Return to the Cloudy Nights Telescope Reviews home pageAstronomics discounts for Cloudy Nights members
Get a Cloudy Nights T-Shirt Submit a Review / Article

Click here if you are having trouble logging into the forums

Privacy Policy | Please read our Terms of Service | Signup and Troubleshooting FAQ | Problems? PM a Red or a Green Gu uh, User

Equipment Discussions >> Mounts

Pages: 1 | 2 | 3 | 4 | 5 | (show all)
Starhawk
Space Ranger
*****

Reged: 09/16/08

Loc: Tucson, Arizona
Re: Open Source Mount Control Suite new [Re: Starhawk]
      #5985760 - 07/23/13 10:13 AM

And as for COM dongles, they are a bit rude.

-Rich


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Post Laureate
*****

Reged: 08/10/09

Loc: Singapore
Re: Open Source Mount Control Suite new [Re: Starhawk]
      #5987965 - 07/24/13 03:10 PM

Found out a few things.

1) the Rikomagic and its RockChip processor has dubious open-source credentials; the Linux kernels running on the RockChips aren't sync'ed with the mainline, and RockChip does not comply with the GPL. Hence from a true open-source standpoint the Rikomagic boxes are verboten.

2) a Korean company called HardKernel sells this - http://www.hardkernel.com/renewal_2011/products/prdt_info.php

$89 including shipping. It uses the same Exynos processor in my wife's Galaxy Note 2. Which is quite a beefy processor, 1.6GHz quad core. The Odroid U2 comes with a heat sink for a case, and you can buy it with a fan. Reports are that running all four cores full blast will crash it, unless you add the fan.

3) CubieBoard2 has dual Cortex-A7 CPUs and is in the $60 range.

All of these options are more powerful than the BeagleBoard or RasPi. I still lean on the BeagleBoard because it is truly open-source (including the hardware design and PCB layout).


Post Extras: Print Post   Remind Me!   Notify Moderator  
pfile
Post Laureate


Reged: 06/14/09

Re: Open Source Mount Control Suite new [Re: orlyandico]
      #5988291 - 07/24/13 06:49 PM

i missed this thread.

for a couple of months i have been idly thinking about a variant of this. i had not considered mount control at all, since, well, all the mounts i own have their own controllers.

but what i was interested in was a small computer to replace my laptop. if the computer were small enough, it could sit on the OTA where my focuser controller and usb hub sits. if it were powerful enough, it could run lin_guider (or PhD now that it's open source), be a wifi client and do what wifiscope does. perhaps also it could replace the focuser controller, be a USB-serial dongle for the mount, and fuse and distribute 12V/5V as well.

with the client wifi, you could just control the mount with something like skysafari - outsource that to the iPhone or iPad or whatever, and don't reinvent that wheel.

to this end i ordered a raspberry pi and a beaglebone black.

i think the beaglebone is a better choice due to it's slightly less janky power situation (the pi pulls power from a usb port, the polyfuses sometimes have high impedance, and the board suffers brown-outs).

i had been thinking that a custom cape for the beaglebone that had a bunch of serial ports, a bunch of USB ports, an input for a temperature diode, and a focuser controller would be something good to design.

i have experience designing digital boards, but i've always used my employer's tools (schematic capture, board layout/gerber generation) and employer's contacts for PCB fab and assembly. i really have no idea how to get that stuff done at home for cheap - seats for stuff like cadence or mentor are hugely expensive. however, due to all the "maker" type hardware design activity that's been going on, there must be free or low cost cad software that people can use.

anyway i got busy with something else and never started working on this. it sounds like it might be useful.

but the software situation kind of scares me. most "stuff" has ascom drivers, and ascom is never going to get ported to ARM/linux. INDI exists but it seems like about 5 people really use it. maybe it's a chicken and egg problem - if the cape existed, perhaps people would get interested.

also as noted above, camera control is also a big problem. since i have the STT, i can write a perl script or something that drives the http interface on the STT. but anyone who has a USB-based camera with no linux-compatible drivers is SOL here. not to mention that a linux-compatible driver is very likely to be an x86 binary rather than source that can be compiled for ARM. that right there probably disqualifies these little "smartphone" style boards, which is too bad because they have small footprints and really good performance/watt numbers.

i spent some time trying to figure out the focuser control stuff, and as far as i can tell, these things have some kind of PIC in them that accepts ascii commands, and translates it to PWM on a set of 4 analog pins. i assume the USB-based ones simply have a USB-serial chip tacked on. in order for that portion of the board to be useful, it would have to emulate robofocus, moonlite and any other protocols that people use, and that seemed to me probably a difficult task of reverse engineering. this, just to preserve compatibility with existing software if you decided you did not want to use this board "natively" with it's own software, but as a sort of super-USB port for your telescope rig.

so anyway, i feel like this could be designed, but the user base would probably be really small.


Post Extras: Print Post   Remind Me!   Notify Moderator  
pfile
Post Laureate


Reged: 06/14/09

Re: Open Source Mount Control Suite new [Re: pfile]
      #5988438 - 07/24/13 08:32 PM

hmm, i just looked at rts2. seems promising. does that use indi or is it its own thing?

Post Extras: Print Post   Remind Me!   Notify Moderator  
OzAndrewJ
scholastic sledgehammer
*****

Reged: 11/30/10

Re: Open Source Mount Control Suite new [Re: pfile]
      #5988526 - 07/24/13 09:32 PM

Gday All

Been reading this thread with interest, but have noted with much mirth that all the current discussion is evolving about "platform" to run on.
In the last few days, it has been pointed out that "this is better than that", but in 3 years ( or maybe even 3 months :-) ) will these options just be another "com dongle"?
ie surely the initial discussion should be around the specifications an "open mount" controller needs to satisfy, ie like an ASCOM interface, but for the low level motor driving etc, not the planetarium to scope.
Until that is really done properly, the rest is just moot.
Thoughts??

Andrew Johansen Melbourne Australia


Post Extras: Print Post   Remind Me!   Notify Moderator  
Whichwayisnorth
Pooh-Bah
*****

Reged: 07/04/11

Loc: Southern California
Re: Open Source Mount Control Suite new [Re: orlyandico]
      #5988642 - 07/24/13 10:36 PM

Quote:

Whichwayisnorth - I believe I mentioned early on this thread that iAstroHub by Anat uses one of those mini Android PC's - the Rikomagic Mk802 - to do mount control, plate-solving, autoguiding, and image capture.

Anat is using a QHY5 for guiding and a QHY8L for capture.

I am currently trying to get my QHY8 working on my BeagleBone. The drivers for the QHY series on Linux are all over the place - the CCD software that Anat is using with his QHY8L doesn't officially support the QHY8 (non Pro, non L) but I scrounged the QHYCCD forums and it seems the original QHY8 is supported if you can get the firmware - which is also on the QHYCCD forums.

So I was (again) wrong to assume that only SX and (to a lesser extent) Apogee are Linux-friendly. QHY is as well, although much more hackery is required.





That is what I get for just skimming the thread.

I favorited this thread to see where it goes. Looks interesting.


Post Extras: Print Post   Remind Me!   Notify Moderator  
EFT
Vendor - Deep Space Products
*****

Reged: 05/07/07

Loc: Phoenix, AZ
Re: Open Source Mount Control Suite new [Re: Whichwayisnorth]
      #5988656 - 07/24/13 10:43 PM

I always like this kind of stuff too. I'm always looking for something that can be put together as a product people can use. I even have two great big mounts (MC500GE from the former AstroDream Tech) that would probably benefit from some kind of new generic controller so that they remain useful when unsupported since they have great mechanics. I'm sure something like the SiTech would do just fine, but if the electronic gurus here can come up with something I would love to try it out (if I don't sell the things first).

Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Post Laureate
*****

Reged: 08/10/09

Loc: Singapore
Re: Open Source Mount Control Suite new [Re: Whichwayisnorth]
      #5988660 - 07/24/13 10:46 PM

Me, I am boggled by all the complexity that goes into things like INDI.

Why not just a simple program that grabs one frame from your camera of choice, and writes the file out to a FITS?

Then another program that can send guide commands to a mount.

And another that can order the mount to go to a specific RA/DEC, or sync the mount to a specific RA/DEC.

With these three applications (which I hope are command-line, no GUI) you can guide, capture, and plate-solve.

So long as these commands were consistent (and hide the hardware-specific stuff behind them) then anybody could script anything using nothing more than UNIX shell scripts.

OpenCV is another very useful piece of software that can be integrated. The CCD library for QHY cameras that's floating around has some sample code that uses OpenCV to take the data from the camera. But why stop there? OpenCV can also do things like do filtering, find the convex hull, etc. All very useful when you're plate-solving or guiding.


Post Extras: Print Post   Remind Me!   Notify Moderator  
pfile
Post Laureate


Reged: 06/14/09

Re: Open Source Mount Control Suite new [Re: orlyandico]
      #5988710 - 07/24/13 11:24 PM

well i have read the spec for indi and it's all very nice protocol design. i think like any project once you start thinking about how to do something very generic that encompasses all kinds of disparate hardware, you end up with something like ASCOM or INDI.

but i do agree that you could "go back to basics" and forget about nice protocols to talk to all these devices and just do as you suggest with a shell script.

the SBIG HTTP interface is pretty clean. it could easily be driven from any language that supports libcurl. it would be nice if more cameras went that way.


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Post Laureate
*****

Reged: 08/10/09

Loc: Singapore
Re: Open Source Mount Control Suite new [Re: pfile]
      #5988943 - 07/25/13 02:44 AM

Hmm. Is this HTTP interface available on the STF? Or only the STT?

Post Extras: Print Post   Remind Me!   Notify Moderator  
Hilmi
Post Laureate
*****

Reged: 03/07/10

Loc: Muscat, Sultanate of Oman
Re: Open Source Mount Control Suite new [Re: orlyandico]
      #5988981 - 07/25/13 04:08 AM

STT & STX range only to my knowledge

Post Extras: Print Post   Remind Me!   Notify Moderator  
MisterBill
super member


Reged: 09/18/10

Loc: Geezerville aka S.FL, MILE 264
Re: Open Source Mount Control Suite new [Re: Starhawk]
      #5993308 - 07/27/13 05:00 PM

I don't think Rikomagic will connect to SkyFi.

Post Extras: Print Post   Remind Me!   Notify Moderator  
Starhawk
Space Ranger
*****

Reged: 09/16/08

Loc: Tucson, Arizona
Re: Open Source Mount Control Suite new [Re: OzAndrewJ]
      #5993468 - 07/27/13 06:59 PM

OzAndrew,

Gday! Now, unfortunately I must do what I can to suck your mirth away.

The discussion has gone where you are talking, and the following points seem to be the hinge for this:

(A) ASCOM is windows based, and therefore linked to that platform's fortunes. That's not good.

(B) COM is a pain to get happy with itself, anyway.

(C) Raw computing power, modern communications architectures, and robotic elements are finally plentiful and cheap. It would be great to migrate to them.

So, these lead to a really big problem though (and everyone please throw darts if I'm screwing this up):

If we left PC/ASCOM, where should we move to? It needs to have legs, some realistic ability to maintain it, be really portable (the software portable), and well behaved.

If, in comparison, we try to migrate to Windows Mobile or some micro windows box, that approach offers very little since it just repackages the current problem. Maybe if it were set up so some sort of app on a regular PC would semi-automatically configure the mobile box. But, at best, this is a kludge and everyone knows it.

So, the current discussion is more along the lines of where to depart from. Orly's voyage with the arduino on the CGEM encoder is a lesson in starting with something better if you can, so it seems like time well spent.

I think many of us are imagining something sort of like an Astro Physics CP3 box on the mount with a bunch of sockets to talk to lots of stuff. Ideally it would link from the mount to the outside world wirelessly. Whether there might be several of these with specific missions (e.g. one for the mount, one for cameras), well, that would seem to come back to depending on what the device was capable of.

It does seem apparent we should expect to talk to tablets and smartphones and use them as the control display for much of the use. Those of us who have been using SkySafari have seen how that can be pretty neat, and some really basic hardware is all you need to enable the mount to display its position that way, though controlling it outright is more complex.

If you've got suggestions, we are definitely interested. Your LX80 progress definitely showed you have the skills.

-Rich

Quote:

Gday All

Been reading this thread with interest, but have noted with much mirth that all the current discussion is evolving about "platform" to run on.
In the last few days, it has been pointed out that "this is better than that", but in 3 years ( or maybe even 3 months :-) ) will these options just be another "com dongle"?
ie surely the initial discussion should be around the specifications an "open mount" controller needs to satisfy, ie like an ASCOM interface, but for the low level motor driving etc, not the planetarium to scope.
Until that is really done properly, the rest is just moot.
Thoughts??

Andrew Johansen Melbourne Australia




Edited by Starhawk (07/27/13 07:06 PM)


Post Extras: Print Post   Remind Me!   Notify Moderator  
OzAndrewJ
scholastic sledgehammer
*****

Reged: 11/30/10

Re: Open Source Mount Control Suite new [Re: Starhawk]
      #5993532 - 07/27/13 07:55 PM

Gday Rich

Quote:

Now, unfortunately I must do what I can to suck your mirth away.



Now now

Quote:

ASCOM is windows based, and therefore linked to that platform's fortunes. That's not good.




Fully agree, but i did say "like an ASCOM interface".
The full church of ASCOM is ( IMHO ) way too narrowly focussed on platform and functionality and is controlled by a limited no of people. I surmise this is done ( a bit like Apple ) to ensure it stays stable, but it does create limitations when wanting to do more than simple tasks, and have those tasks availabe as a "std operation" vs a driver specific addition.

I merely wished to point out that its concept of a common interface is a good one, and that interface should be designed first for a mount "controller", before the topic of what hardware or O/S to use gets into too much detail.

Andrew Johansen Melbourne Australia


Post Extras: Print Post   Remind Me!   Notify Moderator  
derangedhermit
Pooh-Bah


Reged: 10/07/09

Loc: USA
Re: Open Source Mount Control Suite new [Re: OzAndrewJ]
      #5993735 - 07/27/13 10:00 PM

The current ASCOM guy who responded, Chris, expressed willingness, in fact desire, to do a cross-platform version; they just don't know how. Given the use of Windows Common Object Model, I don't doubt they had problems if they really want one version. But the ASCOM client API, extended to include the multi-client add-on (I forget the name) has the functionality needed to offer to client apps; so the solution is to make the same functionality accessible in a cross-platform API.

My preference would be to avoid the device-specific drivers currently used in ASCOM on the PC, and put a standard protocol over the network, and require the devices to have any proprietary translation internal to the device / device control box.

I make the interfaces platform-independent. That means a network, and since remote access and control is interesting to many people, as are standards to developers, you start with an observatory LAN carrying TCP/IP over ethernet running http, ftp, ssh (for your command line interface ), etc as needed, and including WAN of course as an option, from the beginning.

My router and NAS, as almost all of them do, shows one way to confiure devices: you go to an address in your client browser using http, the built in web server throws up a number of config pages, and off you go. The near-real-time protocols need to work another way, but again, over TCP/IP.

My NAS box sends me e-mail when it has a software update it needs me to approve.

The other COM mentioned, RS-232, has nothing to do with Windows, and all to do with the devices to be controlled. Instead of the serial port, each device needs an ethernet port and a TCP/IP stack. The place to do that conversion is at the device, not at the back side of some computer.

WiFi can be a device option, perhaps, if the device is low bandwidth, high latency, and can connect to a standard hosted WiFi network (someone else's), so it can share the bandwidth over the air efficiently; this is an issue specifically at star parties. Southern Skies has run into this issue, and is addressing it (or has done so).

I think I could have a lot of fun developing tiny embedded controllers running a stripped-down version of Firefox OS. Then everything but the hardware is open source, web standards/HTML5, and it's appropriately designed for low-power embedded devices (low-end phones). Firefox OS is built on the Linux kernel and is open source, javascript-oriented (as opposed to Java on Android). I think there's a canned version for Raspberry Pi to play with.

Lee


Post Extras: Print Post   Remind Me!   Notify Moderator  
Alph
Carpal Tunnel


Reged: 11/23/06

Loc: Melmac
Re: Open Source Mount Control Suite new [Re: derangedhermit]
      #5993948 - 07/28/13 12:18 AM

Quote:

The current ASCOM guy who responded, Chris, expressed willingness, in fact desire, to do a cross-platform version; they just don't know how.




A dynamically or statically linked library written in plain old C would do the trick. Most programming/runtime platforms (except JavaScript within a browser, and VBScript) would be able to use such a library. VBScript would require a COM wrapper and JavaScript would require JSON over HTTP and a server process.


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Post Laureate
*****

Reged: 08/10/09

Loc: Singapore
Re: Open Source Mount Control Suite new [Re: Alph]
      #5994067 - 07/28/13 02:30 AM

I am also of the opinion that something like JSON over HTTP is a better choice than yet another XML encapsulation.

That said.. the current state of software development on non-Windows platforms is simply horrific. With ASCOM you have a small group of people controlling the direction, with UNIX you have nobody. A disparate group of people are all writing their own software, in many cases duplicating each other's work...

and much of the existing software only aims to ape the Windows functionality.. including its dependence on a high-resolution screen, mouse, etc. Not very useful for my desire for a headless controller.

What I want is something that was done at UC Berkeley 15+ years ago - http://astro.berkeley.edu/bait/public_html/baitman/protocol/telescope.html

That also said, it seems that the more telescope equipment vendors start exposing their devices via HTTP - such as the SBIG STF/STT - the better. QHY also already has one such model, the IC8300. This does add cost (you're basically putting a small BeagleBone-class processor on every device to handle the TCP/IP and HTTP stuff).


Post Extras: Print Post   Remind Me!   Notify Moderator  
cn register 5
scholastic sledgehammer


Reged: 12/26/12

Re: Open Source Mount Control Suite new [Re: orlyandico]
      #5994169 - 07/28/13 05:32 AM

The fundamental reason ASCOM uses COM is to for the scripting support, otherwise, as Alph says, dlls would do.

We would love to have more people actively involved in developing ASCOM but it's very difficult to get them. We can't even get people to test beta software.

We value stability very highly, maybe too highly, but being backward compatible is probably one of the strengths of ASCOM

Orlando's idea of a series of separate programs all communicating with the mount hardware has the challenge of getting multiple applications to connect to the same serial port at the same time. What you end up with is a single mount-specific process that handles all the mount control and exposes a common interface for other applications or scripts to use. That's all that ASCOM does.

It would be quite easy to map the existing ASCOM specifications to some text based protocol that could go over any connection that supports strings with terminators. My preference is for a CSV format, something like:
send "get,RightAscension\r"
receive "get,RightAscension,15.765\r"
This sort of thing can map directly onto the existing ASCOM protocol.

Each platform could write shims that converted the programming interface of their choice to and from this line protocol.

The discovery and selection of drivers over a network will be a challenge.

It's the state of development and support on non-windows platforms that puts me off getting involved in it.

Chris


Post Extras: Print Post   Remind Me!   Notify Moderator  
orlyandico
Post Laureate
*****

Reged: 08/10/09

Loc: Singapore
Re: Open Source Mount Control Suite new [Re: cn register 5]
      #5994590 - 07/28/13 11:49 AM

Chris - you're right. The fragmented nature of development on non-Windows platforms is a complete

Of course it is unreasonable to expect equipment vendors to support non-Intel platforms; the fact that they support Linux Intel is already a significant concession on their part.


Post Extras: Print Post   Remind Me!   Notify Moderator  
Alph
Carpal Tunnel


Reged: 11/23/06

Loc: Melmac
Re: Open Source Mount Control Suite new [Re: orlyandico]
      #5994663 - 07/28/13 12:33 PM

Quote:

Chris - you're right. The fragmented nature of development on non-Windows platforms is a complete

Of course it is unreasonable to expect equipment vendors to support non-Intel platforms; the fact that they support Linux Intel is already a significant concession on their part.




I am not sure what you mean by all that. You can use Eclipse on Windows, Linux, Android and other OSs. Millions of apps for Android and iOS and millions of web applications prove that it is possible to develop without Visual Studio, VB6, and/or .NET. I have a feeling that ASCOM guys are stuck in a time warp. They will have to bite the bullet and learn new techniques or ASCOM will be doomed if it is not already. Look at the latest browser statistics and trends. Microsoft is just one of the players and it is not the most important one anymore.

http://www.w3schools.com/browsers/browsers_stats.asp


Post Extras: Print Post   Remind Me!   Notify Moderator  
Pages: 1 | 2 | 3 | 4 | 5 | (show all)


Extra information
3 registered and 28 anonymous users are browsing this forum.

Moderator:  Dave M, richard7, bilgebay, iceblaze 

Print Thread

Forum Permissions
      You cannot start new topics
      You cannot reply to topics
      HTML is disabled
      UBBCode is enabled


Thread views: 4310

Jump to

CN Forums Home


Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics