I've purposefully set aside mechanical issues. These seem to vary greatly from mount to mount, manufacturing date, luck of the draw or whatever!?

1. Make the 1024 sep encoder the default option for the Dec axis. This instantly improves Dec guiding by at least 0.25"

2. Update the G2 software to read and use the existing encoder index to synchronize PEC. It's possible some early mount's encoder wheel are not indexed, but then they could offer to sell new ones as a trivial upgrade. I think they under $15.?

That's it folks. These two things would put the GII into a new class where total rms guiding is at a minimum under 0.5" out of the box

Peter

How hard is it to upgrade the encoder? Is it an semi easy install then just change settings in the Gemini II. Will it work with the newer tucked in motors?

Has anyone here tried the TDM telescope drive master on their G11?

On Wed, Sep 22, 2021 at 02:44 PM, David Malanick wrote:

How hard is it to upgrade the encoder? Is it an semi easy install then just change settings in the Gemini II. Will it work with the newer tucked in motors?

David,

It's a very easy upgrade. The cover at the end of the motor unscrews, and then there? is another cover that is removed exposing the optical encoder disc and reader. You need both the disc and the reader. The reader is socketed. They are bought as a pair. As far as the tucket motor, if they are the same motor as the old mounts then yes. It's a settings change in G2.

"Has anyone here tried the TDM telescope drive master on their G11?"

I looked into this in detail. It's expensive. Using this you are seeking to eliminate the lower frequency errors, which in my case are 80s, 240s and 2600s. Really it ought to be possible to eliminate these via PEC. You might be interested in the discussion here?/g/Gemini-II/topics?PEMPro is what I'm looking at. I've really not had much luck with it reducing these error however. I've not completely given up. Brian is writing a user's guide, which I think will help a lot go through the steps correctly. But I'm not convinced PEC actually works. I feel it looses phase with the worm after you park the mount. It needs a lot of looking into and analysis and I'm not sure I want to lose imaging time, as my guiding is around 0.4" total rms. when I started out on the quest to improve my guiding I averaged around 1.35" or higher.? With the encoder upgrade my Dec is for the most part around 0.12" with good polar alignment and average seeing.?

I'd start with the encoder. Perhaps we could team up on the issue of getting PEMPro/G2 PEC to work. Scientific testing is needed here.

Peter

开云体育

On Thu, Sep 23, 2021 at 01:28 AM, David C. Partridge wrote:

Peter,

?

I don’t know where you got the idea that either the G1 or G2 can use the index marker on the motor encoder wheels.

?

Pin 2 of the encoders isn’t connected to anything and you need fair number of changes to exploit the index marker signals:

?

1. An extra wire from the motors to the controller for the index signal
2. Encoders that actually do output the index signal
3. Motor PCBs that routed this signal to a 7 pin connector
4. Changes to the PIC software that manages the motor to make use of the index signals
5. Changes to the protocol used for the SPI messages between the PIC and the ARM
6. Changes to the ARM code to make use of the index information.

?

The early servo motors (and maybe also the Maxon motors) use HEDS-9100 encoders which don’t have index marker detection/output on pin 2 (you need HEDS-9140 for that).

?

While the High-Torque motors now use US Digital encoders they probably don’t use the variant with the index marker output as that would cost more and there’s no way to use it.

?

David

David,

I found out from Michael shortly after I posted this idea that the signal is not brought into the controller. It's a shame and? I think a missed opportunity to future proof the design.

Thanks for the History details,

Peter

Peter - I know you "purposefully set aside mechanical issues" but I would like to see these addressed at some point.? Specifically the ...

• DEC axis backlash - mechanical rotational play (My GM811G DEC axis had the spring loaded worm installed)
• Mounting of the drive motors into the 'plastic' gearbox housing
• The grub screws coming loose on the spur gears
• Possibly a better motor to worm coupler
• Indexed flat surface on the worms for better grub screw contact surface

------------------------
Jim W
Phoenix, AZ. USA

Losmandy GM811G, NINA 1.11, ASI2600MC Pro, Sky-Watcher Scopes

David is correct there is an index marker output from the heads 9100/9410 and USDigital encoders. ?Even the old versions?

It’s pin “I”. ?I was playing with it to count motor revolution speed. ?Part of testing the Gemini electronics I do when repairing them

i have to ensure sidereal rate is in actual fact “sidereal rate” and using this pin to count period is an easy method. ?In reality I went s different method I used motor pinion chopper disc snd opto coupler. ?The motors jitter and I can program to time it over a number of revolutions to get a better picture of accuracy……This way I can test other motors etc…but, that’s just my method FWIW

David is spot on. ?The index is useful information especially wrt PEC. ? ? Well kinda. ? Would be better to have an encoder on the worm. ?

I guess the philosophy is to keep it simple and cheap as you can. ?Adding a worm index has benefits and costs and maintenance or user setup issues. My guess is why do all this if the system is already accurate enough in counting pulses throw in worm index will need done error recognition/correction?
--
Brendan

On Thu, Sep 23, 2021 at 12:37 PM, Brendan Smith wrote:

My guess is why do all this if the system is already accurate enough in counting pulses throw in worm index will need done error recognition/correction?

Brenden,

I think you meant to say accurate enough in the last century when the mount was used for visual.

I really think they ought to open source the Gemini II software.? I think if they do not, then they are compelling people to move to OnStep. Really why keep it proprietary? Imagine the possibilities! Rene are you listening?

As an engineer I'd love to use that index marker and totally rewrite PEC.? Make the whole thing a synchronous process with its own built--in FFT analysis function.?

Daydreaming again!

Peter

Hi!

I really think they ought to open source the Gemini II software.

As a fan of open source and long time user of Ekos/Indi and Stellarmate, I totally agree with this. I believe open sourcing Gemini software would open for huge development.

Best,

Magnus

My take...FWIW....and its not worth much

The Gemini is an industrially produced system, designed by experts, produced and backed up by experts who also back their product with precise mounts and spare parts.?? What's the cost of a G2 computer?....$1K USD.? Considering what you get........this is excellent, considering all the support.?? This is what you pay for.

The system needs good techs to fix it.? Your average backyarder cannot do 144pin QFN R&R.?? Imagine the Q&A on answering how to load the bootloader then firmware....How to fix torn up micro PCB traces.

Yes you can do better....but not in this price range of mounts.? Also not these days......anyone in electronics knows there is a huge shortage of ICs.?? I mean really bad!? It will last for the next 2 yrs minimum probably longer as most pasts lead times are now past 2 years estimate. ?? I do know Losmandy has purchased thousands of the required parts ...enough for years or production and support.??? They anticipated this.?? This is the hidden support.?? I had to spend quite an amount to buy up the last available ARM processors, H bridges, DSPicc33's, Displays....yada yada.

Send it open source.....your mount will be also outsourced from you know where with no support.?

ASCOM is the Open source for it.? Again...FWIW.......
--
Brendan

Brenden,

I have to be careful what I say here but I think it's true to say the controller sits in the hands of really one, perhaps two people. Look at a project like NINA. It's vibrant, stuff is happening. It's not in maintenance mod, it moves forward with fantastic new features that now put this software in a class of its own. You may be surprised by the engineering talent in this group. Regarding messing with electrics I agree, but people do make mods if there is compelling reason to do so. Look you can send your iPhone away to replace a cracked screen or do it yourself. Look at all the Onstep controllers being built. ?Further the software could be ported to a more powerful platform with more menorah, CPU power, etc. In fact, I would argue this is the way to go. There is no servo based open source controller available. Opening up Gemini would I think drive increased mount sales. One has to move with the times. Folks like me want to be engaged at all levels of the hobby, including the engineer side of things.?

Peter

On Fri, Sep 24, 2021 at 01:22 PM, Peter Boreland wrote:

There is no servo based open source controller available. Opening up Gemini would I think drive increased mount sales. One has to move with the times. Folks like me want to be engaged at all levels of the hobby, including the engineer side of things.?

Which is probably for a reason as controlling PID in a servo loop is an intensive thing especially for coordiante control. ? That said...I'm a fan of OS the ONSetep is decent well worth putting on a Losmandy mount.?? It was originally known as a tinkers mount.?? There are a growing number that want or would like OS but there is also a growing number who would prefer off the shelf factory.

I think the amount of Losmandy/Gemini units sold.

What's best, servo/stepper??? Steppers have come a long way in the past 10 years as with encoders indeed they may be the future.??

As for changes to the mount.?? Yes there can be improvements like a worm index.?? The change would necessitate incorporations that would not be backwards compatible with older mounts. ? Something the mfgr has been trying to keep.? Your 20 year old mount can pretty much be upgraded to the latest with factory parts, standards. ? That's something owners love.

Yes I think there can be improvements but they will need to introduce a new mount to probably do so.?? I'd buy that as well!
?
--
Brendan

Just my opinion, but open source works best when there are a good number of capable developers to move the project forward. Gemini is a hybrid of software and specific hardware that requires knowledge and skills, low level, real-time programming, along with some celestial mathematics and control theory. There are not as many individuals here, as one might imagine, that will be willing or capable of programming such a device. Much easier and much more accessible project, Gemini ASCOM driver, has been open sourced for more than 12 years. Guess how many others have jumped in to help move the software forward?

If you want an open source controller and software, you may want to start with a much more open hardware/software platform that is easier to program and that is more accessible to others. Gemini doesn't seem like a good candidate to me. Again, just my opinion.

Regards,

? ? ?-Paul

Paul ... Yup, given my 30+ years experience in control systems software and having been in the position of hiring such staff, I would +1 your comments.? <rant> The closer you go in software stack to the bare instrumentation and actuators, the less viable an all-hands-on-deck open-source melee is. To be successful, you are still going to end up with a few key domain experts to manage the architecture, algorithms, timing, blah blah. Code stability is often much more important than functional perfection.? </rant>

--
Edward

There are many reasons why nuclear power plants are not run using open source software. Many of the same reasons apply here.??

--

Chip Louie Chief Daydreamer Imagination Hardware?

? ?Astropheric Weather Forecast - South Pasadena, CA?

I've worked on areospace safety critical systems, and can tell you that there is nothing wrong with open source software being used. It just needs to go through the right process.?

Customers often press for open source so they can fix simple bugs themselves.

How sure are you that nuclear reactors don't use open source software?? Many proprietary products are built with gcc, which is open source public domain.? Some embedded vendors sell their own compilers but not many customers use it, they go with the tried and tested solution.? But hardware vendors don't provide proprietary compilers, they go with gcc because it provides the best support, they are not in the compiler business.? For instance I work with a Freescale toolchain, it comes with a gcc and Eclipse based IDE, all open source.? It is often hard to avoid.

The advantage of open source and free is that many eyes are scrutinizing the code, and a huge customer base to shake out most bugs.

I worked on a secure product where everything had to be coded from scratch.? I was wondering, what is the benefit of me writing a secure checksum algorithm from scratch with the risk of introducing bugs if there is elaborately tested open SSH source code around with an equivalent solution that is almost guaranteed bug free and secure?

About telescope controllers, IMHO it is best done by one main developer because it is simple enough to be designed and programmed by one person.? That may still be open source for instance OnStep so others can help out with simple tasks like BSPs and bug fixes.? Might not be a bad idea for Gemini.? I don't think anyone would run with the IP, there are already other codes around with the same scope.? Explore Scientific also open sourced their G11 controller, I think that aspect was received well.

开云体育