User:Steven J. Koch/Notebook/Kochlab/2009/07/29/Feedback tweaking, noise, FTC
Steve Koch 17:03, 29 July 2009 (EDT): Ant is floating the table. I am trying to fix the scaling issue with the software (need to scale by voltage, not integer) and noticed some noise issues. Sum signal is very sensitive to table vibration. More on this later. Currently, the noise level on sum signal is about 0.098 V to .105 V, when the laser is at 10.33 mW (current is 11.18 amps, modulation control is 0.06 V).
- At 50.55 mW is 0.593 to 0.600 V; a 7 mV noise level. Interestingly, this is the same level of noise at the lower power. I would have expected it to scale with power.
- At 198.1 mW, it's about 2.425 to 2.435 or 10 mV noise.
When you knock on the table, the sum signal is noisy, but not as much as the X, Y (scaled). This would imply to me that it's noise on the input beam position on the objective aperture.
DAQ Gain Not The Same
Discovered that the DAQ gain is not the same. The only valid gains with the new boards are:
- 0.5
- 1
- 5
- 25 (which is currently not an option)
I do not understand how the gain is stored in config file, so I am not going to change the gain choices now. It will be up to the user to do the right thing, unfortunately.
200 mV offset
The Feedback96 program is now giving similar bit values to the oscilloscope program. However, there is a strange 200 bit offset. Since the oscilloscopes give same answer on each computer, this is still likely to be a programming error. Perhaps related to the fact that I'm reading unscaled numbers? NO: I tried reading voltages and same thing happened. What is different between how oscilloscope DAQ and this DAQ is set up? Settling time?
- No, I do not think it is settling time. (Leaving it set at 6 microseconds.)
Things improved a lot
- fixed bug in VC (if power control went out of range, stupidly stopped moving stage)
- data files 23 and 24 in Ant's SDM data directory have probably some dsDNA stretching curves. The bead escapes the trap.
- power maxes out at 1 W (on the modulator screen), even with a 2.5 W input. This probably means I need to adjust something if we want more power. I didn't want to mess with it, though, so we could possibly try DNA unzipping. Things are still noisy as all hell, but I think DNA unzipping would be distinct enough that we could see it. At least I think it's worth trying? The reason I say this is because it would be so nice to know what the force is, and also because it would be a great motivator and foothold
