Tektronix XP400, Video capture issues with VGA Grabber, Sync/Contrast anomaly

[DiTBho]

Tektronix XP400 X-Terminal: Video capture issues with Epiphan Grabber, Sync/Contrast anomaly

Hi everyone,
I'm currently restoring a vintage Tektronix XP400 X-Terminal, dating back to mid-90s, MIPS-based architecture, and I've run into a bizarre analog video signal issue when trying to capture its VGA output.



I am using a professional capture card, an Epiphan VGA2USB LP, to record the output.

Interestingly, this grabber works flawlessly with my other vintage terminals, including a Tektronix XP217 and an NCD NC900.
However, when connected to the XP400, the grabber constantly loses sync—dropping the signal and displaying a continuous "Tuning" loop.

Here is the kicker: If I connect the XP400 directly to a modern LCD monitor, the picture is absolutely perfect.
What I have tested/ruled out so far:

• Resolution & Refresh Rate: I forced the terminal to standard VESA modes (640x480 @ 60Hz and 800x600 @ 60Hz) via the boot NVRAM configuration.
  The behavior did not change at all
• MPEG/Overlay Coprocessor: The XP400 has an optional MPEG-1 hardware decoder board containing a C-Cube CL480 and a Cirrus Logic CL-PX1070 overlay controller. I completely pulled this board out to isolate the system down to the bare motherboard GFX path. The capture issue remains identical, meaning the problem lies strictly within the motherboard's main video output/RAMDAC stage.
• Impedance & Loading: I suspected a signal loading or impedance mismatch issue on the RGB lines ($75\ \Omega$). I tried running the signal through a high-quality powered/active VGA splitter (acting as a buffer/distribution amplifier), but even the output of the active splitter causes the Epiphan grabber to go into the same "tuning" loop with crushed colors.

My Hypotheses:
Since standard LCD scalers are very forgiving but the professional Epiphan digitizer is highly rigid about timing and DC references, I suspect something is electrically out of spec right behind the DB15 connector on the motherboard:

• DC Offset / Black Level Clamping: The crushed/high-contrast colors point toward a failed black level clamping reference. I suspect the AC coupling capacitors on the R, G, B lines (likely 10uF–47uF tantalum or electrolytic caps near the RAMDAC) might have degraded or high ESR, causing a DC voltage offset that saturates the grabber's ADC.
• Sync Pulse Edge/Deformation: The H-Sync and V-Sync lines are at the expected 5V TTL levels, but the continuous tuning loop suggests the grabber can't cleanly latch onto the rising edge of the pulses. Could aging passive filtering components (ceramic caps/pull-up resistors) be smoothing out the square wave into a slope that modern digitizers reject? Or did Tektronix employ an unusual composite sync configuration on these high-end workstation motherboards that tricks the auto-detection?

Before I go ahead and blindly shotgun-recap the entire video output section of the motherboard, has anyone encountered this type of input-stage rejection with professional Epiphan grabbers on 90s UNIX workstations or specialized hardware?

Any insight on typical Tektronix RAMDAC video output configurations or specific filtering anomalies of this era would be highly appreciated!

Thanks in advance 

[DiTBho]

[DiTBho]

Anyone?

[Zoli]

Anyone?

Check the VGA sync's for fidelity; find a way to restore, if necessary(internally or externally).

[DiTBho]

Anyone?

Check the VGA sync's for fidelity; find a way to restore, if necessary(internally or externally).

how?

I think I should use a DSO to "see" what's happening.

[Zoli]

...
how?

I think I should use a DSO to "see" what's happening.

Or any oscilloscope(analog) should do it; look for timing(pulse widths), pulse amplitude, DC level, fronts; worst case you can try to regenerate the signals. Or worst case you can use a logic analyzer

[DiTBho]

has anyone encountered this type of input-stage rejection with professional Epiphan grabbers on 90s UNIX workstations or specialized hardware?