Cooling the Meade DSI IIc with a Peltier Chip

28 May 2014.  I tried to get an image last night, but the attempt was foiled by persistent clouds and wind: the Clear Dark Sky forecast was not very accurate last night.  I did get to play with the “new and improved” chiller setup for my DSI IIc camera, so I thought I’d report on that.

I wish this idea were original to me, but it certainly is not.  A couple of years ago I decided to follow others’ leads (I googled Peltier cooling of Meade DSI cameras, and I hit on a web site that gave me some ideas (http://www.backyardastronomy.net/peltier_cooling.html)).  This guy used an inexpensive Peltier kit with a computer CPU fan to cool down his DSI.  I decided to try it.  I milled off the back of the heat sink so I’d have a flat surface where the Peltier could be attached, right above the “cold finger” that cools the back of the CCD.  I put a salvaged CPU cooler fan (~$5) on top of the Peltier, to carry heat away from its hot side.  Thermal paste is applied to all surfaces that need good thermal conductivity (CCD/cold finger, heat sink/Peltier, and Peltier/fan heat sink). Last fall I added insulation around the outside of the camera and switched to a smaller Peltier chip (TES1-12704, ~$5) that is a bit better matched to my power supply (~$80 for a current-regulated supply).  The temperature (as reported by Nebulosity’s DSI control code) vs. current performance is shown below.  That got things cold enough that I was having issues with dew and frost, but my big mount failure and winter intervened (I don’t cool the camera when Utah nights keep it cold without any artificial intervention).  To combat the moisture (which really isn’t bad at my site; after all, all of Utah is desert) I loaded 4 small desiccant packs (about $0.10 apiece) inside the camera body, and after watching numerous rubber bands rot and fall apart, I used cable ties to secure the cooling fan and Peltier chip to the camera (I’ll have to add a picture of that later).  Anyway, the bottom line is that for less than $100 this rig lets me cool down to about 15 °C below ambient and significantly decreases the noise on those hot summer nights.  

The milled-off heatsink back of my DSI IIc camera

Peltier chiller performance.  I tried a couple of different heat sinks, and the larger, higher capacity sink worked better.

Setting Up for MacBook-controlled Astrophotography Using DSI Cameras

An Astrobin user asked me about using my ST-80 scope with DSI cameras and controlling it with a MacBook.  I spent enough time replying that I thought I’d post my answer, edited a bit, here.  A picture of my setup is shown here:

In the picture, I have the imaging camera mounted on the 9x50 guide scope and the guide camera on the ST-80 refractor, but normally I have the cameras swapped.  It just depends on what field of view I'm shooting for.  

Optically, I use either a Baader Fringe Killer or Baader UHC-S filter on the front of the DSI IIc imager, because neither the ST-80 nor the 9x50 is apochromatic.  The filter helps suppress violet fringes, and the UHC-S (which I prefer of the two) also blocks IR and enables sharper focus.  I sometimes also use an Antares 0.5x focal reducer threaded into the DSI IIc nosepiece (the filter goes on top of that); again this is to achieve the field of view I want.  The reducer causes distortion around the edges of the field, but because the DSI IIc's chip is fairly small, it isn't too bad.  

To connect the equipment, I use a powered USB-2 hub at the mount into which I plug both DSI cameras, and, until recently, a Shoestring GPUSB for controlling the mount.  My mount is the very old CG-5, to which dual axis motors with a hand control have been added; you have to wire an ST-4 guide port jack into the hand control to connect the GPUSB, and instructions for doing so are on the Shoestring site.  The hub connects to a 20 m active USB repeater cable (like the powered hub, I bought the cable at Monoprice.com), and I run the cable into my living room and connect it to my MacBook.  I control the imaging DSI IIc with Nebulosity 3, and I control the DSI Ic guide camera with PHD2.  Until recently, I used the Mac versions of everything.

A couple of months ago, after frying the mount controller (by plugging it into 12 V instead of the 6 V for which it was designed), I built a PicGoto Simplificado mount controller and have been using that rather than the GPUSB+ST-4+Dual Axis Controller.  Because the PicGoto server is Windows software, that required me to start using VMWare and Windows 7.  I had long resisted using VMWare/Windows, because it was slow to launch and just clunky, but in January I replaced the hard drive in my MacBook with an SSD.  The SSD makes loading VMWare lightning fast, so now I can use VMWare with no pain.  I currently control the mount using PicGoto server + Cartes du Ciel, and love it for how easy it makes finding targets and for how much more quickly I can slew the mount (30x sidereal, with the same motors as before, rather than the 8x I could get before I broke off the switch on my old hand control).  I don't do long slews, because even 30x is pretty slow; I use a Telrad finder to find a bright star near my target, sync on that, then let the PicGoto + Cartes du Ciel find the target; this works to get the target in the DSI IIc's field of view every time (so far).  

After experimenting a bit, I also switched to the Windows versions of Nebulosity 3 and PHD2, just because I find that to be simpler than switching back and forth between Mac OS and Windows.  In addition, the Windows versions support many more camera choices (I suppose that's a moot point if your cameras are supported under MacOSX, as DSI's are), they connect to the DSIs more quickly than the Mac versions, and they enable ASCOM pulse guiding, which seems a bit more accurate (but that's just an impression; I don't have hard data to support it other than I think my images look better since I started using it).

Here’s one of the better images I’ve taken with this setup (prior to using the PicGoto Simplificado, so this image was all Mac-native):

Date: 8 Jun 2013 

Subject: Markarian’s Chain

Scope:  Orion ST80+Antares 0.5x telereducer

Filter:  Baader Fringe Killer

Mount: CG-5 (Synta motors)

Guiding: 9x50 Finder/Guider + DSI Ic + PHD 1.14.2

Camera: DSI IIc (cooled at 2.25 A, 10.0 V, 12 °C)

Acquisition: Nebulosity 3.1.6

Exposure: 32 x 300 s

Stacking: Neb 3, bad pixel map, bias included, normalize first, trans+rot align, 1.5 SD stack.

Processing: StarTools 1.3 Crop; Wipe:Color & Brightness; Develop 84.95%; HDR:Optimize; Sharpen; Track; Color:Bottom 2.0, Top Full, Sat 600%; Denoise; Magic:Shrink 2 pix; Life:Heavy.  CS6+Astronomy Tools increase star color; deep space, space NR; some clone stamp to remove an edge artifact; crop again; less crunchy more fuzzy; AstroFrame.

Things that Might Have Led to Improvements

One of my fellow Astrobin users recently commented that he thought my images had gotten better recently, which led me to think about why that might be (assuming it is true!).  I think maybe the postprocessing software (primarily StarTools, with a little Photoshop CS6 at the end) has improved, which made me improve along with it. The StarTools color module is definitely better than it used to be. Artistically, these days I'm tending to opt more for smoothness in my images (i.e., less noise) than for sharpness, but I'm using layer masks to try to get the best of both worlds (so I retain the sharpness where there is some detail in the image, but decrease the noise where there is no detail to preserve).

Acquisition-wise, things have definitely gotten better since I built a PicGoto Simplificado mount controller. It has made finding targets trivially easy, I think it tracks with smaller steps than the original Synta controllers did, it uses ASCOM pulse guiding, and I no longer fear re-acquiring a target after a meridian flip so it has become easy to image longer (now if that darn tree were not in the way in the west!).

 I also added a solid-state drive to my MacBook, which has made using VMWare/Windows 7 very fast (Windows launches almost instantly), so now I'm not afraid to use Windows software and I'm no longer restricted to using Mac-native software. Mac-lover though I may be, I must admit the cheap/free astrophotography software available for Windows is generally better than the Mac-native stuff.  I don't have good, hard, empirical data, but FocusPal seems to control my FCUSB better (less backlash, finer control) than Shoestring Focus (which I also like), Nebulosity (Windows) definitely finds my DSI cameras faster than does Nebulosity for Mac, PHD2 for Windows seems a bit more refined than the Mac version, all of this software supports a lot more hardware under Windows than under Mac, ASCOM (Windows only) works extremely well for running almost anything and multiple devices simultaneously, SharpCap (Windows only) is very good for lunar/planetary webcam control, PicGoto Server is Windows-only, and the combination of PicGoto Server with Cartes du Ciel under Windows is wonderful for finding things.

Croc's Eye Galaxy, M94

3 May 2014.  I chose this target because I was not very happy with last year’s image (it was quite noisy).  Conditions were again pretty good, though there was a bit of breeze.  I located M94 using the PicGoto starting from Cor Caroli; I have become completely spoiled by the ability of the PicGoto to find things easily.  I had guiding issues, especially early, because I didn’t carefully balance and weight-bias the mount.  After doing a meridian flip, I paid attention to balance (I tried west-weighting again) and guiding was much better.  However, I had nearly 6 hours of subframes, from which I selected the best 200 minutes to stack.  Processing this stack was more difficult than usual; the core of M94 is quite bright, so the dynamic range between the core and the outer ring is very high and required some fiddling.  Still, in the end I like this image better than the overprocessed one I took last year.

Date: 2 May 2014

Subject: M94, Croc’s Eye Galaxy

Scope: AT8IN + HPS Coma Corrector

Filter: None

Mount: CG-5 (Synta motors, PicGoto Simplificado)

Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.2.2 (Win 7 ASCOM)

Camera: DSI IIc no chiller T = 15 °C)

Acquisition: Nebulosity 3.2.0, no dither

Exposure: 40x300 s

Stacking: Neb 3, bad pixel map, bias included, normalize first, trans+rot align, 1.5 SD stack.

Processing: StarTools 1.3.5.279 Crop; Develop 81.72%; Color:Scientific, 200% sat.; Wipe, 75%; Develop 84.41%; Color: Scientific, 250%; Deconvolute auto mask 3.0 pix; HDR:Reveal core; Sharpen; Life:Moderate; Track RNC 6.37%; Magic:Shrink 1. CS6 Astronomy Tools Gradient Xterminator with layer mask; increase star color; Healing brush; Deep space noise reduction; Space noise reduction; Less crunch more fuzzy; layer mask the denoising; Astro Frame.

Guiding Problem Solved? M106

2 May 2014.  I think maybe I’ve finally figured out why my guiding is always bad after a meridian flip!  I decided to image M106 because conditions looked very good and my attempt at M106 from last year is pretty noisy.  Once again, finding M106 with the PicGoto was just a matter of centering a bright star (in this case Phad) and pressing the "goto" button.  Conditions were excellent, but once again I had nasty guiding issues, especially after performing the meridian flip.  Things would go along smoothly for a while, then it seemed as if the RA tracking would just stop for a while, then start up again.  After thinking about this and fighting it late into the wee hours, it finally occurred to me that what is probably happening is that the mount guides fine until it bumps the legs after passing the meridian.  At that point I go out and do a meridian flip, but in the process of bumping I'm driving the mount against the tripod leg and that loosens up the clutches (thank goodness; otherwise I'd strip everything!).  The loose clutches then result in erratic behavior after the flip.  Once I tightened everything up last night, the guiding was excellent.  I was trying the unorthodox approach of making the scope west-heavy rather than the usual east-heavy, and it seemed to help.  I backed off from the 600 s subframes I have been taking lately and went with 300 s.  That seemed to work fine, and I probably had to throw out fewer subframes this way.  I need to take a new set of darks and a new set of bias frames, but I’ll do that when it is cloudy or there’s lots of moonlight.   

Date: 1 May 2014

Subject: M106

Scope: AT8IN + HPS Coma Corrector

Filter: None

Mount: CG-5 (Synta motors, PicGoto Simplificado)

Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.2.2 (Win 7 ASCOM)

Camera: DSI IIc no chiller T = 13 °C)

Acquisition: Nebulosity 3.2.0, no dither

Exposure: 49x300 s

Stacking: Neb 3, bad pixel map, bias included, normalize first, trans+rot align, 1.5 SD stack.

Processing: StarTools 1.3.5.279 Crop; Wipe 75%; Develop 87.1%; HDR:Reveal core; Color:Scientific, 300% sat., cap green to brown; Deconvolute auto mask 2.1 pix; Life:Moderate; Track RNC 49.51%; Magic:Shrink 1. CS6 Astronomy Tools increase star color; Healing brush; Deep space noise reduction; Space noise reduction; Less crunchy more fuzzy; layer mask the denoising; Astro Frame.  I could probably make the dust lanes stand out more with an unsharp mask, but I elected not to.
Here's what it looks like with a different set of processing choices, mostly designed to further sharpen the core. I think I like this version better. I've done similar processing as above except I used "cap green to yellow" instead of "brown", I've deconvoluted more (2.5 pixels), I've done one pass with wavelet sharpening, and I've used an unsharp mask with a layer mask in Photoshop to bring out the dust lanes.

NGC 4244, the Silver Needle Galaxy, and Getting Better Slowly But Surely

1 May 2014. One of the things I like most about astrophotography is that even though the learning curve is steep (and I like that too), it is easy to see improvement from year to year. I decided to try NGC 4244, the Silver Needle Galaxy, last night.  I’ve only imaged this once before, 2 years ago, and back then I was still using my Meade 4501 4.5” F/8 reflector.  I had sort of resisted going back to this target because it seemed to have little structure and a silver streak in the sky was just not that interesting.  On the other hand, I reasoned there was a good chance I could make significant improvements over the prior image given what I know now, so I decided to give it a shot.  Conditions were excellent: crystal clear skies and no moon with little or no wind.  I used Cor Caroli as my sync star and easily got to this target using the PicGoto.  Guiding started off very well.  The weather is warming up, so the camera sensor started at 13 °C and fell to about 11 °C with no active cooling (I guess I’ll need to start that up soon).  I stopped every hour to refocus using Nebulosity's fine focus feature.  I performed a meridian flip, but my guiding was much worse after the flip; this seems to be a recurring thing.  It was worse on both axes too.  Most of the subframes after the image flip were blurred.  I need to figure out why this is happening; likely it is a balance issue, because I know balance is critical with my mount.

Anyway, here’s the processed image from 2 years ago.  It’s not terrible given my level of skill and extremely inexpensive equipment at the time.

Date: 23 Apr 2012

Subject: NGC 4244, Silver Needle Galaxy

Scope:  Meade 4501, 0.5x Antares telereducer

Filter:  none

Mount: CG-5 (Synta motors)

Guiding: Orion ST80 + 0.5x Antares telereducer + DSI Ic + PHD

Camera: DSI IIc (cooled at 2 A, 3.0 V)

Acquisition: Nebulosity 3.0.3

Exposure: 15 x 300 s

Stacking: Neb 3, bias included, normalize first, align, 1.5 SD stack

Processing: Digital development & levels, unsharp mask in Neb 3, a bit of levels & Astronomy Tools star color enhance in CS5.

And here’s the new image, processed.  Note that the orientation is not the same; the new image is rotated approximately 180° relative to the old one.  The AT8IN is a much better scope than the Meade 4501 I used in the earlier image, my imaging and processing skills have improved, and the software I’m using now (StarTools) has given me big improvements.  In the new image it is possible to see some subtle internal dust lanes in this edge-on spiral galaxy.

Date: 30 Apr 2014

Subject: NGC 4244, Silver Needle Galaxy 

Scope: AT8IN + HPS Coma Corrector

Filter: None

Mount: CG-5 (Synta motors, PicGoto Simplificado)

Guiding: 9x50 Finder/Guider + DSI Ic + PHD 2.2.2 (Win 7 ASCOM)

Camera: DSI IIc no chiller T = 11 °C)

Acquisition: Nebulosity 3.2.0, no dither

Exposure: 21 x 600 s

Stacking: Neb 3, bad pixel map, bias included, normalize first, trans+rot align, 1.5 SD stack.

Processing: StarTools 1.3.5.279 Crop; Wipe 79%; Develop 79.57%; HDR:Optimize; Color:Scientific, 250% sat.; Develop: 80.11%; Deconvolute 1.7 pix; Life:Moderate; Track RNC 5.15%; Magic:Shrink 1. CS6 Astronomy Tools increase star color; Healing brush; Astro Frame.