Where can I tap a sync signal for s-video?

[lastcallhall]

Please bear w/ me - I'm currently one handed and on a lot of pain meds coming out of shoulder surgery...

Last week I decided to build a s-video/composite circuit out of a CXA1645 like I've done at least a dozen times before, only this time I planned on sticking it into a TG-16. I built the circuit, attached the RGB IN locations on the chip to the RGB OUT locations on the TG-16 EXT connector (23A, 23B, and 23C), along with the SYNC IN on the chip to the SYNC OUT on C22. everything is properly attached to their respective components and grounded. In addition, I tapped the Audio from 1A and 1C, and grounded them to the board, as well. There's also a reset button and region chip installed, as well. I powered the console on in s-video and I get audio just fine but no sync on s-video. I try it with composite, same thing. I test it with my turbo booster and I get good video and audio. So, I figure my problem is with the sync pin. After some researching I find that it's actually labeled composite sync, but that shouldn't make a difference, right? Sync should be sync, as far as I understand it. I'm hesitant to take the system apart again, now that everything is soldered and screwed into place, so is there anywhere on the bottom surface of the board that I can tap into?

Thanks!!!

[SignOfZeta]

The only thing you can get from a PCE is composite sync, and it's not the best example of composite sync out there. If it doesn't work for you...sorry, I have no advice as I've had nothing but trouble with it.

[lastcallhall]

Hm. Ok, thanks for the info. I think i'm going to try and tap directly off of the C6260A and see what happens...

[ogre]

To get a good sync for the CXA use a LM1881N. Put the composite video through this circuit and it will output composite sync.http://www.national.com/images/pf/LM1881/00915001.pdf This is needed as the systems own c-sync does not put out a useable signal.

[lastcallhall]

To get a good sync for the CXA use a LM1881N. Put the composite video through this circuit and it will output composite sync.http://www.national.com/images/pf/LM1881/00915001.pdf This is needed as the systems own c-sync does not put out a useable signal.

Fantastic, thanks!

[Keith Courage]

So does this mean by using the said IC and using that signal along with the typical RGB outs you can run it to an S-Video connection? I have also been wondering how to do this myself for a while now.

[lastcallhall]

That's what I'm hoping for...

[lastcallhall]

So, I have a quick question: I understand the schematic except for the RSET pin. What is it, and what is it used for? More importantly, do I need this pin for what I want to accomplish? Thanks!

[bust3dstr8]

Rset=Reset

[lastcallhall]

Rset=Reset 

That's what I thought, but I never assume anything in electronics, LOL.

Also, I was looking through the data sheet and saw that there's an application circuit diagram below the connection diagram. I'm thinking I should use the app circuit instead of the connection circuit?

Reference: http://www.ti.com/lit/ds/symlink/lm1881.pdf

[grahf]

The connection diagram on page 2 is all you need. I used this chip for my Svideo boards I was making a few years ago. Also, you didn't mention anything about the 3.579 MHz frequency input the CXA1645 needs. I guess you picked up an oscillator? If you don't have that hooked up the circuit won't work. They are available in half size dip packages, like this:

[SignOfZeta]

Question: if I have a JROK or NeoBitz s-video encoder can I just wire that directly to the internal RGB in a PCE or does it need amplification?

[thesteve]

the sync on the TG/PCE is a good signal, its just weak (needs amplification)
i use it on the ad722/ad724 chips for my TE/GT screen mods

[lastcallhall]

The connection diagram on page 2 is all you need. I used this chip for my Svideo boards I was making a few years ago. Also, you didn't mention anything about the 3.579 MHz frequency input the CXA1645 needs. I guess you picked up an oscillator? If you don't have that hooked up the circuit won't work. They are available in half size dip packages, like this:

Yeah I have a half can oscillator on there.

Thanks for clearing the diagram up - I was going crazy trying to find a bi-polar 510pF capacitor.

[lastcallhall]

An update...

This is cross-posted from Sega-16.com, so I apologize in advance for any repeats in information.

Long and short of it is, is that I bought a Duo-R a couple months back to replace my TG-16/Turbobooster combo I've had for a little over a year. Well, the Duo is in need of a CD lens replacement, so in the mean time I thought I'd try and mod the TG-16 for s-video and audio out, and see what comes of it.  I was able to look at the documentation shown over at GameSX.com, and from there took the proper signals and fed them into the CXA chip, composite signal included, on top of adding a region chip (Obey chip from BlueBMW and TheSteve from pcenginefx) and reset button. Once I had everything hooked up, I powered it on and got this:

As you can see, the sync signal seems to be trashed. I've done a little research and found the following:

1) The sync signal from C23 on the EXT port is labeled composite sync, as is pin 44 on the 6260 CPU of the TG-16
2) Knowing this, I purchased a LM1881 Video sync separator, and fed the a) the composite signal from the EXT port, b) the composite sync signal from the EXT port, and c) the composite sync signal from the 6260 into the chip (all at different times, obviously), hoping to get a clean sync signal out of any one of them, but no such luck. This was actually suggested by someone over on the PCEFX forums.
3) I have several 4558D OP AMPs lying around from an earlier CCAM mod attempt, but I'm not sure if I can use them to amplify this type of signal. even so, I don't have the slightest clue as to how to get it done. 

I'd really like to get this up and running (and sold...), but I seem to have hit a wall on this. I'm half tempted to take it to school so I can look at it all in a lab with proper test equipment, but if I could solve it prior to that, it would be even better. So what do you all think? Any help is always appreciated. Thanks!

[thesteve]

the 4558 is way too slow (tried it)
a transistor amp can be built that works well.
the 1881 also works, but must have a capacitor input, duo to the offset in the chip

[lastcallhall]

the 4558 is way too slow (tried it)
a transistor amp can be built that works well.
the 1881 also works, but must have a capacitor input, duo to the offset in the chip

Hm. I put a 0.1uF cap on the 1881 composite pin and it didn't change anything.

I think tomorrow or sometime soon, I'm going to rebuild the circuit from scratch. This can't be that complicated.

I just hate having to undo everything and start all over again... >=(

[thesteve]

a good way to confirm the sync sig is to hook it to a tv.
the tv will see the sync as a black picture (not rolling)

[lastcallhall]

a good way to confirm the sync sig is to hook it to a tv.
the tv will see the sync as a black picture (not rolling)

So, here's what I did, and my observations:

I took the csync signal from the 6260 and added it to the composite out jack on the system and the image stabilized, albeit in B/W, which leads me to believe that the sync signal is good on the system, I just need to rework the CXA1645 circuit - something must have gotten messed up somewhere.

Parent/teacher conferences tonight, so I doubt I'll get to this until morning, but I'll figure it out...

[thesteve]

the 1645 needs the sync to be TTL levels (0 to 5V switching)
you have about 180mV off the chip

[lastcallhall]

the 1645 needs the sync to be TTL levels (0 to 5V switching)
you have about 180mV off the chip

OK, so what would you suggest I do? Is using an op amp for the sync going to be enough? Should I use the RGB/SYNC shown here:



then feed it into the chip?

I'm kinda lost right now and not sure which way to go.

I'd appreciate the help.

[thesteve]

nope that circuit wont work at all

[SignOfZeta]

OK, seriously, there are a lot of s-video modded Duos out there. Someone must know how to do it.

Also, my question got buried back there: will a Neobitz work directly from the Hu chip or do I need something else? I also have a JRok.

[thesteve]

ok i dont know anything about neobits...........checking......................nope
not enough info on the jrok.

most have used the 1881 sync stripper on the composite output.
the TE uses a common emitter transistor amp, which causes +sync of TTL levels.
the best way is comp sync from the chip--->common emitter transistor amp----->logic inverter------>encoder ic

[grahf]

SignOfZeta, I've always heard that the JRok kits don't work for whatever reason, and the guy who makes them doesn't know why. Never tried one myself, though. I have no idea about the Neobitz.

lastcallhall, seriously, stop jumping around all kinds of hoops and just go with the LM1881. It's pretty much the standard way to do this, and it works. Everyone uses it for everything. Here is a picture of a circuit from the gamesx page:

Original link:

https://www.gamesx.com/misctech/lm1881.htm


It's incredibly easy, and you only need like 50 cents worth of components. You are ONLY going to use pins 1, 2, 4, 6 & 8.


Composite out is pin 1, which is what you're after. This goes into your CXA1645.

Composite Video goes into pin 2, NOT composite sync from the HU6260. The easiest place to get composite video is off the back of the original AV out plug. Just solder a wire to the right place on the inside of your Duo. You need to put an 0.1uf capacitor in-line, which looks like this:   and will say "104" on it.

Pin 4 goes to ground. Make sure the LM1881 and your CXA1645 circuit share the same ground. The grounds for everything should end up connected to the ground on your Duo's motherboard.

Pin 6 is the "set" pin. Not reset. You need another 0.1uf capacitor, the same one as above. You also need a 680kohm resistor. Make sure the value is correct. You hook both of these parts to pin 6, then the other side of them goes to ground. The same ground as everything else.

Pin 7 is 5v input. You can use any value up to 12v, but it's easy enough to get 5v off of the Duo motherboard, and I personally use 5v.


Good luck, and if this still doesn't work please post pictures of what you have currently made.

[lastcallhall]

Good luck, and if this still doesn't work please post pictures of what you have currently made.

Good deal, that sounds workable. Thanks!

Quick question, though - I'm doing this on an american TG-16, not my Duo. Would the video signal from VIDEO A23 on the EXP port suffice for pin 2 on the LM 1881?

[thesteve]

Quick question, though - I'm doing this on an american TG-16, not my Duo. Would the video signal from VIDEO A23 on the EXP port suffice for pin 2 on the LM 1881?

yes

[lastcallhall]

 

Still didn't work.

Here are some closeups of what I'm looking at:

Here's the LM1881. Solid green is video in from pin A23 on the EXP port, striped green is sync out, solid blue is 5V in and the two striped blue are tied together and are connected to ground.


Here's a shot of the voltage regulator. You can see the striped blue wire coming from the LM1881 to the center pin which is ground, along with another ground wire from the CXA1645, as well as the solid blue 5V wire from the LM1881, and the 5V line from the CXA1645 (Will show shortly) connected to voltage out, which should supply a 5V voltage to all components.


Here's the CXA1645 itself. Pin 1 is grounded, pins 2,3,&4 are RGB respectively, pin 5 is No Connection, pin 6 is connected to pin 3 of the oscillator (shown above the IC), pins 7, 12, & 19 are all tied together, and fed into pin 4 of the oscillator, which is connected to the voltage out on the regulator (shown in the previous pic). Pin 8 is no connection, pin 9 is grounded, and pin 10 is sync in. Pin 11 is no connection.


Bottom half of the 1645. Pin 13 is connected to a 47KOhm resistor and a 0.1uF capacitor, which are then each ran to ground. Pin 14 is connected to a 10uF capacitor connected to ground, while pins 15 and 16 are connected to a 75ohm resistor and a 220uF capacitor in series each, connected to chroma and luma respectively. Pin 17 has no connection. Pin 18 is connected to a 20kohm resistor out to ground. Pin 19 has been discussed earlier, pin 20 is also connected by a 75ohm resistor and 220uF capacitor in series out to composite out. Pins 21, 22, & 23 are not used for this circuit. Pin 24 is connected to ground.


Here's a shot of the ground wires all bundled together and fed into a single wire for organization's sake (note: my finger is on the composite out wire; the striped orange wire ABOVE my finger is connected to the bundle of wires running to ground:


So... here we are.

[thesteve]

check if the output from the 1881 chip will give a raster (sync no-pic) on your monitor.
another note when you have multiple grounds tie them together at the chip

[lastcallhall]

check if the output from the 1881 chip will give a raster (sync no-pic) on your monitor.
another note when you have multiple grounds tie them together at the chip

OK, I'll take a look shortly, thanks.
Also, I tied together several groups of ground @ the chip - there's something like 12 grounds on the unit and I got them down to four before tying them up to one on the way out...

[thesteve]

you really want the grounds tied as close to the chip as possible, as you can see in this pic i have them jumpered across the chip.



BTW that chip outputs composite and S-Video

[lastcallhall]

No raster signal when the output is applied directly to the composite out - it still flickers.

I even pulled the video signal from A23 on the EXP port and fed it straight to the composite out - no change.

[thesteve]

i just pulled S-vid strait off the huc6260.
i needed to amp the Y output and filter C out of composite

[grahf]

No raster signal when the output is applied directly to the composite out - it still flickers.

I even pulled the video signal from A23 on the EXP port and fed it straight to the composite out - no change.

lastcallhall, your descriptions of what you have going to each pin looks good. I compared it with the CXA1645 datasheet, and everything is right. The only possibility is that one of your chips is fried, which isn't that likely, or you mixed up a wire somewhere.
You said you were getting composite video signal off of pin A23 on the expansion port. That's the wrong pin. Composite video is on pin A22. See this for the pinout:
http://www.gamesx.com/misctech/pcebp.php

[thesteve]

ya could go simple, and just pull S-Video right off the huc6260



the cap is 0.001uf
the transistor i tested with is an S9013
40 is pin 40 of the huc6260
comp is anywhere you can get composite video (I tapped the vid pin on the RF modulator)

[lastcallhall]

No raster signal when the output is applied directly to the composite out - it still flickers.

I even pulled the video signal from A23 on the EXP port and fed it straight to the composite out - no change.

lastcallhall, your descriptions of what you have going to each pin looks good. I compared it with the CXA1645 datasheet, and everything is right. The only possibility is that one of your chips is fried, which isn't that likely, or you mixed up a wire somewhere.
You said you were getting composite video signal off of pin A23 on the expansion port. That's the wrong pin. Composite video is on pin A22. See this for the pinout:
http://www.gamesx.com/misctech/pcebp.php

Oh, crap.

Someone on the Sega-16 forums said that something was probably overlooked. I'm tapping into the wrong signal from the EXP port...

I'll play with it now and see what I can do.

[lastcallhall]

OK, it still doesn't work, but I have it narrowed down to the chip...

From another member on Sega-16, I constructed this circuit prior to running the composite signal into the LM1881:

Connect them in series and wire them like this:

Code Select Expand

Gnd --|===|--#--|===|--- CSync
       680   |   10k
             |
             |---||---» LM1881 CVideo input
                .1uF

LOL @ ASCII schematic :V

Take care, I edited this message like 4 times :p

[/FONT]

I purposely left the leg on the 10k resistor connected to composite uncut and when I connected it directly to the composite out, it stabilizes the picture, but the colors are all out of wack. I tried to take a picture and video, but it doesn't come out the way I see it in person. Anyway, I'm fairly sure that the CXA chip is the culprit, as I can get a solid sync signal from the EXP port, the C6260, and from the composite out on the EXP port, as well. So, what I'm going to do is I'm going to order a new CXA1645 from a stateside dealer and just redo the entire circuit. This may take some time, so I'll check back in once I get that done.

[thesteve]

skip it, you dont need an encoder, as the TG16 outputs S-Vid almost native.
you will need a driver amp for the Y (luminance) and an isolation cap for the C (chroma)
see simple schematic above

[lastcallhall]

What the hell, I'll give it a shot.

[ApolloBoy]

What the hell, I'll give it a shot.

Let me know how it goes, I'd love to give that mod a shot if it works!

[thesteve]

i just tested it before posting

[lastcallhall]

What the hell, I'll give it a shot.

Let me know how it goes, I'd love to give that mod a shot if it works!

Will do. Looks like I'm going to have to order the S9013 - All I have are 2SC945 and 2N3904 transistors lying around (unless one of those will work, but I'm guessing not).

Off to Digikey...

Also, @ TheSteve: which pin did you tap into on the RF modulator for video? I took my modulator off to make room for the other mods. Thanks!

[ApolloBoy]

Looks like I'm going to have to order the S9013 - All I have are 2SC945 and 2N3904 transistors lying around (unless one of those will work, but I'm guessing not).

You could try a 2N3904, it's an NPN transistor like the S9013 is and it usually comes up as an S9013 substitute.

[thesteve]

any npn should work fine
rf mod pin2

[lastcallhall]

You could try a 2N3904, it's an NPN transistor like the S9013 is and it usually comes up as an S9013 substitute.

any npn should work fine
rf mod pin2

You're both awesome.

It works! None of that fancy encoder stuff, you can pull it right off of the chip! As soon as I find my video capture card, I'll toss up some comparison pics. However, I can't tell if the chroma signal looks slightly washed out, or if that's just the way the games are meant to be seen. Either way, I'm sure there's a method to adjust it.

Thank you all so much!

[thesteve]

the picture intensity can be adjusted by changing the resistor value (less resistance= brighter picture)
the transistor chosen will effect contrast just a little.

[grahf]

That's pretty cool, I've never heard of that before now. I'm really interested in seeing how it looks.

[thesteve]

i tested it on 2 monitors last nite using a 47ohm resistor.
1 was perfect, the other a bit clipped.
they both looked good, just my Y signal was super hot (too much for 1 of the displays)
the 100 was somewhat a guess.

[ApolloBoy]

What would the signal be like with a 75 ohm resistor?

[thesteve]

solid, might be too hot depending on the transistor