Composite amp circuit

[TurboXray]

 Taken from a US Duo system:



 CVBS out still has a voltage offset applied to it further down the line, so you can get negative voltage sync levels. But the levels are the same.

 A quick check: The 'blanking' levels of the direct vce pin output are Y is ~2.1v (bottom of sync is ~0.9v), R-Y and B-Y are ~3.25v (with polarity at +/- ~1v and less for the carrier waveform), CB is 3v (with +/- 1.5v polarity of the carrier syncro waveform).

[Charlie]

Nicely done. 

The first transistor (902) and the third transistor (904) are MMBC1623L6.
The second transistor (903) is MMBA812M6.
The VCE is U903, the HUC6260.

I show some slightly different resistor values:
The second resistor is shown as 6.3k ohms, I show it as 6.2k ohms.
The 1.36k (going to ground on the first transistor) as 1.5k ohms.
The 2.7k ohms (between the first two transistors) as 3.3k ohms.
The 3.5k ohms as 13k ohms.
The differences between these two setups may simply be a circuit update.  My data is from '93.

The output "CVBS" goes, of course, to the AV port #JK-601 pin 3, via C962,R930, and L603.

Charlie

[TurboXray]

The first transistor (902) and the third transistor (904) are MMBC1623L6.
The second transistor (903) is MMBA812M6.

Oh, thank you Yeah, 'L6' and 'M6' were the only labels on the transistors.

I show some slightly different resistor values:
The second resistor is shown as 6.3k ohms, I show it as 6.2k ohms.
The 1.36k (going to ground on the first transistor) as 1.5k ohms.
The 2.7k ohms (between the first two transistors) as 3.3k ohms.
The 3.5k ohms as 13k ohms.

Those are the values I got from measuring them directly. It's entirely possible that they might be out of spec (some bad resistors). I do know that even with a 75 ohm load applied when viewing it on the scope, the CVBS levels I'm getting are to high (max white is above the 0.7v spec). Where did you get your data from?

[Charlie]

==>Those are the values I got from measuring them directly<==
Aha.  Your numbers may very well be what you measured "in circuit", which means you are seeing the effect of the other components.  Or did you actually remove each part before measuring it?  Either way, good job.

As I said, got this data back in '93.

 
Charlie

[Game-Tech.US]

Wait, the Duo has component video outs?

[SignOfZeta]

No, composite.

[TurboXray]

Wait, the Duo has component video outs?

Yes, but not amp'd (only composite output is built and amp'd). S-video amp circuit can be made from the 4 outputs (without a separate RGB encoder). But if you want to make component output video, you need to demodulate the two chroma channels (bridge rectifier etc). Luma is correct for component (Y for YCrCb on NTSC starts at IRE 0). Not sure simple demodulation will be in the same color space as YUV(YIQ), so you'd have to adjust for that too (YUV to YCrCb conversion).

[SignOfZeta]

So in other words, "no".

[Game-Tech.US]

Wait, the Duo has component video outs?

Yes, but not amp'd (only composite output is built and amp'd). S-video amp circuit can be made from the 4 outputs (without a separate RGB encoder). But if you want to make component output video, you need to demodulate the two chroma channels (bridge rectifier etc). Luma is correct for component (Y for YCrCb on NTSC starts at IRE 0). Not sure simple demodulation will be in the same color space as YUV(YIQ), so you'd have to adjust for that too (YUV to YCrCb conversion).

Think i'll just stick to trying to get component from and RGB+sync -> SonG component circuit...
BTW if you know of such a circuit please share, I have one, but untested so far.

[TurboXray]

==>Those are the values I got from measuring them directly<==
Aha.  Your numbers may very well be what you measured "in circuit", which means you are seeing the effect of the other components.  Or did you actually remove each part before measuring it?  Either way, good job.

I'll cut the traces and then remeasure them.

As I said, got this data back in '93.

From a schematic or from you (or someone you know BITD)? Just curious, 'cause I hadn't seen anything online before.

 Also, what's the part number for the inductor right before the connection to DIN (L603)?

Think i'll just stick to trying to get component from and RGB+sync -> SonG component circuit...
BTW if you know of such a circuit please share, I have one, but untested so far.

I have like 4 or 5 circuits saved (the are a few RGB to YPbPr circuit designs floating around for Amiga 15khz RGB to 15khz YPbPr). As a challenge, I wanted to build my own instead. The maxtrix conversion itself is pretty simple; I just want to see if I can design the conversion myself in the analog circuit domain. But to be honest, I'm don't care too much about RGB to YPbPr (analog YCrCb). I'm more of a purist/nut and rather get the real/actual colors of the TG16 into component video (which grabbing/converting from RGB lines won't do) either by demodulating the two component lines of the VCE out, or ADCx3 of RGB with the VCE pixel clock out -> index to corrected RGB values in a rom -> then back out to a RGB2YPbPr circuit. But baby steps first

[Duo_R]

so is this from HU6260? Also this circuit is completely missing from the Turbo Express, but is it possible to build it as you listed to get composite out from a TE?

[TurboXray]

so is this from HU6260? Also this circuit is completely missing from the Turbo Express, but is it possible to build it as you listed to get composite out from a TE?

Correct. But as Charlies mentioned, I need to cut the traces to isolate the resistors to make sure some of the values in question are what they are (and at that, they still might be bad). But otherwise, yes. You definitely don't want to just tap the pins directly, even though they could drive composite or svideo by themselves. You want to protect the VCE (6260) in case something happens on the output lines.

[thesteve]

the circuit has now been built
http://pcengine.freeforums.org/express-gt-composite-t179.html

[soop]

This should work fine in a regular white engine right?

[thesteve]

shouldnt be needed

[Duo_R]

so lift and ground the pins you mentioned, and pins 20, 26, 33, 40 hooked up with the resistors per the diagram?

[thesteve]

thats it
done 2 now
just remember the pins your grounding are hooked to +5V stock and connected to each other in the chip

[soop]

shouldnt be needed

Well actually...  I have a plan to make an LT from one of my PC Engines.  Which involves finding a 4:3 LCD of about 4.5" and moving some crap around....  Maybe even include a battery.

[thesteve]

for a core graphix the composite circuit is prebuilt by NEC, you wont need to build it

[soop]

for a core graphix the composite circuit is prebuilt by NEC, you wont need to build it

Oh snap you're right.  I remember the MMMonkey guide now.  Sorry, it's been a while

[Duo_R]

That isn't difficult. The PS One screen is a good candidate and you can even hook it up with RGB. I have one I did posted somewhere.

[HercTNT]

You beat me to it duo. The ps one screen works great, and can be bought brand new for around $45. That sounds like a pretty slick project you're thinking off soop. If you get around to it, make sure you post pics. I would love to see it.

[soop]

That isn't difficult. The PS One screen is a good candidate and you can even hook it up with RGB. I have one I did posted somewhere.

The PS One screen is too big (I think even with work it won't fit inside the PCE's dimensions - but I do have a decent screen saved somewhere on eBay.

I'd love to see your work though if you can find it!

[Tatsujin]

Damn, am gonna mod all my spare white cores nao

[turbokon]

Tat, why go composite when you can go component with this:)