Barn door tracker

[Ags]

I fancy using a stepper motor for fine control... Given that I am using a straight drive bolt I need to speed up the bolt by the tan() of the angle. But I currently can't see how to drive the bolt, maybe by a belt...

[Mr TamiyaCowboy]

I fancy using a stepper motor for fine control... Given that I am using a straight drive bolt I need to speed up the bolt by the tan() of the angle. But I currently can't see how to drive the bolt, maybe by a belt...

hehehe thats fairly easy. 

direct drive, what happens is the motor is screwed to the bottom plate of the barn door. 

the threaded bolt is attached direct to the motors shaft. a NUT is sunk and glued into the top plate.

as the fixed down motor turns the bolt, the fixed in nut screws up and down the threaded rod.

a simple arduino and stepper motor driver board and your up and running. 

more awkward but smoother is using a geared drive system, belt drives are ok 

but can be prone to slippage, something like radio control cars have a pully drive system and diffenet sized pully wheels.

they usually run on a layshaft layout ( drive gear and pully wheel on a shaft driven by motor and pinion gear).

more harder to get sized up but a better smoother working system.

[Ags]

I considered this design, but while the bolt and nut will start out square, after an hour there is a 15 degree difference which I imagine would be a problem. Am I missing something?

[Mr TamiyaCowboy]

maybe have the motor fixed to the bottom plate with a hinge.

as the bar rotates and the camera angle changes the the motor is allowed to tilt

keeping all free and smooth ?

other way is buying a curved rod and using some kind of torque slip wheel system.

a rubberised wheel rests on the curved bar, when running it grips bar and move sit up/down.

when stopped it holds the bar via slip grip from the rubber surface. this method would

be somewhat governed by weigh on the top plate, to much weight it wouldnt grip just slip.

[Ags]

Gary Seronik's design may be mechanically the simplest:

http://www.garyseronik.com/?q=node/52

EDIT: Forget that. Here is the solution:

http://www.cloudbait.com/projects/bd_drive.jpg

[Mr TamiyaCowboy]

was then going to come back and say

TWO HINGE POINTS mount lol , but you beat me to it . 

yes that would work perfect , the dual pivot point lets the drive rod 

tilt as its rotating and the angle increases/decreases. 

[Ags]

It's also not too bad from a maths point of view... an isoceles triangle is just two right angle triangles back to back.

[Ags]

Anyone know what is the name of the widget that connects the threaded drive rod to the motor shaft? In the cloudbait design it looks like a tube with two lock nuts...

[JamesF]

Anyone know what is the name of the widget that connects the threaded drive rod to the motor shaft? In the cloudbait design it looks like a tube with two lock nuts...

If you mean the tube with two grub screws then it's probably just called a coupling or coupler.  It'll just be a hollow (aluminium, probably) tube (the holes aren't always the same size at both ends) with a couple of grub screws threaded into it.  I'd imagine that there's a flat on both the motor output shaft and the threaded rod that the grub screws press on so they can't spin relative to each other.

James

[JamesF]

This sort of thing, for example: http://www.technobotsonline.com/inline-coupling-5mm-to-5mm.html

James

[Ags]

If you mean the tube with two grub screws then it's probably just called a coupling or coupler.  It'll just be a hollow (aluminium, probably) tube (the holes aren't always the same size at both ends) with a couple of grub screws threaded into it.  I'd imagine that there's a flat on both the motor output shaft and the threaded rod that the grub screws press on so they can't spin relative to each other.

James

Thanks. also for the link in your next post. I think I will get a plain threaded rod, but making the flat section on one end for the grub screw to press on will be quite simple for me I think.

EDIT: I Love the site you linked to!

[JamesF]

Yes, shouldn't be a problem to file or grind a flat onto the end of a piece of threaded rod if required.

James

[happy-kat]

http://www.liv.ac.uk/~sdb/Astro/K2/

This just got shared by a different member another version does long time

[Ags]

Naw, I'm looking for integration times of an hour or more. The K2 design manages 15 minutes before tracking errors. That is within the range of what I can get with my my current manual design (in fact I could go longer, but manually nudging a bolt for longer than that would be punitive :-) ), although I am limited to 50mm or less because of MPE (manual periodic error). I correct tangent error manually by using an app that gives me an accelerating clock. Assuming I've got my maths right 

I'm really keen on raspberry pi-driven system, as I'm a programmer by trade. A digital stepper is always going to beat an analog potentiometer design I think. All I need to do on the software side is to adjust my accelerating clock app to stepper pulses. Where I fall down is the practical side - soldering, woodwork and so on.

[Mr TamiyaCowboy]

i have a rotor drone brain here lol

its basic a 328 uno arduino , but populated with 3axis gyro - 3 axis accel - 3 axis mag compass and a barometer.

it can also run two servos as pan tilt ( reading the boards sms) ftdi and Uart and I2c ( if i solder the lines in)

i also got a arduino uno untouched to just waiting on something to come along worth while building.

couplers, they can be a pain biggest i know of in the hobby areas is a 5-8mm to 5mm,

if we want bigger say something for a 26cc engine we tend to mill our own from alloy.

flattening motor/bar shaft is easy use a grinder/dremel just use tape to cover any vents on the barn motor

stop any filings getting into the carbon brushes and comunictaor.

rubber couplings may help but the have that inhernet flex/give unlike a solid metal coupler.

if you want to kind of cheat a bit you can get a U-joint, we use them in drive trains. just google

" traxxas U-joint, they are also used to make cheap mans steadycams to. could be an idea to couple the rod to

the motor without the hassle of finding the barrel nut.

would U-joints work both ends plate end and base end to help with that dangle angle 

[Ags]

Here is a screen capture of my barn door clock app. Not all the buttons and settings are functional yet, it is all hardcoded for my barn door. I also need to make all the widgets night vision friendly.

[happy-kat]

oooh it is Android.

I am enjoying this thread as it gives inspriation that a big heavy unweildy EQ mount is not needed for a simpler approach to imaging of less demanding objects.

[Ags]

I'm writing the app using AIDE - with AIDE you can write a Java app directly on an Android device. I works surprisingly well...

[Ags]

The clock is mostly finished... Here's how it looks now (the text fields are now all used in the code, so it can be applied to any straight-bolt barn door tracker configuration).

[Ags]

Well, I've been doing more work on the clock (curse these cloudy skies) and the challenging thing is avoiding incremental errors that can cause quite a drift over an hour or so. My first 'complete' version drifted by 0.25 degrees over an hour which was really appalling, but now I've tightened the maths up a bit and it is down to 2% of that - or 1 pixel trailing in an hour long sub at 50mm with my 1100D. That is probably still a bit too high (for a so-called clock ), but I can't see how to drive the error down further just yet.

[Mr TamiyaCowboy]

maybe you have pushed it down with maths as far as you can.

i think now its the plates and fixings, check if these are throwing in errors.

a slight slip of the hinge when is working ( hinges tend to be sloppy and free they wiggle)

you could be seeing some type off error here and thats impacting your maths errors.

i am more a builder than a tech maths dude, i would have pulled it apart by now and

started rebuilding with different parts while cussing swear words and looking for a hammer and screwdriver.

that saying maybe some speed tweeks on the motor, if your running arduino maybe some kind of 

hall effect sensor could be thrown in to give more fine tuning/data

[happy-kat]

An hour is a long time, wont that mean the drift say over a 15 minute window would be invisable, or it is that you could take 4 subs of 15 minutes if there was no drift over that hour. Wont you need a super long bolt to achieve and hours worth of tracking?

[Ags]

You would need roughly 6-7cm (+ wood thickness) for an M6 bolt and 22.85cm radius to get an hour's tracking. At a realistic 3 minute sub, the error is unmeasurably small, even if you go to 200mm lens. But I am a software developer and I'm not putting a clock on Google Play that can't keep time accurately

Cowboy, I'm testing my app against real time, not against the stars. So this is all at the software level. I've tested the hardware at up to five minute subs and got good tracking, just a bit of periodic error caused by my poor bolt nudging technique. So the manual version of my tracker is done. I will move to the motorized version later in the year - that is a big leap for me so I need to do more learning first. Also I am saving up for a 200mm dob, so I don't want any more expenditure until I've got that !

[JamesF]

I work that out as an error of about 20 arcseconds per hour.  That seems a lot given the bare numbers, but then people talk about getting five minutes unguided subs with a HEQ5 Pro and DSLR at 600mm focal length as being pretty good and that's pretty much in the same ball-park I reckon, perhaps not even as good.

I don't think anyone should be too distressed at making a barn door mount that tracks as accurately as a HEQ5 Pro  I'd say that's quite impressive.

James

[Ags]

Er... my five minute test subs were are 50mm! So I am far away from HEQ5 territory :-) If I can nail my technique I could probably get the same results at 100mm, but then the limitations of manual bolt turning become insurmountable. To push to longer focal lengths and a higher sub count, a stepper motor driven at a variable rate is essential.