Super retro trio tear down - Taking apart the SR3.

[sean]

I might be wrong but, since the sheet said VIL 4V to 6V that was doing with the voltage it runs at

[Jorge Nuno]

there's no 4V or 6V on that quote. The samsung and the mitsubishi chips ARE 3.3V

[MaxWar]

I see VIL ( Input low voltage) 0.4 to 0.6v, looks like you missed one decimal

Anyway, I interrupt this program to inform you that I just made great progress in eliminating the S-video Jailbars on the genesis side of that thing. I studied and probed the circuit board some more and realized why I failed to control the jailbars earlier.

Composite is actually mixed entirely from Chroma and luma out of discrete parts right on the board. It is not output by the ASIC. The asic only outputs Y and C.
So I removed one resistor to prevent chroma from being sent to the Composite mixing circuit.
BAM! 80% s-video Jailbar reduction. Of course your composite is now Black and white but your S-video is much improved.

To get this result you need to remove R7. ( Edit: Btw, this is a 1k resistor, for the records. )

[sean]

I feel dumb but, to my defense I am tired to day.
I don't know if this would work but maybe you could lift the Voltage input legs and add a 5v to 3.3v step down converter like this http://www.ebay.com/itm/DC-5V-to-3-3...item417060fd4b

[Jorge Nuno]

No. The system wont run properly at 3.3V. And the rams can't be put on 3.3 while the rest is kept at 5. I said this earlier: definite solution is to get 5V RAMs, alternate shitty solution is to stepdown voltage to 4.something. 3.3 is just too low

[sean]

How, about replaceing the sony cx for a nintendo s-rgb the sony has ok composite video but, the s-rgb chip is way better.

[MaxWar]

How, about replaceing the sony cx for a nintendo s-rgb the sony has ok composite video but, the s-rgb chip is way better.

But the CX has nice S-Video so I don't really care.

[Ace]

How, would I make the unit work with pal games ?

That requires moving some solder around on some jumper points and in some cases, like with the NES board, a complete change of the master crystal.

It's the year 2014 and they could have wired the region to the switch they have for the genesis region select.

Umm... that 4-position switch on the front of the console actually IS for changing the region of the Genesis side of the Super Retro Trio. BUT, the Super Retro Trio has a fault that even the older RetroGen Adapter has: in any of the PAL settings, be it PE (PAL Europe) or PA (PAL Asia), the TCT-6801 still runs in NTSC! While this works fine with some PAL games, sometimes, you might end up with a disastrous mess like what I got on the RetroGen Adapter here with Sonic the Hedgehog running in PAL (and if you're wondering, yes, I was using an actual PAL Sonic the Hedgehog cartridge for this):



I can only imagine how messy PAL-optimized games will be on both the RetroGen Adapter and Super Retro Trio in the PAL settings.

I was able to find a datasheet for the KM68U1000CLG-8L at http://www.digchip.com/datasheets/pa...1000CLG-8L.php
And it seems that they run at 4 to 6v just fine.

NOWHERE in that datasheet do I see any mention of the RAM running at 6V. In fact, let me quote the recommended voltage ranges for the KM68U1000C and KM68V1000C in that datasheet:

Power Supply Voltage:
KM68V1000C family: 3.0~3.6V
KM68U1000C family: 2.7~3.3V

The KM68U1000C is supposed to run between 2.7V and 3.3V, and the absolute maximum ratings are here:

KM68U1000 absolute maximum ratings.jpg

The chip should never be run over 4.6V, and even then, Samsung notes running the chips at that voltage could affect their reliability.

Composite is actually mixed entirely from Chroma and luma out of discrete parts right on the board. It is not output by the ASIC. The asic only outputs Y and C.
So I removed one resistor to prevent chroma from being sent to the Composite mixing circuit.
BAM! 80% s-video Jailbar reduction. Of course your composite is now Black and white but your S-video is much improved.

To get this result you need to remove R7. ( Edit: Btw, this is a 1k resistor, for the records. )

Well this is new. I always though the TCT-6801 output both Composite and S-Video from the GOAC.

[bustedstr8]

No. The system wont run properly at 3.3V. And the rams can't be put on 3.3 while the rest is kept at 5. I said this earlier: definite solution is to get 5V RAMs, alternate shitty solution is to stepdown voltage to 4.something. 3.3 is just too low

Not necessarily....we won't know for sure if we don't try. LV memory can work in a 5V system, some Snes repros are built with LV Flash and a LM7833 and work just fine.

Another reason I think these Rams have a good shot of working on 3V is that there are two unused IC spots by each pair of rams . If we look at U10 it seems like these boards where made to accept a SOT-23 3V regulator to power the rams. Someone with a unit can probably figure out what resistors/caps need to be moved to get the rams running on 3V with the added regulator. I can't see U9 well, but my guess is the unused output pad goes to Vcc and CS2 on the other two rams.

This looks like typical shitty Chinese quality control, they let the fabricators use LV rams and could of cared less to have the boards properly reworked

I wonder if they actually shipped any boards with the proper LV ram power circuits?

[Jorge Nuno]

Not necessarily....we won't know for sure if we don't try. LV memory can work in a 5V system, some Snes repros are built with LV Flash and a LM7833 and work just fine.....

Yes LV memories CAN work. BUT, those chips will burn from IO overdriving, unless they are specifically "5V tolerant", or otherwise protected from that with buffers or series resistors + clamping diodes.

[MaxWar]

At first glance it does not look like there are many resistors on the IO lines of those chips.

I am curious as to how long those things will actually last. I think I will play some snes on my retro trio

[Ace]

I see no such 3.3V regulator on the Super Retro Trio I have here. Personally, I don't think any of them are running the RAM at its recommended specs.

Also, while playing around with the audio circuit RetroGen Adapter, since its audio setup matches that of the Super Retro Trio's Genesis side and both clones share the same TCT-6801 GOAC, I discovered the TCT-6801 has a Stereo headphone amp built-in! Have a look at what I've done with it on my RetroGen Adapter in the Mega Amp thread: http://www.sega-16.com/forum/showthr...d=1#post669178

As the Super Retro Trio has the same audio circuit as the RetroGen Adapter, with the SMD versions of the resistors and capacitors shown in that post, you can turn the existing Super Retro Trio audio circuit into a Mega Mixer.

[bustedstr8]

Ya...I thought about that too. They rate the IO on the ram for Vcc+0.5, so about 3.8V. We won't know for sure what VOH is from the other chips until someone sticks a scope on them, but it might be within spec of the ram...or pretty close.

Swapping to 5V rams is clearly the best option, but it's a lot more expensive and changing TSOP40 chips is probably out of skill range for most collectors.

For $3 or so I think it's worth trying the LV regulators. Worst case they still burn up from overdriven IOs and need to be replaced like the chips being run on full 5V.
Best case is the chips last for years from running close to/in spec, and installing a SOT-23 is probably something anyone who can change a cap could do.

[Ace]

I could probe the pins with my oscilloscope, though right now, I need to deal with a really crappy probe. It should do the trick, though.

[Jorge Nuno]

VOH on a CMOS chip like everything today is 99% VCC or pretty damn close on an unloaded condition.

In case of an 5V ASIC IO connected to a 3.3V powered RAM what will happen is that the ASIC VOH will drop to 4V-ish, because there's an ESD diode inside the ram that will drain the IO line to the RAM's lesser VCC (3.3V + 0.6/0.7V diode forward voltage drop). This diode will be under stress when the IO is high, and when it blows, the ram IO will survive, but not much longer as there's nothing to drain the excess IO voltage and that will breakdown the oxides on the input. A series resistor in here protects the diode by limiting it's current quite a bit, letting the ESD diode survive (and the IO thereafter)