[RASMB] counterbalance functions

[John Philo]

Bo, that is indeed an interesting idea for checking the counterbalances.

I should perhaps clarify for everyone that while changing to a different
rotor or counterbalance can't directly alter the calibration of the optical
systems, simply taking any rotor in or out of the instrument can shift the
drive a little relative to the optical components. That is, both calibration
points on the counterbalance will shift in or out by the same amount as the
axis of rotation moves relative to each optical system (and the shift seen
by each optical system is not in general the same). 

Thus in principle one could argue that the radial calibration should be
re-done for each and every run. However in practice these shifts don't
really have a significant effects on the results. Sedimentation coefficients
(the thing we can measure most precisely) depend mostly on radial position
differences, and thus are independent of such shifts (to first order). Even
a huge shift of 1 mm would only change the apparent sedimentation
coefficients by ~1.6%. Masses from equilibrium do depend more directly on
absolute radius values, but even this huge shift of 1 mm would only change
the apparent M by ~3%. As I recall these run-to-run shifts of the drive
relative to the optics are actually < 0.1 mm, so I doubt AUC users will ever
see significant errors arising from them. 

John

-----Original Message-----
From: Borries Demeler [mailto:demeler at biochem.uthscsa.edu] 
Sent: Friday, January 07, 2011 2:22 PM
To: jphilo at mailway.com
Cc: 'Xin-Ping Xu/FS/VCU'; rasmb at server1.bbri.org
Subject: Re: [RASMB] counterbalance functions

> 
> Xin,
> 
> What Virgil said is mostly correct, but it is only part of the story. 
> The counterbalance is also required for "delay" calibration of the 
> absorbance optics (to set the timing of the lamp pulses). This 
> calibration happens every time you start a run or when the rotor speed 
> is changed by more than 2000 rpm.
> 
> Radial calibration needs only to be done intermittently (I recommend 
> doing it only when the service tech has done something to the 
> instrument). Since the optical components are entirely independent of 
> which rotor is used the radial calibration is independent of the 
> rotor, so I have to disagree with what Virgil said about needing to
recalibrate for each rotor.
> 
> If the counterbalances are machined accurately then they all should 
> give the same radial calibration also. Again, remember the optical 
> components do not move or change when you change counterbalances. If 
> different counterbalances give different calibrations then it is 
> impossible to say which one is more accurate. In that situation I 
> would recommend adopting one of them as 'correct' and sticking to its 
> calibration. Otherwise your results are potentially different from one 
> experiment to another depending on which counterbalance you used for
radial calibration.
> 
> If you are not using the absorbance optics then in fact the 
> counterbalance is not required and an additional interference or 
> fluorescence sample can be run in that hole.
> 
> John

Yes, that is a chicken and egg sort of conundrum, if you cannot trust your
counterbalance you cannot trust your radial calibration. However, you can
get a little closer to the truth with some software I wrote.
One way to check if the counterbalance is correctly calibrated is with the
UltraScan3 rotor calibration program. While the main purpose of the program
is to calculate the rotor's stretching function, you can also use it to
"measure" the counterbalance and cell housing and centerpiece combinations. 

The way it works is that you would scan the counterbalance (or
centerpiece) in intensity mode at multiple speeds, generating a sharp
boundary at each edge. The edge would move with each speed, as the rotor
stretches. If the edges are truly centered at 5.85 and 7.15 cm, then the
center should be at 6.5 cm, which is how the rotor has been machined.
The program will take the differences between the edges at successive speeds
and extrapolate these differences with a second order polynomial to zero
speed. The positions at rest should be 7.15 for the bottom then and 5.85 for
the top, and the center (which is also calculated by the
program) should then be at 6.5 cm. If the center is not at 6.5 cm, then that
is an indication that something is wrong with the counterbalance.

In addition to the counterbalance dimensions, you will get the rotor
stretching function from this calculation. By scanning centerpieces in
different cell housings you will also get the true bottom of the cell this
way for each cell housing/centerpiece combo.  This is helpful if you want to
do mass conservation calculations in equilibrium experiments (the stretching
function will predict the offsets at each speedi). See the recent posts on
this topic.

Regards, -borries