Ronin-S already has some pretty amazing features for timelapse and panorama photography. Has anybody ever thought of using a gimbal like Ronin-S as a star tracker for astrophotography? Would it even be possible?

Thomas

Unlikely. The absolute precision is not high enough. Just imagine how high the precision needs to be to keep one star on one sensor pixel for an extended amount of time.
You will be better off with multiple pictures and stacking.
And for deep field photography you need a polar mount.

So depending on what you want, it might help but just a little. Image stacking is the better solution. for such simple cases.

That would be my take at least.

WernerD Posted at 2018-7-4 10:17
Unlikely. The absolute precision is not high enough. Just imagine how high the precision needs to be to keep one start on one sensor pixel for an extended amount of time.
You will be better off with multiple pictures and stacking.
And for deep field photography you need a polar mount.

Werner, thanks for the reply! You are probably right. Just wishful thinking from me I guess, after looking at solutions like iOptron SkyTracker Pro

I think it's very possible.  The question is whether the gimbal motors are capable of very-VERY slow and consistent movement.  
Active tracking isn't necessary at all if the gimbal's software program would be set to function moving the motors the same way a star tracker works.  A star tracker's motor moves in one direction at a very specific and continual speed.

sky_makai Posted at 2018-8-26 15:41
I think it's very possible.  The question is whether the gimbal motors are capable of very-VERY slow and consistent movement.  
Active tracking isn't necessary at all if the gimbal's software program would be set to function moving the motors the same way a star tracker works.  A star tracker's motor moves in one direction at a very specific and continual speed.

Thanks for the reply. Yes, it would be a pretty incredible and unique feature to add to an already awesome gimbal. I'm at least going to submit a feature request for this.

The Ronin-S is indeed able to make very subtle movements. Let's see what the DJI engineers think about this.

What is actually the best way to submit a feature request though? Is there a dedicated DJI page for feature requests?

Thomas Berglund Posted at 2018-9-30 13:41
Thanks for the reply. Yes, it would be a pretty incredible and unique feature to add to an already awesome gimbal. I'm at least going to submit a feature request for this.

The Ronin-S is indeed able to make very subtle movements. Let's see what the DJI engineers think about this.

This is an interesting topic, whether the gimbal motors are capable of as much accuracy as a step motor.  Step motors are extremely accurate for slow increments, and are what's used in polar mount trackers.  

I'm not sure if/where there's a feedback channel for DJI.

sky_makai Posted at 1-2 19:17
This is an interesting topic, whether the gimbal motors are capable of as much accuracy as a step motor.  Step motors are extremely accurate for slow increments, and are what's used in polar mount trackers.  

I'm not sure if/where there's a feedback channel for DJI.

I never did Astrophotography, but what about create a track motion and the A and B points would be the exactly points for the movements of the stars and you set like hours to run this entire track and use the intervalometer to take the pictures instead of the Ronin S?

X2 for this!!!!  If the motors are capable of very small continuous movement, this would be a killer application. It would still be necessary, of course, to find a way to align the unit with Polaris, so the tracking would be synchronized with the earths spin.

On a related note, has anyone used multilapse on the ronin S to do a series of stop motion long exposures of the Milky Way?  Waiting for a clear night to try this.   Thanks!

Easy to find out.
For star tracking you need three things: Precise absolute position, smooth movements and long term exposure.
The most simple thing to test is the long term exposure. So position the Ronin to a static object and do a 30sec exposure with a 70mm lens. Is the picture sharp? I say it is not. There is always a slight vibration

WernerD Posted at 2018-7-4 10:17
Unlikely. The absolute precision is not high enough. Just imagine how high the precision needs to be to keep one star on one sensor pixel for an extended amount of time.
You will be better off with multiple pictures and stacking.
And for deep field photography you need a polar mount.

Do you think its possible to use the ronin app to get it to track slowly? I havent personally tested it, but with the timelapse mode, I wouldnt be surprised if you could get it to work

I think we need to clarify the use case.

For a wide angle lens like a 16mm or even a 35mm you don't need a telescoping mount. You take multiple images over a long period of time and stack them together on the computer. All the alignment is done during that post production.

The opposite of that is deep space photography. The lens is a 2000mm and with a 30 second exposure time the image is all black still and because of the large zoom even the slightest move would mean the stars are not sharp. So here you need to keep the star precisely at the same pixel for the whole time while the earth is turning. That is the sweet spot of telescoping mounts. For these extremes the Ronin will have too many vibrations in the motors. Not to mention you cannot mount such large lenses.

Somewhere between those two extremes is our use case. I have the feeling that even a 200mm lens and 30s exposure time will have so much shifting and tilting in the motors that even a static object will not be sharp. But that is to be tested, as said. If that has been proven, the next test is to follow the star. For a wide angle lens the precision can be good enough. For a longer lens probably not.

Don't get me wrong, I would love to do that. And in fact I am working on it. I am currently trying to debug the Bluetooth communication protocol between the Ronin app and the Ronin-S in order to send commands of my own. DJI is not very supportive here, to say the least, and the protocol would need to support certain functions I doubt it does, e.g. absolute positioning. The app does support relative movements only: Pan with a speed of x% to the left. We would need the speed value to be with units, e.g. 10°/sec. And best would be to say "pan to position +10° and tilt to +20° relative to zero".

WernerD Posted at 1-6 23:40
I think we need to clarify the use case.

For a wide angle lens like a 16mm or even a 35mm you don't need a telescoping mount. You take multiple images over a long period of time and stack them together on the computer. All the alignment is done during that post production.

Hello all,

this looks like a really cool idea if we can get it to work.  I did a stability test with a Panasonic GX8 and a 12-35 lens (24-70mm in 35mm full-frame equivalent) with 30 seconds exposures.  The camera has to be really well balanced to get it as stable as possible.  The results are somewhat variable, tough I cannot really put my finger on the exact reasons.  In about 1/3 of the exposures I could get a perfect result without any wobble.  

It would be interesting if some DJI engineer could provide input on the technical questions such as precision and smoothness of the motors, plus the API and its precision.  My wild guess is that some firmware tweaks might be necessary to achieve the level of precision required for star tracking.

Best regards,

Karol

Hmm, interesting! Would love to have this as an option on the Ronin-S!

Basically, there are 2 things to be solved to compensate Earths rotation:
-in case you create an equatorial kind of mount, You need to be able to set the pan axis pointing to Polaris (ok, those ~0.5° distance to real N wouldn't be a big issue by using not so big focal lengths and long exposures). The benefit is, just one axis motion needed for tracking. Drawback - the easiest way to get to this is just to buy an astro mount (by far most of amateur-enthusiast mounts are equatorial). And, futher, if You dive in astrophotography, the mount is the most important investment. But, for just a SLR caliber camera, there are quite affordable options .
-in case You use Ronin as Alt-Az mount (which, basically, it is), the challenge is to be able to precise set the horizontal component of rotation speed (in Ronin App UI, there's no option to manually put in speed in degrees/hour or smth like that. Second task would be to set the vertical (which is even harder – depends both on direction and heigth above the horison etc). Tehnically, its doable (there are beginner/kids-targeted alt-az Go-To mounts in ~250 price range
-and the hardest part would be to compensate the FOV rotation, if You want to use longer exposures.

So, in short: Ronin-S is preciselly targeted tool, which is meant and engineered for stabilisation. Yes, there is some minor potential for astrophotography, but take it in creative plane – You basically can use exposition long enough for Milky Way and similar wide- or mid-angle astrophotos w/o need to track anything, and Ronin Timelepse mode can add some movement to Your Nightscape timelapse, which is great (the easy way is to divide 500/f to get the max spot star exposure for particular focal length).
But, if You are looking for deep-space/big focal length astrophotography, there's no other way, than to go for real astro-mounts, OTA's, autoguiders etc.
And, if You go further, specialized cameras u.o. stuff.

WernerD Posted at 2019-1-6 23:40
I think we need to clarify the use case.

For a wide angle lens like a 16mm or even a 35mm you don't need a telescoping mount. You take multiple images over a long period of time and stack them together on the computer. All the alignment is done during that post production.

Let me know how you go with the debugging. What language does it use? Java?

raphael_bellizzi Posted at 2019-1-2 20:04
I never did Astrophotography, but what about create a track motion and the A and B points would be the exactly points for the movements of the stars and you set like hours to run this entire track and use the intervalometer to take the pictures instead of the Ronin S?

I think this might be worth a try for short (60-120 second exposures).

This is an old post, so apologies for the up.

I have been doing astrophotography for several years and the required precision of your star tracking depends of your zoom (focal length + sensor pixel size). Without any tracker, you have the "rule of 500" that tells when you start having star trails. Example: 500 / 100mm = 5s exposure without star trails (for high pixel sensors, it's more like 300 / focal length). A number of star trackers are not very precise (eg Omegon LX3) but very portable; they don't promise a perfect tracking but by improving the tracking even just a little, it's often possible to achieve 30s or even 60s at around 200mm.

In other terms: yes, any gimbal could be used as a star tracker, even if they don't use step motors. Just don't expect to do 2mn+ exposure at 200mm+ as this requires a completely different type of equipment (eg equatorial mount of 10kg+ with DC12V at 5A). My guess is you could achieve a precision with a gimbal about the same as the LX3 which promises a rule of 6000 / focal length of exposure without star trail ( eg 6000 / 100mm = 60s exposure).

Without entering into too many details, you then have two ways of tracking: alt/azimuth (up/down + left/right) or equatorial (pointing towards polaris with a RA and DEC axis). Alt/Az, I guess you could just use the active tracking on a star and see how it goes, the cons is you would observe field rotation after 15-20s. Equatorial would require a software update / another app for your gimbal, as essentially you would have to orientate the camera towards polaris and have a permanent roll rotation around that axis, while the two other axis could move to point towards the object you want.

Anyhow, that would definitely be fun to try, but my guess anyone with a DJI gimbal should be able to increase their exposure time by just tracking a star (eg no tracker 100mm = 5s, with gimbal + active tracking, I guess you could do at least 10 or maybe even 20s). If you are in a dark sky area, that's sufficient to capture plenty of nebula and galaxies. If you are in cities (bortle 6/7+), this won't work as you would need filters and with them longer exposures (2mn+).

I agree, FOV rotation is the bottleneck.
Even-though it tracks, there would need a camera rotator that kills the party unless you are shooting from the equator.
OR …
An equatorial base would be needed that tilt the gimbal to pole direction that also makes the whole exercise super complicated with polar scope, calibration of the polarscope and all the rest of it.

I was thinking the same and came to this thread. I think other than tracking, you have to lock the camera first to a pole star or a referene point and then the software has to calculate the rotation/movement for solar/lunar and other objects.