Background
Observatory Automation
Vellman 8055 interface board and TI azimuth sensor bench testing
04/ 10/2006- The installation of the Lesve Dome interface is complete!
Background | Description | POTH | AP mount peculiarities | ACP | Conclusion
I purchased my 10' Home Dome with DC motor drives and the supplied Technical Innovations power supply. Being a DIY'er, I had every intention of developing my own dome drive system. In fact I found that a simple Ir transmitter/receiver based on this circuit worked wonderfully for stepping the dome along with the scope. Similar methods could also be implemented using electromagnetic induction and Hall-effect sensors. Still the Infrared tractor beam as I would call it was not capable of reliably positioning the dome on a long telescope slew and did require some manipulation by the operator for each target. My goal is to be able to operate the dome at least semi-autonomously so that I can sleep while the telescope gathers astrometry and photometric data on multiple targets. TI has an Ir dome tracking system that uses four sensor/transmitter pods spaced around the telescope ota. Based on my results with my Ir tractor beam I didn't think their device capable of unattended long slewing. The most elegant solution would be to have a computer program or dedicated micro-controller determine the correct dome slit orientation. This already exists in the form of Technical Innovation's DDW package as well as various systems availavble from other dome vendors. These solutions are not inexpensive however.
Soon after installing my dome I also decided that I would not be able to tolerate the screaming sound of the oem DC drives with my observatory in the middle of a subdivision. My neighbors have been very accomodating and I did not want to repay their courtesy by waking them at 3am with a dome slew. The original TI DC motors are loud enough to wake the dead! Think of 10 LX200 drives slewing simultaneously in high speed-then amplify that sound several times. The acoustics of the dome geometry amplify everything inside considerably. I tried the sound covers and found the reduction in noise minimal at best. In my opinion this is the design's major fault. The actual power transfer mechanism is brilliant however. ** I understand that the original drives have now been replaced by newer quieter units- I have no experience with these so cannot comment** Donn Starkey also recently installed a TI dome and I learned he ordered the TI drives sans dc motors and built a very quiet drive system using AC gear motors. I followed this approach and retrofitted my drives with a pair of Grainger PN 4Z275 PSC type, reversible, right angle, gear motors. These are incredibly quiet. During a slew I can hear the dome relays clicking away and you can now carry on a pleasant conversation with someone inside the dome.
With the sound issue solved I was ready to start developing a system for automating my observatory. I discovered Pierre de Ponthiere's ASCOM compliant Lesve Dome Driver and Interface. His design is built around an off the shelf Vellman USB interface card. Needless to say I was inspired and immediately ordered the parts and started construction. I found this a fun project to undertake. It was also nice to actually save a fistful of cash in the process. Most of my diy projects usually do not include that added benefit. <G>
Below are photos and description of my installation. ( Click for larger image )
Mounted
on the North wall of my observatory; AC motor switching cabinet (
left ), TI DC power supply (center), the Lesve USB interface mounted
in a RadioShack project box with lexan cover. With my oscilloscope
I discovered the TI dc supply voltage to be very noisy so I added a 12
volt regulator and a few filter capacitors to smooth things out. 12 volts
supplies the control power to drive the actuating relays and power the
azimuth sensor. LED's on the interface box indicate power available, S2
position, as well the state of the K1 and K2 relay windings. The Vellman
interface card also has onboard LED's that can be used for diagnostics.
Not pictured is the manual control switch S1, that I have mounted in a
handpad. S2 can be seen on top, near center of the interface box. This
is used to permit CCW movement of the dome when manually slewing
with the handpad. When the PC is in control the switch is thrown toward
the right and a red LED is illuminated. You can find my basic wiring IO
schematic here. I have left off the various cable connections
and extra LED's. My design is slightly different from Pierre's
since I am using K1 and K2 to control the AC motor slave relays.
My azimuth circuit is identical to Pierre's design.
The
AC switching cabinet and relays that control the dome motors. These motors
are PSC type, so the capacitor is always in the circuit. The motors
and capacitors are wired in parallel. LED's on the front indicate the state
of K3 and K4 DPDT slave relays. K3 initiates rotation and K4 is used
for direction control. Flyback diodes are placed across the coils to inhibit
emf spikes. The enclosure is made of heavy gage steel that limits
RFI generated by switching the AC load. I also should note that I have
two
electrical systems in the observatory. The AC supply for the dome rotation
motors is supplied from the AC drop to the observatory which powers all
the AC outlets and traditional 120 volt appliances. All vital loads ( CCD,
mount, computer, etc. ) are fed from a branch that is first filtered by
an APC UPS. This ensures that any relay switching noise that might propogate
in the AC wiring is eliminated before getting into the sensitive power
supplies for the computer and CCD.
**Update** Shortly after I implemented this
system I discovered an intermittnet problem where the observatory computer
would randomly reboot while Maxim was running when the dome would slew
or jog position. Through trial and error I would discover the solution
to this issue would be to place the AC motor drive power on its own UPS.
This actually acts as an R-C snubber that further reduces the switching
noise. I also note that CCD bias frames appear unremarkable as well. This
has worked very well for over a year now!
I ordered an azimuth sensor from TI. This
works really well. Center is a security switch mounted on the base
ring . I purchased this at Radio Shack and it is used to sense the "home"
position for the dome. The actuating magnet rides on the dome (right)
One
of two dome drives. Grainger
PN 4Z275
The motors rotate at 173 rpm. I machined
a coupler/adapter to mate a timing pulley identical to the size found
on the TI DC motor gearbox to the output shaft of each AC motor.
This reduction with the stock drive system yields a dome rotation rate
~1rpm. Pierre adresses potential problems with dome rotation speed.
Using his spreadsheet my hole-no-hole transition time is only 93 msec.
To date I have not had problems with this setup even when operating many
cpu intensive programs simultaneously. As long as my computer doesn't hiccup
and XP operates as it should I should be fine :-) My Observatory pc is
a Compaq Armada M700 ( PIII 750MHZ, 40G HDD,512M running XP Pro)
This works very well even with TheSkyPro 6 running in the background
Positioning a hemispherical dome to align with a german equatorial mounted telescope isn't as straightforward as slaving the dome aperture to the telescope's azimuth. They have different centers and or offsets. This is also non trivial as the telescope is most often staring through the dome at an oblique angle. Using the ASCOM compliant POTH it is simple to slave the dome slit with the telescope. POTH has a dialog box in the dome setup for specifying mount geometry in relation to the dome. The dome sync code in POTH and ASCOM dome are based on a work by John Oliver with algorithms developed by Toshimi Taki and Chris Lord. You can find out more about Dome Sync here.
This
image is from the Dome Sync website. Using this diagram I measured my mount
offsets from a point where the right ascension and declination axes intersect.
My pier was approximately centered N-S when I installed it however it was
~2.5" West of center. Due to the sheer size of AP1200GTO mount the pivot
point is displaced some 8" North of center.
The measurements (converted to mm ) are:
X = N/S ,(+N) = 200mm
Y=E/W ,(+E) = - 60mm
Z= U/D and (+U) =125mm
r = GEM Axis Offset =650mm
Dome Radius =1.52m
With these parameters entered in the POTH setup
I was puzzled why the dome incorrectly positioned. I did learn that
at least one other individual decided to swap the X and Y values as he
decided the implementation of the POTH code was also swapped. Doing so
my setup looked like this:
These numbers worked! I did have to tweak the
settings a little. Notice positive value for the measured West offset that
is placed in the N/S box. This should've been negative however I did notice
that dome slaving was just a little off low in the northeastern sky.
I tweaked the number until the issue was resolved and I still maintained
good performance in all quadrants. Does this mean the POTH code is incorrect?
I don't know. It could be with my geometry and measurement errors together
this combination works for me. This should be something to investigate
should the dome positioning fail to operate correctly. I have not investigated
this with Ascom Dome however I think the coding is a little different in
that program.
The AP mount's meridian delay
feature is really handy. With the telescope on the west side of the pier
looking east you can slew to an object west of the meridian without having
to do a flip. This can be helpful when slewing to a bright star nearby
for refocus or when looking for a standard field near the region of interest.
I usually enable a 2 hour delay however I noticed that when changing this
setting the dome would reposition itself west of the target depending on
the magnitude of the delay entered in the keypad. I also tried this with
an east delay setting and verified the same behaviour in the opposite direction.
This is related to how the AP mount controler accomplishes the delay function.
I know of no way to address this with POTH. Everything works well with
a setting of zero.
I took advantage
of the Spring 2006 special discount offer Bob Denny had on his
latest version of ACP.
This is an expensive piece of software, especially when you consider that
POTH is free and works well. If you just want to slave the dome to the
scope and control your focuser POTH is all that is needed but for automation
software ACP is the cat's meow! I have used several of the other softwares
and this one is my favorite...I usually setup remote desktop to operate
the observatory computer from inside the house over a 100Mbps LAN.
See the trial image of M51 produced on a humid moonlit night below. This
was executed while I slept just before the mount was automatically flipped
to shoot a dozen minor planet targets on the other side of the meridian
for astrometry observations. When I awoke I found the telescope safely
parked and the camera shutdown. Current operating practice is to generate
a script using ACP planner and TheSky Pro 6. I then go to bed while the
telescope gets to work. Now I have reams of data to analyze on cloudy nights.
Since the images are plate solved on the fly , astrometric reduction with
Pinpoint is quick and effortless.
M51 acquired with 0.36m Meade SCT and Optec f/5 reducer during dome
testing.
LRGB 60:30:30:30 ST10XME @ full resolution ( 1 x 1 binning )
04/10-04/14 2006
( very bright moon )
The Lesve driver and interface work
very well. I can send the dome to any azimuth position with accuracy and
excellent repeatability. The system has proven itself reliable for over
a year now with long unattended observing runs spanning the whole
night. I have examined several different control systems over the
past and none come close to Pierre's simple robust, approach. This design
can easily be duplicated by others with off the shelf components. There
are no lookup tables to generate, no complicated electronics or micro-controllers
to program.The Vellman board is even available already assembled and very
reasonable. I already had some spare parts laying around but would estimate
this project is easily under $200.00 for the basic design and that's with
a pre-assebled Vellman interface card and azimuth sensor from TI
!
The best thing about all of this
is that I can now gather multi target data for research and pretty pictures
while I sleep!
The only thing remaining is to design a system to operate the shutter... more to come.
Last updated 6/2/07