Rifast collimation

The Rila is a new scope for me, mine’s a 400mm f3.75. I’ve previously collimated an SCT and of course my refractors haven’t require any changes since I’ve brought them.

I’ve got to say the manual from OS looked like it was designed for an RC but the guys in their support are really helpful and know what they are doing. Cris at Astronomy Alive was a God send, not only putting me in touch with the OS guys but also other amateurs who have had to work through similar configurations.

The Rila has a number of possible adjustments.

1. The adjustment of the spider that holds the secondary mirror
2. The adjustment of the collimation screws on the secondary
3. The adjustment of the collimation screws on the primary
4. The adjustment of the tip/tilt adaptor that attaches the imaging train to the back of the scope.

So all up its 13 degrees of freedom. Be prepared to spend some time.

First of all if your collimation is out the cause is most likely your tip/tilt adaptor. Play with this first. Check out your corners.

The equipment setup:

photo 4



I had a replacement collar (the red bit connecting the focuser to the scope) sent out from Italy so I could attach the collimator and the tip/tilt was very different.  I should have marked where everything was first. Next time.

Laser attached:

photo 2

Don’t bother to remove the primary baffle, it makes no difference. Don’t bother using the laser with circles because the rings cannot be seen because of the size of the secondary. See photo  below.

photo 1


But the rings are clearly visible on the secondary.


photo 3

OK, the steps:

1. Measure the distance from the secondary mirror holder to the external octagon. It should be possible to get the four equally spaced measurements from any section to the secondary within 0.5 mm. I used a vernier calliper all measured up to 140.72mm.

2. Loosen the locking screws on the primary mirror. Screw down the collimation screws on the primary and secondary mirrors so they are fully done up, but not super tight.

3. Put the laser in the back of the scope with the provided collimator bracket and adjust the secondary until the laser hits the center of the circles in the secondary.

4. Use the Tak collimation scope and put the dot in the center of the secondary mirror circles. The Tak doco is pretty good on this. However you cannot see the spiders and you cannot see the reflection of the edge of the primary mirror. You can see the shadow of the secondary but it should like up pretty well because of step 1. The tak causes an additional shadow which you can see around the outside of the secondary mirror shadow. Ignore it. Mine was not quite a circle and I discovered it move when I rotated the Tak anyway. When you have the Tak at the right focus you can see the central dot and what looks like diffraction rings around it. using these you can get a very accurate positioning of the dot inside the secondary circles, adjust the secondary’s collimation screws to ensure the dot is in the center. Just beware that this is still mechanical alignment.

5. Pull off the collimation adaptor and put your imaging rig back on, defocus so you can see the diffraction rings clearly, bigger stars say >20 pixels across is fine, we can make them smaller later, start taking say 2 second frames with your imaging rig and put the star smack in the center (ok plus or minus 20 pixels is fine) and make sure your tracking is working well. Point it at a not to bright a star say mag 5 and check out the star shape. This is typically what you do with SCTs but here you are playing with the primary mirror collimation screws only. You will need to back each of them off say 1/2 a turn to start with so you have some room to adjust. The basic idea is to move the star, using the screws. The star should move in the direction of the squished diffraction rings. The idea is to get a nice circle. Equally spaced diffraction rings on each side. You will need to play to determine which screw moves which angle and if tightening or loosening moves the star so that the diffraction rings spread out or move closer together. This is standard collimation stuff and there are lots of examples on the net about how to do this. You will need to do this with a star that is well out of focus, and then closer to focus. The closer you get the smaller the movements need to be. This star has to be kept in the center of the image all the time so be prepared to move your scope after each adjustment. This is in my opinion the longest and most frustrating step. Ignore the star shapes on the edges of the frame. Take longer shots up to about 30 seconds to see how you alignment goes as you get closer to focus.  This means you end up with something like this, one sub 1200 seconds, Ha, stretch only, nothing else:

M16(en)-003Ha E11p4 E1p0 E1p0


The central stars are pretty close but check out the corners!  The top left and top right are miles out, with the bottom left OK – but not brilliant – and the bottom right almost as good as the top right.

6. The final stage is to make sure the stars are the right shape across the entire image. You do this by making very small adjustments to the tip/tilt plate that attaches your imaging train to the scope. Mine is a red collar with small hex grub screws. undo the larger grub screws a very small amount and try tightening the smaller grub screws fractions of a turn. Check out the edges of you image while you make these adjustments so you can see if the star shapes in the corners are headed in the right direction.

Good luck its taken me days.

Guiding v2

After many months of waiting I finally have some new gear (more later).  I’m back to problems with guiding.  PHD worked perfectly on the Tak FS 102 NSV with my SBIG ST402ME (9 micron pixels) and my Planewave Ascension 200 HR mount.  The the focal reducer I was running gave me an arcsec/pixel ratio of about 3.  Deviation from zero was max 0.4 pixels on a bad seeing night.

Enter the new gear – OfficinaStellare 400mm F 3.8 RILA.  Focal length is 1520mm.  The guidescope now sits on a MMOAG with my Proline.  The guiding was lousy!  More like averaging 3 pixels variation.  Now the stars in the guider where not very round – they where lousy not even oval.  First step adjust the angle of the prism/mirror on the MMOAG. One small piece of folded over tape later (under the back edge of the blocks that hold the prism) and my stars now look like ovals.  Tracking improved a lot.  Average deviation from the mean is < 0.5.  Now I use PHD 2.0 because it’s easy and I’m dumb.  Alas it does not seem to support binning – come on Craig.  The SBIG on this rig has an arcsex/pixel ration of 1.22 – not good in Sydney sky’s.  So I went back to Maxim DL for guiding and binned 2×2 (arcsex/pxiel now surprisingly 2.44).  My guiding even with all the high cloud last night was great – deviation from average max 0.2 pixels.  Yeah!

Training the scope for guiding in Maxim is still a pain – some times it misses the star altogether, sometimes the stars become stripes etc BUT regardless it guides really well.

Arduino Based ASCOM Rotator

One to many cloudy nights has spurred me into action. My imaging setup is getting closer to automated but I’ve had a distinct missing link, the ability to choose a guide star from within my astronomy applications. Since adding the MOAG and updating my guide camera I’ve added a Takahashi CAA rotator and it provides excellent manual rotation.

The reason I selected the CAA was price and aperture. I did look longingly at a number of electronic rotators.

Automation of the rotation was another issue. There seems to be lots automation for mounts and focuses but nothing for rotators. Google was not my friend. So I decided to write something. This is very much a work in progress but so far I have adapted the code from the SGL Automation guys and attempted to make it rotator specific. It supports ASCOM v6.

The ASCOM interface for rotators is reasonably simple, Move, Position, etc and the ASCOM documentation and templates are excellent. Well done standards guys, shame us health professionals can’t get our act together as we’ll as you have.

First go was to get the basics operating: Arduino, USB to serial driver, stepper motor, easy driver motor controller a couple of belts and a pulley. Check out the photos below.

I’ve only tested it with Maximdl using my bench setup as yet, but it seems to work well.

The bench setup.


The arduino (note the resistor is for the reed switch, keep reading).

The reed switch and magnet.


Reed switch and magnet.


I was having trouble this weekend determine the number of steps to make 360 degrees so I dug up an old reed switch, found a very small magnet and introduced homing and a step counter into the software, as of Sunday night it all appears to work.

The Arduino Project and VB in the solution - ES - works but its still a little messy to release, but here you are anyway.

Code in action below

Sample rotator screen shots with debug



If you’ve been reading my blogs you will know that I’ve been struggling with tracking. It’s been months now and last night the problem finally became clear. I’m just a very slow learner.

My ST 402, is a great little guide camera, hanging off my MOAG I never fail to find a star. The problem was that I was getting pixel movements during guiding of up to 1.8 pixels in X and Y. Obviously the resulting images where poor. I was reading again about sampling and came across the old 2 arcsec per pixel rule. I run my FLI Proline 16803 at 2x binning to get a ratio of about 1.7 arcsec per pixel (as reported by Maxim). The ST 402 has pixels of 6.9 by 4.3 so the ratio when hanging off the back of my MOAG is about 0.36 (an average of the two dimensions). I changed the binning to 3×3 giving a ratio of about 1.1, and surprisingly the tracking variation is now reduced to a max of 0.4 pixels in either direction, average deviation is less than 0.15. The pictures look heaps better as well.

Now I have to go back and fix my auto focus, FocusMax is causing me grief.


26/01/2013 – I’ve worked out FocusMax and what a wonderful solution it is.  Perfect focus every time!  The trick is choosing a dim enough star (in my case around 7 is perfect) and following the instructions.  They one major factor in this was counting the circles on the profile where the line began to diverge from straight, to determine the Near Focus HDF.  After I got this right it worked like a charm.  See the pictures below.

New web server

I’ve just spent all day with my son migrating my web and MySQL server to a Pandaboard. It’s ARM based dual core 1.2 GHz, with 1G of RAM using 64G SD card for disk. Very low power consumption. Our VM stack with iSCSI back end is no more, they are all RasberryPi’s, less power again.

So far it seems to run really well.

Dew heater

My SCT fogs up most evenings. I’m a little reticent to use a hair dryer and not sufficiently rich to buy a proper dew heater, so I played around to make my own.

I dug out some nichrome wire, cut it to the length of the diameter of my scope (about 1m) and played around with voltages until I had a small increase in temperature (comfortably warm enough to hold onto with my fingers). It must only be a few degrees. Just be warned its VERY easy to fry you fingers so start with low voltages and work up.

I had a Jaycar kit MOSFET power supply hanging around, slight overkill I know. I dug up a relay I’d used for something else, added a DHT22, an LCD display along with an Arduino and behold!


The power supply is turned on by driving the relay. The Arduino code reads the DHT22 every 5 seconds, calculates the dew point temp, displays the current temp, the dew point temp and relative humidity. If the current temp is within 4 degrees of the dew point it turns on the power supply.

BTW the knob is only used to control the contrast on the LCD as it gets to bright when the scope is taking pictures.

Note the insulator on the nichrome as well. My scope conducts!


An now turned on


More experiments with collimation – an alternate approach

Can anyone tell me why this shouldn’t work? I brought a Hotech SCT collimator a week or so ago, I was sick of not being able to accurately collimate using the out of focus star method. The Hotech was quite easy to use but my final collimation was worse than before, just poor operator I suspect.

I started looking at 60 second shots and the stars were clearly badly collimated, the longer the exposure the worse the star looked, and it’s not the tracking. Out of sheer frustration I started taking 1 second exposures continuously and began twiddling the secondary mirror knobs. The star began to get smaller, and smaller and smaller and more evenly shaped. It ended up with a perfect circle at 400%. The star was out of focus, but not so much that it looked like a doughnut.

So next step run the autofocus routine, the star (it was mag 5) shrunk still more and was still perfectly shaped. Collimation fixed. Now when I’m on on either side of focus or in focus the star is a nice circle. Why can’t we do collimation this way all the time? It was nice and easy and produced a great result.

M16 v2


After ages away from the scope I had some time to snap M16 in a wider angle.   This is taken on my Tak FS 102 and QSI 583 WSG, in LRGB.

the 9 subs are 5 minutes each of LRGB at -12C, using Astronomik 2C LRBG filters.

Guiding using Maxim dl

See update below.

I’ve been trying for weeks to get guiding to work perfectly in Maxim DL. I’ve just upgraded to v5.22 in case it was a bug.

Here’s the scenario. I’ve recently added an 8 filter wheel extension to my QSI 583. I also added the guider port. I’ve attached my Orion autoguider (with lense) to act as an off axis guider. I’ve had real problems getting Maxim to calibrate the guider using the stars I’ve selected. With my combination I’ve found that with pixel intensities in the star of greater than about 160 then the little red L would appear reliably every time.

Over the last couple of nights nothing I did would get it to calibrate, even on the brightest stars. I tried replacing the USB cable (I was getting desperate), playing with the previously good settings but to no avail.

I went back to PHD and it worked first time on much fainter stars with the old cable. I know when I’m beaten PHD it is.

21/11/12 I’ve just acquired an ST-402 and got it working as a guider. Guess what all the problems experienced under Maxim dl guiding are gone. In short buy a better guide camera.

Temperature focus compensation FLI Atlas

My FLI Atlas is a lovely focuser however it does not have temperature compensation.  After many frustrating nights ducking out to fix my focus after one or two images I decided to do something about it.  Enter Maxim DL Scripting.

I figured I could read the temperature of the focuser store it in a file on disk and then at the end of the next image read the temperature again and compare it to the one I wrote to disk.  If the difference was substantial (i.e. some number in my case 0.2C) I could move the focuser an specific amount.  In my case for the LX 200 it measured up to be about 0.19mm per 1  C.

Maxim dl scripting is pretty straight forward except when it comes to the Focuser.  If you create the object directly you can end up with two instances of the focuser object so what I did was use the Generic Hub and pointed Maxim to this and the Hub I pointed to the FLI Focuser.  The code is setup accordingly.

I run it at the end of each image (configure under Autosave->Script).

Code attached as a vbs so it only works under windows.  Download it from here it’s only a few lines.