MOD GUIDE - Universal RGB-to-YPbPr/Component Circuit & Mod (8/24/2014)

[thesteve]

just corrected error on schematic

[turbokon]

I notice the 220uf at rgb inputs.  Is 22uf ok to use instead.

[thesteve]

Should work fine, just better white level, with bigger caps

[thesteve]

Please update first schematic again

[thesteve]

C1 C2 C3 are optional
may help with some systems, but not needed on PCE

[thesteve]

block diagram

[NightWolve]

So, I was updating this, saw you made quite a bit of changes, and then realized I don't understand the arrows to the two 10 KOhm resistors for mixing the -Y into Red and Blue that you did... What the heck is going on there ? Will have to simplify that as fewer people will understand that notation. Is it 5 K to the left, 5 K to the right, and the connecting point going to the base of the transistors ?

[turbokon]

So, I was updating this, saw you made quite a bit of changes, and then realized I don't understand the arrows to the two 10 KOhm resistors for mixing the -Y into Red and Blue that you did... What the heck is going on there ? Will have to simplify that as fewer people will understand that notation. Is it 5 K to the left, 5 K to the right, and the connecting point going to the base of the transistors ?

That's for the pots I believe.

[thesteve]

that is a standard notation for a 10K pot
it shows a resistor with unfixed center tap

[NightWolve]

Ah, I get it, so the 2 pins on one side of the 10K pot would connect to the left and right points (incoming signal + inverse Luma/Y and DC) and the single, center pin would connect to the transistor's base.

What is the default setting to initialize the pot with that's best ? Like 6.8K was the previous solid resistor choice.

Well, lemme know if you're fully done with the design, I might just work on making it look pretty soon.

[thesteve]

start right in the center, and you should be close
set the pots for clean white
in this version as the sync is separate, the Y pot adjusts contrast perfectly

[NightWolve]

in this version as the sync is separate, the Y pot adjusts contrast perfectly

I was wondering about that, did you move the Luma/Y pot before the Sync line because of my pointing that out earlier, as the old design attenuates both at the same time, or were you intending to do that all along ? Just wondering if I helped again in any way. ;) Also, I take it that when a 10K pot is centered, it is 5K of resistance on the pins that are on the same side from the center pin ?

Anyway, I'll probably start up my prettified schematic.

[turbokon]

My boards already uses the pots.  I will eventually revise the board layout for the additional sync circuit.

[thesteve]

Injecting sync after the Y pot was to improve stability (eliminate a sync drop on white screen)
The 10K pots center is 5K per side

[turbokon]

Here is the new revised component board layout.  I still need to triple check it again.  Note, original C1 & C2 are removed.  C1, C2, & C3 are at RGB or resistors can be used in its place for more universal application.

[thesteve]

change R18, R15, R11 to 500ohms
increases sync current and picture brightness range

[NightWolve]

Major update is now live!! Steve's circuit design is up-to-date as of 8-23-2014 and hopefully is in what's considered a much easier to understand image form for the masses. This is a free universal RGB-to-YPbPr encoder circuit by a fan for us do-it-yourselfer fans that works with many other consoles as well! I will continue to improve the OP, bring the parts list up-to-date and so forth in the coming days or so.

[thesteve]

change R17 mine, R12 NightWolove to 10K

[thesteve]

better black level, more consistent sync level
fixes picture washout on Genesis

Lastest:

[thesteve]

note it should work fine with 22uf ceramics on the input, instead of the 100uf im running now

[thesteve]

SCART to component converter

Lastest:

[thesteve]

do note that its good sense to put a large capacitor across the power supply input + and ground
i have a 470uf cap on mine

[crans]

@TurboKon

What components are required to add into your existing rgb-comp pcb? Or is it imposable to mod and existing board i have?

[turbokon]

@TurboKon

What components are required to add into your existing rgb-comp pcb? Or is it imposable to mod and existing board i have?

Doesn't look like it would be practical to mod the existing board, too many new components and require some new traces.  The redesign will increase the size of the board significantly unless I go with smaller SMD components which would be harder for me to work with when assembling the boards.

[turbokon]

Another revision to the board layout.

[thesteve]

check it again
the sync amp is wired wrong

[thesteve]

more like this

[thesteve]

try this one

the designators have changed
R1 AND R2 10k POT
R3 100 OHM POT
R5 10K
R17, R19, R20, R21 75 OHM
R18, R12, R6, R16, R11,R10  500 OHM
R13 15 OHM
R14, R15, R7, R8, R9,R4 1K

C1 22UF
C2, C3, C4 100UF
C5, C6 470UF

[crans]

Thank you turbokon. Seems I'll have to push the current ones I got real quick. One I can use but two others idk.....I can drop ship for you from California ?

[thesteve]

new designators to match board layout

Lastest:

[crans]

@turbokon

Looking forward to your report how this is better or worse then your current mod sold.

[thesteve]

the signals are more stable
higher clarity
better black level
better green purity
more consistent brightness over different sources and displays
higher contrast range (adjustable)

[crans]

I might have to take a trip to all electronic and make one. Now only if i can find a clone store to copy my self 10X to keep up with work.

[thesteve]

if you have the previous handy it can be upgraded much easier then building new (thats how i proto tested)

[crans]

i have two ready to be upgraded.  ill have to dig through this thread and see whats changed and work on it

[thesteve]

wow, i just looked at mine, i have a few resistors marked wrong
i had R8,  R9 2K not 1K
switched R9 to 1K and the color got even better

corrected values

2014 Lastest:

[secnine]

i picked up a rev 1 board from turbokon, but hes out on vacation atm,  so i wanted to see if anyone else had the same issue im running into...

I am using a PEXHDCAP capture card, and the picture does show up albeit dark , however as soon as i start any game play it seems it flickers and looses signal , like im draining the power as i hit the controller buttons.. i am currently testing with one game blazing lasers.

The software I am using is amerec and stream catcher and they seem to have the same issue. a strobe affect that cuts the picture off..

My settings on the capture device are 720x240 YUY2 59 FPS

I checked the voltages and seems to be low compared to some of the posts ive read.. so i dont think its a compatability issue..

I wonder if the new circuit design will help here.

heres my DMM Input readings

Sync = 4.94v
G  = 4.16v
R = 4.15v
B = 4.2 v
5v = 5.00v

Output
Y = 901mV
Pr = 2.88v
Pb = 2.61v


I am using this on a north american TG-16, after crafting my own crimp housing with jumper cables to the expansion port, so im wondering if theres any improvement in the new design that might help me here.

I don't have any other CRT to test except VGA, and a composite upscaler, so im looking around for someone who might have one to test.  I am also wondering if its a length issue of my component cables.

[thesteve]

ok yes the new design does address this issue with some games/tele's, havent tried a capture card

[Keith Courage]

try adjusting the Luma gain potentiometer. Some TVs don't like it being down too low or too high. Gotta mess with it a little to get that sweet spot so the pick doesn't flicker on and off.

I have to turn down the gain slightly or I loose picture using the TV in my living room.

[thesteve]

lens?
luma perhaps
also the new board has much less trouble with it

[Keith Courage]

Sorry I meant Luma. So used to talking to people about cd read issues that I guess my brain automatically turned Luma into Lens via my typing fingers.

[cjameslv]

......

I am using this on a north american TG-16, after crafting my own crimp housing with jumper cables to the expansion port, so im wondering if theres any improvement in the new design that might help me here.

I don't have any other CRT to test except VGA, and a composite upscaler, so im looking around for someone who might have one to test.  I am also wondering if its a length issue of my component cables.

He's back home now. I got a v.2 board on its way in the mail. I'll be sure when i install and test it and let you know. I have a capture card that doesn't get much love so i am looking forward to trying this out.

[Helder]

So I built a through hole PCB based off the latest schematic on page 1 and for the most part it works great except that there is too much green where there should be black. For example the TMS intro screen has a strong green outline and I've played with the pots and nothing fixes that green outline. I didn't use the same resistor values but it was close enough like 500ohm I used 512ohm and the 15ohm I used 12ohm. Thanks for any info.

[NightWolve]

Hi Helder, first off, thanks for going the DIY route instead of buying a board which was the intent of this thread, to not have to buy some JROK or similar board and learn to build it yourself. Very few participants have tried to do so and everyone else just resorted to buying turbokon's boards. I found that disappointing, but I understand, not enough has been done to reduce the first impression complexity of the circuit. The OP is very unfinished and more work on better presentation and guiding is needed which I hope will happen over time.

Anyway, I have a quick experiment for you to try. I dunno if it'll help, I wanted to be set up with my scope before getting there, but I am behind on it.

So, for relevant background, the formula for producing a proper Luma/Y signal for YPbPr when the resolution is 240p/480i/480p is something like this: Y = .299 (Red) + .587 (Green) + .114 (Blue). It's different if the resolution is 720p and up and the TV's circuitry would expect a different mix as a result.

Given that formula for Y, those first 3 resistors (R1-R3) for the RGB lines are what condition our Luma/Y mix to get to that standard which most TVs will expect. Now to translate it to precise resistance values is the trick/thing!! The way I've seen it done was based on the first feedback resistor to the first amplifier component which inverts the signal (thus, gives you a 180 degree out of phase signal, hence "-Y" and why the P [for Phase shifted] notation is used for P_blue and P_red). The properly selected feedback resistor is 324 Ohms in that case, so for R = 324/.299 = 1070 Ohms, G = 324/.587 = 549, and B = 324/.114 = 2940 Ohms.

Well, Steve's chosen feedback resistor is 500 Ohms and so if blindly using that same formula, then:
1) Red = 500/.299  = 1672 Ohms
2) Green = 500/.587 = 852 Ohms
3) Blue = 500/.114 = 4386 Ohms

So to get to the point, for R1 on Steve's schematic (page 1 of thread), I want you to try 1672 Ohms, R2, 852 Ohms, and R3, 4386 Ohms. R2 going to 852 Ohms will weaken the green by a bit from the 500 that it's at now, but since this calls for weakening the others, still couldn't guess if that would bring the 3 signal levels to the proper balance that the spec calls for. You see, I don't have the special insight when it comes to the S8050 transistor and how that relates to the chosen feedback resistor of 500. Steve does, but he didn't closely follow the spec and went by his eyes in this area I believe. The circuit that I learned this math from was using an op-amp and the feedback of 324 Ohms was like a special sweet spot for optimal operation of it and what not... I am blindly applying it here with transistors so no guarantees obviously.

For thread reference/research, here is said circuit/schematic, but note that it's not complete as they do not inject H/V Sync (unlike Steve's circuit) to produce a working Luma/Y (left up to the user):


Well, if trying my idea, you'll obviously need like 2 solid resistors in series or 10k pots to get those non-standard values (like 4386 Ohms) for all 3 signals, but I wanted to offer those values up for a test since I haven't done it myself. The short answer is though you can simply increase the resistance value of R2 which is for green obviously. This assumes you did the whole circuit correctly as it is tricky and one soldering mistake will seriously throw things off!!

[Helder]

Thank you for the speedy reply! I should have given a little more info in the previous post, I'm generally trying this on a Sega Genesis Model 2 on a SD TV so it should take 240p without issue as I have done so in the past with a different circuit. The board I made is actually a fabricated PCB I made so it would be real simple to mod and alter as needed so that I can later shrink it smaller using SMD parts. If anyone here wants a simple DIY board to try this circuit I can sell the remainder (9) for $5 plus $2 shipping if in USA.



So I guess as a safe bet in trying the values you posted is to use a 5k pot on the 3 input lines in place of the 500ohm and 1k resistors, and adjust them close to the values you mentioned and take it from there. I will post back the results tomorrow after using the pots, thank you again.

[thesteve]

another thing i found running geni is the black isnt near as good, due to a wider sync pulse
the 15ohm brought it close to the PCE, but a slightly larger resistor could serve you better (or a resistor to ground from sync input)

[NightWolve]

So I guess as a safe bet in trying the values you posted is to use a 5k pot on the 3 input lines in place of the 500ohm and 1k resistors, and adjust them close to the values you mentioned and take it from there. I will post back the results tomorrow after using the pots, thank you again.

Aha, I thought so, you have 1K for the blue line! That was a mistake I caught for Steve after studying the SDTV formula for Luma/Y. If you look at the schematic again, you'll see it was changed to 2K per Steve's judgement and my insistence that 1K had to be wrong because blue is supposed to be the weakest signal in that mix. But yeah, I'm a perfectionist and want to know the math to compute the exact resistance values for what those 3 input resistors should be, which is why I shared my research and suggested updated values. For now, the official circuit/schematic photo has Steve's generic choices, 1000 Ohms for Red, 500 for Green, and 2000 for Blue.

That's a nice looking layout on these boards of yours BTW and looks like your pricing is pretty fan friendly. Thumbs up! You could probably stuff it in a regular envelope, and slap 2 forever stamps on it for shipping in the US! Heh. I bought a SDRAM module for $3 bucks on eBay, he charged $2 for shipping and that's what he did, sent it to me in an envelope with 2 forever stamps! So shipping cost him .75 cents or whatever, the sneak! Anyway, trouble here is Steve may resdesign things again down the road, but at least this issue is just a matter of resistor choices.

[cjameslv]

My v2 board arrived today. I normally build alot of my mods but this one was so slim and sleek i decided to get the pre-made and forgo the building stage and go straight to the installing one. I'll post my thoughts on this board once i'm done installing & reviewing it.

UwtCfg.jpg

[turbokon]

My v2 board arrived today. I normally build alot of my mods but this one was so slim and sleek i decided to get the pre-made and forgo the building stage and go straight to the installing one. I'll post my thoughts on this board once i'm done installing & reviewing it.

/5bd0kz.jpg

Awesome, notice R19, R20 and R21 are empty.  Those are for other systems application.  Thesteve already confirmed 75ohms there will work for the SNES and Genesis.  We still need to test for other systems.

[cjameslv]

Awesome, notice R19, R20 and R21 are empty.  Those are for other systems application.  Thesteve already confirmed 75ohms there will work for the SNES and Genesis.  We still need to test for other systems.

We have some great guys on this forum devoted to develop and figure out these mods. I really appreciate it!