Some hints dealing with Magnet Configs in the presence of LER Tune FFWD

HVS 8/30/01

mod: 5/02 HVS I found a bug in this algorithm see note at bottom to get it right.

So let's say you want to re-load a magnet configuration into the PEPII LER but you are confused about what the LER TUNE FFWD is going to do once you have done that. Presume for the sake of this discussion that you are starting with no beam.

Stop for a moment and reflect about what the feedforward does. It changes the tune magnets based upon a change in current.

So here's the problem: The configuration you like was saved at some high current, and if you just load/activate the config and fill, the FFWD will take you someplace you don't want to go.

Here's the way out. I've done this scheme MANY times and it works great.

One: Load and activate the configuration you like.

After you have done that the tune magnet settings will be at a place which would be great if you had the same current as was there when the config was saved. Trouble is that you don't have that current...you have zero current

Two: Manually dial the tune knobs the amount FFWD would have if the beam had just aborted.

You have to figure out how much current was there when your favorite config was saved. HSTB or config title usually works. At this point re-assign the tune knobs so they read zero. Now figure out the what the FFWD coefficients are. SCP path is Index -- Feedback -- Slow Feedback -- PEPII TUNE FFWD -- Test Run Panel -- Diagnostic Panel. Down at the bottom is a number called LER coef. There is one for the LER_NUX loop and one for the LER_NUY loop. Take these numbers and multiply them by the amount of current (important! see modification at the bottom of this page) that was present in the config you liked (units to use are milliamps). This product is the amount you want to move each tune. Be careful about the sign. If the coefficient is positive, you want to move the tune negative. (Think this through yourself...you are mocking up what FFWD would have done on an abort. An abort is a negative change in current).

Three: Reassign the knobs and fill...

At this point you want to again reassign the knobs to make them read zero. What you have done is to put the machine in EXACTLY the same condition it would have been in had there been an abort just after your configs were saved. With the FFWD running, go ahead and fill it up, knowing that at the top of the fill the tune knobs should want to be at zero.

Easy as that; One, Two, Three!

Important Modification:

I had trouble making this work right until I figured out about the "cropping" parameters. This stuff is explained in "Display Group Help". Basically the software has (user specifiable) limits on the currents which will be used. They are called "current cropping parameters" and can be seen on the touchpanel for the FFWD loop select. Presently the "minimum LER current" is 400mA. That means if there is an abort at 1700mA, the FFWD will not move the tunes by LER coef times 1700mA, but rather by the amount LER coef times 1300mA. So in step two above, don't use the current value which was present in your favorite config, use that current minus this "minimum LER current".