UKC

has anyone built/used one of these for astrophotography?
http://en.wikipedia.org/wiki/Barn_door_tracker

thinking of trying this simple design: http://barn-door-tracker.co.uk/

any advice/example photos welcome...

In reply to malk:

You can track for longer if you account for the fact that the angle does not change linearly as you turn the handle & mark out scales accordingly.

A friend of mine was working on a software controlled one. Don't know if it ever got finished, but that was a consideration.

In reply to davidbeynon:

I built one earlier in the summer. Probably seriously over-engineered. Arduino controller, stepper motor, lots of aluminium sections, PTFE bushings. Initial tests indicated may minutes exposure with 50mm lens and no tracks (without tracker you can get about 5 seconds). But, and this is still an open problem for me, the tracker is only as good as the polar alignment and for a non-permanent installation it's a bit of a challenge. I've currently got a home made alignment scope with a calibrated 3/4 degree offset (NCP to Polaris) which coupled to knowledge of the siderial time might be good enough. If you are patient, there is something called the drift method, but it takes a while. Anyway, I'll put some pics up somewhere of what happens when you let an ex-engineer loose with too much time on their hands...

In reply to malk:

I know it's cheating, but I've had great results with an Ioptron Skytracker. It really is quite portable and I've had much better results with my camera and decent lenses than I have with my camera and Nexstar 6SE telescope.

I looked into a barn door tracker, I think it'd work with short focal length lenses but for long lenses it looks quite prone to vibrations when adjusting it.

In reply to Marek:

OK, I've put the pictures here...
http://www.zenadsl6044.zen.co.uk/bdt/

bdt-01.jpg - Overall view, main bearing and alignment scope on the left, stepper motor assembly on the right (inside the aluminium section) and the Arduino-based controller (+batteries) at the bottom.

bdt-02.jpg - Close up of the lower main bearing. There's a hardened steel shaft that runs through both top and bottom bearings, clamped to the moving arm. Lateral support is by PTFE bushings tightened to remove any play. Longitudinal support is the single ball bearing at the end of the shaft. There's pretty much zero play and zero stiction in this bearing. You can also see the wires going to the illuminated reticule in the alignment scope - surprisingly hard to make!

bdt-03 - End view of tracker showing the stepper motor and 0.8mm pitch drive shaft. The shaft is long enough for just over a half-hour exposure. It also shows the mount altitude.

bdt-04 - Close up of the motor mount gimbal (PTFE bushings all round again).

The Arduino controller does all the maths in real time to convert the geometry of the tracker into a rotation. In theory accurate to a couple of pixels with a 300mm lens over an hour. In practice, error in my polar alignment technique are far worse until I can figure out how to do it quickly and accurately.

In reply to Marek:

Nice work! It's a bit more sophisticated than the traditional two bits of wood, coach bolt and a hinge.

In reply to richard_hopkins:
That looks both very cheaty and very good. Can you mount that on a standard tripod?

Also, how big and heavy is it? Is it the kind of thing that could be taken backpacking?

In reply to Marek:
That's really good.

I had an idea for a 3 axis tracker that uses a secondary camera and star tracking software to avoid the need for polar alignment but it's probably a bit too complex to be worth building.

In reply to davidbeynon:

I thought about the closed loop tracking approach but thought it unsuitable for a portable system since the tracking camera/lens would need to be almost as good as the main camera. Another open loop single camera approach I might try is to take two exposure a say 10 minutes apart, analyse the drift between them and then use the drift parameters to adjust the tracker alignment for the proper long exposure. I've hacked a prototype Android app that does the drift analysis and which can measure sub-pixel linear drift (it measures drift of up to 20 stars in the field of view), but that's as far as I've got.

In reply to richard_hopkins:


Presumably you still have the polar alignment challenge with this?

In reply to Marek:

Yeah. I put together some similar software for auto alignment a couple of years ago. I'm not sure that the camera sensor needs to be as good as the main cam for the closed loop system as you can get away with a much smaller field of view.

In reply to malk:

I thought this would be a thread for a device which records how I fall off most regularly.

In reply to Marek:

The sky tracker just fits straight onto the tripod. On my tripod you unscrew the ball head which exposes a 3/8" thread which fits straight onto the sky tracker base. You then fit the previously removed ball head into the tracker and put the camera on that.

For polar alignment, there is a little spotting scope with illuminated graticule. There is a simple app you can get to show where to align the pole star on the scope depending on your location and time.

It can hold about 2.5 kg, which is enough for my 300mm f4 and camera and tracks well.

I think the best open loop tracker is regarded as the astrotrac which I believe also supports closed loop tracking with a 2nd camera but is still limited by your declination alignment.

These devices really win for portability, you can be up and running within a couple of minutes.

In reply to Marek:

I got the tracker out last night in the garden for a quick run with my new technique for polar alignment.

300mm lens, ISO400, 4 min exposure shows no obvious tracks...
http://www.zenadsl6044.zen.co.uk/bdt/IMG_8350-noNR-pixel.jpg
CR2-->TIFF via RawTherapee with exposure compensation only.

Not bad for an early moonlit night in the middle of Macclesfield - lots of LED street lighting now
Realised afterwards that I had misaligned by 15 arc mins (I forgot that my reticule mark was 1 degree rather than 44 arc mins), so I think I could be getting about 2 pixels drift in the picture. Without tracking the star streaks would have stretched about a quarter of the width of the whole picture.

Looks promising given some darker skies and less forgetful operator.
No prizes for guessing the subject matter.

In reply to Marek:
Very nice!

What sort of scope did you have in front of your camera?

In reply to davidbeynon:


Scope? None. Just my cheapo Canon EF 75-300 lens set at 300mm f8.
A quick bit of browsing shows that the star on the left of the pixel-peeping window is mag 13.95 and the faint one just to the left of the one in the bottom-right is 15.3. With a bit of imagination I can also identify a couple of stars at about 17 just about above the noise floor. I might have to start trying out the deep sky stacking software soon.

In reply to Marek:
very impressive- your overengineering was worth it did you start off with a simpler model?
think i might start with adapting a flower press for portable wide angle tracking similar to the OP link. i could prob get sufficient polar alignment for wide milky way shots with a spotter scope? vibration is a worry though..

In reply to malk:


I knocked up a zero cost manual one from some bits lying around in the garage to start with. It convinced me that more robust and computer/motor driven one was worth it. Or at least fun to build.


It all depend on the field of view you want to use. For my 300mm lens (3 arc seconds per pixel) I worried about stiction and play in the main bearing and flex in the arms. It doesn't take a lot of force to bend a 1 foot length of anything by a few arc seconds. For wide field use - say 50mm - you can afford a lot more tracking errors. Vibration may be a problem. I use microstepping in the motor to minimise it, but I can still hear it so it's there. Some rainy day I might try any measure it. Environmental vibration (e.g., wind) is probably a bigger problem. So far I've only used it on windless nights. I actually found that the greatest source of flex from ground to camera was the ball-head camera mount, so that may be something to think about. In terms of design it's quite tricky to ensure you can point it anywhere in the sky without the camera fouling something and keep everything compact and stiff.

In reply to Marek:

Do you think there is any mileage in setting up a windbreak around the tripod & mount? Obviously not so hot if you are aiming close to the horizon, but...

In reply to Marek:


Okay, I have absolutely no idea what I'm looking at there, but I like it!

In reply to davidbeynon:


I suspect you'll just generate more vortices. My approach at the moment is to keep the whole thing as low to the ground as possible (about a foot or two). Means rolling around on the ground to set up the alignment, but it's simple and keeps the tripod as stable as possible.
I'll add an appropriate experiment to my 'rainy day list'.

In reply to Fraser:


... The result of an injury which kept me off climbing/running/cycling for months plus rose-tinted recollections of being an physicist and engineer instead of a 'strategist'.

In reply to malk:

https://www.bhphotovideo.com/bnh/controller/home?O=&sku=843972&gcli...

This is a much simpler alternative, less faff and reasonably cheap. Super results so far.

In reply to Marc Langley:


That's a bit like saying "But you can just walk round the back!" (to exploit some climbing context).

If you want less faff and cheaper you can look up SDSS-III on the internet. Even better results!
BDTs are frankly more about DIY engineering than astrophotography. "Faff" is what it's all about.

In reply to Marek:

I'm going to sell my camera and use google image search for all my photography from now on.

In reply to malk:

Barn door trackers are cool for shorter focal lengths, especially if you can negate any vibrations induced, but other than that, I'd really be looking for something automated. I use an astrotrac, the old one, when I travel, and limit my FL to around 200mm. At 200mm I can manage around 5 minute exposure with no trailing on an APS-C sensor.

In reply to andi turner:


I'm expecting better than 10 minutes at 300mm (<2 px drift, my longest lens). I'd have to resort to the labourious drift method for polar alignment to get better than that and it's probably not worth it unless I can access some properly dark skies. Or unless I start messing about with narrow-band filters but that's another cost game.

Vibrations - I just tried taking a video of a back-lit pinhole while the tracker was running. As far as I could see there was not more than 1-2 pixels (300mm lens) vibration correlated to the motor (micro)stepping. On the 'return-to-zero' phase when I drive the motor with as fast as possible with full steps it's a lot worse (~4-5px).

In reply to Marek:

That's a great result on the Andromeda galaxy. Really good tracking, makes your pictures especially rewarding since you made the tracker

Definitely look into Deep Sky Stacker or something similar. One of the advantages of stacking is that the unavoidable small errors no longer add up and if there is a duff frame due to some wind, a passing satellite or other effect then less data is lost as you can use short frames.

I think with my skytracker on a tripod, the largest source of bad frames is wind shake and the ground moving.

I invested in a cheapo remote shutter / intervalometer control (10 quid on ebay, exactly equivalent to the £100 Nikon item) and I set up long sessions on the remote then walk away and leave it to it. You can program in a start delay, the frame duration, gap between frames and total number of frames. The part number is MC-36B and there are flavours for Nikon and Canon.

Even walking around near the tripod causes the ground to move a bit and spoil the exposures, so I set it up, leave it tracking for a while to settle and for all the backlash to go then set it going. Hanging a heavy weight from the tripod might help as well if it doesn't swing about and sway the image.

In reply to Marek:

What camera are you using? Ten minutes is a very long time for anything non-cooled....It's also a very long time for anything un-tracked in my opinion (atleast in the UK) as seeing can begin to play an important part in the clarity of your image.

Needless to say, that's an awesome piece of DIY!

As for the filters, I'd definitely recommend some form of LP suppresion. Even where I am up in the hills the skyglow begins to become an issue after about 5 minutes.

In reply to richard_hopkins:


I've had a quick play with DSS a while ago, but got the impression that it's something to use after you've got the exposure time as long as you can in the 'traditional' manner - I had to stack 10 x 1 minute exposures (ISO800) to get something approaching the quality of a single 2 minute exposure (ISO400). Obviously the 'dark', 'flat' and 'bias' compensation is a significant bonus, but again, you have to get a lot of frames not to adversely affect the noise floor. I think I tried 4 darks - it got rid of the hot pixels, but overall the noise was worse. Perhaps I didn't do it right.

In reply to andi turner:


Canon 550d (it's what I've got). I've also performed the full-spectrum mod on an old 300d, so may give that a try, but it's a really old technology sensor.

>Ten minutes is a very long time for anything non-cooled....

Perhaps.

>It's also a very long time for anything un-tracked in my opinion (atleast in the UK) as seeing can begin to play an important part in the clarity of your image.

Not sure I understand. Yes, the seeing is poor (especially in the middle of Macc), but I don't see (ha ha) how it limits long exposures.


Proper ones seem to get rather expensive and I though their usefulness was limited given the shift from sodium to LED lighting? I have the 'poor man's' version - aka Redhancer - which is waiting for a trial. Better than nothing?

Any hints on good(ish) spots in the Peak for dark skies? My local default is Goldsitch Moss at the back of the Roaches. After that it's a trip out to Llyn Brenig in north Wales.

In reply to Marek:

With DSS you need to use the histogram sliders once all the images are stacked.
10x1 minute at iso 800 should be similar to 1x5 minute at iso400, except with the stacked images you can now tolerate 5 times worse tracking error.
When you have lots of stacked images you can really focus in on the small range of intensities where the image is.
With DSS I increase the colour saturation up to 30 or so and then adjust the histogram to focus on the narrow hump where the image data is.
Here is a result of 58 stacks of 30 second exposures
https://www.flickr.com/photos/77393937@N02/15783960508/

That was taken from near Guildford, so not dark skies at all and I didn't have any filters. Nikon D5300 so a modest APS-C camera and 100 pound 200mm F4 ebay lens.

I don' think that image used either a flat or a bias frame but there were probably about 8 darks.

In reply to Marek:

Personally I wouldn't waste my money getting narrowband filters for a colour dslr. Fair enough, it'll darken the background, but you'll be looking at greater than 4x the exposure length to capture the same signal, by which time your sensor will be super hot and your tracking will be stretched. It's worth seeing how your camera copes at ten minutes, I imagine it'll be really hot, especially by the third or fourth exposure, even if you're leaving a good interval. If I'm doing anything with a DSLR, then I have an IDAS filter, makes a huge difference. I use it in all but the absolute darkest sites.

Needless to say, I've seen some excellent images done when te photographer has gone for quantity rather than length of exposure (Seems to be the case with the CMOS sensors folk have started using now too), so going for 50 x 2 minutes is giving a better result than 10x 10 minutes.

As for the seeing, well, you could put your camera on a Mesu, and drift align to the nth degree of perfection, but without tracking the stars will still wander around a bit due to atmospheric conditions. Guiding smooths this out hugely (Mine will make an adjustment every 2 seconds), AO further helps, so there's a limit to what I would attempt on an unguided mount, but then, as you've alluded to earlier, that's part of the fun.

As for dark sites, yes the Roaches is quite good, or anywhere around there, but beyond that, yes, Snowdonia or the Lakes and beyond that, the North York Moors and Northumbria are better still. And despite travelling to quite a few places around the world to dark sites, I still think one of the best skies I've seen was in Scotland one time!

In reply to andi turner:


I obviously need to give DSS a proper trial.


Hmm, sounds like the low frequency component of 'seeing' is lower than I expected.


But not in June which was the last time I was up north!
Thanks.

In reply to richard_hopkins:



Obviously worth my persevering with DSS.

In reply to Marek:


Surprise view car park is one, 'officially' measured, back of Millstone.
Parsley Hay and Minninglow are 2 others, not sure they' be better than the Roches though.

They in the PDNP car parks.

http://www.peakdistrict.gov.uk/looking-after/projects-and-partnerships/dark...

In reply to malk:



They'd be OK for night sky shots, much less so for astrophotography.

You need a good mount for doing tracking, so how good is your tripod?

An alternative cheap and cheerful way in would be to get something secondhand, you can often resell for almost what you paid, so it doesn't cost much to try it for a year.

There's one of these for sale for £60 ATM (about average price and not mine!) and it's the steel tripod version, so pretty solid. OK for short exposures and stacking. Easy bodge to put a camera on it. Probably cheaper than an equivalent tripod. Plus you could add an OTA later if you wanted.

http://www.nightskies.net/scopetest/mounts/supatrak/supatrak.html

IIRC you can add goto and autoguiding too if you use a big enough hammer.

There was one of those ioptron skytracker mentioned above, on a steel tripod with ball for £200 secondhand a few weeks ago.

Acuter Merlin is another similar thing, steel tripod again £190 new, rare secondhand.

In reply to malk:

This is mine stacking five minute exposures, but this is guided.

https://flic.kr/p/hR9mhb

In reply to andi turner:



How many exposures/darks/etc?

In reply to andi turner:

Andy wins the prize for the best Andromeda picture
To be fair though, a $5k Tak isn't in the same league as a £100 quid eBay lens!

I've followed your super images for a while now, Did you submit photos to the astro photo of the year this time?

In reply to richard_hopkins:



Same again for the camera!

Great stuff though.

In reply to richard_hopkins:

Ha, well, yes, fair enough.... Although the Tak has now gone, destroyed by a single night's dewing, I won't have another.

No, I wasn't shortlisted this year, I'd not really done much and the standard was exceptionally high (although I wasn't that blown away by the winner to be honest), maybe next year....

I think my point is that even if you develop a device which can track at exactly the same speed as sidereal rate, it's still going to be limited by the atmospheric turbulence, the sensor of a dslr, the curved field of a standard lens, the wobble of a tripod, the passing of clouds or planes or satelites, light pollution etc

For me, a barn door tracker is a device for using with a wide lens to capture widefield photographs. I can still remember being totally bemused by this bloke who bought an RC off me last year. He had a mount which is literally worth 10k, yet he was using it unguided. I looked at his pictures and they were really poor, couldn't get my head around it!

I think the device you've made is absolutely amazing, like truly amazing! My advice would be to go to a dark site and do loads of short exposures (2 minutes?), making sure your histogram is hitting the 20% mark. Don't skimp on callibration frames, and plug your results into DSS. Looking at what you've got so far, you're going to produce some astonishing images.

In reply to andi turner:

It's Marek that's done the hard work and deserves the praise, I'm the cheat with the Skytracker!

That's really unfortunate about the Tak. what have you replaced it with?

In reply to richard_hopkins:

A Borg, then a couple of RC's, then a newt, but seem prettt happy with the current set up of an Altair 115 edt and a little Borg 67 fluorite jobby. Back to using a Kaf8300 sensor too, bit low on the QE but makes up for it with well depth and real estate

I'd love to be able to make stuff like that tracker!

In reply to Marek:

Wonderful

In reply to andi turner:

Oops, I meant your andromeda galaxy photo was wonderful, andi turner. The 'wonderful' wasn't just a general observation on life addressed to Malek !

In reply to Marek:

Update: Dark sky above Boot in Cumbria plus a bit of stacking (4x5min) got me this:
http://www.zenadsl6044.zen.co.uk/bdt/Eskdale-M31-4L5D8B-cr.jpg
I should have dragged the tracker a bit further from the bright lights of Boot (yes, there are) and I still need the get the polar alignment more accurate (I think I was off with my sidereal time) and focussing isn't spot on, but it's showing promise.

In reply to Marek:

Very nice.