RGB modding CRT TV - JVC AV-20N8

[Zebidee]

My latest modding victim was a JVC from early 2000's. This time it's RGB!

The TV had obviously lived a hard life, but the tube is still good.


 

 


   


Can't really tell the model number from that sticker! Seems like "AV-N2088". Whatever it really is, it seems an *almost* perfect match for JVC AV-20N8 (BK). Schematic and jungle datasheet attached.

In case you're curious, the Thai part of the label contains some warnings: unplug TV after use (sensible!), don't immerse it in water (there goes my idea for a "Swimmer" cab!), and don't remove the back cover (Oh nose, no warranty for me!). Then there is an address, interesting that a Japanese company houses itself in the China-Thai building.

The RGB mod itself is straightforward. The jungle IC is a Toshiba TB1226EN, which has external analog RGB input pins at 23, 24, 25. The analog RGB blanking (Ys) pin is 22, it requires at least 0.5v for RGB. These pins were originally unused, completely unconnected, except for the Ys (blanking) which has a 10k "pull-down" resistor to ground.

No need for hacking the OSD signal here, So don't have to worry about OSD RGB MUXing stuff.


 


So I made up a 6-pin JLT connector with 30cm wires, and fed them through a convenient hole in the middle of the board. From there I carefully soldered the R,G,B wires directly to the relevant jungle pins. For blanking (yellow), no need to remove an existing 10K terminator on Ys, as it will have negligible effect on the input voltage level, and the solder pad provides a good place to solder the blanking signal wire.

For the sync and ground (white/black wires), I connected them under an existing connector for external video input (to right in pic), close to where it meets the AC coupling input capacitor (10uF, C233). Left the cap as it is. There is a 75 ohm terminator resistor somewhere too (not pictured), so don't have to worry about that either.

Next, we need some RGB termination resistors and AC coupling caps, and a couple of resistors on the blanking voltage. 
To make connecting inputs easy, and to make some space for the input caps & resistors, I made up a mini-PCB with some breadboard. Similar to what I've done previously in component TV mods.


   


I tweaked the mini-PCBs a lot, so multiple versions. I utlimately used the bottom one, with the 68R resistors, except I bypassed the large resistor (470R, 1/4W) you can see on the sync input (I decided it was better to put the resistor on the VGA input wire instead).

The 68R resistors (in series on the RGB inputs) are because this jungle chip expects RGB inputs at 0.5vpp @ 100 IRE, which is a bit less than the standard 0.714vpp. In practice, I found that if the signal is too high then it causes certain distortions (blooming and funky vertical lines) to become noticeable/worse.

So, I did some experimenting by putting a gang of three 500R pots on the RGB inputs and measuring the voltage output with a plain white screen (@ 75%) from the Nokia test program. I've used these pots in other projects. By adjusting the pots and measuring changes in voltage I estimated that there was roughly 75R impedance on the RGB outputs from the video card. This is important for the math to come.


Maths!

If math bores you, skip to the end.

There are already 75R terminators, and plugging these numbers into a voltage divider formula shows that the video card normally outputs RGB at 1.428vpp @ 75R impedance to 75R terminators, ultimately bringing the voltage to 0.714vpp. The math then says it needs 68R to reduce 0.7v to 0.5v. In practice this turned out to be almost spot-on. For those that care about the math, solve for R1, and then Rin:

Vout = (Vin*R2)/(R1+R2))

0.5 = (1.428*75)/(R1+75))

R1 = (Vin - Vout)/Vout * R2

R1 = ((1.428 - 0.5)/0.5) * 75

R1 = 139.2 (R)

We could also ask AI to get the same results.





Because we now know R1, we also know Rin:

R1 = Rin + impedance = 139

Rin =  R1 - impedance = 139 - 75 = 64 ohms

I also confirmed this by doing some testing (with pots) to see what looked best. Went with 68R for the permanent resistors as was the closest value I had.


Coming soon: Rear inputs

[Zebidee]

CONNECTING REAR RGB INPUTS (VGA+SCART)


For connecting the inputs to the mini-PCB, I used female VGA heads. The kind with little cups (for soldering the wires into) are the best. I ordered VGA heads with the cups, but the seller send me VGA heads with little pins instead 

So I ordered some more VGA heads (cups) from a different seller, but didn't want to waste more time. So I initially made some female crimp connectors for the heads with pins. These were great except that the wires tended to become loose. In the end, and after the VGA heads with solder "cups" came in, I made three different versions of the VGA input. One with pins, one with cups, and one with cups that is paired with an alternate SCART input.

The first two pictures below are VGA with pins. The second two are VGA with cups. The VGA-SCART combo uses VGA with cups. The cups work much better


   

   


For the SCART header, I originally used female crimps to connect the wires. Eventually I soldered these on to make them more secure.

With the mini-PCB for inputs installed, I can choose/swap different input sets for different occasions. Mostly I use a simpler VGA-only input when testing things on the workbench, because then I can move the back cover out of the way.

The last picture of the VGA-SCART combo input (above) also features a gang of three pots (500R) on the RGB. I keep the pots handy for projects like this. The pots were only temporary, connected while I was testing/tweaking RGB input resistor values to bring voltage p_p levels down to ~0.5vpp (discussed in earlier post). Later removed and replaced with permanent resistors on the mini-PCB.

The VGA input has a 510R resistor inside the H-sync wire, and there is 150R inside the blanking input wire (from both inputs). These are already noted in the schematic above (earlier post).


   

 


So, that covers both the basic theory and practical elements for this RGB mod. Pretty straight-forward, yeah? Hope you find this useful.

In reality, this TV was on and off my work table over about 2 months, for both tweaking the RGB mod and performing various repairs. As mentioned earlier, this TV has had a rough life (manufactured ~2003 I think), so lots to do. I did much of the fiddly work featured above while waiting for parts to arrive. For brevity I skipped over most of the repairs.

So, if you are still reading and keen to continue, I'll cover some repairs in a moment.


Coming soon: TLC and repairs

[Zebidee]

REPAIRS

Fun stuff!  Here is some of the things I fixed/improved, or at least attempted to.


NO BLUE

Once I got the basic RGB mod done, first thing I noticed was no blue. Blue gun dead? I checked by shorting the blue cathode (BK) to ground, and the screen was flooded with blue so obviously the gun was OK.

Then I checked my inputs with a DMM - blue signal was coming through fine.

Next stop, colour drive transistors. I checked with my DMM and the blue transistor gave me funny readings. So I jumped online and ordered some new transistors. then I took the chassis off and had a good look under my magnifier light and I saw...





Ah well, maybe I should have looked closer before ordering! In any case, I reflowed the solder on the entire neckboard and blue came back!

Sometimes repairs are simpler than we assume initially


CAPACITORS

First repair was to replace an obviously bulging capacitor (1000uF, 35v) near the flyback. Picture below, top right





Many other capacitors look dirty or rusty on top, but mostly not too bad, considering their history.

I went through with my ESR meter and changed many electrolytic caps, though not every single one. Most of the electro caps on this TV are Rubycons, and most are still in fairly good condition despite age, rough life and difficult climate. My rule of thumb was if I I had a better replacement (ESR, capacitance), then I would replace. Sometimes, the original Rubycon was still better than my replacements, or I didn't have the right type (handful of bilpolar caps). Regardless, If I thought a cap might be going bad, I replaced it anyway, ordering new caps as needed.

I swept over the board several times, in case I missed something.

Following this philosophy, I replaced about 80% of electro caps in the power supply, horizontal and vertical deflection, flyback, video processing and micom sections. Given that I mostly intend to use this TV as a gaming monitor, I didn't really bother much with the IF (tuner) or audio sections unless I found an obviously failing cap, though I made sure to check any such caps connected to power supply rails.


BLOOMING

Next challenge was dealing with excessive blooming and some slightly squiggly lines. Both of these issues are difficult to notice unless the right conditions are present. It may be that the TV has always had these issues, but that they were not apparent in general use with AV inputs. These improvements are small, but collectively reduce the effects somewhat, though not completely.  They include:

- Reducing the signal input voltage from 0.7vpp to 0.5vpp
- Replacing aging electrolytic capacitors generally
- upgrading ceramic capacitors on neckboard colour drive transistors' emitters (from 330p/330p/390p on RGB to 390p/390p/470p). This improves the colour response slightly, but raising the values past this creates colour bleed.
- adding a 104 mylar cap in parallel with 100uF cap already present on 9v input to neckboard.


BACKWARDS CAP, BLOWN FUSE

This one I did to myself! I put an electrolytic cap in backwards by accident. This blew a special fuse on the 14v output from the main power transformer. So, I had to order some of these fuses and wait. Also replaced the cap (assuming it was also damaged). In the picture below I've marked the bad parts with red dots. I'd already replaced the capacitor before snapping the picture below. The fuse looks exactly like a TO-92 transistor, except with only two legs. It is designed to fail quietly, without explosions/flames etc., so nothing to see. Had to use DMM to find the fault. Pictured below are old and new parts.


 


VERTICAL ROLLING, LOSS OF VERTICAL DEFLECTION

This was one of those intermittent problems, therefore difficult to diagnose. Sometimes the picture would roll, occasionally the TV turned itself off completely! The picture would disappear into a horizontal line, which made me fear it was vertical collapse, maybe a problem with the IC. Sometimes it was all good for a long time, even if you bumped the TV around a bit, but then it would come back.

I tried reflowing the solder across the entire board, but this did not elminate the issue. Some tapping seemed to suggest something was failing in the vertical deflection, but I never found anything beyond a slightly suspect solder join.

Eventually I found that tapping the PTC thermistor (for the degaussing circuit), even very gently, seemed to cause rolling. This was when I realised I'd never heard the degauss coil engaging. So I removed the thermistor completely and this improved things, but rolling still persisted.

I ordered some new thermistors (3-legs) and waited, but the seller sent me the wrong ones (2-legs)    This is the second time for this mod! This time I demanded a refund (about $2-3!) and sent the wrong thermistors back, they pay. Totally not even worth the effort, but you can guess I'm getting tired of online retailers sending me the wrong things.

In the meantime, I pulled an old thermistor off some old TV junk I have lying around, and used that for a while. Ordered some new thermistors too, and eventually they arrived and so I finally chucked out the old NOS one (it was working but falling apart) and put in a new one.

Old, dead thermistor on the left. On the right are the new PTC thermistors. The one right up the back is the NOS one, from an donor TV, that I used temporarily for a couple of weeks. I had to solder one of the legs back on, and use some tape as the plastic case was brittle with age and cracked.


 


I also found a 56uH inductor, apparently "factory" fitted under the jungle IC, which was coming loose. When I tapped on the jungle it went crazy. I've already shown it earlier, but here it is again (under the jungle on a bed/blob of glue):


 


I removed all that crap glue from underneath, sheathed the inductor in some clear heat-shrink, and re-soldered it back, a bit more snugly against the board. Now it sticks out less, and will be less prone to damage from "bouncing" if/when the board rests on top of it. Interesting story is that there is no spot/footprint on the board for this inductor, but it is in the schematic! It has been retro-fitted to a wire bridge. Sometimes I wonder what would happen if I removed it completely and replaced with a wire. Maybe less than I worry about.

So far, no more random vertical rolling or switching off


REMOTE CONTROL SENSOR

One of the first things I did with this TV was to try an get into the service mode. This requires an original remote, but I didn't have one! I had a "universal" remote that was supposed to be compatible with JVC TVs. It worked for changing channels and entering the main menu, but would not open service mode.

So I found a supplier of remotes on ebay. They claimed to make remotes for specific models of TV, and included the exact model of remote that I needed (RM219)! Picture looked right too (there is a picture of the remote in the service manual). So I ordered one of these, and around two weeks later it arrived.

To my shock, the TV didn't respond at all! I checked the new remote with my smartphone camera, it was definitely outputting a signal. I tried the other remote I have, which worked before, but now nothing. Everything else checked OK (voltages etc.), even the IR sensor did not reveal any obvious fault with the DMM, but it was very rusty. Checking everything, and applying Occam's Razor, I came to the conclusion that the TV's IR (infrared receiver) unit sensor was faulty.


 


So, I pulled an old IR receiver from a junked "China" TV, noted the pinout (3 pins - 5v, signal, ground), and connected it to the TV with a short cable to the correct holes on the board. Nothing.

Then I rationalised that the original IR unit may have a resistor or diode or something built into it, and instead replaced it with the entire IR board (small PCB about 2cm x 2cm, with an LED, cap and 3 resistors) from the China TV. This again had 3 input wires, and I mapped them over to the corresponding holes on the JVC TV board. Amazingly, it worked!


 



The new remote turned out to be the real-deal, and I was able to use it to access service mode. Hurrah! In particular I used service mode to reduce the vertical size and improve horizontal centering.

Eventually I found and ordered a proper replacement for the IR sensor unit, with right pinouts and internal resistor. Once it arrived, I popped it in (I just had to clip about a millimetre of metal on one leg for it to sit at the right height) and it worked beautifully. Now I can put the TV back together! Looks much neater, but I still miss the coolness factor of my improvised replacement stuck on the side with blu-tac.


 


WIDTH CAPACITOR

This TV had too much overscan. It is a common issue for CRT TVs used as arcade monitors, but in this case it was quite excessive. Service mode allowed some control over vertical size, but none for horizontal width.

I managed to eliminate much of this by changing a large polypropylene cap near the width coil connectors/flyback/HOT. I swapped the existing 474J for a 564J, and finally went up to a 684J. I could have gone further, but I feel the potential gains have already been made and this gave me what I needed - roughly 1.5cm extra picture.

If you're going to "try this at home", be sure to use the same type of capacitor and same or better voltage rating. This is the capacitor I changed, C574. Forgot to take a pic of the initial cap, it was 400v rated iirc.





Compare these before (left) and after (right) pics.


 


With the Nokia test screen initially (474J), you can only read "okia" on the left side, only part of the "o" visible, and on the right side it only barely gets to the "a". None of the solid circles containing "Nokia" are fully on-screen. With the 684J in place, those solid circles are now all fully on-screen (please excuse the moire!). Anyway, see for yourself!


 


The above examples show the Attract-Mode menu screen. There have been no modification to the frontend's layout. The before shot (left) does not quite fit into the screen properly, especially the window "frames" which are partly off-screen. The text at bottom is partly hidden. The after shot (right) shows the frames "pulled -in" to fully fit on-screen and all of the text at bottom is readable.

Note that these screens are displayed at 720x480i, a desktop mode which I normally expect to be partly off-screen. Most gaming modes (e.g. 240p) now actually fit on (Mostly!). For example look at the in-game screen shots (e.g. Double Dragon, Metal Slug) provided earlier.


RESISTORS

Finally, I also replaced a bunch of resistors that were looking chipped and worn, exposing the metal film or wire beneath. They all tested good, but looked horrible, like they had been deliberately bashed with a screwdriver. Oh, and someone has definitely been poking and distorting the wire bridges, WTF!!?      Below are just a few examples.





If you have made it this far then congratulations! I am impressed with your fortitude. If you enjoyed this thread, please make a post to let me know!