Things have changed a lot since the last time that I ran Focusmax. Among other things, I have built up a whole new astro-PC, and I never really understood FM well enough to just make it work out of the box, so I was essentially starting from scratch last night.

It was a gorgeous evening, dark sky because the moon is 3 days past full, and clear as a bell until midnight. I’ll get to that later.

I had characterized my focuser before sunset, so I knew that Pumpkin was near focus, and that the ends of the focus run were set properly.

As an aside, I also discovered “Configurations” in MaxIM. Very cool. It remembers which cameras, which exposure type, etc. were up so you can swap from “guider only = PEC” to “main cam only = Focusmax” to “main/guider = DSO”… nifty!

So I needed to set several parameters very carefully in Focusmax. I figured out several of them, but it took hours, and a lot of experimentation.

Here is how the night went.

I tried a “first light wizard”, which is supposed to “characterize” the focuser. The problem was, I couldn’t get it to stop scrolling out. It kept defocusing the star more and more until either the star was too big for the subframe, or the star was too dim for FM to detect.

I learned some things.

First, a 100×100 box is more than large enough for FM to work with a fairly out-of-focus star (anything that’s less than 1500 steps out of focus should fit, on Pumpkin). 50×50 is actually a pretty good size; it keeps “double stars” from creeping in, but you have to have the focuser in a pretty thin band of focus (within about +/- 500 steps), or you end up doing a lot of “it’s out of the box, increasing the size” stuff at the ends of the V-curves. Works, just twice or 3x slower.

You want to set a exposure length that is long enough to get a hit even when you’re at max defocus. I started with 3sec, went to 5sec and then 8sec. I think that 4 or 5 is probably the right answer now that I can restrict FM to the area near focus. Maybe lower, but again, you run into “can’t find a star, increasing exposure time” stuff, which just eats time.

I never got a first-light to complete. The “characterization” of the focuser never finished, because FM ran the star until it was bloated beyond belief and kept going until it lost the star. Here’s what I think was happening. According to something I read on the ‘net, there is a “max increment” or something like that, it’s in the focuser driver settings. I always thought this had to do with the “jog” function; ie “don’t jog more than ‘max increment’ in one go”. Something to do with overheating the motor or some such. I am now led to believe that this is “how many steps out of focus will FM go during the first light run”. So I need to change this from 10,000 (current setting; focuser never jogs more than 3500 anyway) down to about 1,500 or so. That might get the first light wizard going.

OK. So. I never got First Light working, so I just decided to force a V-curve. I did the “set the beginning, set the end, set the step size” type first, got a so-so curve, then swapped to the “set the middle, set the half-width, set the # of steps” type and ran off 3 good V-curves. With these 3 curves alone, the differential thingy (PI?) is at 26 (steps?) or so, about 50µm apart. Doing a quick calculation of focus tolerance:

Focus tolerance (depth of field) = 2 * f * d
where f is the f-ratio of the scope and d is the Airy Disk diameter, which is:
d = 2.44 * lambda * f
where 2.44 is a constant (why), lambda is the wavelength of light (bluer light has less tolerance than redder, so Ha at 656nm is easier than “white light”, which peaks near 555nm), and f is again the f-ratio.

Focus tolerance for Pumpkin in “white” (really “green”) light is 63µm. Veronica is about the same, 65µm.

So 26 steps (I already calculated the focuser to be doing ~2µm steps) is enough to hit the focus band, but getting it a little more dialed in would make me feel better. More V-curves next time out.

Now that I kind of understand what I’m doing, I need to start running lots and lots of V-curves. I’ll need curves for every setup; Each combination of scope, camera, and filter needs work. At some point, I will be able to figure out how much offset to use when the Ha filter is in; that will save time. But there is a lot of work ahead on Focusmax.

For Pumpkin, with the IDAS LPS and no Ha filter, I used exposure time of 8s, bounding box of 100, half width of 500, and step size of 25 (for 40 steps on the curves), and had pretty good success.

That’s a good starting point for V-curves for future scopes. I can crank it down to 3-5s exposures once I’m comfortable with the focuser.

I played around with different star fields. The area near M44 (but not the cluster itself; too many stars) is pretty rich with medium-brightness stars that FM seems to like.

I had to shut things down when the scope hit its safety limit and the clouds rolled in, both within a few minutes of each other, near midnight.

http://www.cloudynights.com/ubbarchive/showflat.php/Cat/0/Number/1691614/page/413/view/collapsed/sb/9/o/all

I decided to move forward with the strings concept before trying to do a total rework.

Things ended up being kind of expensive anyway; the parts to finish off the strings turned out to cost $110 or so.

Part of this is because I made a design decision while staring at hardware at the store; instead of adding lower strings, I am going to cover the distance with large turnbuckles instead. This meant buying 6 turnbuckles instead of 3, and bigger ones, which are a little more spendy. Also a couple of extra bits and bobs to get them connected up.

But I think this might help me in 3 ways.

First and foremost, I will be moving forward with the design. That’s key. The scope has sat around for too long gathering dust.

Second, the turnbuckles will be easy to install and very very stiff.

Finally, in a serendipitous convergence, the turnbuckles will add much-needed weight to the tail of the telescope. Currently, with the non-motor focuser and minimal camera rig at the upper cage, I need about 5# of extra weight at the tail to counterbalance. The turnbuckles will, I hope, help with this.

I am so glad that I made a trip to the hardware store. Now I have all the parts on-hand (and a way forward), I realize I’m 6 drill holes and about 45 minutes of installation away from being able to test again. Cool!

These are some MaxIM logs from the past few runs.

After 10 days of fighting Veronica, I’d had enough. Tonight, I swapped the camera back to Pumpkin, rebalanced, and tried my luck at two targets.

With Cassiopeia showing above the house, I decided that I wanted to try for IC 1805 and IC 1848, The “Heart” and “Soul” Nebulae that lie between Cassiopeia and Perseus. Note that it’s July, not October. But, since Pumpkin seems to do pretty well (owing in no small part to the 5”/px and huge FOV and huge guidescope), and since I like to avoid running the scope into the pier, I find that I choose objects that are “a little low” and let them rise all night.

But these guys were still below my “horizon” (read: The wall of the observatory ) yet, so I needed to shoot something else for an hour or so, until Cassiopeia rose a little bit.

So, I decided to continue my widefield survey of Cygnus/Cepheus, and did an hour (back to 20m subframes — I love that li’l orange scope!) on Sh2-119.

The Sharpless nebulae are opening whole new vistas in my target list, and I’m *really* enjoying going after these objects in full knowledge that I’d probably get nothing but noise and pain if I tried this without the Ha filter.

Sh2-119 in Cygnus

I like Cygnus. It’s really good to have Pumpkin back. I have nothing more to say about this object.

Now it was late enough that I could start getting some time on Cassiopeia. Some clouds threatened to roll in, ate my guidestar for awhile, etc. Pumpkin just powered through all that. I only lost one frame out of 13 (to clouds). Since IC 1805 was laid in “horizontally”, I missed the nearby IC 1848. That’s for tomorrow night. This is a full 4 hours of Ha from Pumpkin on an object that I have no right to be shooting for another few months.

IC 1805 The Heart Nebula in Cassiopeia

In a word, Wow. I’m back on track again. I am having fun with my widefield shots in a way that I haven’t been having fun for awhile now. I want to buy an OIII filter and go back and shoot all this stuff again. I am giddy when I see the cool stuff show up in each subframe.

The guiding is a little rough. But the results speak for themselves. I am ready to shoot my way through the rest of the month. Yay!

I stacked the 16 frames shot last night into a single GIF animation. There is a very strange “retrograde motion” thing going on during the course of an evening.

I don’t know what’s causing it, although candidates now include:
- balance (been messing with this for 3 days with no perceiveable difference in the problem)

- pointing model (is there something bad built in?)
- motor speed (ie runs too fast/slow)
- tracking speed (ie Sidereal vs King?)
- guide magnification (add barlow)
- tube flexure
- bad spots on the RA gear (am I hitting the same section of the gear every night?)

I’m sure there are others.

Here’s the animation. I didn’t include it inline because it’s 500K and who wants to download a thumbnail that big?

If you have seen this before, please let me know.

OK, this is getting silly. Either I have no idea how to balance a telescope properly (which is not outside of the realm of reason), or balance is not the problem with the rig. On the other hand, I get this very distinct oscillation that’s worse at either end of the run and OK in the center. grr.

So, as usual, I only saved 5 out of 16 images.

However, they came together OK. Here’s The Cocoon Nebula, IC 5146, in Cepheus:

Here is a cropped view of the object (1500×1000->900×600 as usual)

The large image is slightly cropped from the full frame as well, but at lower right, you can still see the “Star Wars” effect that I’m getting because (I believe) the spacing between the MPCC and the CCD is being thrown off by the Ha filter (it goes camera, T-ring, Ha filter, MPCC, focuser, and the MPCC wants to be 55mm from the camera, but the filter is extending that slightly). I’m going to ignore this for now, since it’s a small quibble when I’ve got a bigger problem.

My next trick is to add a Barlow (telextender) to the guidescope, in hopes that guiding at a higher magnification will help the scope get rid of this crazy oscillation. Also, I’m going to make an animated GIF of the images so that I can analyze the oscillation more carefully. Stay tuned.

Cepheus is still a little low in the sky to get a full treatment. But I wanted to see what would happen if I shot an object that was low enough that the scope didn’t collide with the mount during the night.

I learned something very interesting. The mount has tracking problems outside of a certain “good” range. I had been starting my evening in this “good” range, and thus only getting a few good images at the beginning of the evening, with junk after that. Last night, I had some junk at the beginning of the evening, getting progressively better for awhile, then worse at the end of the evening. Very interesting. At least I have something to work with.

In any case, this time, instead of getting 4 or 5 out of 16, I got 8 out of 16 images (because I got the whole “bell curve”.

This is NGC 7635, The Bubble Nebula. The bubble is being created (from my understanding) by a Wolf-Rayet type star, which is apparently cranking out some serious solar wind and making a bubble in the surrounding nebula.

I just think it looks cool.

Here’s a closer view of the nebula core:

In each of these images (and most other images on my website, I followed my normal resizing pattern:
- for a “full frame” image, I crop to the largest 3×2 aspect ratio image I can

- for a “cropped” image, I crop to a 1500×1000 area
- either way, I then resize down to 900×600

900×600 is a nice size for the web, usually produces astroimages that are ~100k in size. Resizing down a little helps to “tighten up” the image, makes things look a little sharper, de-emphasizes the noise in the image, etc.

What I’m saying is that, generally, shots from Veronica can be as large as 1.3°x0.9°, but often I’m taking a 0.6°x0.4° chunk out of that. So far, with Pumpkin, I’ve always done full frame, because that’s sort of the point of a widefield scope. Pumpkin’s field is about 4.2°x2.8°. But because I do some cropping and resizing, the scale from image to image is not an exact 1:1 match to other images. The 1500×1000 -> 900×600 type images are all the exact same scale, of course.

My goal, instead, is to make the object as large as possible in the final image, while maintaining some kind of concept of scale.

I’m having a tracking problem with the mount. This is all eerily familiar. Things run fine for a couple of hours and then the wheels fall off.

This results in only being able to use a fraction of the frames that I shoot in a given session. Very irritating.

I adjusted the counterweights last night, making the mount slightly heavy to the east, rather than being in perfect balance. I am told that this keeps the RA gears meshed, which helps the tracking. Perhaps I didn’t find the sweet spot of east-heaviness. Perhaps the whole concept is hooey. All I know is, I again lost 2/3 of my frames; what follows is a stack of frames 3,4,5, and 7 out of 16.

I adjusted the CW back to near balance when I shut things down.

On a slightly positive note, the mount hit its safety limit and stopped tracking before touching the pier this morning. This is a function of where the object is in the sky, but joy anyway.

Last month’s shot through Pumpkin shows, as usual, much more context around the target.

NGC 6888, The Crescent Nebula in Cygnus

Another nice Ha image, another night where I’m a little disappointed in the mount’s performance. I may have to tweak the balance a little more.

Saturday was cloudy, but Sunday night was clear, so I set up the scopes for another night of photography.

I decided to shoot Sh2-101, The Tulip Nebula. I am still working on figuring out what Veronica’s FOV is compared to Pumpkin’s. This nebula turns out to be nicely-framed in Veronica’s FOV, big enough to show detail without cropping, small enough that it’s not difficult to cram in. As usual, Veronica is the workhorse telescope she’s been for me for 5 years now.

I am having some kind of strange error on the mount — it works fine for about 2 hours, then goes nutty and all the images have trailed stars after that. On Friday night, where it was pretty cloudy, I was willing to blame the clouds for this. But I didn’t see any clouds on Sunday (which is not to say they didn’t appear after I fell asleep). I worry that I’m having a cable-pulling problem or some kind of imbalance thing going on. This will take further investigation, but I’m quite frustrated that essentially I am wasting 2/3 of my shots (and all of the ones after I turn in). Again, this image is 6×20m, a full 2 hours, but I threw away another 9 or 10 images that were on target but with trailed stars. I’ll play with the balance and see if that does anything.

Anyway, such as it is, here is last night’s image:

Compare to my previous shot. It’s a different kind of “feel” between the two images — Veronica highlights the brightest part of the nebula, and produces lots more detail because of the extra magnification. Pumpkin gives much more context to the area, and I certainly never would have known about all that wispy dust lane detail if I’d only seen the Veronica image.

It seems that the two scopes match each other in resolution pretty well. I’m looking forward to more Cygnus hunting with Veronica this month!