THS8134B

Larry Blumer

Other Parts Discussed in Thread: THS8135, TFP410, TFP401

Hi,

I'm using a THS8134B to output an RGB VGA signal. I have tied M1, M2 to GND and SYNC# and SYNC_T to DVdd. What I am seeing is an offset (300-500mV) on the green output. I've tried a few things but the offset still remains. I have noticed that the 8135 or 6 would seem to have another mode called generic that may eliminate this issue but for now I'm locked into the 8134.

Any suggestions would be helpful.

Thanks,

Larry

over 15 years ago

0 Larry Taylor over 15 years ago

TI__Genius 16810 points

Hi Larry,

The THS8134B offset is apllied to the green channel to allow room for sync insertion and cannot be disabled. This is typically not an issue with video or graphics applications where the outputs are AC coupled prior to the monitor's DC restore.

You are correct that with the THS8135/6, the generic mode removes this offset.

0 Larry Blumer over 15 years ago in reply to Larry Taylor

Prodigy 120 points

Hi,

Thanks for the resonse. I had tried to AC couple the green signal, I had used a 10nF cap, not quite sure waht value would be best.

What I see when I do that is a violet tint. I suspect now I'm not getting enough green drive?

Thanks,

Larry

0 Larry Taylor over 15 years ago in reply to Larry Blumer

TI__Genius 16810 points

How does it look if you DC couple the green channel? 10nf is too small. The monitor should be able to handle a DC coupled input with the offset, since it problaby also supports sync-on green.

Color inbalance could also occur if:

- RGB mode is not enabled in the THS8134B. To ensure RGB mode, you may need to pulse the /Sync input so that M1 and M2 are interpreted. An input clock must be present when you pulse the /Sync input.

- 0-255 code range is not being generated by the digital source, or if code 0 is not gnerated by the source during horizontal blanking.

- The horizontal sync is not positioned properly, resulting in the monitor doing DC restore during active video.

With the DAC ouputs terminated with 37.5 ohms, the DAC outputs should swing 700mV p-p when a 100% white field test pattern is used. The green output should range from 350mv to 1050mV, the red and blue ouputs should range from 0V to 700mV.

0 Larry Blumer over 15 years ago in reply to Larry Taylor

Prodigy 120 points

Hi Larry,

Thanks for the response. When I DC couple there is a green tint, this varies with the display used. I'm driving a Dell VGA monitor and another small display.

The datasheet is a little unclear, I'm setup for RGB mode M1,M2 = GND Syncs = H. Currently I can't pulse the syncs. In table 4 this seems to say I'm in GBR with sync on green? For the monitor syncs, this is being used like the VESA standard where H,V sync are digital (H sync delayed 8 clks for the 8134 pipeline delays), not part of the analog video stream. One thing I should mention on the data bus is a TFP410 this displays the video signal quite well.

Signal levels seem OK, I'm not terminating as suggested just into 75ohms except I've tried 37ohm on green. R,B= 0 to 800mV, G=~300 to 1V.

Thanks,

Larry

0 Larry Taylor over 15 years ago in reply to Larry Blumer

TI__Genius 16810 points

Larry,

Sounds like you have been careful with HSYNC position. HSYNC and VSYNC polarity, length, position are all important for digital monitor format detection and correct clamp or DC restore position.

Are you using the /Blank control input to force blank or black level?

If you look at RGB at HSYNC time are all channels at there lowest level?

When DC coupled with no bandwidth reduction, peak-peak RGB should be better matched than watch you are seeing. The DACS are curent soruces, so the RGB outputs must have matching loads. The DACs are designed to provide 700mV when doubly terminated with 75 ohms (37.5 ohms).

Some of the control inputs are latched after /Sync edge transitions, so pulsing /Sync is recommended. This may not be your issue, however.