Alert -- Pluto Science Data/Image Return to be Slow

originally posted by: choos

originally posted by: tanka418

In 2005 I purchased a Fujifilm DSLR...it is 6MP. New Horizons appears to be 1MP...everything is bolted to the chassis and can not be positioned without turning the entire spacecraft...there's some engineering eh?

engineering also involves consideration of reliability.. bolting it to the chassis is the smart choice.

the more moving parts the more that can go wrong, im sure you know this.
and we are talking about moving parts that would have been exposed to extreme cold temperatures for many years.

^^^^^ THIS x 10

Remember what happened with Kepler's reaction wheels? And that was launched in 2009.

The same company which made Kepler's doomed reaction wheels is responsible for reaction wheels which have failed or were deemed too unreliable to be used, on NASA’s Far Ultraviolet Spectroscopic Explorer in 2001 and Japan’s Hayabusa mission in 2004 and 2005. NASA’s Thermosphere, Ionosphere, Mesosphere Energetics and Dynamics (TIMED) satellite experienced a single reaction wheel failure in 2007.

They recently laid of a bunch of employees (hmmm wonder why?)

Also don't be surprised if the Dawn spacecraft at Ceres has a reaction wheel failure. It has ones of the same type made by the same company which designed the other failed wheels.

Designing spacecraft for deep space is far different than designing a system like tanka's robotic telescope. There is always a trade off between reliability and capability. Deep space is a harsh environment. Harsher than anything on Earth. Sensitive electronics must withstand extremes in hot and cold and be protected against radiation.

The more moving parts the more likely the chance one will fail.

originally posted by: Jonjonj
Reaction wheels...but they only change the attitude of a vehicle right? I mean in no way can they actually change trajectory, they can only spin about the centre axis, as in left /right, up /down, tilt left/ right correct?

Plus, have you guys ever seen the shake down tests they do for space launches? They are crazy and extremely violent. That was, I am sure, a restriction on what could be utilized on the probe.

Exactly.

I'd like to invite Tanka to look at some engineering test and data pipeline discussion memos regarding a NASA spacecraft which will launch in 2017 to look for transiting planets around nearby stars.

Perhaps he can see just how involved all this stuff is. Things down to individual capacitors are scrutinized. The people in charge know what they are doing and have in some cases been involved with planning or testing various components for different spacecraft for 40 years!

As others have said Tanka, there is a big difference between your project and something like TESS. Most notably, if you need to fix something you can just go out and fix it. That's not something we can do at present with a spacecraft in even low earth orbit much less one 34 AU at Pluto.

And no, I did NOT listen to N-Sync Jadestar!!!

originally posted by: choos
a reply to: tanka418

hard to say if its a poor engineering decision..

how reliable are these wheels as compared with thrusters??

Answer: Sometimes.....Not Very.....

But they are looking to make them more reliable in the future.

originally posted by: choos
but anyway.. this is sidetracking the topic.. it doesnt matter how smart you are and how much more technologically advanced you would have made it.. its been made with what it has.

Indeed...

its download is slow given the huge distances and the hardware it has.. you are a computer guy you should know.

Yep that what they are saying...it is the vast distance that requires a connection on par with a 1200 baud modem. Except the distance has nothing to do with it. That distance affects ALL radio signals the same, regardless of frequency or, modulation rates and type.

It is kind of funny...they necessarily had to use a rather high frequency radio set...the size of antenna they could pack is very limited...so they necessarily were using microwave frequencies. yet they used audio frequency modulation...something they didn't need to do, and could have easily pushed the communications limit to at least 10MHz (DSL speeds).

This would have allowed them to send back vastly more data. IF they had used a moveable antenna they would not have to reposition the device for communication.

originally posted by: JadeStar
I'd like to invite Tanka to look at some engineering test and data pipeline discussion memos regarding a NASA spacecraft which will launch in 2017 to look for transiting planets around nearby stars.

You don't want me reading that kind of material...I might find:

2.0 Design Trade Decisions

We have decided to write as much of the POC MP system in Python as possible. This was based on the knowledge and experience of the team and the availability of suitable FLOSS packages.

-- snebulos.mit.edu...

Python is an interpreted language that was designed to be simple and easy for a human to read. This also mean it is not a very efficient language, and will be entirely unsuitable for Realtime operations. It may be okay for background computations, but, in general will cause its processor to be a bit slow.

There is also the comment about how Python was chosen because of the "knowledge and experience" of the team...That is truly sad.

It almost seems a bit odd that serious scientists won't take the time to learn about computers and software design and implementation. But insist that by using their inappropriate, though workable, solutions using inferior languages that their system is a "good system".

One of the things I have noticed is that almost the entire "space science" community wants to use modern technology, but end up crippling it with their uninformed, uneducated technological selections. I've noticed this in virtually every aspect.

Astronomical datasets where data is represented by an inappropriate data type, not really a big deal...if you don't mind a complex conversion process. The use of Interpreted language where something a wee be different would work better. Say a language that is at least compiled into "p" code...like C#...though I wouldn't use C# for realtime either...that is C+'s and assembly's domain. Course now, I do understand them not wanting to use assembly language...even I don't go there anymore.

I suppose I should cut them some slack, after all they are physicists, astronomers (sorry), etc. not computer scientists...but ya know even a computer scientist needs a decent job. Though someone of my ilk won't provide a software system that a physicist can read and understand the code...what we will provide is a fast, efficient, system that requires absolutely the least amount of storage for it to run...it will also be one of the most cryptic things you will ever read...(major software function on a single line)...I hate it, and do it all the time...because it is the best way to code...and if you "speak" the language it becomes clear as day...

My primary language has been "C" or one of its dialects since 1980. Today I use a lot of C#...it runs natively on Windows. Can run on linux, android, Ios, and of course Unix...though I'd almost bet that the "space science community" doesn't use any of those operating systems...at least not on a spacecraft...even though a couple of them are ideally suited for the application.

Anyway, I am well aware of the engineering involved...I've been doing it for over 40 years.


Reaction Wheels are a serious disappointment! They do seem to have a short life span...and if you compare that to a modern hard drive...it gets even worse. I've owned several hard drives that have operated continuously for more than 10 years...it's platters comfortable rotating at their 7200 rpm, 24/7/365...maybe NASA should get their reaction wheels from Seagate!

There is another interesting thing I've noticed. My approach to this observatory I'm building...I'm very sure if you were doing the design, it would appear as though a "Real Astronomer" was building a telescope. If someone were to "look into" your thought process it would reflect that you are an Astronomer. With me...I'm designing and building a remote / robotic data acquisition system...the fact that it has a telescope associated is almost a side note...which leaves me a bit handy capped when trying to describe an astronomical instrument...

originally posted by: tanka418

This would have allowed them to send back vastly more data. IF they had used a moveable antenna they would not have to reposition the device for communication.

thats an IF.. they used a fixed antenna to save weight and increase reliability.. you need to live with their choices regardless of what you want to believe..

so is the communication rate at about 1kbit/sec or not, given the hardware? if it is, wheres the conspiracy?? if its not what should it be?

originally posted by: tanka418

One of the things I have noticed is that almost the entire "space science" community wants to use modern technology, but end up crippling it with their uninformed, uneducated technological selections. I've noticed this in virtually every aspect.

sometime its down to a trust issue..

new technology has no track record.. proven technology has a good track record.. which is generally why it is chosen..
it seems you refuse to acknowledge this logic..

oh and new technology would have many hundreds/thousands lines of code.. pretty sure that would have its own headaches, as well as unseen bugs.. which is again a reliability issue..

a reply to: tanka418

Their choice of CPU was wholly inappropriate.

Sorry but their choice is based on reliability not speed.
The R3000 micro was introduced in 1988.
They chose one that is radiation hardened and proven for space.
You cannot use stock CPU's and put them in a metal box. It doesn't work that way.
Hence the space program uses VERY old computer systems.

originally posted by: samkent
a reply to: tanka418



Hence the space program uses VERY old computer systems.

unless its in a space located satelite, operating in the ghz!

a reply to: tanka418

With regards to you Fuji 6mp DSLR the ccd in use on new horizons would out perform your DSLR because of light levels out there.

It may only be one mega pixel but it's 13x13 mm in size larger physical pixel size leads to better quality even more so in low light.

The data size is smaller and they can also stich image's together if required.

Similar type ccd's have been used in many missions with great success.

Also your telescope set up looks interesting I am following your progress.

originally posted by: tanka418

originally posted by: JadeStar
I'd like to invite Tanka to look at some engineering test and data pipeline discussion memos regarding a NASA spacecraft which will launch in 2017 to look for transiting planets around nearby stars.

You don't want me reading that kind of material...I might find:

2.0 Design Trade Decisions

We have decided to write as much of the POC MP system in Python as possible. This was based on the knowledge and experience of the team and the availability of suitable FLOSS packages.

-- snebulos.mit.edu...

Python is an interpreted language that was designed to be simple and easy for a human to read. This also mean it is not a very efficient language, and will be entirely unsuitable for Realtime operations. It may be okay for background computations, but, in general will cause its processor to be a bit slow.

There is also the comment about how Python was chosen because of the "knowledge and experience" of the team...That is truly sad.

You do realize that Python is the the language most used in astronomy (and beyond that science in general) right? As since you know, the PI (Principal Investigator(s) and others who will be working closely on TESS come from the astronomy and astrophysics wouldn't it make sense to, idk us a language they are familiar with and use on a daily basis.

FYI most of the Hubble Space Telescope software is also written in python.

It almost seems a bit odd that serious scientists won't take the time to learn about computers and software design and implementation.

Because python just works. And the stuff you bring up is an apples to oranges type criticism. It's like saying to someone who composes EDM music, "you know, you really should use Linux as it is more secure and has advantages over Windows and OSX".

But insist that by using their inappropriate, though workable, solutions using inferior languages that their system is a "good system".

Because it -IS- a good system. These are research scientists, not computer geeks.

One of the things I have noticed is that almost the entire "space science" community wants to use modern technology, but end up crippling it with their uninformed, uneducated technological selections. I've noticed this in virtually every aspect.

Space exploration (and astrophysics) is not about being "bleeding edge". It's about efficiently delivering useful data to be analyzed and evaluated by as wide a community as possible. And that community spends a LOT of time working in python.

I suppose I should cut them some slack, after all they are physicists, astronomers (sorry), etc. not computer scientists...

Exactly.

originally posted by: JadeStar
You do realize that Python is the the language most used in astronomy (and beyond that science in general) right? As since you know, the PI (Principal Investigator(s) and others who will be working closely on TESS come from the astronomy and astrophysics wouldn't it make sense to, idk us a language they are familiar with and use on a daily basis.

FYI most of the Hubble Space Telescope software is also written in python.

Yes Jade...Python is currently the 6th most used programming language. Right behind Java, and every dialect of "C".

Again, Python was designed to be easily understood by Human, non-programmers, and little to no consideration was paid to the storage of the source code, which, in an interpreted language, such as Python, MUST be present at "run time"...so, all those 10's of thousands of lines of code, even the "blank" ones are there, in memory on every spacecraft that uses Python. On the other hand; "C" is a compiled language, meaning that those 10's of thousands of lines of code have been reduced to what we call "native code" meaning that it is now a binary block (ones and zeros) and only occupies a few kilobytes, as opposed to 10's of megabytes. This represents in some cases 100's of pounds of extra and unneeded weight.


You should not have mentioned your "PI"...which is about the same as a "principal architect", or "principal engineer" (I've been both in my career). You see, it isn't really necessary for the "Principal" to understand every small aspect of his project, after all, he is more a manager than a designer (the reason I don't like the job). Listening to you gives me the impression that, outside of the computer/serious technology areas where I have been isolated for so long, nobody has a clue how or what to do...sorry.

It seems that perhaps you should add a wee bit of computer science and project management to your curricula...seriously, it would probably be well worth the extra effort.

Because python just works. And the stuff you bring up is an apples to oranges type criticism. It's like saying to someone who composes EDM music, "you know, you really should use Linux as it is more secure and has advantages over Windows and OSX".

Yes, Python works, I didn't say it didn't...but, it is not the best choice for the application. And no, it is nothing like Apples and Oranges, not anything like Apples and PC's either...just simple real world software engineering.


Oh, by the way; did you know that Linux, OSX, IOS, and android are the same operating system...Linux, which is a small computer implementation of Unix...that got ole Linus a PhD.

Because it -IS- a good system. These are research scientists, not computer geeks.

Well, not really as "good" as you think, but, it des work, and that is a prime requisite. You all would do very well learning a wee bit of "Computer Geek" stuff, seriously, it will help!

Space exploration (and astrophysics) is not about being "bleeding edge". It's about efficiently delivering useful data to be analyzed and evaluated by as wide a community as possible. And that community spends a LOT of time working in python.

You're right Space exploration, etc. isn't the place for "bleeding edge" technology, neither is your "corner business", nor any other place we typically see technology...the reality is the "bleeding edge" isn't rolled out as soon as it is first built. I my career I've worked on "bleeding edge" projects...the PC was one of the early ones...look at what happened to that, and there was a project for General Motors in the mid 80's...today virtually all cars, world wide are built "in the dark". Both of these projects took decades to mature into what you take for granted.

Wen I criticized the selection of technology, I did it with a very deep knowledge of what was available at the time, and for New Horizons; they failed...plain and simple.

There were many better technological choices available; but, as it seems is the case with most sciences, they somehow "know better" that the scientists and engineers that originally developed the technology they want to use. So...no need to bother that processor architect from Intel, his choice wouldn't be very smart...no need to get an electrical engineer for their radio sets, he might want to use modern technology.

One other thing using real engineers would have done; no need to restart the system now that it is at its destination, and 'needs" to work now.

originally posted by: wmd_2008
a reply to: tanka418

With regards to you Fuji 6mp DSLR the ccd in use on new horizons would out perform your DSLR because of light levels out there.

It may only be one mega pixel but it's 13x13 mm in size larger physical pixel size leads to better quality even more so in low light.

The data size is smaller and they can also stich image's together if required.

Similar type ccd's have been used in many missions with great success.

Also your telescope set up looks interesting I am following your progress.

Well truth be known, the FUJI isn't really a very good device...it works, takes nice pictures, but, technologically...it is 2005, and doesn't work well for photometry development...my newer Galaxy S4 is much better...

I used that as an example of what was available around 2005 when New Horizons was being built. Plain and simple they used late 80's technology in a circa 2005 spacecraft, it is no wonder it underperforms so badly.

While it wasn't available in 2005, my robot will use a Kodak kai-10100. A CCD sensor that is capable of taking a 10.7MP image, or scaling back it resolution, and pixel size, through a technique called "binning" to 0.67MP and of course a pixel that is now 16 times larger than original.

Ya know, one of the problems with my project is that you can't really "see" what I'm working on, much of what I'm doing right now runs in the background...I'll have to see about putting together a simulation of some sort...then you can "see" the "Field of View" Objection recognition working, you can see the process of building a "FITS" file, and some of the processing available...

originally posted by: tanka418

Well truth be known, the FUJI isn't really a very good device...it works, takes nice pictures, but, technologically...it is 2005, and doesn't work well for photometry development...my newer Galaxy S4 is much better...

While it wasn't available in 2005, my robot will use a Kodak kai-10100. A CCD sensor that is capable of taking a 10.7MP image, or scaling back it resolution, and pixel size, through a technique called "binning" to 0.67MP and of course a pixel that is now 16 times larger than original.

An old 6mp sensor if aps-c size will still be better than a Galaxy S4 don't believe the phone camera hype!

Your Kodak sensor may be 10 mp BUT it's physical size is not much bigger than the new horizon 1mp sensor.

Imaging sensor Kodak KAI-10100-CXC with microlensing, color Imaging sensor size 17.86mm x 13.49mm

Larger physical pixel size on a sensor helps in low light situations.

Just look at the Sony A7s 12mp FULL FRAME mirrorless camera no other DSLR can touch it in low light!

originally posted by: wmd_2008
An old 6mp sensor if aps-c size will still be better than a Galaxy S4 don't believe the phone camera hype!

Well...I don't know what the "phone camera hype" is, so... But, I do know what kind of images are produced by my S4, and how useful they might be.

Your Kodak sensor may be 10 mp BUT it's physical size is not much bigger than the new horizon 1mp sensor.

Imaging sensor Kodak KAI-10100-CXC with microlensing, color Imaging sensor size 17.86mm x 13.49mm

Larger physical pixel size on a sensor helps in low light situations.

Just look at the Sony A7s 12mp FULL FRAME mirrorless camera no other DSLR can touch it in low light!

The reason I've stayed with the Kodak kai 10100 is because of it's variable pixel size, ad t's ability to scale in both resolution and sensitivity. This feature gives me the ability to do a rather fast data acquisition, thus it is suitable for both Astronomy and Astrophotography.

Sorry...can't use or control a DSLR...the device I've chosen is designed specifically for use with telescopes. And, thus, attaches to all the other "crap" that will be hanging off the instrument as well (off-axis guide, Filter Wheel, etc.). Integration of all the components is as important as the components themselves...