RGB->Component Schematic

[Segasonicfan]

Here's a great RGB->Component converter schematic but I notced there isn't a sync input (?!)  I was google searching this and someone made a mention that this would be for a sync on green input...but do game systems output this?  Or is there a way I can encode sync into the green line?  Any help with this is much appreciated!

-Segasonicfan

[bk2099]

maybe this will help

http://www.raphnet.net/electronique/sync-o...on-green_en.php

Code Select Expand


            100 uF
         ! !
Green ----! !---------------------------------
        -! !+                                !
                                             !
                                 680 ohms    !
                                             !
HSync/CSync ------------------------/\/\------o---------  CSync on Green


[Segasonicfan]

wow, that's perfect- exactly what I was looking for.  I'll build the circuit now and hope for the best!

-Segasonicfan

[phreak97]

let me know if this works, i'd LOVE an rgb to component adaptor.. i love my tv, yet it has only composite and component.

[viletim!]

This circuit has no video clamp, it won't accept an AC coupled video input.

[Segasonicfan]

viletim, what if I juse clamp the RGB lines using a CXA1645?  Would that make it work?  

Also, please check my other post (the NTSC encoder one) as I was asking for your help there too...

Thanks!!

-Segasonicfan

[viletim!]

Well...the  CXA1645 does indeed have a clamp in there but it would be wasetful to use the chip just for it's clamp :). Taking a look at the circuit again I can see more problems...From what I can see (where's the artical that goes with the circuit?) it'll only accept 75 ohm, DC coupled, sync-on-green, RGB video. I guess it's for converting video from a computer (w/ sync-on-green) to something that can be displayed on a high defintion TV (Plasma, LCD, etc).

Fortunately someone has designed a circuit to do exactly what you want - convert AC coupled RGB + Video to Component. http://elm-chan.org/works/yuv2rgb/report.html see the bottom of the page. I've not built it but it looks good to me. If you already have some MAX4451s then you can use them in this circuit.

[Segasonicfan]

Thanks for the reply

I don't mind using a CXA1645 to clamp it for right now...I have plenty to spare and I'll be combining this inside a supergun with a CXA1645 anyway so the clamp outputs are readily available

As for the other problems, the previous post shows a schematic for combining the sync with green, should that work?  As for the 75ohm DC coupled output, I think the CXA1645 RGB (clamped) outputs provide that given I throw some 75ohm resistors in there, right?  Let me know if this will work, it seems like it might....


As for the other circuit, I already built that one (what a HUGE pain in the ass with those obscure resistor values) and I couldn't get it to work.  I have a ton of MAX4451s around so you think I could drop them in for the same circuit?  Ugh, I'm just not looking forward to working with it again as the other circuit is SO much simpler.

Also, the NPN transistor in the later schematic is obsolete so I used a NTE2356...do you think that may have caused the problem?  I think (but I am not 100% sure) that the two are complimentary.  My knowledge of transistors is very limited  

Thank you SO much for your help, I have been trying to build a component converter for a very long time and I really appreciate the assistance

-Segasonicfan

[Segasonicfan]

also, I forgot to mention, here is the full article for download (see the word file):
http://files.filefront.com/rgb_converter_i...;/fileinfo.html

It's blocked on the web so I had to upload it.  I found the instructions to be terribly unhelpful though :/

-Segasonicfan

[viletim!]

The article is about an RGB > component circuit (which I've actually read before in the printed magazine). The component > RGB circuit/article must have been published in a different issue.

The clamp inside the 1645 isn't really suitable because it clamps at about 2v and the circuit wants video clamped at 0v. You've also got the problem of having no sync on the green. Bk2099's circuit isn't good enough, you'll have to clamp the green signal (you need to clamp an AC coupled signal to do pretty much anything to it) and put it in properly with an opamp. This circuit really is beyond hope for this application.

I think you should revisit the ELM circuit. The resistor vlaues in the multiplexer stage are made by putting two standard values in parallel. See the table to the right (// is the symbol for 'in parallel with'). The RN2202 is a 'digital transistor' which is basicaly a standard transistor and two resistors in one package. You can make an RN2202 equivilant with a BC558 (or your favorite PNP), a 10k resistor connected between the base and emitter, and a 10k in series with the base. Though it seems the NTE2356 is a direct replacement. The max4451s should work fine - they have the bandwith and can run from a +-5v (10v VCC-VEE, just below the max of 11v ) supply. Make sure you put a power supply decoupling cap near each chip or you could end up with a fast oscillator.

[viletim!]

well...I actually sat down and built that ELM circuit for myself, with MAX4451s no less (I was going to use them for a macrovision remover but never bothered). It seems to work, something resembling component video comes out the other end but i'll have to wait till the middle of the week to try it on an actual television. One problem I did notice was that the '4451s doesn't like having a high Z (impedance)  on it's input. In this circuit, when there is no input signal, the Z in of the opamp is pretty much infinate. This causes the output to swing to the positive supply rail, wastes power, and gives some strange outputs (there's +5v on the Y output :/ ). Maybe I can fix it but i'll see if the picture on the output looks ok first.

[Segasonicfan]

great to see that you're working on this  If the MAX4451s aren't working maybe you should try the LM circuit in the schem?  That's what I used but I think I screwed up the PS cause I couldn't get a decent picture out of it.  Also, do you know what voltage the output should read if it is working properly?  That would help.....thanks

-Segasonicfan

[viletim!]

haha...you actually built that power supply?? It generates -5v from a +5v supply so you have +-5v. It's a switcher, and switching power supplies (esp ones built out of discreete parts) are best avoided unless you've got a good reason to use one. I just built a simple linear supply (see pic). I used a transformer with a single 9v secondry because I can get them really cheap as surplus.

Unfortunately, the LM6172 isn't available localy as far as I know. I think I can tame the '4451.

[Segasonicfan]

oh I didn't mean I built a full switching power supply from scratch lol.  I used a 7905 and 7805, but I think I screwed something up with the power b/c I've checked all my work over and over and I can't find any problems.  I'll just end up building it again with the 4451 chips.  Let me know when you make some headway on this, I don't want to spend more money on parts till I know I can actually get the friggin thing to work  Thanks,

-Segasonicfan

[viletim!]

I tried out the ELM circuit on a plasma today. It works quite well except for some noise in the picture which I beleive is due the 4451s being unstable in this type of circuit. If you try this one again the I reckon you should avoid the MAX4451 it's not as good as it appears :/.

[Segasonicfan]

damn....how bad was the noise?  maybe the power lines just need to be regulated better?  just trying to be hopefull

So should I just go with the LM6172?  I can order some from national, or do you recommend I look for a better opamp?  

Also, can I use the clamped RGB outputs from a CXA1645 instead of the 74h04053?  and what did you use for RN2202?

I'd also be interested in pics of your work if you could possibly post them

Sorry for all the questions, but you're the only person that's helpful at all with this.  It's much appreciated

-Segasonicfan

[viletim!]

right....it seems the plasma I tried out the circuit on had a faulty component video input - the same noise is present when a DVD player is connected. Today I tried it on an LCD TV and got a perfect picture .

It still goes a bit spastic when there's no input which i'm sure is due to the 4451s (and i'm sure it's easily fixable).

I dunno about using the CXA1645's clamp - the idea sounds silly to me. You can't control it's clamp voltage so clamping at ~2v may cause a DC offset on the output. I used a standard cmos 4053 in my circuit in place of the 74HC version, it seems to be fine.

[Segasonicfan]

wow thats great news  Please post some pics if you can, it would really help me to see a finished board (for building my own).


-Segasonicfan

[viletim!]

Here's a picture from the top andbottom.

btw, an RN2202 is a transistor and two resistors. Get a datasheet for it from datasheetarchive.com

[Segasonicfan]

*whew* finallllly got this sucker working...or at least 90%+ working....

I had to tie C sync directly to the output to get a picture....which is kinda strange.  My Compnenet TV won't show anything without a sync signal present so this was the only way around it.  Do you think the NTE2356 was a bad call?  Maybe that's why the Sync didn't mix into Y.  

Either way, the picture is great minus the colors being ever so slighlty off (a little bit of a yellow tint).  But I'm pretty sure some 500pots on the RGB lines will fix this  There's also a *tiny* bit of interference (some non-moving very thin lines in the background) but this is probably due to my horribly unregulated power supply.  I'll make these changes and post here with the results.

Thanks so much for your help!!

-Segasonicfan

[Segasonicfan]

okay, I added the 500ohm pots to the RGB lines and regulated the power supply....

Now it's perfect  100% beautifullllll Component.  I'll be posting some pics here and on my site of the board later.  Still seems strange that I have to tie C Sync to the Y output, but hey, as long as it works, right?  If you know why this is though viletim, please let me know

-Segasonicfan

[Segasonicfan]

Okay pics are up on my site....I added all my little touches like usual.

I tried the encoder on some more video-finicky systems and I got some undesirable results.  My other arcade board (Raiden 2) doesn't show video output at all and neither does my Sega Genesis.  The Sega CD shows output on some parts of the game....(wtf?)  
Help Viletim... :unsure:

-Segasonicfan

[viletim!]

Fuck! that's a lot of solder!

You're main problem (no sync) seems to be with the part you've put in place of q4. Have you tried the equivalent circuit in it's place? The off colours could be caused by connecting the sync signal to the video (make the green go funny) or it could be caused by you using carbon resistors instead of metal film type.

On you website under 'application tips' you describe the quescent voltages present in the circuit. Because you used the '4451s too your circuit is behaving like mine, not how it's meant to. When there's no signal present you should measure 0 volts at the inputs to the opamps and 0 volts on all video outputs. Nothing goes in, nothing comes out. Instead, what is happening is the buffering opamps (first stage) are sort of self-biasing themselves and swinging their outputs to the V+ rail. When you hook up an input, clamp circuit keeps the input from floating off.

I haven't had a chance to look at it too much but thinking about it now, i've come up with two solutions. First is to increase the value of each coupling capacitor from 100n to 1u and add a 47k resistor between each input and ground right before the input to each opamp buffer stage. Second idea is to detect the absence of a composite sync signal coming from the LM1881 and force the back porch output low. This will make the 4053 clamp all the time, essentialy shorting the inputs of the opamps to ground. A circuit like this should do. It made from a one-shot that's retriggered everytime a sync pulse comes a long, if it doesn't get a pulse for 120uS then it'll 'time out' and turn on the transistor, which shorts the lm1881's back porch output to ground (it's safe to do this due to the nature of the lm1881's outputs).

[viletim!]

oh, and the IC in the picture is the 74LS123.

[Segasonicfan]

Thanks for the super-informative reply (as always)

Yes, I tried an equivellant circuit in place of the RN2202...I put a PNP wiht 2 10k resistors like the datasheet shows and it worked  I guess something was just funky with the NTE replacement.

The colors have always been good though (I know my site implies otherwise).  I know I was wrong about the +5v on the output on my site too, I have to update it to say only with the MAX4451's.  

After some more testing on a different TV (actually a projector) I found that the board works fine with a Genny signal  My TV is just a piece of crap.  

The image was FAR better using the projector because I was no longer working with my grabage RCA TV.  Everything is crystal sharp, there's zero interference, and the colors are great.

My only complaint left (and it's a small one), is the strong luminosity in the signal.  I noticed that Component tends to wash out details within bright (white areas) and this board does the same.  I think this is just a flaw in the format, but I'm thinking of putting 220uf caps on the output to try and fix that.  

As for the circuit you showed me...what does it do exactly?  Is it just to remove the picture noise when there's a lack of input?  Or does it do something else?

-Segasonicfan  

[viletim!]

I put the 74LS123 addion onto my circuit and it behaves much better now. The power regs are cold in the quiescent state and only slightly warm when there's a signal on the input. As for what the addition does...it's fairly simple...

The 4053 behaves as a switch, switching the input (through the clamp capacitor) to ground during the back porch period. This looks to see if there's a composite sync signal coming from the LM1881. If it's not there then it hijacks the LM1881's back porch output, forcing it low. The 4053 switches on (it's been uninhibited) and shorts the input to ground utill the LM1881 spits out a valid sync signal and recovers everything.

It's not an ideal soultion, but a definate improvement. If anyone else is thinking of reading this and thinking building the RGB>YUV circuit then go for the LM6172 - they're (probably) a lot more suitable.

[Guest]

ahhh i see, you wanted to cool down those regs huh?  yeah they do get damn hot, you noticed I had to put a heat sink in there.  

This is very helpful...any chance you could simplify that schematic for me so I can try it?  I just need to know what pin you're using on the 4053 and what the heck that 120u is lol...if you could make a simpler one for less-advanced peeps like me it would be much appreciated

As for using the 4451's...it's a shame they aren't better suited cause they're free (samples)  My local electronics place (Jameco) doesn't sell the LM6172, do you think the LM6171 would work?

-Segasonicfan

[viletim!]

Yeah, the circuit is a bit rough...
There's only one connection between the LM1881 and the 4053 on the ELM circuit...it shouldn't be too hard to suss out. The box with 120uS written on it is one half of a 74LS123. It's a retriggerable one-shot, see the datasheet for pinouts. The 120uS is time duration of the one-shot with those specific values of resistor and capacitor. The upward pointing arrow is a connection to +5v and /CSYNC comes from pin 1 of the LM1881.

From philips' website LM6171 looks just like LM6172 except there's only one device per package so you'd need twice as many of them.

[Segasonicfan]

oic, the 4053 connection is just the ELM connection (I thought you added a new one).  I see how the circuit works now =)

What are the pins on the 74LS123 though?  I can guess from the datasheet but I'd like to be sure... pin 1=CSync, pin 2=+5v, pin 4=Q to the BPO line, pin 6=10nf to pin 7, pin 7=33k to ground?

That's just rough guessing from your drawing and the datasheet...thanks,

-Segasonicfan

[viletim!]

Well...there are two devices per package....pick one of them. If you want to use device 1 then its Rext/Cext connect to pins 14 and 15. Both upward pointing arrows  indicate a connnection to +5v (Vcc). Also remember to connect Vcc and Ground (pins 8 and 16). And while I think of it, CLR (pin 3 if device 1) should be tied to Vcc too.

[kschafer2598]

Here's a great RGB->Component converter schematic but I notced there isn't a sync input (?!)  I was google searching this and someone made a mention that this would be for a sync on green input...but do game systems output this?  Or is there a way I can encode sync into the green line?  Any help with this is much appreciated!

-Segasonicfan

Sorry to bump such an old topic, but does anyone still have a pic of this schematic?

[Segasonicfan]

There's an updated schematic on my website (I think viletims too): http://www.freewebs.com/vgmods/arcadeworks.htm

[jamisonai]

I've looked over the updated schematics on your website and I'm confused about how this thing gets power.  I assume the ICL7660 provides power, but I don't understand how since its not connected to the rest of the circuit.   Can anybody explain that to me?  Also I see it requires 5V, I suppose its my option how I get 5V into it using a combination of AC Adapters and power regulators? 

[kschafer2598]

There's an updated schematic on my website (I think viletims too): http://www.freewebs.com/vgmods/arcadeworks.htm

Thanks a ton! Sorry for the late reply, I've since stumbled upon this one (http://elm-chan.org/works/yuv2rgb/rgb2yuv.png), but yours looks easier to build.

Karl