On 02/07/2014 23:41, Jim Thompson wrote:
On Wed, 02 Jul 2014 21:26:44 +0100, Mike Perkins
wrote:

On 02/07/2014 20:36, Jim Thompson wrote:
Just did a video DC restorer function (in CMOS) for a UTC camera chip,
and realized that it would be trivial to do the same function with
bipolar discrete devices.

On the S.E.D/Schematics page of my website check out...

Video_Restorer_Discrete_NPNs.png

...Jim Thompson


I have used a similar idea before, but its general usefulness is limited
by the black level being dependent on the sync height,

Everything is ratiometric in "IEEE Units". In my particular case the
video is fixed amplitude from sync tip to UTC (Coaxitron) peaks.

[UTC (Coaxitron) are the control bits for the camera.]

and to a small
extent, the average video level.

Wrong

I see you have now removed the offending circuit from your website.

You "had" a 100k resistor down to ground to bias the circuit. The
current flowing through this, and the clamp will be dependent on video
level.
The clamp will have an impedance and so there will be a small variation
of clamp level with video level.

Ordinarily you would need a buffer after this to ensure the clamp is not
affected by anything downstream.



I'm surprised, given this video signal is of known origin, why would you
need such a circuit? Plus you can hardly integrate the 100nF on the CMOS.

Video is commonly AC-coupled. A DC restorer "clamps" the sync tips to
ground, from which all measurements are referenced.

Agreed, but why would your customer specifically introduce AC coupling?


Who said the 100nF was integrated?

...Jim Thompson


--
Mike Perkins
Video Solutions Ltd
www.videosolutions.ltd.uk