Saturation when using THS7316

Hello Dhaivat,

There are two configurations shown. The first (RGB) has inputs terminated with 150 Ohm resistors and the second (HV) has no termination at all. Was this intentional? Do you know what the input signal voltage is with this termination configuration? How long is the signal path? You may be getting reflections. Video signals are typically terminated with 75 Ohms to match the video cables. Are you using cables or board traces?


Have you measured the output signal amplitude with an oscilloscope? It's possible that without the termination resistors the signal from the THS7316 is saturating the monitor input.

Also, the THS7316 was designed for video up to 1080i (not 1080p). Is it possible your signal bandwidth is over 36MHz?

Another possible cause for saturation in the AC coupled (RGB) lines is that the DC voltage of an AC coupled signal can drift as the coupling capacitors collect charge and drift towards the average value of the signal. Have you compared the voltage of the video signal on both sides of the coupling capacitors? It is possible that the saturation is occurring in your video monitor and not the amplifier.

Regards,
Loren

Hi Loren,

Please see my reply as below.

I have used the termination at the input of the THS7316 as it is required by the COM Express Module (refer to the PICMG reference guide for COM Modules). There is no requirement for the termination for the HV sync signals and therefore these are not terminated with 150 ohms. However, the HVsync signals are 3.3V tolerant output. I am attaching the description from the PICMG reference guide as below.

The signals are routed on the PCB  with the distance from the COM module to the THS7316 IC being about 100mm. These signals are surrounded by ground plane.

I am not sure about the signal bandwidth being more than 36MHz. The DAC runs at speeds up to 350MHz in the COM module. I do not have information about the DAC as it is in the COM module. 

My aim is to provide a dedicated buffer to the VGA signals originating from the COM Express module. I am not sure the signal is 1080i/1080p. I am just connecting to standard monitor (not TV) at the output of THS7316. And also We are not using very high resolution as well.

I saw the signals on oscilloscope long time ago so cannot actually remember. But did not see anything unusual, I will check again and let you know.

Best regards

Hi ,

I have now fixed the RGB side of signals and hence the saturation is now fixed. Initially the the output of AC coupled RGB signals was more than 0.7Vp-p at the output (roughly 1.5Vp-p). By increasing the resistance to 150 ohms and removing the second 10pF cap the saturation issue was fixed.

I am now struggling to fix the sharpness/fuzzy/spread/micro lines issue. I am sure this is because of the Hsync & V sync. These signals are 5V at output and square wave at 48kHz & 60Hz respectively as I am using the computer monitor at 1024x768 @ 60Hz. I have replaced the 22 ohms at output with caps removed and tried 150ohm & 75 ohm.

Apparently at 75 ohm the screen was sharper than the 150 ohm. But I don't seem to understand why this unsharpened and is there a better way of improving the sharpness. Do you think I should use a 75 ohm termination at the input and 75 ohm at output.

Under careful observation on the monitor, I was getting vertical micro lines at 75 ohms whereas the spacing was more when I used 150ohms, but at 150 ohms the alphabets were slightly experiencing slight spread/blur.

Yes, I use VGA cable.

Hi Dhaivat,

Video systems use 75 Ohm cabling.  It is best to terminate the video cables with 75 Ohms.  The mismatch from 150 Ohm termination would cause reflections that can cause fuzzy lines.  The output of the THS7316 should be terminated with 75 Ohms.

I can't help you on the input side.  You may want to contact the vendor of the module that is generating the video.  In general, though, video systems are 75 Ohm and long signal paths that are not properly terminated will cause ghosting problems. 

Regards,

Loren

Hi Dhaivat,

I would recommend using a 75-ohm termination at the input and 150 ohm (75ohm series + 75 ohm load) termination at the output of the THS7316 to get the right sharpness. The THS7316 has a gain of 6-dB which allows the signal to be divided down at the detector using a 150 ohm (75ohm series + 75 ohm load) termination at the output. You can refer to Figure 1 in the THS7316 datasheet.

Regarding the vertical micro lines at the monitor screen, I suspect that the sync signal is not timed correctly with regards to the actual RGB signals. Normally, there is a Schmitt trigger in the H & V sync signal path which is responsible for noise immunity on this path. Is there a reason for using the THS7316 instead of a Schmitt trigger on this path? The THS7327 or THS7347 video buffers could also be used in your application which have dedicated H & V sync buffer outputs.

Also, the 5V output on the THS7316 is close to the output rails. The rise and fall edges of these sync signals close to the output rails could affect the timing with respect to the RGB signals. Could you reduce the amplitude of the H & V sync signals by using a resistive divider, so that THS7316 outputs are not saturated on the H & V sync paths?

Best Regards,

Rohit

Hi Rohit,

Yes, I was thinking on those lines. I will let you know the result.

Many thanks.

Best regards,