Forums

what does the Q2 do in this CCD preamplifier?

Started by Unknown October 11, 2012
http://instrumentation.obs.carnegiescience.edu/ccd/gcam/images/PRE_SCH.jpg

looks like Q2 is a clamp reset to zero voltage reference ,i.e. the ground.
However, the ccd signal is changing rapidly , for example, within 1us the ccd has a signal like ILX511B (http://www.oceanoptics.com/technical/ILX511B%28E%29.pdf) and within this time constant the
capacitor didn't ground due to the long time constant.

so this clamp reset doesn't work for 1us ccd?
Looks to me like a "black level clamp" to set the DC level prior to each scan line. That make sense? The CCD data sheets may offer more clues. 

--sp
On Oct 10, 11:41=A0pm, sis...@gmail.com wrote:
> http://instrumentation.obs.carnegiescience.edu/ccd/gcam/images/PRE_SC... > > looks like Q2 is a clamp reset to zero voltage reference ,i.e. the ground=
.
> However, the ccd signal is changing rapidly , for example, within 1us the=
ccd has a signal like ILX511B (http://www.oceanoptics.com/technical/ILX511= B%28E%29.pdf) and within this time constant the
> capacitor didn't ground due to the long time constant. > > so this clamp reset doesn't work for 1us ccd?
I don't know much about CCd's. Do you need to reset the cap before reading each pixel, or just before a whole scan? (I think that's what Speff said.) George H.
On 11.10.12 5:48 , George Herold wrote:
> On Oct 10, 11:41 pm, sis...@gmail.com wrote: >> http://instrumentation.obs.carnegiescience.edu/ccd/gcam/images/PRE_SC... >> >> looks like Q2 is a clamp reset to zero voltage reference ,i.e. the ground. >> However, the ccd signal is changing rapidly , for example, within 1us the ccd has a signal like ILX511B (http://www.oceanoptics.com/technical/ILX511B%28E%29.pdf) and within this time constant the >> capacitor didn't ground due to the long time constant. >> >> so this clamp reset doesn't work for 1us ccd? > > I don't know much about CCd's. Do you need to reset the cap before > reading each pixel, or just before a whole scan? (I think that's what > Speff said.) > > George H.
Once for the scan, if the time constant is long enough, or, for a multiline scan, once for each line. The point is that a CCD has a pretty large and drifting DC baseline, compared to the video signal. The purpose is to charge a reference DC level into the cap during the interval between the scans and hope that it stays well enough to give good video out. -- Tauno Voipio
sisuaq@gmail.com wrote:
> http://instrumentation.obs.carnegiescience.edu/ccd/gcam/images/PRE_SCH.jpg > > looks like Q2 is a clamp reset to zero voltage reference ,i.e. the ground. > However, the ccd signal is changing rapidly , for example, within 1us the ccd has a signal like ILX511B (http://www.oceanoptics.com/technical/ILX511B%28E%29.pdf) and within this time constant the > capacitor didn't ground due to the long time constant. > > so this clamp reset doesn't work for 1us ccd?
That looks like the makings of a chopper amplifier, used in cases where supper sensitivity is required. The CCD reference most likely will include the pulses, but at a rate faster than expect and thus can be filtered later down the line. Looking at C2, it looks like the integrator needed to remove this or reduced BW. That's just a guess... Jamie
On Wednesday, October 10, 2012 11:41:10 PM UTC-4, sis...@gmail.com wrote:
> http://instrumentation.obs.carnegiescience.edu/ccd/gcam/images/PRE_SCH.jpg > > > > looks like Q2 is a clamp reset to zero voltage reference ,i.e. the ground. > > However, the ccd signal is changing rapidly , for example, within 1us the ccd has a signal like ILX511B (http://www.oceanoptics.com/technical/ILX511B%28E%29.pdf) and within this time constant the > > capacitor didn't ground due to the long time constant. > > > > so this clamp reset doesn't work for 1us ccd?
If the circuit is driven by an unbuffered CCD shift register then because the output impedance is very large, Q2 does effectively GND the amplifier input. It is not necessary to charge C1.
On Thursday, 11 October 2012 13:32:23 UTC-4, Tauno Voipio  wrote:

> The point is that a CCD has a pretty large and drift > DC baseline, compared to the video signal. The purpose > is to charge a reference DC level into the cap during > the interval between the scans and hope that it stays > well enough to give good video out.
The time constant with the switch 'off' is 1 second (so it won't change much in 100 or 200msec). With the switch 'on' it is about 40usec (so it will reset pretty well in a few hundred usec). Here's the original article, which mentions the part under discussion in passing as a "DC restore". http://instrumentation.obs.carnegiescience.edu/ccd/gcam/preamp.html There are actually two "reset" switches-- one is Q2 and the other is inside the CCD chip. The latter switches during each pixel time between a reference voltage (perhaps several volts from ground) and the video voltage (offset from the reference voltage). Q2, in conjunction with the 100n capacitor, gets rid of the several volts offset mentioned above (with its attendant drift and effect on signal to noise in subsequent processing). --sp
On a sunny day (Wed, 10 Oct 2012 20:41:09 -0700 (PDT)) it happened
sisuaq@gmail.com wrote in
<d4a3f5c1-0cc5-4bf3-b839-358905ccd0b0@googlegroups.com>:

>http://instrumentation.obs.carnegiescience.edu/ccd/gcam/images/PRE_SCH.jpg > >looks like Q2 is a clamp reset to zero voltage reference ,i.e. the ground. >However, the ccd signal is changing rapidly , for example, within 1us the ccd has a signal like ILX511B >(http://www.oceanoptics.com/technical/ILX511B%28E%29.pdf) and within this time constant the >capacitor didn't ground due to the long time constant. > >so this clamp reset doesn't work for 1us ccd?
Normally the horizontal clamps work for a few us at the start of each line. There is no need to completely discharge a full capacitor in one clamp action, it just slowly changes the charge so black is at the right level. You are just removing bits of charge every time the clamp is active. The whole reason for the clamp is LF restoration. After equilibrium your clamp only corrects the little charge imbalance (from DC) added by each scan line, and that is sufficient.