I have been playing with a copy of the PEMPro (?) program.  Unfortunately I

do not agree with its premise (at least for the cheaper-smaller toothed Meade

worm gears) that you can use FFT transforms to model the behavior of the

gear error terms.  There seem to be too many high frequency components

of a non-sinusoidal nature in the RA centroid motions when I just study the

drive system (PEC turned OFF and no guiding).

  I get fairly reproduced types of errors when I collect data over 3 different

HA zones (I use a 'locked down' permanent mount where I never change

the section of the RA gear actually used).  The data set I now have was

taken with seeing about 3" arc and a well balanced scope and expose-

read cycles of 2-3 seconds.  I'm not exactly 'synched' to the phase of the

worm detector - but it's sort of (weird) fun to try to phase match the

data [like an odd variable star with a known period but small dwarf-nova

outbursts].

  There are some well-known issues with PEC "update" actually being a

form of 'new data set averaged with ALL the old' .  I'm going to trust

what the guider centroids tell me - but take care of the loading of the PEC

table through the COM interface rather than the GUIDER input pulses.

I like to see the raw data (PEMPro good) but massage it myself (PEMPro

restrictive).

  Summary- I'll massage and study the data I have - and retake it after I

get an "East-weight" (I'll use a pulley and cable).  Simple CCDOps and

Excell work can do just fine for this data (as others have found).

Steven Nelson

PS

  PEC local background.  The first PEC system was conceived and birthed

at Lick on the 3m Shane reflector by Bob Kibrick, a programmer and several

mechanical engineers.  It was needed (late 70's) after electronic auto-guiding

determined the extent of wear-and tear on the 20 year old giant worm drive.

An Apple II type processor (6502) ran the whole thing - after programming

information was collected by the auto-guider.

PPS

"Fighting" autoguider and PEC - it seems the phase of the correction is critical

when there are high frequency error components. This is why I'd rather study

the system driving with no guider feedback working (i.e. unstable servo system)

-----Original Message-----
From: Ivor Barker <ibarker@...>
To: norcal_ai@yahoogroups.com
Sent: Wed, 24 Dec 2008 9:05 am
Subject: RE: [norcal_ai] LX-200 PEC training expert?
0A