Mod a CRT to increase its TVL

The place for all discussion on gaming hardware
SamIAm MkII
Posts: 20
Joined: Sat Feb 13, 2021 3:54 am

Re: Mod a CRT to increase its TVL

Post by SamIAm MkII »

I'm glad that people are interested. :)

Irfanview is a great little program. By using its batch-processing functions, I was able to create a mega-pack for you guys in a matter of minutes. In the archive, you'll find subfolders with the following vertical resolutions:

240
480
540
600
720
768
900
1080
1200
1600

If I'm not mistaken, the numbers in the lower left should still be proportional and everything. Just let me know if you want any others. The table in my post on the last page is still applicable as far as total lines and TVL are concerned, but the pixel clock and analogue frequencies will be different for each set.

It looks like pushing up to 3840 or even 5120 might really be possible for a lot of us, at least at lower vertical resolutions. At 240p, if the pixel clock is the only limiting factor, then even 10,240 horizontal pixels is not out of the question. I'll have to do some more experimenting later.
buttersoft wrote:I'd put in a vote for 1080, my KV-HR32 does that interlaced :)
Hey buttersoft, do you happen to have a modeline handy that produces 1080i with all the timings close to broadcast standards? That's something I'm going to want to hunt up (or figure out) soon. I'm also interested in getting 540p going. This is for that Ikegami monitor.
Last edited by SamIAm MkII on Wed Sep 22, 2021 4:32 am, edited 1 time in total.
User avatar
matt
Posts: 567
Joined: Mon May 09, 2005 5:46 am
Location: Honolulu, Hawaii

Re: Mod a CRT to increase its TVL

Post by matt »

I tried this on my mitsubishi 2070sb yesterday (I just set the resolution to 2560x480x60hz and used the images at half height).

Apparently, my monitor is at least 960 TVL at 31khz, but the 1920 TVL pattern is beyond the limit of the aperture grille. I'm curious how it holds up at higher resolutions as the bandwidth increases.
SamIAm MkII
Posts: 20
Joined: Sat Feb 13, 2021 3:54 am

Re: Mod a CRT to increase its TVL

Post by SamIAm MkII »

matt wrote:I tried this on my mitsubishi 2070sb yesterday (I just set the resolution to 2560x480x60hz and used the images at half height).

Apparently, my monitor is at least 960 TVL at 31khz, but the 1920 TVL pattern is beyond the limit of the aperture grille. I'm curious how it holds up at higher resolutions as the bandwidth increases.
Cool, man. :) If you'd like another resolution besides the ones in the set I posted above, just say the word.

I just made a second set of the fifteen images that are 3840 pixels wide. I didn't adjust the stretch on the numbers because they're only a little squished in 4:3, and you should still be able to read them just fine.

The splits are a little different from 2560. (I'll format this into a table later).

Pixel width / Total lines / TVL

1 3840 2880
2 1920 1440
3 1280 960
4 960 720
5 768 576
6 640 480
7 548 411
8 480 360
9 426 320
10 384 288
11 349 261
12 320 240
13 295 221
14 274 205
15 256 192
SuperSpongo
Posts: 315
Joined: Sat Mar 17, 2018 2:49 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by SuperSpongo »

How exactly did you go about displaying the image in the super resolution?
If I remember correctly, CRT_Emudriver squishes Windows desktop down to 4:3 when enabling the super resolution, but the Arcade_OSD is still usable.

My plan is to do the following:

- Open the image using Windows Picture Viewer (or whatever it's called) with a standard resolution like 720p.
- Place the Arcade OSD into the quickstart bar so I can open it by pressing Windows key+1 or similar
- Open up Arcade OSD, switch to the super resolution and use it as Desktop resolution
- Alt Tab into the Picture Viewer, Press F11 for full screen

Am I missing something or should this work?
User avatar
buttersoft
Posts: 383
Joined: Sun Jul 24, 2016 7:49 am

Re: Mod a CRT to increase its TVL

Post by buttersoft »

SamIAm MkII wrote: Hey buttersoft, do you happen to have a modeline handy that produces 1080i with all the timings close to broadcast standards? That's something I'm going to want to hunt up (or figure out) soon. I'm also interested in getting 540p going. This is for that Ikegami monitor.
I'll give you my timings, from the EDID in the HDFury2 i use, but you may have to play with them a bit. The thing with 1080i is that every damn set out there that could handle it, handled it differently. And trying to make modelines gets really messed up because GPU's just don't seem to know how to produce it. And when they do, they mislabel it or scale the fuck out it and do other weird things.

Direction/Active/Front_Porch/Sync_Width/Total_Pixels
Hor/1920/88/44/2200
Ver/1080/4/5/1125
This is reported by my nividia drivers as having a refresh rate of 30Hz progressive at a scan rate of 33.75kHz with a pixel clock of 74.2500MHz.
SamIAm MkII
Posts: 20
Joined: Sat Feb 13, 2021 3:54 am

Re: Mod a CRT to increase its TVL

Post by SamIAm MkII »

SuperSpongo wrote:How exactly did you go about displaying the image in the super resolution?
If I remember correctly, CRT_Emudriver squishes Windows desktop down to 4:3 when enabling the super resolution, but the Arcade_OSD is still usable.

My plan is to do the following:

- Open the image using Windows Picture Viewer (or whatever it's called) with a standard resolution like 720p.
- Place the Arcade OSD into the quickstart bar so I can open it by pressing Windows key+1 or similar
- Open up Arcade OSD, switch to the super resolution and use it as Desktop resolution
- Alt Tab into the Picture Viewer, Press F11 for full screen

Am I missing something or should this work?
That sounds like it will work. Shortcut keys are a great idea. At 2560x240, I can still just barely use the GUI, and so these last few times I kept the first image loaded in the background like you described and simply double-clicked it after closing out of Arcade OSD. If we successfully move up to 3840x240 and beyond, though, I'm sure I'll need another approach.

I hope the Windows program doesn't do anything like add a tiny side border and scale the image to fit. If you take any of these images, especially the high-line-count ones, and scale or filter them even a little bit, they get severely messed up. It kind of ruins the test.
buttersoft wrote:
I'll give you my timings, from the EDID in the HDFury2 i use, but you may have to play with them a bit. The thing with 1080i is that every damn set out there that could handle it, handled it differently. And trying to make modelines gets really messed up because GPU's just don't seem to know how to produce it. And when they do, they mislabel it or scale the fuck out it and do other weird things.

Direction/Active/Front_Porch/Sync_Width/Total_Pixels
Hor/1920/88/44/2200
Ver/1080/4/5/1125
This is reported by my nividia drivers as having a refresh rate of 30Hz progressive at a scan rate of 33.75kHz with a pixel clock of 74.2500MHz.
Thank you! I'll definitely be keeping this around, and I'll let you know what happens. :)
conango
Posts: 2
Joined: Thu Sep 23, 2021 3:42 am

Re: Mod a CRT to increase its TVL

Post by conango »

Hi, so exciting mod I've ever imagine.
it would be very helpful if you could attach some photos to each step.
Especially for how to connect a 250v dc part.
Many thanks.
SuperSpongo
Posts: 315
Joined: Sat Mar 17, 2018 2:49 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by SuperSpongo »

Alright, so I did some testing yesterday, here's a couple of impressions:
I used a CRT-Emudriver build going into an Extron 192 to combine the sync. First thing I immediately noticed was that the picture was a little wobbly which was a problem introduced by the interface. I disabled DDSP and it got a little better.
The 240p image is slightly tilted which I hadn't noticed before. I don't know if I'd have to tilt the yoke or if the D32 supports rotating via service menu. But when displaying the higher TVL images, this resulted in a noticable moiré pattern which is due to the aperture grill (is my best guess).
The D32 claims a resolution of 1000TVL in both, 16:9 and 4:3 mode, so the 4:3 test patterns should have been discernable up until, say, 3 at least. This was definitely not the case, so I fiddled with the settings a little. The video below is with 16:9 stretch an underscan activated. Underscan should somewhat compesate the 16:9 stretch.

https://youtu.be/hXN5v4PKxhU

Like I said, the picture was not as uniform as directly above the number display. I had noticable moiré patterns and also discoloring (although that could have been from magnetic interference). I also think that the tube is not as sharp in certain areas.
SuperSpongo
Posts: 315
Joined: Sat Mar 17, 2018 2:49 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by SuperSpongo »

Ah okay, that makes sense. This is also the case for most TVL specs that I've seen so far. If I revisit this topic using a different Extron device, I'll be sure to capture footage from the center. Maybe it would be beneficial to move the numbers to the center of the screen as well? I found them to be a nice reference point when filming.

Just had a look at the service manual and it does support rotation.
SamIAm MkII
Posts: 20
Joined: Sat Feb 13, 2021 3:54 am

Re: Mod a CRT to increase its TVL

Post by SamIAm MkII »

That was really interesting to see, SuperSpongo. I wish I had a camera that could capture the detail of my HTM-2050 as well as yours captured your D32.

Both your 3 and your 2 were significantly better than mine. Even on 2, it's still possible to count the white dots on yours, and on 3, they were quite clear. Really, all of your numbers were better than mine, if not quite to the same degree as 3 and 2. They all had less white eating into black.

Whether your D32 passes 3 or even 2 all just depends on how you want to define the cutoff point for TVL. Is it when you can't see the black lines anymore? Or is it when white gets so muddled that you can't confidently count the dots anymore?

By the way, what you see around the numbers is simply shading applied by the font. I checked out a few purely monochrome options, and they were all pretty bad. As for their position, I put them where I figured it people would find them the least obtrusive. If people really want them more centered, I guess I could make that change.

Anyway, thanks for that. It's great to see how what was arguably the must deluxe CRT ever made handles a TVL test. Are you thinking of trying 720p at any point?
Last edited by SamIAm MkII on Thu Sep 23, 2021 3:51 pm, edited 2 times in total.
SuperSpongo
Posts: 315
Joined: Sat Mar 17, 2018 2:49 pm
Location: Germany

Re: Mod a CRT to increase its TVL

Post by SuperSpongo »

Yeah, it's an interesting test for sure! I also have a couple of other monitors of interest such as the F520 and FW900. I assume once ElBartoME is back from vacation, he'll want to join in on the fun as well :D

At the moment I don't have a 720p modeline installed (I kinda ran outta time yesterday evening) but once I'm able to spend a little more time with this, I can add a 2560x720 resolution and test again.
SamIAm MkII
Posts: 20
Joined: Sat Feb 13, 2021 3:54 am

Re: Mod a CRT to increase its TVL

Post by SamIAm MkII »

Just got 3840x240 working. 5120 and up will present at least an added layer of challenge.

Anyway, I tried the 3840-wide TVL image set. By the standard I used before, 4 (720 TVL) passed. 3 (960 TVL, same as 2 in the 2560 set) naturally did not pass.

Overall, and unsurprisingly, this is better than the 2560 set. I'll expand it to 20 in case anyone wants to use it for lower TVL monitors.

Now to get 540p/1080i working...
conango
Posts: 2
Joined: Thu Sep 23, 2021 3:42 am

Re: Mod a CRT to increase its TVL

Post by conango »

LukeEvansSimon wrote:During my reading, I bumped into this patent from Intel on CRT design that describes what the G1 biasing grid does, and it made me think of this thread. So I thought I'd post an update.
Image

I haven't touched this project since May. I fell down the rabbit hole of non-solid state TVs, that is, TVs that use more than just one vacuum tube. I have been studying classic CRT research articles, such as the collected works of: RCA's Zworykin, RCA's Grob, and RCA's Schade. Non-solid state CRT displays are significantly more simplistic than solid state CRT displays, and I am developing a from-scratch all new old stock vacuum tube YPbPr video monitor that uses a new old stock "M-class" fine dot pitch CRT and several vacuum tube diodes, triodes, and pentodes. It is very satisfying to get down to the core circuits of a CRT display like this. Solid state CRTs are great, but they use very complicated circuits with many black box "jungle" ICs. Vacuum tube rectifier, amplifier, and oscillator circuits are very simple and elegant. It is like an entire circuit built out of little CRTs.

Back to the G1 mod: I realize that I did not post pictures of the final mod, so here they are...

Image
I ended up adding RLC filters before and after the potentiometer. These filters only made minor improvements to the video quality, but the oscope showed a very good improvement in G1 voltage stability, so I left them in the final build just in case. The G1 to cathode voltage is more stable after the mod than before! Note: R4 and R5 simulate the wires that connect the G1 control PCB to the neckboard. They are not discrete resistors!

Image
The final G1 control board is tiny. If anybody is interested, I will post a parts list. I used 3M adhesive standoffs to fix the board inside the TV. The pot has an extra long dial so you can safely adjust it without an extra tool.

Image
The neckboard already had solder holes for a bus wire that connected the G1 trace to ground. So I removed the bus wire and replaced it with two quick disconnect terminals. This way the mod can be unplugged from the neckboard via wires that have quick disconnects on both ends.

Image
The 27-inch TV I modded has lots of empty space inside. The G1 controller board easily fit in the empty space next to the flyback transformer. This allows the G1 G2, and G3 potentiometers to all be easily accessible in the same area of the back of the TV, with 3M adhesive standoffs gluing it in place.


Image

Image
I found this side-by-side comparison on my phone from May. The pictures of Alucard from SotN are not in the same area of the map, so ignore the background pixels. You can see the spot size is much smaller after the mod, but the camera photo picks up a moire effect. The moire effect is hardly noticeable in real life, but for some reason when zoomed in with my iPhone, the moire effect is exaggerated. I have noticed that my iPhone exaggerates the moire effect of other CRTs, even ones not modded. However, this does bring up a real limitation of this mod. As spot size gets very small, moire effect becomes more and more noticeable. However, you can use the potentiometer to adjust the spot size to a point where moire is not noticeable in real life, yet the spot size is sharp. If you make the spot size too small, then moire becomes noticeable, even in real life. The fact that aperture grille CRTs do not have a horizontal masking of the phosphor like slot mask means the moire effect on an aperture grille will be less pronounced.
So glad you made it a practical PCB!
Waiting for the parts list.
Could you please also write a more detailed installation guide for someone like me with very limited circuit knowledge?
Like solder A point on PCB to B point on the C board?
Many thanks!
SamIAm MkII
Posts: 20
Joined: Sat Feb 13, 2021 3:54 am

Re: Mod a CRT to increase its TVL

Post by SamIAm MkII »

It's starting to look like 5120 and higher will be less easy to enable and/or not as widely compatible as 3840.

Combining the 2560 and 3840 image sets, we get the following TVL test patterns between 192 and 1000, which you might say is our primary range:

960
720
640
576
480 (640 lines total)
411
384 (512 lines total)
360
320
288 (384 lines total)
274
261
240 (320 lines total)
221
213
205
192 (256 lines total)


That ain't bad. Of course, it'd be nice to have a few more in the upper range - especially something around 800 - and it would be extra nice if we could view them all at once in a single ultra-high resolution. OTOH, though, I bet this is already going to tell most people everything they really want to know. When you run through either or both of these sets, you get a very good idea of where your monitor's limits lie. Also, anyone who wants to mod a TV that was under 500 TVL to begin with should have no trouble finding a pattern with which to demonstrate improvement.

If someone really wants a specific number, the best way to do it is to create a custom image and modeline. 800 and 600 TVL test images, for example, would be useful to confirm the ratings of certain pro monitors, and they're both conveniently available through 3200x240. Making the images for this by truncating the 3840x240 images is so easy, I did both of them just now.

Thoughts?

EDIT: Just got 540p to work on the HTM-2050 (except for a tiny warp at the top), and holy crap is it better. 2560-3 (640 TVL) looks great. Black lines are highly visible.

EDIT 2: Calamity, the fellow behind CRT Emudriver and Groovymame, personally did some testing and confirmed that going above 4096 horizontal resolution is not possible. 3840 gives better splits anyway, so I guess that's that. Maybe I'll make a few more variants, then bundle everything up and call this done.
antorsae
Posts: 7
Joined: Tue Jun 15, 2021 3:42 pm

Re: Mod a CRT to increase its TVL

Post by antorsae »

Hi - I did the mod to one KV-25C5E (Sony 25" European consumer Super Trinitron)

Before:

Image

After (-150VDC):

Image

I suck at taking CRT screenshots but this is the first time I can get very defined scanlines @ IRE 100 on all screen. PVMs can do it so it is great to be able to do this with Sony consumer sets (very bright P22 phosphors and dark/black tubes... 8) ).

Other (after):
Spoiler
Image
See full imgur gallery for RTYPE video: https://imgur.com/a/s9aO204

I mounted ElBartoME's board in the neck board drilling a small hole.

As noted before by LukeEvansSimon, once I started lowering G1 I got thinner scanlines and got red/blue bleeding. I then adjusted G2 to compensate and it went away. I have yet to tune red/blue cutoff.

I also noticed focus uniformity improved significantly with this mod.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: Mod a CRT to increase its TVL

Post by Josh128 »

Looks like your hori static convergence is off quite a bit-- have you tried the hstat pot to adjust it?
antorsae
Posts: 7
Joined: Tue Jun 15, 2021 3:42 pm

Re: Mod a CRT to increase its TVL

Post by antorsae »

I believe the first major obstacle to increase TVL at last in consumer Trinitron is the Jungle. The one Im modding (KV-25X5E) uses the CXA2060AS which is not documented, but similar chips list 10 Mhz bandwith.

I bought a few 130 Mhz LM1208N which have drive/cutoff controls via potentiometers.

I have started designing a circuit to "bypass" the Jungle chip in my consumer Trinitron:

Image

The idea is:

1. Intercept RGBS right after the 75 ohm termination and protetion diodes (connector J2), feed each color to the LM1208N
2. Use a LM1881 sync extractor to get the clamp pulses necessary to restore black level.
3. The jungle would still work for deflection, etc. so in order to get exact blanking signals, extract the original signals from the jungle (connector J1) and use an analog comparator to adjust blanking level (via potentiometer) based on eg the output of the R channel (which already has blanking).
4. There's two output connectors: J3 is the output to the neck via LM1208N, and J4 is the original jungle output. This way it's easy to A/B test the board.

Im still debating how to get 12V and 5V from the Sony board. I think I will attach connector to "LOW B" voltage in the Sony board (this is where +9V and +5V are generated in the Sony board), and use LM7805 and LM7812 to get +5V and +12V.

I also would like some guidance as to what type of capactitors to use.

This is obviously a big WiP (Im learning Kicad as we speak) so constructive feedback very welcome.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

I got addicted to pre-solid state TVs, which I know sounds insane to the people here that use BVMs with multiple CPUs in them. So I haven’t checked this thread in months. It is great to see this starting to become a community effort as that is the only way for the final goal to be achieved. Proper TVL test patterns is great to see, and will have applications outside of this modding effort. Will be great to see the marketing claims for broadcast monitor TVL put to the test.

Having others try the mod will help us learn how it varies from CRT to CRT.

Exploring how to bypass low bandwidth jungles is also important to figure out, and it also opens the door to a universal RGB mod that can be applied to any CRT display as opposed to relying on CRTs with jungles that have RGB OSD pins.

Now I will disappear for a while, as I tinker with really old CRTs and vacuum tube video amplifiers. Peace!
atohmdiy
Posts: 57
Joined: Wed May 06, 2020 9:48 pm

Re: Mod a CRT to increase its TVL

Post by atohmdiy »

Hello,
I plan on doing this mod in two tv, one philips 28PT4458/01 and one thomson 33MS24E.
The service manual of the philips is fine and i managed to get the info i need for the mod.
But for the thomson i really feel dumb.
Their is service manual for the 33MS24E and the ICC17 chassis but both doc do not seems to give the schematic of the crt board. There is occurrence to the "CRT-Board" in the manual (you can find it in archive.org) but no schematic. I browsed it several times and i don't find anything.
There is a part number in the crt board, 10533820.03 but google do not give anything else than this :
http://www.sklepy.abc24.pl/default.asp? ... top_sklepu

The G1 anode is on the pin 5 of the socket in the philips, but i am not sure at all this is the case for all sockets, so i am asking here if someone have an idea, i don't want to make a mistake especially in a crt :roll:
wren
Posts: 2
Joined: Sun Jun 12, 2022 2:02 am

Re: Mod a CRT to increase its TVL

Post by wren »

ElBartoME wrote:
I now did that mod on a good lot of TVs and I also went away of using the additional winding on the CRT. I took your advice and used the same design you use now, minus all the filtering. But I designed a PCB now and added some filtering and I shamelessly took the last stage of your filter and added it to my design. I hope that's okay. :mrgreen:

[...]
The resulting PCB is very small (40mm x 25mm).

Would you be willing to open-source the PCB files?
User avatar
Initial-d
Posts: 20
Joined: Thu May 31, 2018 7:13 am
Location: Not Japan sadly

Re: Mod a CRT to increase its TVL

Post by Initial-d »

Will this mod work on a Wells Gardner K7000?
I have a 19” and wonder if this would be possible?
User avatar
WrenlyBewick
Posts: 2
Joined: Thu Dec 01, 2022 12:14 am

Re: Mod a CRT to increase its TVL

Post by WrenlyBewick »

Initial-d wrote:Will this mod work on a Wells Gardner K7000?
I have a 19” and wonder if this would be possible?
Within my limited knowledge of the innards of a CRT, it is my understanding that the principles apply to virtually any CRT. With that said, you would have to figure out the gory details of power supply types and what to use and how to wire it into the K7000 chassis. This mod isn't for the faint of heart, or so I understand.
Rapper_skull
Posts: 1
Joined: Thu Feb 16, 2023 10:45 am
Location: Italy

Re: Mod a CRT to increase its TVL

Post by Rapper_skull »

ElBartoME wrote: I modded a 29X5D so I think it should work without a problem.
I want to mod a 29C5 and it looks almost identical to the 29X5. Did you use pin 10 (the one marked on the schematic as 27v) of the flyback?
ElBartoME wrote: Since you are in Europe I can send you one PCB.
Do you happen to have a board to sell or give away?
k1t3
Posts: 2
Joined: Sun Dec 03, 2023 10:53 pm

Re: Mod a CRT to increase its TVL

Post by k1t3 »

Anyone know how well this would work with a PVM 2010QM (EU ver of 1910qm)? It only has 350 TVL so it could really use that boost..
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: Mod a CRT to increase its TVL

Post by Josh128 »

k1t3 wrote: Sun Dec 03, 2023 10:56 pm Anyone know how well this would work with a PVM 2010QM (EU ver of 1910qm)? It only has 350 TVL so it could really use that boost..
Contact LukeEvansSimon on the Reddit CRTgaming subreddit.
LukeEvansSimon
Posts: 128
Joined: Mon Aug 17, 2020 7:28 pm

Re: Mod a CRT to increase its TVL

Post by LukeEvansSimon »

perrocontodo has some great photos on Reddit of his implementation of the G1 mod.
User avatar
Josh128
Posts: 2123
Joined: Thu Jan 16, 2014 9:01 am

Re: Mod a CRT to increase its TVL

Post by Josh128 »

Looks great. Appreciate you updating here.
Post Reply