I read the entire thread before jumping in. I stay out of mod threads being a non-modder but I didn't appreciate the Reddit post in crtgaming boasting about boosting the TVL. Reddit knowledge level is far lower than here. That's also disingenuous to compare the spot size on Mega Man using non-PVM to PVM and imply your mod makes the image look like the PVM's with its higher quality phosphors.
SamIAm explained in different terms. As impressive as your analysis is to solve red color bleeding by adjusting the cutoff voltage in the service menu, you don't warn anyone on page 1 that they'll need to do this and not all CRTs have that ability.
Josh128 straight out told you that he didn't count more TVLs in pre-mod and post-mod pics and you defended yourself by re-defining "TVL" from the definition actually used to sell PVMs and adopted by everyone in retro gaming who then re-sell PVMs. Addressability vs resolution is interesting and scientific but you mention neither when start by saying you increase TVL and show calculations of how 1000 TVL is achievable. We had heavyweights
mikejmoffitt and
6t8k weigh in expertly as well and no one scrolling to the last page will see that. Last thing I want here is someone selling a "TVL boosting" mod service on an already rare and expensive PVM and damaging it.
The explanation of increasing CRT lifespan confuses me. You're making the screen less bright with negative G1 voltage and increasing G2 voltage to compensate. Image still looks dim so you're turning the brightness and contrast knobs up to compensate? Doesn't that increase the cathode voltage and work of the RGB amplifiers? But G1 < 0 reduces the level needed for cutoff for true black so that is an advantage. You don't mention the cutoff voltage changes for R, G and B. I think the average cathode voltage level would be higher since there is more luminance than darkness in most video games.
I don't blame you for showing spot size in green, the most luminant color. Red is second highest and blue the least. Would be more fair to show each color, or at least one image each high in R, G and B to verify bleeding is minimized and convergence is good. That is one thing you and
ElBartoME very fairly point out that convergence issues are more noticeable with smaller spot size. It just doesn't get edited onto page 1 for everyone to see.
CRT flybacks have an exposed core. If you loop a wire around that exposed core about 125 times, then the voltage between the ends of the wire will be 250 volts. The current is extremely low, but that is a good thing because the G1 does not need a high current.
Making a -250 voltage source that way, [voltage] = [current] * [resistance] = -[area perpendicular to the magnetic field] * [change in magnetic field with respect to time] => emf = i * R = -N * dB/dt. As you add a turn N, for more area on the way to 125, you increase R by the equivalent cable length, thus current is constant despite more voltage. You increase voltage with constant current and therefore increase power linearly. Power for G1 and higher voltage and therefore higher power for G2 don't come from nothing. Are you not straining the flyback transformer and reducing its lifespan and heating the nearby components? Running G1 and G2 higher than what they're rated for? I don't know how the deflector part of a CRT works but the electrons hitting it have much more force by than they're supposed to by F = q * v X B, where X is the cross product and the higher G2 increases v, the velocity vector.
I would think then that you're reducing the lifespan of the CRT more than increasing it.
I'm done being mean.
------------------------
I do give you credit for mentioning to add the capacitor to reduce ripple voltage. Ripple voltage is not DC and is lost as heat. I could show the math to calculate and estimate it and the minimum capacitance needed since I don't think everyone attempting this mod is going to have an oscilloscope or know how to use LTspice.
I also give you a great amount of credit for constructing SPICE simulations of a mod. Sets a new standard when I can't think of another mod that does this. I sometimes wonder where new resistor and capacitor values come from. Thin air or first thing that works?
I actually think the math on page 1 is fascinating for theoretical minimum TVL and I learned quite a bit. I want to contribute one source I found on CRTs that starts on page 3:
https://www.scribd.com/document/4948290 ... nformation
My understanding from this mathematics book for programming excerpt is that the negative G1 voltage reduces the number of electrons that come out of the electron gun. More negative, fewer electrons, so lesser intensity, so lower brightness. I get that from a geometric argument an outer ring of negative voltage pushes the electrons closer together as they approach the iris but then wouldn't the positive voltage G2 ring, of much greater magnitude, spread them apart? I think the reduced electron count is far more significant for the smaller spot size. Not that it changes anything.
The prominent scanline look is something gamers are interested in.
I have a different opinion. We know the classic divide was increase RF + composite blurring in consumer CRTs to reduce flicker and increase scanlines on professional CRTs so broadcaster could judge tendency of flicker.
In my childhood RCA CRT, I could see the faint scanlines and they ended up being extremely important in one SNES game to count the value of the player's reputation. I don't need thick or thicker scanlines, that minimal amount is enough and LCD smoothening between horizontal lines looks like hot garbage in comparison. My 20L2 has unsurprisingly thicker scanlines than my consumer JVC 20F703, except I noticed JVC's are thicker for S-Video. Maybe I need to adjust settings but that took me by surprise.
For horizontal resolution, I didn't see anyone talk about accuracy of the image. SNES native resolution is 256×224 (8x7), right, with a few games using 512x448 as someone mentioned. CRT stretches that. My 600 TVL PVM gives a more accurate (4:3) scaling than consumer low TVL and I see that borne out in the images here. Making smaller dots doesn't help in and of itself. Maybe you can take a CRT that uses 2 triads / RGB groups per pixel and change G1 and G2 to yield 1 triad per pixel and achieve better scaling. No black line separating the 2 for higher TVL by PVM salesman definition but I don't think you'd encounter much resistance if you claimed it. What I'm not sure is if 1.2, 1.5, etc. triads per pixel down from 2.0 is actually better for video games that never went above 480 / 525 horizontal resolution.
Also curious about PAL. Abomination that it may be for 50 Hz, the alternating phase of the lines, thus the acronym, on half the chroma signal reduces noise in theory. Shouldn't composite and S-Video look better and therefore benefit more from this mod?