AUTOSTAR INFORMATION Last updated: 16 December 2001

Calibration, Training, Percentages: what they are and what they do

Sent:	Monday, June 18, 2001 20:53:06
From:	rseymour@wolfenet.com (richard seymour)
There has been a lot of discussion about what training and other
steps are required for proper Autostar operation.
I don't think i've ever seen the various "tuning" factors discribed
in one shot, so here's my attempt at it.

Some of the information (the Calibration LED adjustment) is derived
 from Meade's patent (US # 6,304,376). 

The Autostar has  five "tuning" parameters which can affect how it
controls position and tracking.

These are: Calibration, Ratios, Training, Percentages, Tracking Rate.

Here are a few items of background info/definitions you need:
 (a) motors: the motors in the telescope base.
 (b) encoders: mounted directly on the motor shafts are little
      gear-like (or "fan-like") objects. Their 36 teeth (or blades or
      vanes) pass between a light source and a light sensor.
      Each time the blade interrupts the light, that's called a "tick".
     By counting ticks, the Autostar learns how many far the motor
     shaft has spun.
 (c) motor circuit cards: at each motor is a small pile of circuitry
     which drives it.  Those circuits accept commands from the Autostar
    and feed voltage to the motors for short bursts of time.  It's
    easier to control time than strength, so slow speed is achieved
    by feeding full voltage to the motors for a -very- short time.
    More real speed is achieved by feeding the voltage for longer
    bursts.
 (d) Gears: between the motor and the barrel are a number (train) of
    gears.  The gears meade uses have a fair amount of looseness in 
    the way they mesh.  That looseness shows up when you reverse 
    the direction they're spinning... they all have to "take up the
    slack" before the motor force actually pushes the barrel in the
    opposite direction.  The amount the motor has to spin to -do- that
    is called "backlash".
  (e) Status display: press the [mode] key for 3 seconds. Release.
     You should now see an RA= Dec= display (if you have an electric
     focuser, you'll see that... press Scroll Down once).
      Now press scroll down once ... you will see the Alt/Az display.
    (if you've never been here before, keep pressing scroll down to
     see the other items available... you'll return to this screen)

That's the hardware, now the tuning:

The first item is Calibration.  

Calibration performs two operations: first it sets the strength
 of the light sources (LEDs: Light Emitting Diodes) used by the
 encoders on the two main motors.  It does that by adjusting the
 electrical current through them as the motors spin, until the
 encoders "see" a half-on/half-off condition (50% duty cycle, in
 geek-speak).   The encoder wheels are semi-translucent plastic,
 and have irregular edges.  If they were crisp, opaque, metal,
 this LED-tuning step could probably be avoided. But it's quick.
 The second operation i believe calibration performs is to
 measure how quickly the encoders move (motors spin)
 given a known voltage(or current)/time waveform to the motors.
 It's used to adjust the -speed- the system will yield when the
 Autostar says "speed 4".   Ye Olde Lore used to say: recalibrate
 if you change power source (battery/AC-adapter) or "load" (hanging
 an unbalanced camera out back).  It's a *speed* adjustment, not
 position. The Autostar can perform this setting without needing you.
It feeds a burst to the motors for a fixed time, and counts how many
 encoder ticks go by.  If you are using an ETX90/125 "default hand
 controller", that initial jog it perfomrs on every power-up is a
Calibration setting.  With the Autostar, you perform a calibration
after every firmware update, just in case the new firmware stores the
value in a different location than the previous firmware. 
When you power up (Initializing... beep) the Autostar sends the
 previously-recorded calibration value -to- the telescope base to
 tell the motor circuit cards how long to feed power to the motors
 to achieve the desired speeds.
 I haven't run a series of tests, but i could certainly believe that 
  low voltages (or restricted current) power supplies would affect
  the validity of calibration data.  
 Changing Calibration data (by doing a "Calibrate motors" ) should
  **not** affect training or any other parameter (except making 
   Tracking Rate more honest).

Ratio is the straight gear-ratio conversion between encoder "ticks"
and expected degree of barrel motion. It *should* always be whatever
Meade set it to.  The only exception to this should be if you build a
 new telescope mount and couple meade motors to it.
 I have heard of one person who has adjusted one of his ratio numbers
 in order to make the scope GoTo correctly... i suspect he has a 
 severly draggy bearing somewhere (or a slipping gear)
If the Ratio number is incorrect a GoTo will arrive consistently (or 
apparently wildly) at the wrong angle.

Training is to determine the amount the motors have to spin to 
overcome the geartrain's backlash, which is the measure of the total
accumultated tolerances in the gearing between the motor and the 
barrel.  By the same token, it tells the Autostar how many encoder
ticks to -ignore- before it starts changing the RA/Dec readout
when you reverse directions.  Remember that the encoder is directly
attached to the motor shaft, and the telescope is many gears away.
When you train, the Autostar merely moves the telescope 400 arcmin
one way, then 350 arcmin back.  You then bring it back the rest of
 the way, using your eye as a fine position sensor.  Then the operation
 repeats in the other direction, and the Autostar averages the amount
 you moved.

Percentage is you telling the Autostar how much of the training value
to "kick" the motors when you request a reversal of direction.
This is applied at a high speed, before dropping to the slew speed
you are requesting (or that "resume sidereal" is requesting).
The "kick' Is attempting to quickly take up the slack in the gear
train, so that your slow-speed slewing actually moves the barrel,
rather than (apparently) sitting there while the motor whirrs slowly.
Too little "kick" means that you -do- sit there for a few moments 
 while the gears take up the rest of the slack.
Too much "kick" means the barrel jumps ahead, then moves at whatever
speed you are requesting.. but it may have already jumped past your
goal.

Tracking rate is the final tuning value: it is an adjustment to
 change from sidereal motion (when tracking an astronomical target)
 to some other rate.  An obvious one is Lunar (the moon moves 
  slower than the background stars by about 1/2 degree per hour).
 Meade doesn't docunt it, but i believe the "units" are tenths of 
  a percent... so saying "-35" means 3.5 % slower than sidereal.

You can directly input Ratio, Percentage and Tracking Rate as numbers.
Calibration is figured for you.
Training is one you have input on, but no direct ability to -set-
 the numbers.

You can -test- the quality of the various values.

Note that ALL of the below are done in Target > Terrestrial mode.
 (except Tracking Rate).

Training is easy (although eventually tedious):
 Choose a landmark that provides good "i'm centered" characteristics.
 Use Object > Landmark > Add   to ""create" the landmark in your 
   Autostar.  Now [goto] it.  
 Note -precisely- where it is in the eyepiece.
 Now slew at least 15 degrees away in a diagonal direction (up and left)
  Go as fast as you like (i use speed 9)
 Now [goto] to return to the landmark.  
 IF the drive is trained *perfectly*, you should be *exactly* back 
  where you started.  Note discrepancies (too high, too left)
 Now repeat that away-[goto]-back in all four diagonal directions.
 Try doing it again. Try going in the same diagonal direction twice.
 Play.  If you consistently undershoot one axis, that axis is 
  undertrained.  If you overshoot, it's overtrained.
 IF it -does- show errors, try -retraining- and see if it improves.
 We are using a LANDMARK (and Target > Terrestrial) for this to remove
 all backlash effects introduced by sidereal drive.  Only the actual
 gear lash is being tested.
 
Ratio test is also easy:  (again: Target > Terrestrial)
 For Az, precisely aim at a landmark.  Bring up the STatus display,
 and scroll to Alt/Az.  Now slew the Az axis completely around in 
 a circle.  When you are again precisely aimed at the landmark,
 the Az readout should read -exactly- the same as when you started.
  (since it should have advanced 360 degrees).
 Once you've "calibrated" your Az readout, note two targets 90 degrees
   or less apart.  Now tilt the scope sideways and run the Alt axis
   between those two targets.  The Altitude readout should match the
   angular distance between them you saw with the Az axis.
When doing an Az or Alt Ratio checkout, approach the target from the
 same direction you'll be running the round-the-horn slew. That will
 remove backlash effects from the test. (i.e. if you are going to
 be slewing Az 360 degrees clockwise, approach the target for the
 -first- aiming in a clockwise direction.)

Calibration can be tested by choosing one of the "degrees per second"
 speeds and timing how far the barrel really moves in ten (or 50) 
  seconds.

Percentages move into the perceptual realm.
 Slew somewhere. Stop. Choose a slow (2 or 3) speed.
 Press the opposite slew key.  Note how long it takes before the object
 starts -moving- in that new direction.  Stop. Repeat in the original
 direction.  Now the other axis.
 A *perfect* percentage would have just a tiny bit more than zero lag
 before moving.  Too low a percentage has a long lag before moving.
 Too high a percentage *leaps* when you press the key, such that it
 would be very difficult to move a target one Mars-width in your
 eyepiece (should you wish to do so).
 If you are using a 909 APM to attach an Autoguider to an LX90, 
   the percentage -is- a factor used in its guiding.  Too large
   a value will cause the guider to hunt (continuously reverse).
  Too little a value will simply mean that it takes a little longer
   for the guider to re-center the target, which is probably not a
    problem (within reason).

Testing Tracking Rate is "simple"... aim at a star and let it track
 for a good long time (1 hour.. 4 hours...).
Since Tracking depends upon proper Calibration, test Calibration
 if Tracking seems incorrect.  There -is- a tracking rate error in
 pre-22eH versions, but it's very tiny.

So there they are: The BIG FIVE.
 Calibration: easy to do, doesn't hurt anything else.
 Training: see how well you trained.
 Ratios: Believe Meade (unless i say otherwise, of course :-)
 Percentages: As You Like Them.
 Tracking Rate: rarely touched.

Hopefully when things go wrong, the above description and tests will
help you isolate (or exclude) some of the mysterious tuning numbers
of the Autostar.  And, when things go right, maybe what -truly- 
 fixed it up will finally become revealed.

Training is by far the most important in terms of improving your 
 comfort with the telescope... 98% of the "rubber banding" and
  creep-after-beep found in 21eK and 22e(anything) can probably be
  completely removed by achieving proper training.

Not covered in the above discussion are the effects of such things
as off-center loading (nose heavy, etc.) or the double-backlash
effect of "flopping over" as you cross from east-facing to west-facing.

good luck
--dick
(ever "will, will not, should," in the above should include the words
 "i believe that it" before them... but then it would be *really* hard
 to read)

And this from Dick:

The -only- reason i can think of is to load the bearings in a 
as-you'll-use-it orientation.

If you do an Alt train with a level barrel, you're pitching forward
on the down swing, which may cause a flop-over effect similar to the
 across-the-zenith one you can get in Az if you're polar mounted.

The 45-degree-up orientation -lowers- the accuracy of the Az training,
since you're working with a smaller visual angle (sin(alt)) that
your eyeball is trying to split hairs in.
We're all just hoping that it's a small effect. (which can be 
compensated for by increased magnification).

Someday (ha, ha...) i'll have to try Az training in both Alt/Az
and Polar orientation, and read back the results to see how much
 they differ.
--dick

Subject:	Training, Setting Percentages and Rubberbanding
Sent:	Wednesday, December 12, 2001 15:13:01
From:	sherrodc@ipa.net (Clay Sherrod)
Ever since new firmware was introduced by Meade which allows for the ETX
user to compensate for system backlash (in this case the delay in
response of either axis when pressing a (usually slow rate) slew key), I
have been answering many inquiries regarding what appears to be an
outbreak of "rubber-banding" the like of which we haven't seen since
version 2.0eg.

We have emphasized the importance of the sequence:

1) RESET
2) RE-CALIBRATE
3) RE-TRAIN

done in precisely that order to achieve the BEST results in Autostar GO
TO accuracy.

However, one step that probably must be listed in that sequence should
be "PERCENTAGES" or at least the desire for many users to set those
percentages.

This very important option enters very heavily into the above
"three-step-sequence."  However, before discussing this, let me say that
setting your altitude and azimuth percentages is NOT ALWAYS NECESSARY.

I always use the slew rate "3" to judge if increasing the percent of
either axis is necessary.  If - with a rate set to "3" - you can detect
motion in the direction you are pressing a particular slew key BEFORE
you can count three seconds, then there is NO need to increase your
percentage in that axis.

Azimuth and Altitude Percent can be changed by going to SETUP /
TELESCOPE /[scroll down to....] AZIMUTH PERCENT [enter] and key in the
value you desire; once set, press "ENTER" once more and then scroll down
one more to ALTITUDE PERCENT and do the same thing if necessary.  It is
not necessary to change both values....only change what is necessary!

Meade defaults the azimuth and altitude percentages back to "01%" when
doing a reset or a download in which you do not want to retain any
settings (with the new ASU 3.0 Uploader program).  Increasing the value
results in the system mechanically "jump-starting" that axis to
accelerate the time in which residual backlash might otherwise delay a
response in the small gears. The higher the value, the less delay
time....HOWEVER, the higher the value, the jumpier the start will be as
well, resulting a very jerky motion to your telescope when slewing in
that axis.

Consequently, "....if it ain't broke....don't fix it!"

However, if you see long delays when set to "3" slew speed, then
experiment with the percentages, starting at "15" and increasing until
you get the desired results.

What works for you might be too jerky to someone else, so just judge for
yourself what feels comfortable.

Thus, IF you chose to reset from the defaulted "01%" in either altitude
or azimuth percentages as described above, it must be done AFTER all the
other three parameters are established.  Thus, we now should consider
FOUR steps of post-initializing from an upload or reset:

1) RESET
2) RE-CALIBRATE
3) RE-TRAIN
4) RE-SET PERCENTAGES (only if necessary).

Setting the percentages BEFORE calibration and training WILL result in
rubberbanding!  Always establish your percentages after you have
communicated the first three parameters to the Autostar and ETX!

P. Clay Sherrod - sherrodc@ipa.net
Arkansas Sky Observatory
www.arksky.org

Return to the top of this page.

Go back to the Autostar Information page.

Go back to the ETX Home Page.