Thread Tools
This thread is privately moderated by Jack Crossfire, who may elect to delete unwanted replies.
Sep 27, 2012, 07:14 PM
Registered User
Jack Crossfire's Avatar
Thread OP
Discussion

Servos, vision, & telescopes


The latest discovery was converting a servo to 360 usually fries the electronics. Because you want linear motion to be faster than angular motion, you give the motor 12V, which goes into the power rails through the spudger diodes & fries the electronics. Even 5V could fry the electronics from back EMF. The only safe way is a fixed resistor for the pot & PWM control, very slow. There's also cutting the motor traces.

There have been many times when custom servo firmware would be a win. A super servo would run on 12V, have a way to control the motor speed through PWM, time out if the pot doesn't move, increase the update rate to thousands of Hz, increase the precision by averaging more pot values, power down the motor, step the motor using a tachometer or timer. There are probably many such things, using bulky Arduino boards.

Such a servo would have allowed great things, during last year's laser painting, especially averaging pot readings to increase precision. Completely overhauling the microcontroller must not have been in the budget.

The alt/az mount for last year's laser painting was built in winter 2011, for a dead 2.4Ghz camera on a failed ground rover. In summer 2011, it was recycled into a stationary alt/az mount for a webcam that would unsuccessfuly fly Marcy 1 outside. That was converted again to a laser pointer in Oct 2011. Then, the electronics were simplified & it was restored to the web cam which successfully flew Marcy 1 in Jan 2012. It was sold in July 2012.

The current vision algorithm, using a flashing LED, could not fly a monocopter. The rotating airframe would create false flashing lights & a POV display would be a false flashing light.

A simple software change could get it to fly a monocopter. An alternating red/blue LED could be differentiated from a POV display or background light. It would work on both a monocopter & fixed copter.

The current algorithm didn't use a colored LED, because the highest possible framerate with least electronics design & highest chance of success was the goal. 1 bit video has the highest framerate with the current hardware. A bicolor target would slow things down or require an FPGA.

The current guy paying the rent doesn't leave any time for experimenting or growth. Just proofs of concept with the most proven techniques.


There has been a long dream of a self guiding telescope mount that could take exposures of several hours. Longer exposures have given better results than stacking. Hubble can only do 45 minutes. It would use 2 very high magnification scopes + low resolution image sensors to track 2 stars. Very high precision servos would then control alt, az, & rotation.

It would be quite massive. The star tracking scopes would need their own alt/az adjustments of lower precision. There would have to be a high speed mechanism to initialize the main camera. There might be a way to couple a star tracking scope to a main scope, using a very precise accelerometer.
Sign up now
to remove ads between posts


Quick Reply
Message:
Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
For Sale Celestron Explorer Telescope***LOWER PRICE*** snolord1 Non R/C Items (FS/W) 1 Sep 21, 2012 03:03 PM
Question Synthesized 72mhz Tx Telescoping Antenna Legnths wannabe FPV Talk 1 May 25, 2012 04:00 PM
Discussion Telescopes at SEFF Dave Burrow SEFF 2012 29 Apr 20, 2012 06:10 PM
For Sale Goto Telescope RCHSF Non R/C Items (FS/W) 0 Feb 29, 2012 12:56 PM
Wanted 8" Reflector Telescope, or Nice Refractor Telescope RCHSF Non R/C Items (FS/W) 0 Jun 21, 2007 09:44 AM