bars. However, this is not entirely accurate because brightness increases and decreases the light emission of every color, not just the darker colors. Brightness additively shifts up or down, the light emission of all colors, even the colors on the rightside of the color bars in the test pattern.

Image

Mechanisms for Adjusting CRT Brightness
  1. "Brightness" setting in the onscreen display (or service menu): If your CRT does not have an OSD, then this is available as pots. This controls the bias voltage of the CRT's cathodes "K". Color CRTs have 3 cathodes: RK, GK, BK. As this brightness setting is decreased, it means the bias voltage of the CRT cathodes is increased. The higher the voltage of the CRT cathodes, the less cathode ray emissions are released from the cathodes because the electrons that form the cathode ray are attracted to the higher positive voltage of the cathodes. As this brightness setting is increased, the bias voltage of the CRT cathodes is decreased, allowing more electrons to be emitted from the cathodes, making the cathode ray have a higher emission. More electrons emitted into the cathode ray means more light emission from the CRT.
  2. Red, Green, and Blue Cutoff (typically in the service menu): If your CRT does not have an OSD, then this is available as pots. These 3 cutoff settings do the same thing as the "brightness" setting mentioned above, except for a single cathode at a time. That is, "red cutoff" adjusts the bias voltage of the red cathode RK. Similarly for blue and green cutoffs. Due to manufacturing differences and other physical phenomenon, each cathode needs a slightly different bias voltage for proper calibration.
  3. Screen pot on the flyback: This controls the bias voltage of the CRT's G2 anode, which is used for accelerating the electrons away from the cathodes and towards the screen. Turning the screen pot clockwise increases the bias voltage of the G2 anode, which causes more electrons to be pulled away from the cathodes and shot towards the screen. Turning the screen pot counterclockwise decreases the bias voltage of the G2 anode, which causes less electrons to be pulled away from the cathodes. Less electrons emitted into the cathode ray means less light emission from the CRT.
  4. Pot for G1 bias voltage: Most CRTs do not have this mechanism, and instead attach G1 to ground. This is typically only found in higher-end CRTs and higher resolution CRTs. CRTs that do have this functionality send an amplified horizontal sync signal to the G1 anode. The G1 anode is a metal tube that surrounds the cathodes and has a small hole at the exiting end of the G1 anode that the cathode ray must pass through, and the G1 anode has a voltage that is negative relative to the cathodes. So the lower the bias voltage for G1, the more electrons emitted from the cathodes into the cathode ray are repelled from the G1 anode and squeezed into a thinner and thinner cathode ray, which is being sucked through the G1 anode's exit hole by the positively charged G2 anode. Since the horizontal sync signal has its lowest voltage during horizontal sync pulse, the cathode ray has the least emission of electrons through the exit hole of the G1 anode during horizontal sync, which lasts for the duration of the horizontal retrace time period (when the cathode ray is being reset to the start of the next line). This thread describes a mod that adds a control for G1 bias voltage.
Image

Calibrating Brightness
While all of these controls allow for changing brightness, they do not have the same result on the image. As explained in this thread, cathode - G1 voltage acts as an electromagnetic iris that the cathode ray must pass through. Increasing cathode - G1 voltage shrinks the diameter of the aperture of the iris, which more tightly packs the electrons into a thinner, more concentrated cathode ray, which creates a smaller illuminated spot size on the screen. However, this also changes the brightness, so G2 - G1 voltage needs to be increased to calibrate brightness. In addition, a higher G2 - G1 voltage also serves to increase the acceleration of the electrons in the cathode ray, which further improves spot size because the electrons have less time to spread out. A much more detailed explanation is in the next spoiler section. This is the core theory of how to calibrate brightness to have proper black levels, as well as the thinnest, most concentrated, sharpest cathode ray possible without modifying the CRT's chassis. In summary, the steps are as follows:

Code: Select all

Repeat the following steps until right before retrace lines become visible:

1. Decrease the brightness setting in the on screen display menus until column 0, the leftmost block that is the darkest shade of the white bar in the 240p Test Suite color bars test pattern becomes so dark that it is the same "true black" as the black background.

2. Increase the flyback screen voltage pot until column 0, the leftmost block that is the darkest shade of the white bar in the 240p Test Suite color bars test pattern becomes visible again and stop increasing right before the background starts to transition from true black to very dark grey.
At this point, you should have significantly increased the flyback screen voltage and significantly decreased the brightness setting in your service menu or user menu of your CRT (or pot if no OSD). Also, you are maxing out the capability of the RGB drive amplifiers on your CRT's neckboard. Typical RGB drive amps output a maximum of around 200 volts. You can't push your RGB drive amps any higher without clipping the chroma signal, which causes horizontal retrace lines to appear (see "Retrace Lines" section below). If your CRT allows for controlling G1 bias, you can squeeze the cathode ray to an even more concentrated, thinner beam by taking the following steps:

Code: Select all

Repeat the following steps until right before retrace lines become visible:

1. Decrease the G1 bias voltage until column 0, the leftmost block that is the darkest shade of the white bar in the 240p Test Suite color bars test pattern becomes so dark that it is the same "true black" as the black background.

2. Increase the flyback screen voltage pot until column 0, the leftmost block that is the darkest shade of the white bar in the 240p Test Suite color bars test pattern becomes visible again and stop increasing right before the background starts to transition from true black to very dark grey.
At this point, if your CRT allowed for adjusting G1 bias, you are really pushing the limits of your screen voltage potentiometer and you have maxed out the amplifier for the horizontal sync wave form. Pushing the amp any further will cause the horizontal sync wave form to clip and retrace lines will be visible (see "Retrace Lines" section below).

Red, Green, and Blue Cutoffs
Now it is likely that the Red, Green, and Blue color bars in the 240p Test Suite are not aligned. That is, the colors don't transition to black at the same leftmost shade of their respective color bar. One of the color bars should properly transition to black only in the background and the leftmost shade of the color should be barely different than black. For this tutorial, I will assume the color is Blue, but it doesn't matter if your CRT is different. Just substitute your color that properly transitions to black only in the background. Now we need to increase the cutoff value of the other two colors until the leftmost darkest shade of their color bar is barely distinguishable from the black background.


Retrace lines
The higher the cathode - G1 voltage, the blacker the color drawn on the screen. During horizontal retrace, the cathode - G1 voltage must be even higher than the the voltage differential used for the color black, otherwise horizontal retrace lines become visible, as the yoke's electromagnetic field causes the cathode ray to reset to the start of the next line. In the calibration steps above, if you increase cathode bias voltage too high, the chroma signal for the cathode starts to clip and the "blacker than black" cathode - G1 voltage is not possible because the color black is already using the maximum obtainable cathode voltage.
Similarly, if you decrease the G1 bias voltage too far, then the horizontal sync wave form clips and is not able to decrease cathode - G1 voltage so that it is "blacker than black". Repeat after me: horizontal retrace needs a significantly higher cathode - G1 voltage than what is used for the color black. CRTs that allow for controlling G1 bias accomplish this by amplifying the horizontal sync waveform and sending it to the G1 anode. CRTs that attach G1 to ground accomplish "blacker than black" by reserving a voltage range on the chroma signal that is a higher range than the voltage used for black.
R=Horizontal retrace lines caused by the chroma signal clipping when cathode bias voltage is too high R=[url=https://ibb.co/5vrdj83][img]https://i.ibb.co/xG1ygjb/retrace-lines.jpg[/img][/url]
Last edited by LukeEvansSimon on Sun Jan 24, 2021 11:59 pm, edited 6 times in total.
User avatar
treminaor
Posts: 16
Joined: Thu Sep 10, 2020 3:05 am
Location: San Francisco, CA
Contact:

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by treminaor »

Great info! Appreciate the writeup.
User avatar
ASDR
Posts: 844
Joined: Sat Aug 12, 2017 3:43 pm
Location: Europistan

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by ASDR »

From what I understand, a normal TV will either have elevated or crushed blacks. Unless there's some gamma circuitry somewhere in there you'll have to make a trade-off somewhere and can't adjust things so that black is black and near black is visible.

I really appreciate this post. Between brightness, sub-brightness, screen pot and G2 adjust in the service menu there are way too many and confusing ways of applying some kind of offset.
Taiyaki
Posts: 1050
Joined: Fri Apr 04, 2014 11:31 pm

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by Taiyaki »

With crt's it's so confusing as brightness and contrast are often labeled in reverse, as brightness controls black level and contrast the intensity or actual brightness output. I think on modern tv's the contrast tab is generally called picture, brightness, or luminosity. On most Sony CRT's brightness tab definitely adjusts what we know now a days as contrast, and EDIT:does not adjusts the global brightness of the picture. I've seen so many crt technicians have to explain this in detail but no one ever seemed to have explained why this was done this way.
Last edited by Taiyaki on Tue Oct 20, 2020 2:22 pm, edited 1 time in total.
User avatar
ASDR
Posts: 844
Joined: Sat Aug 12, 2017 3:43 pm
Location: Europistan

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by ASDR »

ross wrote:
ASDR wrote:From what I understand, a normal TV will either have elevated or crushed blacks. Unless there's some gamma circuitry somewhere in there you'll have to make a trade-off somewhere and can't adjust things so that black is black and near black is visible.

I really appreciate this post. Between brightness, sub-brightness, screen pot and G2 adjust in the service menu there are way too many and confusing ways of applying some kind of offset.
That's what I've found on standard CRT TVs and my PVM.

It's funny, NTSC used to get a bad rep and was perceived as outdated for setting black level at 7.5 IRE, but this unintentionally allowed for the setting of a deeper black level while retaining all intended shadow detail.

Unlike most consoles, the NTSC-U/C PlayStation has 7.5 IRE setup on its NTSC output, which makes S-Video worth considering over R'G'B' if you're playing on a CRT.
On my Sony X1 there's a gamma setting in the service menu and it doesn't have this black crush issue when setting black to the lowest level. Strange that later Sony TVs don't all seem to have this circuitry / feature.

This 7.5IRE raised black pedestal thing only applies to Composite/S-Video? And JPN consoles don't have it, PAL doesn't have it and even most NTSC-M / US consoles, except PS1, don't have it? I was thinking of getting some S-Video cables for my NTSC-M PS1 anyway, might we worth a try.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by LukeEvansSimon »

Taiyaki wrote:With crt's it's so confusing as brightness and contrast are often labeled in reverse, as brightness controls black level and contrast the intensity or actual brightness output. I think on modern tv's the contrast tab is generally called picture, brightness, or luminosity. On most Sony CRT's brightness tab definitely adjusts what we know now a days as contrast, and adjusts the global brightness of the picture. I've seen so many crt technicians have to explain this in detail but no one ever seemed to have explained why this was done this way.
I wish the service menus just explained which CRT electrode was being impacted (cathode, G1, or G2), and whether the setting was the bias voltage or the gain. “Brightness” typically refers to cathode bias voltage and “contrast” refers to cathode chroma signal gain.

The import thing is to maximize the cathode bias voltage because that also has the effect of concentration of the cathode ray into a tighter beam. It also gives a wider voltage range for tuning contrast without getting crushed or clipped whites
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by LukeEvansSimon »

ross wrote:Assuming white column 0 on that colour bar pattern corresponds to 8-bit level 16 or 6.27 IRE, I'm not sure it's the best pattern for calibrating black level or "brightness".

Every CRT I've come across has a very similar EOTF (at similar contrast/brightness settings), swallowing up very low IREs into black, at least up to 3.5 IRE and possibly up to 6 or 7 IRE, when calibrated for a true black. Calibrating with that colour bar pattern won't show that, if those assumed values are correct, but the lost shadow detail and increased colour saturation is definitely noticeable in practical use.

Interestingly, the 240p test suite wiki page says to calibrate to column 1, which would be level 32 or 12.55 IRE (if column 0 is meant to be visible as shown in the screenshot).
A colorimeter is ideal for calibration. The main point for writing this post was to explain that maximizing cathode bias voltage and maximizing screen voltage is the ideal way to tune brightness as opposed to guides on the internet that claim that “brightness” or “screen” should be set to the middle value and then tuned. That leaves RGB drive amplifiers under utilized, and it leaves the flyback’s screen winding underutilized. My post lets you squeeze out a slightly smaller spot size by maxing out the capability of these components. A thinner cathode ray allows for a better depth of field (focus anode is more effective), and the higher acceleration anode (screen) voltage leaves less time for the cathode ray to spread out before it hits the screen.
daskrabs
Posts: 257
Joined: Sat Oct 01, 2011 1:05 pm
Location: Philly Area

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by daskrabs »

What does it mean if a retrace line in only one color (red, green, or blue) is visible on a black screen?
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by LukeEvansSimon »

daskrabs wrote:What does it mean if a retrace line in only one color (red, green, or blue) is visible on a black screen?
It means the amplified signal for controlling the output of that cathode is “clipping”. The color signal is inverted by the RGB drive amps, so darker colors have a high voltage and bright colors have a low voltage. Because electrons are attracted to a high voltage, the darker colors are made by charging the cathode with a high voltage when it should draw nothing (true black) for that color. Each cathode will have a slighthly different voltage that it needs to hit true black. Then, if the CRT chassis design ties G1 to ground (most do), then the cathode needs to use an even higher voltage than black during horizontal retrace because the magnetic fluctuation during horizontal retrace can make some of the cathode ray leak past the G1 anode. This voltage that is even higher than “true black” is called “blacker than black”. The RGB drive amplifiers have a maximum voltage that they can output. Decreasing brightness actually shifts the voltage of the amplified waveform up! Yes this is confusing, but remember cathode voltages are inverted form of the raw RGB signals. Blacker than black voltage is the peaks of the waveform, and if they get clipped, then cathodes leak during retrace.
Image
daskrabs
Posts: 257
Joined: Sat Oct 01, 2011 1:05 pm
Location: Philly Area

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by daskrabs »

LukeEvansSimon wrote:
daskrabs wrote:What does it mean if a retrace line in only one color (red, green, or blue) is visible on a black screen?
It means the amplified signal for controlling the output of that cathode is “clipping”. The color signal is inverted by the RGB drive amps, so darker colors have a high voltage and bright colors have a low voltage. Because electrons are attracted to a high voltage, the darker colors are made by charging the cathode with a high voltage when it should draw nothing (true black) for that color. Each cathode will have a slighthly different voltage that it needs to hit true black. Then, if the CRT chassis design ties G1 to ground (most do), then the cathode needs to use an even higher voltage than black during horizontal retrace because the magnetic fluctuation during horizontal retrace can make some of the cathode ray leak past the G1 anode. This voltage that is even higher than “true black” is called “blacker than black”. The RGB drive amplifiers have a maximum voltage that they can output. Decreasing brightness actually shifts the voltage of the amplified waveform up! Yes this is confusing, but remember cathode voltages are inverted form of the raw RGB signals. Blacker than black voltage is the peaks of the waveform, and if they get clipped, then cathodes leak during retrace.
Image
Thanks for the explanation. Where in the RGB drive circuits would one start looking to remedy this issue? Caps, VR, etc.?
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by LukeEvansSimon »

daskrabs wrote:
LukeEvansSimon wrote:
daskrabs wrote:What does it mean if a retrace line in only one color (red, green, or blue) is visible on a black screen?
It means the amplified signal for controlling the output of that cathode is “clipping”. The color signal is inverted by the RGB drive amps, so darker colors have a high voltage and bright colors have a low voltage. Because electrons are attracted to a high voltage, the darker colors are made by charging the cathode with a high voltage when it should draw nothing (true black) for that color. Each cathode will have a slighthly different voltage that it needs to hit true black. Then, if the CRT chassis design ties G1 to ground (most do), then the cathode needs to use an even higher voltage than black during horizontal retrace because the magnetic fluctuation during horizontal retrace can make some of the cathode ray leak past the G1 anode. This voltage that is even higher than “true black” is called “blacker than black”. The RGB drive amplifiers have a maximum voltage that they can output. Decreasing brightness actually shifts the voltage of the amplified waveform up! Yes this is confusing, but remember cathode voltages are inverted form of the raw RGB signals. Blacker than black voltage is the peaks of the waveform, and if they get clipped, then cathodes leak during retrace.
Image
Thanks for the explanation. Where in the RGB drive circuits would one start looking to remedy this issue? Caps, VR, etc.?
First, decrease screen voltage until the retrace lines disappear. Then increase brightness to restore the darkest shade of gray that is not black in the color bars.

If that doesn’t fix it, then your CRT may have a heater to cathode short. You can fix such a short using a CRT Analyzer/Rejuvenator.
thumptech
Posts: 6
Joined: Tue May 02, 2017 12:07 am

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by thumptech »

LukeEvansSimon wrote: [*] Pot for G1 bias voltage: Most CRTs do not have this mechanism, and instead attach G1 to ground. This is typically only found in higher-end CRTs and higher resolution CRTs. CRTs that do have this functionality send an amplified horizontal sync signal to the G1 anode.

Retrace lines
... CRTs that allow for controlling G1 bias accomplish this by amplifying the horizontal sync waveform and sending it to the G1 anode. ...
Does this mean when implementing the G1 hack from your other thread, it would be better to amplify the HSync waveform rather than apply pure DC? I assume this means only applying the bias voltage during the actual visible scanline?
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by LukeEvansSimon »

thumptech wrote:
LukeEvansSimon wrote: [*] Pot for G1 bias voltage: Most CRTs do not have this mechanism, and instead attach G1 to ground. This is typically only found in higher-end CRTs and higher resolution CRTs. CRTs that do have this functionality send an amplified horizontal sync signal to the G1 anode.

Retrace lines
... CRTs that allow for controlling G1 bias accomplish this by amplifying the horizontal sync waveform and sending it to the G1 anode. ...
Does this mean when implementing the G1 hack from your other thread, it would be better to amplify the HSync waveform rather than apply pure DC? I assume this means only applying the bias voltage during the actual visible scanline?
CRTs that tie G1 to ground already merge the hsync waveform into the RGB chroma signals so that the chroma signal goes blacker than black during hsync retarce. So in the mod I described in the other thread, it is not necessary to send an amplified hsync to G1. It is only necessary to charge G1 to a constant -250 volts (actual value depends on CRT triode design).
Taiyaki
Posts: 1050
Joined: Fri Apr 04, 2014 11:31 pm

Re: Optimal CRT Brightness Calibration Using Triode Theory

Post by Taiyaki »

LukeEvansSimon wrote:
Taiyaki wrote:With crt's it's so confusing as brightness and contrast are often labeled in reverse, as brightness controls black level and contrast the intensity or actual brightness output. I think on modern tv's the contrast tab is generally called picture, brightness, or luminosity. On most Sony CRT's brightness tab definitely adjusts what we know now a days as contrast, and adjusts the global brightness of the picture. I've seen so many crt technicians have to explain this in detail but no one ever seemed to have explained why this was done this way.
I wish the service menus just explained which CRT electrode was being impacted (cathode, G1, or G2), and whether the setting was the bias voltage or the gain. “Brightness” typically refers to cathode bias voltage and “contrast” refers to cathode chroma signal gain.

The import thing is to maximize the cathode bias voltage because that also has the effect of concentration of the cathode ray into a tighter beam. It also gives a wider voltage range for tuning contrast without getting crushed or clipped whites

My bad, I had to edit to point out that the contrast setting on CRT does NOT adjust the global brightness, but I guess you understood despite my poor explanation. Yes you're right, and the problem is that most users would not be able to achieve this however, which is why most end up adjusting with available controls even in situations where this may not be the optimal solution.
Post Reply