Mod a CRT to increase its TVL

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

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

ElBartoME wrote:900V is coming off the flyback and going to the neckboard. From this voltage the focus and screen voltage is derived.

According to the service manual the voltage range for G2 is between 320 and 440 volts. There is a series resistor to the potentiometer of 680k. I could change that value to squeeze out a bit more voltage like say 550 volts.
Service manual for my consumer TV says G2 voltage is 1000v. I actually haven't measured how much I am increasing it because at the time I did the mod, I didn't have a high voltage oscope probe, only a probe that went up to 500v. According to the literature, gun design is the biggest determining factor for voltage to spot size. Without the datasheets for these tubes that show the spot size versus voltage plots (an example was shown in the Display Labs Inc slides I linked in the OP), we won't know how much spot size will change unless we experiment.

I have a JVC BM-H1310SU (also re-branded as the Panasonic BT-H1390Y) arriving in a week. It is a 750 TVL dot mask, and it is an example of a broadcast monitor that does NOT tie the G1 anode to ground. This was a question raised earlier in this thread as to why monitor designers wouldn't use this trick if it would help achieve higher resolutions. Here is the neckboard schematic, where it is clear that G1 has a negative voltage relative to chassis ground.
Image

Looking at the schematic, this monitor appears to make G1 more negative during horizontal blanking. This is one of two methods for ensuring that the cathode to G1 voltage is "blacker than black" during horizontal retrace. This allows for a better spot size because the RGB drive amplifiers can operate at their peak voltage range when scanning and then during retrace they are at a black cathode voltage and G1 is pulled negative for blacker than black voltage. The other technique, which is older, is the design used in my consumer CRT, where the RGB drive amps are used at a lower voltage when drawing each line and during retrace the amps increase cathode voltage to blacker than black.
Image

Here is the schematic for the consumer CRT's neckboard that I modified. You can see that it simply ties G1 to ground and if you oscope the RGB drive amplifiers that control cathode voltage, you can see that when scanning lines, the cathodes do not operate at their peak voltage range, they only go up to about 200 volts for black... but during horizontal retrace (hblank), the amps increase cathode voltage above 200 volts for blacker than black.
Image

The benefit to the fact that the JVC BM-H1310SU's neckboard controls G1 voltage is it should be even easier to decrease the monitor's spot size by changing the bias voltage for the hblank signal going into the amp that drives the G1 anode. I am not sure how much range there will be to play with though as the service manual provides limited information. I will use my oscope to reverse engineer.

Incredibly Useful Technical Articles
Zahid Rahim published numerous technical articles in the 1990s on CRT design. I linked to one of them in the OP, where Zahid talks about controlling spot size... This one is another good one to read, and this article of his describes how to drive a 27-inch consumer grade CRT with a 0.37 mm dot pitch at 1600 pixels horizontal resolution. This comes out to horizontal phosphor triad to spot size ratio of approximately 1.2 triads per cathode ray spot. This is yet another data point that the dot pitch bottleneck that many have brought up in this thread is not a bottleneck for increasing TVL, and that monitor TVL can be increased with more aggressive driving of the CRT anodes. In the early 1990s, Zahid even demo'ed this large consumer grade CRT driving 1600 pixels horizontal resolution. Here is the picture of his demo from the article:
Image
ElBartoME
Posts: 82
Joined: Tue Aug 07, 2018 5:55 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by ElBartoME »

So I tried it by changing the resistor and it really didn't have any effect. I measured G2 at maximum potentiometer setting and it is over 600V as my multimeter says overload.

I will reverse the changes on the 1442QM and maybe work on the 14N2 sometime later.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

ElBartoME wrote:So I tried it by changing the resistor and it really didn't have any effect. I measured G2 at maximum potentiometer setting and it is over 600V as my multimeter says overload.

I will reverse the changes on the 1442QM and maybe work on the 14N2 sometime later.
Were you able to oscope the G2 voltage to ensure it is not dipping during high cathode ray current?
ElBartoME
Posts: 82
Joined: Tue Aug 07, 2018 5:55 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by ElBartoME »

LukeEvansSimon wrote:
Were you able to oscope the G2 voltage to ensure it is not dipping during high cathode ray current?
I unfortunately do not own a high voltage probe. I have some at work that I could borrow but that will take some time.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

ElBartoME wrote:
LukeEvansSimon wrote:
Were you able to oscope the G2 voltage to ensure it is not dipping during high cathode ray current?
I unfortunately do not own a high voltage probe. I have some at work that I could borrow but that will take some time.
I am going to experiment with a JVC AV-27950 this weekend. It is a late 1990s 27-inch bubble curved consumer slot mask that uses the same tube as the later JVC D-Series. The neckboard design will make the mod easy because there is already a fuse resistor and then a diode in series from G1 pin to chassis ground. I will upgrade the diode to one with a higher breakdown voltage and add a smoothing capacitor in parallel with the diode. Then attach my bench supply in parallel with the smoothing cap. I am curious to see if I can reproduce similar or better results to what I got on the 27-inch consumer Panasonic TV.
Image

If the results look good, then I will experiment with using an added secondary winding on the exposed core of the flyback to create a $1 power supply that I can use for the permanent mod. Since this TV uses the same tube as the JVC D-Series, success will mean the mod will work on a very popular line of consumer CRTs. The JVC D-Series has visible scanlines already, but they are a bit less prominent than what many CRT gamers would want. So this mod could be popular for just tweaking for more prominent scanlines and slightly sharper video on the JVC D-Series.
Last edited by LukeEvansSimon on Fri Oct 02, 2020 2:48 am, edited 1 time in total.
User avatar
mikejmoffitt
Posts: 629
Joined: Fri Jan 08, 2016 7:26 am
Location: Tokyo, Japan

Re: Mod a CRT to increase its TVL

Post by mikejmoffitt »

Ah, okay. This explains something odd I'd seen on the New Net City / Toshiba chassis. There is a potentiometer between Pin 4 on the CR-23 socket and chassis Ground (common with pins 5, 10, and 12). I may hook the chassis up to an A68KJU96X and see if it helps resolve the poor beam shape that I experienced when I used this chassis in the past.
Image
Ikaruga11
Posts: 1454
Joined: Thu Apr 07, 2016 1:32 pm

Re: Mod a CRT to increase its TVL

Post by Ikaruga11 »

I think a lot of consumer CRTs have the potential to become sharper since scanlines were seen as a disadvantage of CRT technology, so the image was probably softened at the factory to reduce visible scanlines. Remember that no display is properly calibrated when manufactured, and only made to look good to consumers on the floor of big box stores like Best Buy with high brightness and oversatured colors.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

GeneraLight wrote:I think a lot of consumer CRTs have the potential to become sharper since scanlines were seen as a disadvantage of CRT technology, so the image was probably softened at the factory to reduce visible scanlines. Remember that no display is properly calibrated when manufactured, and only made to look good to consumers on the floor of big box stores like Best Buy with high brightness and oversatured colors.
Exactly this. The prominent scanline look is something gamers are interested in. The average 15khz CRT TV consumer wanted their 480i television to look smooth and not flickery. Another way to think about this is that a CRT is a vacuum tube that can be driven in many different ways. One way is not better than the other. It depends on how you want to make the tube shine.
Ikaruga11
Posts: 1454
Joined: Thu Apr 07, 2016 1:32 pm

Re: Mod a CRT to increase its TVL

Post by Ikaruga11 »

LukeEvansSimon wrote:Exactly this. The prominent scanline look is something gamers are interested in.
Yes, exactly.
LukeEvansSimon wrote:The average 15khz CRT TV consumer wanted their 480i television to look smooth and not flickery.
Oh yeah, I forgot about that. Less visible scanlines hides interlaced flicker better too.
Another way to think about this is that a CRT is a vacuum tube that can be driven in many different ways. One way is not better than the other. It depends on how you want to make the tube shine.
Yep, since it's an analog display and not a digital display. CRTs can be tweaked in so many ways that aren't possible on LCDs, OLEDs or even Plasmas.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

mikejmoffitt wrote:Ah, okay. This explains something odd I'd seen on the New Net City / Toshiba chassis. There is a potentiometer between Pin 4 on the CR-23 socket and chassis Ground (common with pins 5, 10, and 12). I may hook the chassis up to an A68KJU96X and see if it helps resolve the poor beam shape that I experienced when I used this chassis in the past.
Do you have a schematic that you can share?
SamIAm
Posts: 475
Joined: Thu Mar 03, 2005 1:09 am

Re: Mod a CRT to increase its TVL

Post by SamIAm »

LukeEvansSimon wrote:JVC AV-27950
Just FYI, according to the datasheet for the main video decoder chip in this unit, the Toshiba TA1242N, the -3db rolloff point for its output is 7MHz.

On one hand, the datasheet may not be entirely forthcoming about the true limitations of the chip, and could be just giving a satisfactory-minimum figure. Also, with an RGB mod, it's pretty likely that the limit would go up since there isn't really any decoding going on.

On the other hand, broadcast NTSC bandwidth was about 5MHz for the video portion (a little higher for PAL), so it makes sense that SDTV makers wouldn't bother to give their internals better frequency response. In the 90s, I don't think that high speed op-amps capable of dealing with video were cheap, although I could be wrong.

As a matter of fact, my BVM-2015S, which was made in 1992, has frequency response up to 6MHz (+/- 1db) for RGB according to its spec sheet. In reality, I think it goes a little beyond that based on how good 6.7MHz dot-clock Genesis games look, but I wouldn't bet on it doing Saturn high-res very well.

Speaking of which, if you're curious, here is a shot I took ages ago of that HD-capable 900 TVL monitor doing a Saturn high-res game (Necronomicon Pinball):
Spoiler
Image
It's actually a 480i game, so it's all the harder to get a good shot of the whole screen. This, at least, turned out OK. I can't find a good 1:1 emulator screenshot of it, but here is something to give you an idea what you're looking at.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

SamIAm wrote:...
Is there software for Mac, Windows, or Linux that allows for outputting nonstandard resolutions over 15khz analog video, such as 1000x240? This would be useful for outputting high TVL test patterns for objective measurement of 15khz CRT TVL. It would really help me with this mod, so that I can have an objective measure of how much cathode ray spot size is bottlenecking TVL. All I have now are theoretical calculations based off of phosphor dot pitch and spot size. It would also be interesting to actually see if 750 TVL pro CRTs are really 750 TVL. I bet many are high TVL at the center of the screen where focusing is better and lower TVL in the actual region of the screen where TVL is supposed to be measured because corner focusing is less effective.
User avatar
mikejmoffitt
Posts: 629
Joined: Fri Jan 08, 2016 7:26 am
Location: Tokyo, Japan

Re: Mod a CRT to increase its TVL

Post by mikejmoffitt »

LukeEvansSimon wrote:
mikejmoffitt wrote:Ah, okay. This explains something odd I'd seen on the New Net City / Toshiba chassis. There is a potentiometer between Pin 4 on the CR-23 socket and chassis Ground (common with pins 5, 10, and 12). I may hook the chassis up to an A68KJU96X and see if it helps resolve the poor beam shape that I experienced when I used this chassis in the past.
Do you have a schematic that you can share?
I do not, unfortunately. For this pin, it's really just a variable resistor to ground, with no other business going on.

EDIT: Looking at it now, Pin 5 is grounded, which should be G1. So, this may not be related whatsoever.

Image
Image
Image
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

mikejmoffitt wrote: EDIT: Looking at it now, Pin 5 is grounded, which should be G1. So, this may not be related whatsoever.
For the OP mod, that neckboard design requires cutting the trace between G1 and ground... or desoldering and lifting the G1 pin out of the neckboard. I have a few CRTs that have that style where the trace for G1 is part of one giant ground plane.

There has been much confusion in the CRT gaming community about which anode is used for brightness and which is used for focus. It turns out that each anode impacts multiple characteristics of the picture, but they do it in different ways, and so it is wrong to think of one anode as being redundant or not useful. The CRT neckboard design can be used to control these properties of the picture through separate manipulation of the voltage of each anode. I created a table to make this clear:

R=Anode&Common Name&Description&Spot Cutoff Impact&Spot Size Impact&Brightness Impact&Focus Impact R=G1&Control&Electromagnetic iris aperture&High&High&High&Medium R=G2&Screen&Electromagnetic accelerator&High&Low&High&Low R=G3&Focus&Electromagnetic lens&None&High&None&High


When it comes to creating a tightly focused spot of light, just like in a traditional optical system, this can be accomplished by only using an iris aperture or by only using a lens. However, if both are used together, the optical system can make a more tightly focused spot of light than when using only one of these components. The same goes for CRTs. The G1 and G3 both influence spot size and focus, but the G3 can only get the spot small and focused up to a point... the G1 adds additional control.
SamIAm
Posts: 475
Joined: Thu Mar 03, 2005 1:09 am

Re: Mod a CRT to increase its TVL

Post by SamIAm »

LukeEvansSimon wrote:Is there software for Mac, Windows, or Linux that allows for outputting nonstandard resolutions over 15khz analog video, such as 1000x240? This would be useful for outputting high TVL test patterns for objective measurement of 15khz CRT TVL. It would really help me with this mod, so that I can have an objective measure of how much cathode ray spot size is bottlenecking TVL. All I have now are theoretical calculations based off of phosphor dot pitch and spot size. It would also be interesting to actually see if 750 TVL pro CRTs are really 750 TVL. I bet many are high TVL at the center of the screen where focusing is better and lower TVL in the actual region of the screen where TVL is supposed to be measured because corner focusing is less effective.
Two main options come to mind.

If you put an old Radeon card into a Windows PC, you can download a hacked driver that will allow it to output 15khz with custom "super-resolutions" of over 2000 pixels per line. Low-res games scale into that without visible artifacting. However, nobody has ever made a program to allow for easy incremental adjustments to look for TVL. You may also have to deal with the available super-resolutions not dividing evenly to produce uniform line widths, but I'm not sure about that. Anyway, the thing to google is CRT_Emudriver, which is what the hacked driver is called.

The other option is to roll the dice on something like this and hope it does what you want:
https://www.ebay.com/itm/Leader-430-Vid ... Swkf5fUBml

https://www.ebay.com/itm/Shibasoku-Vide ... Sw~P9an-Or


EDIT: I guess a third option is that if you can program (or if can entice someone who can program) it should actually be fairly easy to get something like a Raspberry Pi to spit out a signal that includes both the necessary sync portions and adjustable sweeps. You only need black and white, which simplifies things greatly.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: Mod a CRT to increase its TVL

Post by Josh128 »

LukeEvansSimon wrote:
SamIAm wrote:...
Is there software for Mac, Windows, or Linux that allows for outputting nonstandard resolutions over 15khz analog video, such as 1000x240? This would be useful for outputting high TVL test patterns for objective measurement of 15khz CRT TVL. It would really help me with this mod, so that I can have an objective measure of how much cathode ray spot size is bottlenecking TVL. All I have now are theoretical calculations based off of phosphor dot pitch and spot size. It would also be interesting to actually see if 750 TVL pro CRTs are really 750 TVL. I bet many are high TVL at the center of the screen where focusing is better and lower TVL in the actual region of the screen where TVL is supposed to be measured because corner focusing is less effective.
Windows - there are two programs you can use--- CRU (Custom Resolution Utility) or CRTEmuDriver. Contrary to this idea floating around the net, you are not forced to use super resolutions with either one. Super resolutions are only required / useful for two scenarios: 1) The video card you are using does not support low dot clocks of normal NTSC 15KHz resolutions, or 2) You want to keep the modes and resolutions created to a minimum and still get satisfactory results that look almost identical in practice to the actual resolutions they replace. Myself, I prefer to use the original resolutions, because they are, well, the original and authentic. My MAME setup only includes 100 or so of my favorite arcade games and the 20 or 30 custom resolutions required to support them is no problem. When you want a full set of 6000+ games, then yeah, super resolutions make sense as CRTEmuDriver is limited to storing only 128 or so unique resolutions at a time.

You can indeed create almost any 15KHz resolution you can conceive with either one. I personally have used both, CRU for creating 240p120 modes in Retroarch, and CRTEmuDriver for creating perfect original resolutions for arcade games in MAME. For example, R-Type and R-Type 2 uses an oddball high res/low refresh 384x256@55.018 Hz. Altered Beast uses 320x224@60.054 Hz. CRTEmudriver allows for your video card to create both of those perfectly, and hundreds if not thousands more. You can create your own modelines and tweak your own dot clocks. You can also make the card output composite sync if you wish. Only catch is, you really need an ATI/AMD HD 5000-7000 series video card for best compatibility, but thats not that big of a deal because they are dirt cheap. I just bought a 2nd fanless ASUS HD5450 this week for $13.50 USD shipped!
SamIAm
Posts: 475
Joined: Thu Mar 03, 2005 1:09 am

Re: Mod a CRT to increase its TVL

Post by SamIAm »

Josh128 wrote:...
Ah, so fine adjustments are possible. That's good to know. In that case, you could probably just make an alternating line pattern in MS Paint and go. You'd just have to make sure that nothing is ever scaling the image.
LukeEvansSimon wrote:I bet many are high TVL at the center of the screen where focusing is better and lower TVL in the actual region of the screen where TVL is supposed to be measured because corner focusing is less effective.
I forgot to mention: Yes, some high-grade monitor specs detail the TVL as being at the center of the screen. The PVM-2950Q manual, for example, says "600 TV lines at the center".

Interestingly, it also says that the -3db frequency response is 7MHz for [composite], 8MHz for S-video and 10MHz for RGB. 10MHz ought to be around 400 TVL. I think it's very likely that you'll regularly encounter a ceiling of 10MHz or lower in SDTVs.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

SamIAm wrote: Interestingly, it also says that the -3db frequency response is 7MHz for [composite], 8MHz for S-video and 10MHz for RGB. 10MHz ought to be around 400 TVL. I think it's very likely that you'll regularly encounter a ceiling of 10MHz or lower in SDTVs.
This is more in the realm of the monitor’s addressability as opposed to the CRT’s resolution. I understand that not having enough bandwidth in the signal path will limit the addressable TVL, but for 240p games other than some Saturn and Amiga games, addressability limitations imposed by 10mhz video bandwidth aren’t going to change the way the games look. However, high resolution of the CRT is the main contributor to the high TVL 240p look.
SamIAm
Posts: 475
Joined: Thu Mar 03, 2005 1:09 am

Re: Mod a CRT to increase its TVL

Post by SamIAm »

LukeEvansSimon wrote:This is more in the realm of the monitor’s addressability as opposed to the CRT’s resolution. I understand that not having enough bandwidth in the signal path will limit the addressable TVL, but for 240p games other than some Saturn and Amiga games, addressability limitations imposed by 10mhz video bandwidth aren’t going to change the way the games look. However, high resolution of the CRT is the main contributor to the high TVL 240p look.
To a great extent, yes.

One thing that might factor in to a minor degree is that the signals coming out of the systems are square-waves (or at least trying to be square-waves) with very fast rise and fall times. If low frequency response in the amplifier circuitry in the TV smooths those into something more like sine-waves, it could blur the image slightly.

But 10MHz frequency response is indeed enough for 240p gaming in general, IMHO.

EDIT:

The manual for the 20F1U and 20E1U BVMs, which respectively are supposed to have 900 and 1000 TVL, say that both have 10MHz +/- 1db frequency response. That's presumably a little better than the -3db rolloff at 10MHz of the PVM-2950Q.

Meanwhile, the HD-compatible D-series are good up to 30MHz.

Actually, some have said that HD-SDI is a better format to use for HD broadcast monitors because 720p and 1080i both have a standard dot-clock of 74.25MHz, and HD-SDI input boards will apparently let that through. Not sure how accurate that is, however.
ElBartoME
Posts: 82
Joined: Tue Aug 07, 2018 5:55 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by ElBartoME »

So I decided to do the same experiment with the 14N2. I ran into the same problem. I can't raise G2 high enough to get near -250V for G1. But this time the cutoff controls actually did something and I was able to dial in the colour at least. But my top for G1 was -165V for good colours.

I grabbed my 14L4 which has 800TVL as a comparison.

Image

Here is a comparison between 0V and -165V. As you can see the spot size does get smaller and shows convergence problems much more than before. I tried to dial in the convergence as best as I could but the rings are really stuck together. Blue is still a little bit off.

Image

Here a high-res image for the interested: https://i.imgur.com/g1VaE2m.jpg

And here an image of the 14L4. The convergence is much better on that monitor but I have to say the 800TVL is not too far away from the 600TVL monitor with -165V on G1.

Image

The only problem I see is that I had to max out Contrast and brightness and the cutoff is also pretty much at the maximum...I think PVMs don't have enough control to allow a big change in the G1 voltage which is a shame. I don't necessarily think it is a good idea to raise G2 even higher without having any kind of x-ray measurement going on.

I also think this mod makes more sense on consumer TVs. I have a Sony KV-X2991D in the basement that I wanted to throw away (the tube is not that great anymore). I will try the mod on that TV.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

ElBartoME wrote:...
The scanline sharpness and thickness is definitely increased. Too bad the PVMs lack the range in G2 voltage coming off the flyback that consumer TVs have so that you could have tried higher voltages and smaller spot sizes.

Also, thanks for pushing this mod forward. You were first to experiment with it on both consumer CRT TVs and on professional video monitors. The more people we can get hacking on this, the faster we will learn how to fine tune it. We are all looking forward to your next consumer TV experiment. Since it is a Trinitron, it will be interesting to see the results. Maybe all Sony CRTs will have limited G2 range? Or maybe consumer Trinitrons end up being a big success for this mod.

It makes sense that a thinner cathode ray would make minor convergence issues more prominent because 3 cathode rays need to be aligned and the thinner the beams, the less they will overlap when slightly misaligned. It is a solvable problem, but tuning convergence rings and magnet strips is the most insanely tedious thing to do. I saw the same issue on my consumer CRT, and didn’t bother tuning the magnets.

Regarding xrays, I am going to buy a meter. My understanding is that xray concerns come from the g4 anode (the one with the suction cup). G4 is the second accelerator anode, and is far more influential on acceleration and xrays than G2.

Any chance you could take pictures of Alucard in Castlevania SotN? The 240p test suite doesn’t have good pixel art that contains many quick transitions in colors in its pixel art. “Dark” games like Castlevania have more variation in pixel color transitions.
ElBartoME
Posts: 82
Joined: Tue Aug 07, 2018 5:55 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by ElBartoME »

LukeEvansSimon wrote:Also, thanks for pushing this mod forward. You were first to experiment with it on both consumer CRT TVs and on professional video monitors. The more people we can get hacking on this, the faster we will learn how to fine tune it. We are all looking forward to your next consumer TV experiment. Since it is a Trinitron, it will be interesting to see the results. Maybe all Sony CRTs will have limited G2 range? Or maybe consumer Trinitrons end up being a big success for this mod.
Of course! Thanks for sharing all the information. It is very interesting to read about it and try it out.
LukeEvansSimon wrote: Regarding xrays, I am going to buy a meter. My understanding is that xray concerns come from the g4 anode (the one with the suction cup). G4 is the second accelerator anode, and is far more influential on acceleration and xrays than G2.
I think so too, but would feel better if there are some measurements behind it to prove it.
LukeEvansSimon wrote:Any chance you could take pictures of Alucard in Castlevania SotN? The 240p test suite doesn’t have good pixel art that contains many quick transitions in colors in its pixel art. “Dark” games like Castlevania have more variation in pixel color transitions.
I do not own a PS1 but I'll setup my Raspberry Pi with RGB and try to take some pictures.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

CRT X-Ray Emission Overhyped
CRT xray production is over hyped. Xrays are produced when the G4 anode voltage is above 10 kilovolts. A typical 27-inch CRT uses a G4 of around 25 to 35 kilovolts. The x-rays are generated when the electrons in the cathode ray impact the screen, and the energy of an x-ray is measured in terms of thousands of electron volts measured in unites of kilo-eV or just "keV" for short. The G4 anode at Y kilovolts can generate x-rays with at most Y keV energy.

So why does every study that tries to measure these x-rays emitted from CRTs never find any x-rays? It is due to physics. CRT G4 anode cap voltages are 35kV or below. X-rays with energy less than 50keV are referred to as low energy x-rays also called “soft x-rays”. Low energy x-rays cannot penetrate through thick materials like high energy x-rays do. The lower the energy, the less penetrating the x-rays are. Since the 1970s, CRTs were made with much thicker glass faces. The main reason was to prevent imploding CRTs. However, the extra thick glass is also too thick for low energy x-rays to pass through.

Does This Mod Make X-Rays Worse
This mod involves increasing the the G2-G1 voltage, which is used to start the initial movement of the electrons in the cathode ray towards the screen. However, at most this mod is increasing the acceleration voltage by only 250 volts to 1000 volts, which translates to an additional 1keV energy for the x-rays. That is approximately up to a 3% increase in energy for the generated x-rays. To be clear, your CRTs are already creating x-rays, but they are so low energy that they are not able to pass through the thick glass on the face of the CRT. A 3% increase in x-ray energy is not going to be enough to get them to pass through the glass.

Natural Background X-Ray Radation
X-rays are naturally in our environment. So a radiation meter will read a very low level of low energy xrays in the middle of a sunny outdoor area. Past studies on modern CRT xray production never find anything other than background xray noise that is not produced by the CRT. If you increase total acceleration voltage from the chassis’s specified 25kv to 30kv and raise it up by 10kv to a total of 40kv, it may be possible to detect low energy xrays a centimeter in front of the CRT... if you have a very sensitive meter. But since the glass is going to absorb most of the energy of the xrays, they still may not be high energy enough to survive passing through the glass and if they do, they will be lower energy by the time they exit the CRT.

Measuring X-Rays
I found an inexpensive Geiger-Muller tube meter that is sensitive enough to detect low energy x-rays as low energy as 30keV, which is just at the boundary of where a CRT would produce x-rays. I will test this mod for x-ray production. As described above, the theory indicates that a at most a 1keV increase in the energy of "soft" x-rays will not turn them into the type of high energy x-rays that can penetrate skin and bone. In fact, the x-rays should still have such low energy that they are incapable of passing through the glass of the CRT.

Citations:
http://arcadehacker.blogspot.com/2018/0 ... s.html?m=1
http://www.industrial-electronics.com/e ... x-ray.html
https://inis.iaea.org/collection/NCLCol ... 030661.pdf
https://www.repairfaq.org/samnew/tvfaq/tvsibwaxray.htm
Last edited by LukeEvansSimon on Sun Oct 04, 2020 2:47 pm, edited 3 times in total.
User avatar
buttersoft
Posts: 383
Joined: Sun Jul 24, 2016 7:49 am

Re: Mod a CRT to increase its TVL

Post by buttersoft »

I've got a cheap ionising radiation sensor that plugs into a smartphone headphone jack. It's documentation notes that it's a guide only, and up to 30% inaccurate. But even so, using a CRT that has no beam current limiter (or, say, bypassing the beam current limit by injecting RGB directly to the neckboard) if you increase the brightness/contrast to about 115-120% but leave the black level the same you get 10x the ionising radiation.

I can't remember any absolute numbers, and that's a backyard measurement, for sure. But i was definitely planning to measure the effect of this G1 mod too, when i finally get around to it. I'm keen to hear what your results are as well.
User avatar
maxtherabbit
Posts: 1763
Joined: Mon Mar 05, 2018 4:03 pm

Re: Mod a CRT to increase its TVL

Post by maxtherabbit »

Just irradiate me, small price to pay for crisp scanlines

Plus you just might get superpowers out of it
qjkxbmwvz
Posts: 29
Joined: Wed Sep 23, 2020 12:01 am

Re: Mod a CRT to increase its TVL

Post by qjkxbmwvz »

Meet the amazing scanline man!
With the power of cancer!

In all seriousness, though, I might try this if I can manage an RGB mod or fix my Amiga 1080.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

buttersoft wrote:I've got a cheap ionising radiation sensor that plugs into a smartphone headphone jack. It's documentation notes that it's a guide only, and up to 30% inaccurate. But even so, using a CRT that has no beam current limiter (or, say, bypassing the beam current limit by injecting RGB directly to the neckboard) if you increase the brightness/contrast to about 115-120% but leave the black level the same you get 10x the ionising radiation.

I can't remember any absolute numbers, and that's a backyard measurement, for sure. But i was definitely planning to measure the effect of this G1 mod too, when i finally get around to it. I'm keen to hear what your results are as well.
Geiger-Mulller tubes are triggered by all sorts of radiation, and we only care about detecting x-rays.
If you use a cheap GM tube to measure radiation, you have to make sure it is not getting triggered by the light emitted by the CRT (which is harmless photon radiation). For example, the popular and inexpensive GMC 300e plus uses the M4011 GM tube, which is known to be triggered by harmless photon radiation because it uses a clear glass chamber. The SBM-20 Geiger-Muller tube is encased in brass, and is not triggered by light, but the paper thin brass case is too thick for low energy x-rays to penetrate, so it won’t even be able to measure the weak x-rays a CRT generates. Now if you increases the G2 + G4 anode to a total of 50kV, then you could get your CRT to emit high energy x-ray that an SBM-20 could detect.

I am skeptical that a modern CRT can emit x-rays, unless it is purposely hacked to use an extra 20 kilovolts of high voltage on top of its 30 kilovolts. The old 1950s CRTs that had such thin glass they were known to implode... I bet those emit low energy x-rays that can be detected a few centimeters in front of the screen.
Last edited by LukeEvansSimon on Mon Oct 05, 2020 5:11 am, edited 1 time in total.
User avatar
buttersoft
Posts: 383
Joined: Sun Jul 24, 2016 7:49 am

Re: Mod a CRT to increase its TVL

Post by buttersoft »

EDIT: the one i have is one of these, so it's a semiconductor sensor - http://allsmartlab.com/eng/smart-geiger/ but it's encased in white plastic not aluminium. It doesn't get triggered by normal LED, incandescent, or flurorescent light sources. It does admit it gets triggered by wifi though, but does a CRT even emit at those bands? I'm happy to say this thing is pretty rubbish, but hopefully i can post up some results and see how they compare to yours.

Sorry thumptech, i must have been editing while you posted :)
Last edited by buttersoft on Mon Oct 05, 2020 12:23 am, edited 3 times in total.
thumptech
Posts: 6
Joined: Tue May 02, 2017 12:07 am

Re: Mod a CRT to increase its TVL

Post by thumptech »

buttersoft wrote:I've got a cheap ionising radiation sensor that plugs into a smartphone headphone jack. It's documentation notes that it's a guide only, and up to 30% inaccurate. But even so, using a CRT that has no beam current limiter (or, say, bypassing the beam current limit by injecting RGB directly to the neckboard) if you increase the brightness/contrast to about 115-120% but leave the black level the same you get 10x the ionising radiation.

I can't remember any absolute numbers, and that's a backyard measurement, for sure. But i was definitely planning to measure the effect of this G1 mod too, when i finally get around to it. I'm keen to hear what your results are as well.


Can you please specify the model and brand of the sensor or provide a datasheet?

EDIT: Thankyou for providing the details of that sensor. I don't believe any measurements from this device can be trusted.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

I am going to try to use the GMC 300e plus to measure if there are any low energy (or high energy) x-rays generated by this mod. Detecting high energy x-rays is easy. The low energy x-rays, which are much less of a health concern, but more likely to be generated, can be detected by this inexpensive Geiger-Muller counter. Due to the GM tube only having a thin glass case, which makes it sensitive to low energy x-rays, it can give false alarms due to harmless light radiation, so I will make sure to correct for that. The tube in this has been tested in an x-ray machine, by an engineer that posted on eevblog. The machine was cable of generating low energy x-rays, which this inexpensive device was capable of detecting.

But again, according to the theory, while all modern CRTs do generate x-rays, they are too low energy to pass through the glass of the CRT, and this mod does not increase x-ray energy enough to cause them to pass through the glass.
Post Reply