Jump to content

Stargazers Lounge Uses Cookies

Like most websites, SGL uses cookies in order to deliver a secure, personalised service, to provide social media functions and to analyse our traffic. Continued use of SGL indicates your acceptance of our cookie policy.

Ajohn

Members
  • Content Count

    783
  • Joined

  • Last visited

Everything posted by Ajohn

  1. The ethernet cable is pretty clever actually. It designed in a similar fashion to the co axial cable that is used to carry radio frequencies such as analogue television signals from the aerial. There is some info on the subject on the wiki https://en.wikipedia.org/wiki/Transmission_line This should talk about distributed capacitance and inductance along the line because that is what makes them work. In other words some values per unit length after a fashion. If this isn't done all sort of strange things can happen. Even the actual resistance of the wire plays a part as it can help damp out the signals travelling along the line. Fourier comes into this where the square waves typically used in digital transmissions are considered. The rising edge is made up of a harmonic series of much higher frequencies. If these aren't preserved there can be overshoot, undershoot or even ringing or nothing at all coming out of the other end. Signals can even reflect back from the part that receives them. The reason solid cored cable can carry signals for greater distances is probably down to controlling the parameters as it's made and more importantly something usually called the skin effect. DC uses all of the diameter of a piece of wire. As the frequency goes up the current levels get more concentrated towards the outer areas, this effectively means that it's resistance increases with frequency. Useful in the transmission line area as it means the wire diameter can be used to control the resistance. Resistance offers a good way of controlling damping - how well a rising edge on a square wave is preserved along the transmission line. I'd guess they under damp. In other words a rising edge over shoots and a falling edge undershoots. I can't find any pictorial illustrations on the web but all of these factors combine and limit maximum cable lengths for some specific design of cable. Me - well I can believe that splicing ethernet cable into a usb cable might increase the maximum length that can be used. I wouldn't take it as fact though. The person that posted this was an amateur radio person - they generally will understand the above and be well aware of the factors involved. On USB 3 it wouldn't surprise me if fibre optic cables become more common over time. John -
  2. I use the cases and enclosures site too. I've had several off them. Nice prompt delivery and for the price decent quality as well. I found them via ebay. I enquired about some from a seller on there, didn't buy and soon after received an email complete with a discount offer. They do that every now and again. I just bought a case of them that should take all of my eyepieces and another for an 80mm APO I bought recently. The fact that they give a full list of sizes helps but remember that some of the dimensions in one direction will include the lid which will be lined with dimpled foam. Another make that can be found on ebay at times is called something like skelington. They are dearer but still not too bad in large sizes. Thanks for the foam link. Haven't checked on that one but I often find that the foam costs more or less the same as the case when bought separately and the ? skelinton ? is in need of a new insert. It hasn't kept up with my camera gear so may as well use it for something else. John -
  3. It looks like there are other INDI like solutions. I was hoping to get some links like this when I originally posted. http://sourceforge.net/projects/iastrohub/ http://maxchen.im/tinyastro/en/ Both of these seem to be extremely low power and very portable - no pc level power needed at all. I was glad to see that on the first one jpg's are downloaded and raw left in the camera. While viewing in FIT's etc may be attractive it's a hell of a load from a dslr. That way download speeds get fast. Many cameras put rather a lot in jpg's these days anyway. That can be increased on some cameras via various settings. I would say the biggest problem with these is that they use a very commercial device rather than 'pi etc. It seems to make a huge difference to plat solving times. Only problem is that I followed a link to one and out of stock. It's probably been superseded and always will be from time to time with different parts. John -
  4. Is that solid or flexible cat 5? Some people on the INDI forum have swapped to an Odroid. Best read if interested. One of them is running the Kstars end in a VM on a Mac book. It could be an Odroid or 'pi on the other end. I ordered an Arduino board yesterday after seeing the INDI focuser. I have had a long break from software after all. Last time I looked at this sort of thing it needed too much kit. John -
  5. I've never tried to do this so much of it is new to me but from habit - work - I have done what I would do there. Want to do this, ok what are the options. I'm reluctant to write any software but I found a couple of sites that may help http://free-electrons.com/docs/ Seems I am now largely what is now called a bare metal person, mostly in assembler in my case so that one is of great interest to me as I feel the need to know what lies underneath Mixed c and assembler on a 'pi and maybe some help with make and configuration files. not sure. My eyes glaze over when ever I look at what is in one of those. I suspect the worst problem is producing the first one in the same fashion as writing a first very early windows application. http://www.cl.cam.ac.uk/projects/raspberrypi/tutorials/os/ And the USB library I mentioned. I'd be a bit surprised if any thing else is used. http://libusb.info/ The other dev tool that can often be of use is an emulator. As I suspected a google for arm emulator came up with some results. It would take a while to see which offers what and what else is needed. Linux is distinctly different to windows in the way it's organised. Linux is the basic software system that runs the machine, often called the kernel. Windows has one too really but it's not so apparent. Above that sits library files which in loose terms are similar to DLL's and services such as X windows and others that look after specific things. The desktop or windows system sits on top of that with currently another graphics add on. I believe most use Qt these days. There are several desktop packages. The most popular of late is probably Gnome, it has been another called KDE in the past. There are also light weight ones. As I understand it packages mostly just rely on them using some version of Qt so that they can run on all of them - best check as I generally have only used one. Next comes distributions. These people package software so that it will run on their Linux. This means a package from one distribution may not run on another and the method of easily installing them can be different. Even the act of installing it all in the first place varies enormously. I would say the 2 that avoid that sort of thing most are Ubuntu/Kubuntu and OpenSuse. The K in front of Ubuntu means it runs the KDE desktop rather than Gnome and that fact plus problems introduced with a complete rewrite of KDE is why Gnome is probably more widely used these days. I've stuck with KDE on OpenSuse all of the time but these days the Ubuntu stable probably has a lot more activity and help around. I some times have to compile code and chase down dependencies - the web will explain what that means. Compiling is generally dead simple and usually stated in the readme file that comes with source code. Chasing dependencies isn't. eg ./configure make su <password> make install make clean ---------- removes some junk left su john On Ubuntu make install would be replaced with sudo make install rather than su which actually means switch user. If no name it assumes the root user. Some think su stands for super user. It doesn't. A root account - same sort of thing as a windows administer account isn't available on Ubuntu - wasn't when I last tried it anyway. The name root user comes from it being the only account that can do anything with the root directory of a disc eg make a directory or write access a number of system files etc. I'd guess that on Ubuntu if some one wanted to run a desktop application such as a file browser it's possible to create a desktop icon that runs sudo what ever the applications is called. I just create one that runs the application as root so I can browse and edit system files as root from the desktop - not much need for that these days. Source files are provided compressed so it's just a case of down loading, right clicking extract here and it will be extracted under a version number directory. Then right clicking on the directory formed and selecting open console here. The console has this thing called BASH, borne again shell. It's like using DOS in an MsDos shell window but has way way more bells and whistles to such and extent that getting fully to grips with it is a real task. I don't bother unless I really need to. It's probably possible to burn dvd's, write and send emails and even browse the web in it. Pass. Where it can be useful is that it's likely to offer more facilities than a desktop application will offer. The commands are documented by a thing called man. As the output disappears when that is exited it's often better to google man command name. If the command name isn't known google linux - what ever you want to do. Both opensuse and ubuntu offer graphical software installers that will contain what people are likely to want Opensuse offers a huge check list when it's installed. These are dead easy to use. They also offer console based ones aptget or something like that on Ubuntu and zypper on opensuse. I might have the spelling wrong on that one. Apt get is a Debian distribution thing, Ubuntu is based on Debian. I have looked at installing that but I use a complicated set up with specific disk partitioning. Looked a lot more difficult on Debian. Most distributions offer various releases. I stick to the stable one. It will have the least number of bugs especially after it's been out for a few months. The latest greatest isn't a good idea on Linux other than for people who want that and can live with the problems they MAY have. I have seen ex windows people get into trouble by assuming all will be ok. The safe option is to let the automated update system do it's job as it will only update from it's official repository - where the software is kept. It's also safer to install software from there. Next best bet is to compile and add any missing library files from the repo. Last is anything you can get your hands on. I have had problems from doing just that. An unofficial repo and later using the official one. The installer took code from both and they wouldn't work together, result no desktop. My answer to that is a quick visit to the distro's forum for help if I can't sort it out myself. I use an iPad for that or some one else's windows laptop. As far as imbedded work goes Linux is all over the place, even in wifi SD cards it seems. The list is endless. At the pro level they are probably using things like Linux From Scratch - build a kernel with just what is needed in it, BusyBox - a light weight shell and similar things plus a little code. Dlink for instance make what they use available but just provide the run code for what they have added. It looks like OS-X developed by IBM and run on Mac's is basically a Linux rip off. Many software packages and idea originate from open source, often stolen for free as I see it. Pity really as people who maintain it often called dev's don't make a decent living which is why they come and go. Many just write or add something for the simple reason that they want it or for shear fun or interest. There are catches. Bugs of one sort or another do crop up. Often fixes can be found on the web. Things seem better of late in that area on stable releases in particular. Drivers for various bits and pieces can be a problem. eg I usually buy Samsung lasers for instance because they do come with Linux drivers. They also fixed one that didn't work in a couple of days. It will now support them itself. HP seem to have added some for their kit - written by HP people in their spare time. Not sure what state this sort of thing is on things like ink jets. On the other hand a bit of web searching and I could make my Samsung printer behave as an Apple printer via my machine for the Ipad's in the house over wifi. Linux supports Apple Talk. Not the most secure protocol, in fact I bug reported a serious flaw in that area and some one started working on it. Distro's will often pick up serious bugs. It basically supports all sorts of things. There can also be odd problems in this area. Nvidia for instance maintain Linux drivers for their cards. At one point they just stopped working following an upgrade. People who used the open source driver didn't have any problems at all. If some one wants to see what it's like - easy - just download one of the bootable DVD's or CD's from their web sites and boot it. As Linux is very memory efficient it can run fairly well that way. These don't install anything. Many can now be booted from USB sticks as well. I keep one called Knoppix about for fixing things if needed. It's a distro intended to be booted this way rather than being installed. I used it recently to reset a windows laptop password. I used Sun's VirtualBox last time I used one. I would probably use that one again if needed. Emulators are effectively virtual machines but some VM's may be more difficult to get up and running than others. Oracle now seem to own Sun but wont currently emulate arm. Dual booting is another option. I understand that MS finally offered a method for doing that some time ago. I'm pretty sure all Linux distro's will also install that way as well. I've never un dual booted a machine but assume it can be done. When Linux is dual booted I would be surprised if there aren't utilities about to allow the windows partition to be read from and written too. Dealing with windows partitions on some distro's has been omitted at times due to copy right worries but the info needed to do this has always been around some where. There is no 32 - 64 bit messing around on Linux but there may be limitations under a VM. Not that this matters all that much really. Must add Linux isn't for everyone though. Forget running the likes of LightRoom and PhotoShop for instance. There are OS packages and they are capable in the right hands but just like PS they have a learning curve. There will be other areas that have similar complications. Linux in a VM is probably the best option then or 2 machines. Linux did run on slow older machines - since the C++ guys got at it particularly in the desktop area it's heavier than in was. Lots of the graphics bells and whistles can be turned off though - in fact KDE will turn some off automatically if it's struggling to cope. Not sure about Gnome. I'd suggest people have a look at Ubuntu or maybe OpenSuse. I've often stuck with OpenSuse as they have a tendency to offer more stable code but feel Ubuntu has caught up in that area now as there are lots of people working on it - at the expense of other distro's. One odd thing about Linux desktops is that 1 click rather than a double is traditionally used. Of late though sometimes 2 are needed are needed in places particularly in packages. Long post that and probably a bit terse in places. Might help some not sure. There is plenty of help out there though from linuxquestions.org and distro forums if it's needed. Probably a lot more help than there is on dealing with it on an arm. John -
  6. t should work as these sort of things tend to use pulse width modulation. In other words the switch at some frequency with a variable on time. You can also get manually operated ones that work the same way. Some people use those for controlling the light intensity when they convert microscopes to LED lighting. The current rating on these is more or less the same so I'd guess the radio controlled ones use the same switching electronics. John -
  7. There are usb to ethernet extenders about. I thought that the link I posted was to one but that turned out to be incorrect. These do convert or maybe just drive ethernet cable usb style - there was a post somewhere that reckoned they had extended one by simply splicing cat 5e cable into an ordinary usb cable that had been cut in half. Other things could be going on though. Most of the extenders state usb 1 some say usb 2 compatible, some usb 2. Personally I'm not sure if usb 2 devices will drop to usb 1 in much the same way as ethernet drops speed if it can't be maintained so I see these things as being a bit dubious. Any built in chip could force USB 1 anyway. If you search ebay for usb extender several are listed. The length of cable they will drive varies. There are also active usb cables that use a chip to receive and boost the signal taking power from the connector, some have a separate power supply. Lengths vary again. While watching a pretty awful spaghetti western 1/2 hr ago I realised that while the powerline ethernet adapters look great they probably don't have gigabit lan drivers in them so data rates may be 100mb/sec in practice. That seems to be the limit for a 'pi anyway. The Odroid does seem to be gigabit and will run the software that interests me personally. Same price more or less but people seem to be having problems with INDI and wifi at the moment. However as powerline could be a private network, nothing else on it, it might turn out to be the fastest option even at 100mb/sec. It should be possible to treat the 4 USB ports on the new 'pi effectively as a hub so there isn't any reason why something shouldn't go in one and be split off into others but 100m USB cables appear to have their problems. Some suggest that the active ones can be daisy chained but if so why do some have separate power supplies. As some one pointed out earlier cross compiling does seem to be the only "dev kit" available in practical terms. Having seen what INDI and Kstars/Ekos can do I don't think I need to write any software at all -. other than I will wonder why they haven't put their remote focus in it as well but that may be down to the number of timers in the processor. If some one uses windows and wants to work in Linux it should be fine working in a virtual machine. I have done that in the past the other way round but not for a long time, My biggest concern in this area is if Kstars/Ekos will run on a decent netbook. I've always intended to run a Linux on it but since iPad it's just been lying around. In this case I don't need a long cable - I only need that to run it all from my desktop. Maybe perusing that via vnc on something else at times. According to the wiki the ethernet cable length limit is 100m which can include some flexible. Like most things I'd guess it could be longer but 100m with decent cable should work 100% and not drop packets. I'd guess that the camera usb ports place another limit on max speed. While they may be usb 2 this doesn't mean that they have to send 480mb/sec continuously, just pulse at a rate they can manage. Bit like what ISP's do to our connections. I haven't downloaded from a camera this way for ages as direct from the card was much quicker. Guess i had better try that and find out. This might seem all a little strange but I want the option to use it all close to the scope and away from home - ie portable. This rules out using watty parts close to the scope for me anyway. John -
  8. I've just come to the same conclusion Mike. 500mb/sec ones can be bought. It should even be possible to use a remote usb hub with these plus the ethernet to usb adapters. There does seem to be some USB2 ones about but I'd guess that would mean actually buying them to find out. I'm after something portable rather than fixed otherwise I would probably just run an ethernet cable. Having seen the UK price of reinforced cable I would probably bury it in toolstation hose pipe if I did that. In real terms I will be using an extension lead plus the powerline units and a dslr which is why the bandwidth concerns me. On that score the Odroid looks more favourable than a 'pi. It's an interesting area to look at. The manufacturers are producing chips to actively boost usb 3 because the max lead length is so short. There are also some fibre optic usb 3 leads about as well but the prices are a bit high at the moment. Even a fast usb wireless standard but no one seems to be doing anything with it. At 10ghz it will probably only work over short distances anyway. Only remaining problem is Altar Astro. I ordered a light weight mount off them specifically for this and expected them to get in touch if it wasn't in stock. Not a peep out of them even after a message on their answer phone. John -
  9. It reads to me that the server side must be Linux based, 'pi etc, but the client side can be windows, linux, mac or android. I became a little curious and found there is a cross platform usb library about that can be run in and used in user space. It's likely to be very compatible with all kernels it can be compiled for. What it appears to do is standardise the usb interfaces across various platforms so the kernels are still being used. It looks like the latest versions will also handle usb 3. The wifi dongles come in 3 flavours, 150,300 and 600 odd Mbits/sec. The later is only available on USB 3 and run at 5ghz. I believe some step back under certain circumstances and that data rates are upped by using more than one channel. If that can't be done due to use or interference the speed drops back. There are some interesting numbers on here, bluetooth might work out faster than some. the latest is supposed to reach 24Mb/sec https://en.wikipedia.org/wiki/Comparison_of_wireless_data_standards I should add that it looks like INDI can be run via wifi or wire but I haven't really ;looked at the detail. Maybe a sophisticated cable reel and solid cored cable is the fastest option. John -
  10. There is an interesting hack post on noise that I have not really looked at closely yet. https://landingfield.wordpress.com/2014/05/13/teaser-nikon-dslr-black-point-hack-for-astrophotography/ Some where around on the same site there is info on modifying the filtering as well. Interesting comments about canon - nikon noise levels. Not sure they are entirely correct as companies make changes from time to time. Eg my E-M5 uses a Sony sensor. Turns out that My E-M1 uses a Panasonic one. Noise levels are similar but the E-M1 seems to have less chroma noise. I believe Nikon use Panasonic sensors from time to time as well. Live view will heat the sensor up - it's being used all of the time. Some mirrorless cameras go to some lengths to only activate it when needed and even warn about that effect when the rear screen is in use. John -
  11. The companies that maintain Linux distro's have a way of ensuring people upgrade - eventually people try to install something and find they can't because things have moved on. I had that happen last week also found I can't update my favourite photo editor -not that I need to do that. Back on topic. There is something about that will link usb over wifi according to one source http://www.raspberrypi.org/forums/viewtopic.php?f=31&t=97179 The person who wrote virtualhere mentions that it's stand alone code rather than using the Linux facilities. That might explain the speed. Not sure. Like most I pull the card out of the camera and plug it into a reader. I haven't used the camera download for a long time but vaguely remember it's slower. It's free to try on one device. Makes me wonder about wifi cards. Interesting demo of remote control here. There are some others on youtube. The INDI telescope end has been ported to a 'pi but there is a 'pi look alike around. Odroid. 1.5khz 4 core and some reports that it's IO is faster. It also has a gigabit lan connection. There is an option to install a different type of flash memory to it as well. Treasure Island Britain makes fibre optics way too expensive as the prices leave a bad taste in the mouth and that's just the cable. John -
  12. Looking a bit more. Cat 5 and 6 cable can be used for up to 100m including 10m of the flexible stuff. The rest needs to be solid cored with kink bend limitations of 4 times it's diameter. In the case of Cat 5 it sounds like it would be better to use 5e. The cable is pretty cheap so a dedicated line or two wouldn't be out of the question. Not much info about on a fibre optic arrangement but this http://www.fiberstore.com/-p-32658.html?currency=GBP&gclid=CJXP8Ir8zMMCFSbHtAodXCsAnA going on this just plugs into the usual connect on a router or what ever http://thefarmingforum.co.uk/index.php?threads/fibre-optic-internet-diy.6846/ Treasure Island GB again, when UK supplier sell them they must laugh all of the way to the bank. There are also active USB extension leads about such as these http://mdrelectronics.com/ProductDisplay.asp?PID=5651 Going on that link it may be possible to just link the usual 20m ones up to each other. Just like the ethernet USB extenders there would be a need to check they don't just really use USB 1. There may be USB 3 ones about. Then there is wifi. If I did that I would want it on a totally different network than the rest of the house. It's pretty busy already. That doesn't sound too difficult to achieve. There are no signs of wifi usb extenders. It might be possible to use a couple of 'pi's for that or just one as a transceiver to another beefier one near the scope. VNC and turn the auto updates off - I just get a pop up telling me some are available. I suppose VNC is ok for people who can't sort other ways out. I don't have any software chuntering about on the internet all on it's own either - can't say the same for other things in the house, mac's, windows, phones, tv, ipads, radio and etc. I've run my current set up for around 3 years, time to upgrade really but it's still as fast as it was when I initially installed it. John -
  13. There is another way but how to use it ? eg, the page may load slowly. http://www.fiberstore.com/10-100-1000base-t-gigabit-ethernet-auto-negotiation-copper-sfp-optical-transceiver-p-20036.html There are also some that use 2 cables. Others look to be even cheaper but not sure how they connect. There might just be a reasonably priced UK supplier. There are some cheapish dual fibre managed ones on ebay but straight links like the above would be better. Not that I know much about them. Looking further they all use 2 fibres. John -
  14. I'll be brief. It's small 220mm but < F3 for a cassegrain. I'm not sure I will ever finish it. I've completed the design including a corrector for it. That took a while. If I do any more work on it it will be via a machine. I have collected a few bits and pieces for that but too busy with other things. I made a ronchi screen but wasn't impressed. I used something with the same basic design as Texereau's but with micrometer spindles to drive it and adapted to to do x-y type knife edge testing as well which I found slightly easier in some ways when things got close. As a final check I knocked up a Dall null tester. With your problem I wouldn't care about what experienced mirror makers thought, just find out for myself. Many use machines in any case. If I did that I would probably see what laps about 4/5 the diameter of the mirror did. Might be an idea doing it by hand as well mirror on the bottom. 5/6 is supposed to maintain the radius with suitable strokes. Not all that much interest in dobsonians now. A 10in put me off a bit some time ago. I made a hexagonal tube from thin ply with glued butt joints - worked well except I found it wasn't entirely square to the mirror a little too late. John -
  15. That's what the remote hub I posted a link to seemed to be when I first looked. Must have been wishful thinking. Some one posted a link to a device of the type you suggest but usb 1.1 and no speed spec. I've looked for better but so far haven's found much. I'm not sure about max distances but having used a networked PC at work it seems to be a pretty long way with just typical pc gear. I got some people in to wire up an office once. I'm sure some of the cables were over 100ft. Later larger premisses and an IT Dept rather than me plus much longer cables but I'm not sure if repeaters were used. They sometimes do this by using ethernet switch boxes. INDI and a Pi seems to be good option as it can take care of the lot on one cable but that needs looking into. I'll ask a few questions on their forum shortly. The 'pi linked too has 4 ports and i/o pins which I suspect could drive the remote focusing they describe on their site - just model servo motors. I can cope with loading up Linux software, I'm just not keen on writing the stuff any more. The habit dies easily and I have the feeling if I started again I might not be able to stop. The INDI instructions look pretty straight forwards anyway. I have found usb extenders with a spec of sorts, ebay 361045474841. 251378758676 Once the distance goes up so does the price. There is also a lot of USB 1 stuff about. John -
  16. I have a similar problem to you Damain on what has turned out to be a long term project. Mixed designing it and making the mirror. The blank was hogged out for me and despite having done it before it hasn't finished up spherical after fine grinding. Not saying you have that problem but if you do it's best to go back to grinding again. I can tell I have a problem with simple lamp test when the mirror is wet. The image point and the distance at which the movement changes is too woolly. The usual way to persuade a basically spherical mirror to polish evenly is to alternate tool and mirror on top. I feel I would still work that way with a floppy thin mirror. Tool on top is an excellent way of getting a good clean edge and that is also probably the best way to finish off a thin mirror. Either way use say a 1/3 overhang max but vary it and maybe a slight W or V. You'll find mirror on top tends to polish up the centre more especially if off centre but again for a sphere not extreme. The other thing is to avoid getting into a rhythm. Vary the steps around and how much what ever is on top is rotated etc. It's the random aspect that makes the sphere automatically. If it's a thin mirror on top make sure you only apply pressure in the centre - a weight maybe. Once the worst if over I'd also try say 1/2 hr polish and then a much longer press but I'm and old fashioned disciple of Texereau other than mirror on bottom for finishing. He does describe a good way of testing for a polished sphere but it needs an adjustable slit. John -
  17. The point was making about think arms rather than platforms is that it's easier to make a rigid pivot. One way for instance would be to fit a ball race into a hole in one, few washers for spacers and and bolt it to whatever it pivots on. The 2nd arm needs a sliding contact - ball race again. Another alternative might be the the bronze bushes bearing suppliers sell. The only problem with those is that they are meant to be driven into holes sized to shrink them a bit. It might be possible to buy say ones with a 6mm bore and ream them out to 1/4 in or some similar idea. Bright drawn metal bar is usually pretty close to the stated diameter. Silver steel these days or drill rod as it's called in the USA isn't as close as it used to be. Some supplier stock ground rod. That will be very close. I would be inclined to use 18mm MDF even though breathing the dust isn't a good idea. I have also routed aluminium plate in the past. To add a camera all that is really needed is a small platform on the end of an arm with a ball head on it. I've been intending to do it for some time but ordered the light weight Ioptron mount earlier this week. I suspect I may get annoyed with the suppliers. If they sell it that should stock it / declare delivery times on their web sites. There are probably some better web pages about on tangent arm drives than the ones I posted. It's an old idea but I would seriously wonder about alignment if really long exposures are contemplated and wonder if the complication of the 2nd arm was worth while. John -
  18. You might do better to think about a tangent arm drive rather than a platform. I've tried to do things with hinges and usually run into problems because they are loose. There are all sort of ways of making bearings and that sort of thing is easier on things that look like arms rather than a platform. This gives and idea of what accuracy the twin arm can give. Use the link towards the top of the page to download it http://adsabs.harvard.edu/abs/1995JBAA..105...65C This might help http://www.mikeoates.org/mas/projects/scotch/hdm/home.htm An alternative is to drive the nut and and bend the all thread as shown here. This one is over the top for just a camera but it should be possible to include the sort of adjustments equi heads have for alignment. Longer possible exposure times need accurate alignment. I suspect that is why many simple trackers have been made just using one pivot and arm. http://www.dv-fansler.com/Astronomy/portable_equatorial_mount.htm Must admit I wondered about a commercial camera platform recently. The max payload is disappointing and they suggest mounting them on a Manfrotto geared head and a hefty tripod. Price - ouch. While I haven't handled a geared head like that I doubt if they allow easy fine adjustment as most equatorial heads do. This is why I drifted away from any form of camera platform but do feel most parts of a tangent drive could be made from mdf or even ply. John -
  19. I had a look at using C on it. It seems running a dev kit on it is a rather slow option. Only problem with that approach is that I retired at a rather early age from a job that involved a lot of electronics and software. The will to write software can fade as surprising as that might sound. Due to that experience I also looked to see if there were any RTOS's available. There are but that would probably turn out to be a major undertaking. I may have found an answer INDI http://indilib.org/ It seems windows versions are also available but I didn't look at that. Looking briefly that does seem to be capable of controlling remotely over wire or wifi. It also looks like the bits needed at the telescope end have been ported to 'pi. On the other hand it should also work via a remote hub. The mention using 'pi USB and suggest adding a powered hub. The new board has 4 and the soldering iron could be used again if the current it can give isn't high enough. The board also seems to have additional port i/o and there is a neat idea for remote focusing of both finder and scope on the site. They also seem to have some sort of autofocus arrangement. I still have a couple of philips CCD webcams. Some seem to have had a bit of trouble using those. I had hoped to use that as a guider and debayering one of them as per the thread on the subject. It's an odd ball on linux and needs a kernel patch to work well or at all, not sure. Sad if they can't be used as from the few I have played with they are still the best. Not much luck with my question on Amazon so far - eg Why don't you just use a 5 port USB hub. Also a link to a usb via ethernet using a long wire to get over USB's distance problems. More useful but USB 1.1. Not sure about distance but I have laid 100ft plus cables in the house in the past without using anything special to drive them. If distance is a problem a 'pi would need a repeater as well. Maybe wifi would be simpler after all. It's just I'm sure some one local has tried to hack our printer every now and again. John -
  20. I found the page relating to the 'pi autoguider. It's here http://sy2000.blogspot.co.uk/ John -
  21. Thanks for the replies. I hope more keep coming. A lot of the 'pi stuff seems to run on SD cards. I would have thought that will slow them down and just wouldn't compete with on board flash and ram. With the cards a lot would depend on the hardware on the 'pi not just the card speeds - just as it does when downloading from cameras or directly from it's card. I have found 1 hub that seems it may have something to offer - I have no connection with the seller http://www.amazon.co.uk/warranty-EasyAcc®-SuperSpeed-Ultrabooks-Raspberry-x/dp/B00IZA09SU/ref=cm_cr_pr_product_top It would need a soldering iron to add local power and just 3 ports so I have asked a stupid question. The reviews suggest that it works with Linux. Not surprising as this sort of thing was added to the kernel when the open source software was developed. I've had my main machine running Linux for at least 15 years now do I know my way round it really - no, I mostly stick with the desktop and delve deeply only when I need to. So if I run into grief I ask on linuxquestions.org adding please be gentle. The various distribution forums can be helpful as well. I'm pretty sure there are a few spare wires in an ethernet cable connection. Might be a simple way of managing remote focus. I've used a sata hub with card reader and usb ports for a long time. Other than one failing after more years than I can be sure of can't say as I have had any problems with it. I use it to load up all of my photo's and to back up my drives at times. As it's close to the screen I also use it for keyboard and mouse comms. It comes with it's own power supply which I suspect helps a lot. The unpowered hub in my monitor is hopeless. I'd guess the 'pi board is this one? http://www.ebay.co.uk/itm/RASPBERRY-Pi-2-Model-B-version1-1-Board-QUAD-CORE-1GB-RAM-The-Very-Latest-/261757168772?pt=LH_DefaultDomain_3&hash=item3cf1f16c84 If there is a decent C dev kit available I might be tempted to have a go at doing something with that in the summer. John -
  22. Thanks Gkec. That indicates that there might be some mileage in what I suspect I will end up doing. There is some OS software around too. More google needed. Nick Morris. A definite option. That is the sort of thing a USB hub server does - just a minimal pc of some sort in a small box that serves usb ports over ethernet. It seems even Maplin sell one. They can be rather expensive but I'd guess those will serve multiple users at the same time. It should be an option on all OS's but I'll bet there are no drivers for Linux which is what I run. On other comments relating to windows home I would have assume that VNC can be installed on anything but the way I would like to do it the USB ports will look as if they are actually on my PC. No remote desktop involved. There is some Pi stuff about but I've not looked that hard and have seen comments about video bandwidth if it tries to do anything with it like autoguiding. There is a demo on youtube somewhere but no details, just autoguiding I suspect. Another set up is shown here. http://www.brahmand.me/astropi/ I forgot remote focus but then thought hang on, dlsr's have magnified live view which might help a little. A 'pi might be able to handle a number of usb ports via a separate hub. Maybe there is a board with lots already. Ethernet can also be sent down a mains cable but I've not looked at bandwidth. There was a reason I posted in DIY. I feel things like VNC and Teamwork belong some where else but maybe not as any suggestions are welcome. John -
  23. I've been nosing around the web looking at things like Rasberry Pi autoguiding and things like that. I'd appreciate any links people have on that subject or anything similar relating to telescope control. There doesn't seem to be single source of links and hours on google bought up the starynights posts and a couple on here eventually but it's slow going. Thinking around the subject - Laptops have the USB ports etc but not wanting to get cold I have wondered about another approach. Basically using a USB to Ethernet Hub Server at the scope. I would be inclined to run a cable to what ever I was using to drive it. 50m cables don't cost much, longer than that and cable and a crimping tool might work out cheaper. If I wanted to be close to the scope all that would be needed is a shorter cable. If I remember correctly even my netbook has an RJ45 socket. Probably another length as well if I drove the car somewhere. Some might favour wifi but I'd guess something could be plugged in to handle that. Anybody tried this or have any thoughts? I've not really tried imaging but as I see it I would need 4 ports. Fortunate as that is what is about at reasonable cost. Webcam, Camera, Autoguide Port and Handset probably with the usual irritating RS232 socket. One other thought was that Linux boxes used to be used purely as firewall routers when people mostly wired things up. Any old PC could be used even a 486 I believe as it didn't have much to do. This could be something a Rasberry Pi could be persuaded to do with adequate bandwidth. I would be ideally connecting up to a Linux box but chances are that any sort of hub might have linux in it anyway or maybe it's some form of ASIC or specifically written software as that would probably be cheaper in the long run. I make some really weird typo's at times. I've checked but ................ John -
  24. It's worth mentioning something about testers that often isn't even mentioned. The book link explains it well. The slit width of the tester is important too. Even if it's a pure slitless type where a single scraper blade etc is positioned part way across a led it can still be adjusted. Best read the book. One of the other factors concerning books and testers is that often they relate to a 6in F8 mirror. It's a fact that a spherical unfigured 6in F8 mirror gets close to meeting Rayliegh's criterion so you might say that precise figuring isn't needed. Even 1/2 of the correction will be ok.. Go up in diameter or use faster F ratios and things change rapidly. On my 1st mirror I didn't bother working to a good sphere 1st. Basically because I couldn't adjust the tester with sufficient finesse. The mirror just blacked out as soon as I adjusted the tilt. I just used the fringes ahead of the knife edge. I then went on to figure the mirror in real terms having not learned to use the tester. I'm sure this wasted a lot of time, It is possible to make a spherical mirror slowly grey out and in my view it's worth people spending some time just making sure they can do that. It's even possible to alter the shadows on near spheres by moving the knife back and forth. Once some one can do this and it may involve adding bigger knobs to their tester or changing bits of it they have far more chance of making accurate measurements later on. The important aspect of working to a sphere 1st is a clean edge region because if it happens to be turned down it will just cause problems later on. The extreme edge is best tested via the fringes ahead of the knife edge as explained in the book but if there is a nice bright diffraction ring round the mirror and the shadows show that it's smooth all is likely to be ok anyway. Biggest problem is that it's a very sensitive test and will show things that don't matter too much. The other maybe sad aspect is if some one finds themselves chasing their tail trying to locally figure a mirror it's often best to bite the bullet and polish it back to a sphere and start again. Like most things there is a learning curve. Also rather than saying that a measurement is 1mm out use Foucault software or sums as outlined in the book to work out what that really means in practice and notice that the figuring steps you use show up in the measurements you take - gives a good idea of how accurate the readings you are taking are. Some take several sets of readings and average them, another way of doing the same thing. Trouble is a post like this sounds complicated and might put people off. All it really means is giving some thought and spending some time on making the tester and using it. The book shows an excellent design for a table to move a tester around with. There is no need to copy all of it just stick the variety you favour on it. It uses metal V's sliding along a bar or tube with the weight of the parts and a spring or even elastic bands to take out the play. The metal V's are a good idea really but other things could be used. Metal - well B&Q sell strips and junior hacksaws. Probably small files to smooth the edges too but the do sell wet and dry emery paper which can be used with a bit of 3in1 etc on it. The table could also be slid by hand and readings taken with a rule etc. Actually I couldn't find a suitable spring so just used finger pressure to take out the play in the axial adjustment screw. Lasers - some might like to take a look at the yahoo interferometer group and see what they get up too. Yes you can build a simple interferometer at home using cheap bits. To get a feel search Bath interferometer wiki. It might interest some people who are keen on making very precise mirrors or just fancy trying it. Some still use a knife and source on it though for some of the work as for certain things it's rather hard to beat. John -
  25. Some people measure the centre, 70% and edge - as near as they can get to that and then check that the mirror is smooth between the zones they have measured by simply drawing the knife back. As Nigel points out a ronchi screen can be used to check for smoothness and is probably easier than judging smoothness with the knife edge. When a mask is used too many zones isn't always a good idea as there will be bright diffraction edges round the edge of each hole. How bad that is part depends on the slit width or pin hole size you use on the tester. Much depends on mirror size and focal length but holes in the range of 1/50 to 1/100th of the mirror rad make readings easier providing they aren't to small. Coudre masks are usually a better option usage wise. They are based on sound maths. This page explains and gives dimensions for several mirrors and a calculator. http://www.atmsite.o.../Carlin/couder/ It also mentions a book you can download from the internet archive. The number of zones needs to be selected to leave a reasonable sized hole at the edge. The web page doesn't bother with a centre reading which is fine with most foucault calculators. That isn't an easy zone to measure but on the other hand if your read the mirror rad incorrectly it can throw all of the readings out due to the way the knife edge movement calculations are used. Perhaps the best answer to that is to measure the rad with a separate mask with a single hole in it. I've found 2 masks useful. One does centre and edge. The other the bits in between for a 5 hole mask. You might find 7 zones better. The faster the mirror the more that are needed really. When using a mask it's best to carefully look what happens to the shadows when the knife is definitely a little in front and behind the correct knife edge position. That way you should pick up accurate shadow reading quickly. With a bit of practice the knife can be positioned to better than 0.05mm or even 1/2 of that if slit and source move together. Some manage 0.02mm with a stationary source. It's not easy to describe what the shadows do so it's best to observe them and deliberately move the knife very very slightly out of position. The tilt adjustment on the tester needs to be very fine too. If it isn't stick a big knob on it. People don't seem to realise that cutting into the beam to produce the shadows needs knife movements of typically 0.01mm or even less. There is really a need to make that level of movement slowly and easily. In days of old when knights were bold they used to press onto the table they used for testing to finally position the knife. That and pin sticks etc take an extra ordinary amount of practice. One word of warning. When things are getting really close this way of measuring can be so sensitive that you can find yourself trying to figure out temperature effects that are down to you last attempt at getting it right. This is why it's best to figure for short periods and try and approach the correct figure rather than going too far and having to work back and even leave glass mirrors for a while before testing them. John -
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.