Here is a project for an automatic anti-glitch and anti-vibration camera shutter. I live near a railway line and a busy road, and vibrations from trains and buses etc
often give PHD guiding heavy weather for staying on track. Passing clouds dimming the guide star also send PHD off target. The chances of
completing a 300sec exposure without a glitch are pretty low, and a 600sec shot pretty much impossible.
Looking at the PHD log and a bit of excel work showed that a shutter closing to occlude the camera during tracking excursions would be effective
in dodging 90% of the > +/- 2arcsec excursions for a penalty of 20% of the exposure time. Which seemed an OK trade if it could be done.
As PHD loses its grip, the RA/Dec corrections get larger before getting back on track - and these pulses are available at the ST4 interface
coming off the guide cam (a ZWO 120MM). Basically the four RA+/- and DEC +/- pins pulse low when a correction is requested - the
greater the correction needed, the longer the pulse. This repeats for every new incoming exposure of the guide cam.
The idea was to get hold of a filter wheel, motorise it, put it in the image train, and have it shuttle between two filter positions - one position
with a 0% transmission filter in it, and the other completely open.
The ST4 pulses - in separate RA and DEC channels - are used to discharge an RC network, which then recharges when the pulse terminates. The
voltage across the capacitor rises and falls. The charge and discharge resistors are separate, so the attack and relax time constants
can be adjusted independently.
When adjusted correctly the RC voltage hovers in normal operation just above a trigger voltage, and when
the excursions become extreme the voltage falls below the trigger point, before rising back above the trigger when on-track guiding resumes.
The passage through the trigger points are used to signal the motor to rotate the filter wheel closed and then reverse back to open. A pulse is
also generated at each close/open to start/stop a clock so that the amount of exposure time closed can be monitored.
Here is a pic of the final assembly mounted on the scope (Skywatcher short tube 6inch F5 refractor, the filter wheel shutter, Celestron 2.5x barlow,
9 position 11/4in filter wheel and Atik 460 cam).
The first video is a Pix Insight blink video from a run of 33 exposures at 300 seconds without the shutter. Target was M63 Sunflower Galaxy, red filter. I have
cut out the worst frames, but the general vibration issue is clear. Next video is a run of 17 exposures at 600 seconds. M63 again, green filter.
The vids are of the raw, un-calibrated frames so the images are pretty CCD so-so, but overall I think it works.
Attached are the circuit diagrams and also some pics of assembling the strip board and the finished control unit.
The shutter filter wheel was motorised with a Hitec 485 servo modified for continuous rotation. The servo is fitted with a 48 tooth plastic
Mod0.5 gear on the servo arm. This engages with the thumb wheel of the filter wheel. Matching gear teeth are fabricated on the thumb wheel
with a needle file - the thumb wheel rim has little grippy indents in it at half the pitch of the gear, so filing the teeth into the wheel is easy,
and you only need enough teeth to drive two filter positions and not the whole thing.
The internal click stop spring of the filter wheel is removed, so the filter holder rotates very easily inside the wheel. Packing washers
are needed over the axle to stop the filter holder wobbling on its mount. Attached are some pics of the modifications to the filter wheel.
Stops are mounted on the inside of the wheel to obstruct and stop the black filter at either ends of its motion open and closed.
Little bits of the furry velcro from hook and loop tape are stuck onto the stops to act as shock absorbers.
Attached are some pics of the filter wheel modifications.
The servo turns the wheel quite slowly - it takes about 3 seconds to open or close. So the circuit has to drive the motor for that amount of time,
(it is adjustable). There is very little torque required to turn the wheel and direct connection of the motor to the shaft would be possible if
a much faster response is needed. The current arrangement doesn't upset the imaging train with vibrations, which could be an issue with
a snappier response.
Here are some pics of the continuous rotation mod to the servo. The 485 is quite a torquey beast and a lower spec one would do.
(NB: The circuit does not provide a servo driver - all it does it switch the forward and reverse servo control signal from another unit. Servo
driver circuits are ten a penny on the net. A single 4538 dual monostable will do the trick and provide forward and reverse).
Now to fix that pesky guide scope flex that no amount of ironmongery has sorted (yet) !
attachments: the 500k limit on attachments has put the curse on these, so only the circuit diagram and scope pic uploaded!
Shutter on scope pic
Pix Insight mp4 - M63 before video
Pix Insight mp4 - M63 after video
Strip board assembly pics
Filter wheel modification pics
Servo continuous rotation mod pics