Hi, the repo for the TCD1304 with linear response is recently updated and now pretty much complete. There will be one more edit on "carry over", and then that's it.
This is also a challenge to all of the linear CCD projects, and for any commercial instrument you use. You should insist on seeing a fluorescent lamp spectrum and linearity graphs as shown in the repo.
Here is the data for mine with a write up on how to do it in detail in the readme.
Looks interesting, but I can't find anything resembling a finalized schematic diagram of the system in the repo. Could you point me to it?
It is there in an indirect way in the the SPICE models, there is one for the analog to digital chain and another for the gate drivers. But, I will post one.
Also important is the layout. There are some special measures to keep the disturbance on the power rails and ground from the pulse drivers from getting into the analog section. I did not mention that in the readme, but plan to add it.
And, there is another piece on mitigating the phantom intensity from the previous frame. That is related to the gate drivers and some physics in the silicon of the chip itself.
We had numerous exchanges with the team responsible for the chip, on all of the above. I still have to translate some of the device physics that they gave me into quantitative terms, but will hopefully get to it this week.
Meanwhile, look for the heading " SPICE models for the 16 bit sensor board"
If you zoom in on the scale in the spice models, you can see that we actually stay within 1 part in 2^18 even for step functions.
Of course there is. That is an absolutely basic circuit design principle, called "power supply decoupling".
Yes, but there is a little bit more to it in this case. We were surprised too.
Will try to post the rest soon, but I am out for the day.
