Jump to content

SkySurveyBanner.jpg.21855908fce40597655603b6c9af720d.jpg

rfdesigner

Members
 View Profile  See their activity

  • Posts

    1,107

  • Joined

  • Last visited

 Content Type 

Blog Entries posted by rfdesigner

  1. rfdesigner

    Getting Better.

    By rfdesigner,

    I've been slightly less than perfectly happy with the guiding, finding it hard to beat 4 arc seconds FWHM in my raw subs. So... I've made a new bracket for the RA motor out of 6mm thick 60mmx60mm aluminium angle.. and finally I've got it guiding tonight with the newly secured setup.
    On a 5 minute sub working at 1500mm focal length on 5.4u pixels I'm getting a FWHM of 4.01 pixels, that's 3 arc seconds FWHM. The focussing image (1 second long) was giving 3.3 pixels FWHM, so that suggests the guiding should have been about 1.7arc seconds FWHM or 0.77arc seconds RMS (PHD agrees). I'll have to check the log and see if PEC might be able to make a difference now.
    Not perfect, but certainly a good big step in the right direction, and leaves me with bigger fish to fry, like getting a coma corrector and filter wheel installed.
    Time for a little happy dance!
  2. rfdesigner

    Where are we?

    By rfdesigner,

    I had a recent message about this blog, good to know I haven't been writing into thin air!.. and of course I've added nothing for too long.
    So.. where are we?
    1. Mount functions:
    Slew works
    Guiding Works
    PEC kind of works.. I have a new PEC algorithum which should improve things.
    Park, needs adding
    GOTO needs adding, software written with this in mind.
    2. Drive needs propper PCBs and dsPIC in the base to allow GOTO (spherical maths, needs sin/cos/tan)
    3. Base unit needs boxing up.
    4. Hand Controller needs the backlight working, not currently wired in.
    5. Hand Controller needs boxing.
    Performance: Primary concern is the performance of the mount itself, so I plan to strip down the RA worm and rebuild with new bearings and new bracket, with a bit of luck things will be much improved, currently its 15 arc second peak to peak over a single cycle plus an additional 15 arc second peak to peak cycle to cycle, and it's rough. It's the roughness that's the biggest concern, followed by the cycle to cycle variation.
    Derek
  3. rfdesigner
    Well it's been a LOONG time since last posting. A house move and other family crises intervening.
    Finally things are settling down and I managed to get out under the stars for the first time proper at our new house. Scope is now in a semi-permnent position (scope permanent, camera, PC etc. not)
    So got the gear guiding again and got out my new and shiny ATIK 383L+ :D
    Thin cloud + full moon so I was never going to get a good shot :mad:, but at least I could get the scope aligned etc.
    New for this session:
    1. Atik 383L+
    2. decently dark sky (in theory, when it isn't cloudy)
    3. semi-permanent position.
    Result:
    About half a dozen stars recorded on 12 x 1 minute frames. FWHM ~= 3.5 arc seconds. stil a way to go on that, target is 2 arc second FWHM.
    Guiding log below (this is at 1.5m and 8.3u pixels, so about 1 arc second per pixel)
    notable issues:
    1. mount needs the self programming PEC re-write adding.
    2. worms need some new ball beaings (in the post) and new motor mounts (also in the post)
    3. CCD iced up a touch, so I mustn't run it so cold for now. (one of the problems on a newt, the CCD can face down, spilling cold air)
  4. rfdesigner

    OAGing

    By rfdesigner,

    Well.. I've now got an OAG (SX-OAG) and after many months of delay caused by other demands and uncooperative weather, I got my first autoguided shot with the mount.
    I knew the DSLR and the guider were not parfocal, so I accepted the inevitably poor shot and had a crack anyway.
    Stacking 30 x 30 second shots with no software registration showed the movement was not discernable. Out of focus yes, but 'sharply' out of focus, not just blobs so the autoguiding is really working.
    Next up: C mount extension tube and DSLR remote control for multiple long images.
    nearly there!
    Derek
  5. rfdesigner
    Next in line is was the second axis. I added the code, debugged it, and got the slew working for it. All relatively simple as it was a cut a paste from the RA axis but without needing to make it track, same error feedback but it keep calculating the same position, sounds a waste but once GOTO and corrections for misalignment and atomspheric effects on position get added it will be changing continuously.
    anyway.. added the circuitry, tested, added the motor drive chip and motor and bingo.. worked first time!.. really!, all the way from the hand controller through the PICs and out to the motor.
    Next: lets stick it on the scope.... now where is that bolt... bother!!
    It's imperial.. whitworth? BSA?.. no idea.. bit like M8 but coarser thread.. any imperial thread gauges anywhere to measure the other one?.. of course not.
    so I'm googling for a solution but it hasn't presented itself yet.
    if anyone has any ideas.. I'm all ears (eye's not being that useful until this is working :-)
  6. rfdesigner

    East a bit..

    By rfdesigner,

    Well after much hair pulling head bashing and general frustration I've finally got my hand controller communicating with the motor driver.
    So you choose 'slew mode' and press the east button, and the motor starts slewing east, let go and it immedaitely resumes its normal tracking, press west and it turns to the west.
    Might seem simple but it's been anything but. Two completly seperate chunks of code in two seperate PICs. The first PIC scanning a keypad for finger presses, interpreting modes double key presses etc., communicating down a 2meter I2C comms link, interpreted by another PIC and then the correct motor instructions sent over the same I2C wires to the motor drivers.
    Not straight forward.. but adding the GOTO commands will be no more than a couple of evenings work, essentially that's how the slewing is working now, it's slewing the target point and the motor is playing catchup. Autoguide port?.. childs-play.
    And I've just confirmed, it's hot plugable.
    Am I happy... of course.:D
    Do I expect anyone else to follow what I'm on about.. :o, maybe not.
    never mind
    Derek
  7. rfdesigner
    On the face of it easy, but in practice less so.
    In my case my drive system has been slightly lagging the stars and I'm trying to work out exactly where the error lies. It looks like the error is around 2% so worm wheels on their own don't account for it, but it could be part of the problem
    I thought it might be that the worm wheels got swapped when the mount was refurbished, in which case the mount would be running a tad slow.
    So how do you count 360 teeth?.. and trust your answer when you're done..
    Solution: reprogram the drive to slew the mount through 360 degrees.... then I know it's done 360 worm revolutions, if the mount has come back to the exact same spot it started at then it's a 360 tooth worm wheel.
    Result:
    359 teeth.
    So it's not that.
    I did however establish that it slews at 2.5 degrees per second... I can live with that :D
    So why is it slow?.. well analysing the various processor signals it looks like it might have been missing an interrupt every so often, and these are used to get the timing spot on. At 40 interrupts per microstep and one missed interrupt per microstep that makes a 2.5% slippage.
    So a bit of assembler to speed up processing and now there's no lost interrupts.. I won't know if it's right until I can star test it again.
    It's bound to be cloudy for the next fortnight.
    Sorry...
    Derek
  8. rfdesigner

    Ouch!

    By rfdesigner,

    Nothing much technical to report from Saturdays imaging session.. alignment is going to have to get better, but that's all.
    Lots of imaging: M42, M1, something as yet to be identified which was supposed to be M81 but wasn't , M82, M51.
    However with the counterweight all black it's not terribly visable.... Not terribly visable as in: bang ouch! :mad:, cut knee through jeans... why do voices carry so much at night?
    What was perhaps more impressive was that the scope never lost track!..:D 1 streaky frame and then everything was back to normal. I suppose that's what you get for the scope+mount being heavier than me.. it wins!
    Oh and my M82 image has just gone into the deep sky imaging section under the novel title of M82.
  9. rfdesigner

    Improved Tracking

    By rfdesigner,

    Well I finally managed to get back outside after what seems like a month of bad weather, poor sleep due to my youngest teething, car problems and colds & bugs.
    The drive system has been adapted with a new shaft coupler.. I'm very pleased with this... all much tighter better algned etc.
    Testing:
    So of course it was a full moon. very very thin cloud and all planets too low, too dim or just not visible.. so where to point.. ah yes M42.
    Well after a bit of aligning I stopped the stars drifting north or south and I got a chance to image with the lodestar.
    a 240 second pic (one full rotation of the worm) produced star trails ~25 arc second long.. pretty goood, certainly much much better than with the old shaft coupler.
    Tried taking lots (450) of 1 second images of the trapesium. Used IRIS to select the best 100 images and the deconvolved out a bit of the remaining noise got down to ~2 arc second FWHM.. not so great. Prboably due to living on a North-South road so I end up with my south view being churned up by the heating systems of the entire road.
    Next: Hand controller communication then second axis then autoguider port.
  10. rfdesigner

    Improved Tracking

    By rfdesigner,

    Well I finally managed to get back outside after what seems like a month of bad weather, poor sleep due to my youngest teething, car problems and colds & bugs.
    The drive system has been adapted with a new shaft coupler.. I'm very pleased with this... all much tighter better algned etc.
    Testing:
    So of course it was a full moon. very very thin cloud and all planets too low, too dim or just not visible.. so where to point.. ah yes M42.
    Well after a bit of aligning I stopped the stars drifting north or south and I got a chance to image with the lodestar.
    a 240 second pic (one full rotation of the worm) produced star trails ~25 arc second long.. pretty goood, certainly much much better than with the old shaft coupler.
    Tried taking lots (450) of 0.2, 1 & 10 second images of the trapesium. Used IRIS to select the best 100 images and the deconvolved out a bit of the remaining noise got down to ~2 arc second FWHM.. not so great. Prboably due to living on a North-South road so I end up with my south view being churned up by the heating systems of the entire road.
    Next: Hand controller communication then second axis then autoguider port.
  11. rfdesigner

    Latest Bug Fix

    By rfdesigner,

    Got the tracking correct, finally.
    What was it?.. a miscalculation.. no.
    mis-specced/counted motor/worm.. no.
    straight lack of understanding.. no.
    pure and simple foul up.. that'll be it.
    The software has a sin lookup table for generating the microstepping, it used to use a 128 sample 1/4 sine ramp to generate it, but I found that a 150 sample 1/4 sine ramp improved theoretical performance. I had changed the lookup table without error, then forgot to change the software that turns that 1/4 wave ramp into sine and cosine waves:BangHead:
    end result was a 'choppy sinewave' running 150/128 too fast.
    At least it was a simple fix. So tried some pics and got Jupiter, M1 and the trapesium. Tracking still dragging in fits and spurts, found the worm was slipping against the shaft coupler, so need a new and better one... but at least the images are starting to work.
  12. rfdesigner

    Hacking the tube

    By rfdesigner,

    This is more of a note to myself
    The focusser was in the origional position (far too close to the primary mirror) so I cut a new hole for the focusser 55mm closer to the top of the tube.
    After collimating images are quite a bit better. No more vingetting at 0.001 degrees off axis.
    Seeing was only ~2 arc seconds, Trapesium stars looked about 1:4 size to seperation.. but at low altitude. Straight up was ok still need to get collimation bang on, not 100% happy right now, although a focoult measurment showed the optics generally were still on the nose.
    Next: electronics boxes!
  13. rfdesigner

    Starting things off

    By rfdesigner,

    A quick potted history is probably due.
    The scope is a Fullerscope MKIV mount with a 12" newt on it. I bought it about 11 years ago knowing it need pulling apart and rebuilding, but the price was low, the mount solid and the optics were good.. The bits that were rubbish I could change.
    The Newt has now been modified with a window so I no longer have diffraction spikes and tube currents are better suppressed. It does however make it rather heavy... I can just about lift it, portable is not the word.
    The mount was refurbished with the paint being stripped off and was then powder coated in white. It has a drive system under development which has to date been through 3 itterations.
    The whole project was started before I took my break from astronomy and so by the time I came back to the hobby a few months ago was mostly obscelete. It was a stepper motor system driven off a PC.. not the excellent Mel Bartels system, but one of my own as I like to challenge myself to new things. These days you don't need a PC, a PIC will do the thinking quite adequately and is easy to programme. The driver chips available now also mean microstepping is a doddle.
    As it stands the Drive is silky smooth despite the stepper motor direct driving the worm. The 150x microstepping is quite sufficient to stop vibrations, and the lack of a gearbox means I have fewer errors to correct. (driver chip : allegro A3972)
    What I now need to do is change the PIC I'm using. A 16F690 is fine for basic timing but no use for anything involving "sin x", tried it and a 32 bit "sin" takes over half the memory for that one instruction. So will be migrating to a dsPIC... but not until I've got the rest of the system good enough to warrant it.
    Which means my home brew hand controller also needs to do the business... at the moment it's not mechanically up to it. So I need some boxes around the electronics otherwise they get wet and stop working.
    Hopefully this Blog will trace the continued development of the hardware and software.
    Derek
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.