Jan 28, 17 / Pis 00, 01 09:27 UTC

Re: How to produce microprocessors, electronic components in the space?  

I wasn't aware of gravity being a required part of the lithographic process. It might matter for the fabrication of the waifers. Additional concerns would be shielding required. I'm sure there's other things I must of missed, too.

Feb 5, 17 / Pis 08, 01 04:11 UTC

Yes, I believe that Shield Radiation is a field that need attention to produce electronics components in the space. Architecture and the Processes of fabrication is already explored on Earth, some changes maybe. Look this article about Juno's hardware and shielding approach https://www.nasa.gov/mission_pages/juno/news/juno20100712.html

The impact of radiation is not restricted to the electronic components, but also in software and human level. I am curious about the radiation solutions that SpaceX's team found for the Mars project.


Mod Edit: Made link clickable - Jason Rainbow 05 February 2017 @ 04:15 pm

  Last edited by:  Jason Rainbow (Global Admin, Global Mod, Asgardian)  on Feb 5, 17 / Pis 08, 01 16:15 UTC, Total number of edits: 1 time
Reason: Made link clickable

Feb 5, 17 / Pis 08, 01 16:19 UTC

It's restricted to the hardware level. It doesn't impact software, but the processor it's running on. Or the medium storing the software, or the data lines. A little spike of electon and the 0 becomes a 1. Assuming the peak isn't high enough to cause physical damages. Humans count as hardware IMHO as they are just machines - a complex stack of simple biological processes, as opposed to a complex stack of simple digital processes, but still machines nonethereless. Instead of flipping bits and damaging components with humans it's screwing their DNA and damaging components. But that's outside the scope of this topic.

Juno is a good example of shielding - but it's not that great. They're having to lift this shielding mass, we'd sensibly be using materials already up there. Where they've used 1CM thick titanium, I'd be using seven. Skinned around a few meters of NiFe. That should keep both Humans and more sensitive electronic equipment shielded. It's likely this can be done a lot smaller, I just prefer maximum overkill.

With regards to Mars, I'm actually unsure - of the variables not just SpaceX's plans regarding. Apparently there's no magnetic field so the solar radiation should be prominent even with it's distance - and it's proximity to Jupiter should be noticable too.

Feb 5, 17 / Pis 08, 01 17:45 UTC

The electronics components are impacted directly for the radiation. The software is a consequence of what on is running. For example, normally critical system runs with memory cache disabled on space, so the processor will run for much more time for the same work/task, and it needs to be on the software arch.. Softwar that are running under radiation must have error/check system thinking in the radiation effect. Of course, the radiation affects directly the physical component. However, the precautions are not restricted to the hardware level.

It is a consequence, It is not clever develop software that will running under radiation without a radiation precaution layer. The system cannot wait for a big failure to has a maintenance. It is not the focus of the topic, but if it would interest, can start checking the safety software develop for medical radiation therapy.

  Last edited by:  Felipe De Freitas Bertao (Asgardian)  on Feb 5, 17 / Pis 08, 01 17:58 UTC, Total number of edits: 1 time
Reason: erros

Feb 5, 17 / Pis 08, 01 19:32 UTC

I'm not aware of any systems - critical or otherwise - that run with memory cache disabled. In most applications this is required for the correct functioning of the hardware. Instead it's more common to find shielding. Especially around chips for or with memory. HD's and things may have the cache disabled but that's more about data integrity due to operations issues than radiation. I was more thinking of sheilding the entire environment the equipment operates in rather than the equipment itself.

It's not clever to develop software systems that are not fault tollerant, regardless of cause. And systems shouldn't wait for failure but instead take metrics which are observed for abnormalities and replaced well before failures actually occur.

Feb 5, 17 / Pis 08, 01 21:18 UTC

Yes, EyeR indeed, the software systems must has tolerance a faults, faults included the equipment physical damages (including from potential radiation damage).

"I was more thinking of shielding the entire environment the equipment operates in rather than the equipment itself." Yes, I thinking in this way too. My point about software impact was like I and you told about the faults, i.e in an arch. software level in some situations. The impact of the radiation on the Systems depends on hardware and software, but, of course, the physical impact need more attention.

About the cache memory disabled on processor, yes, there are embedded applications that are designed for this technique. I believe (not sure) in some aviation system too. Some citation about it , and a way to use cache also: http://ieeexplore.ieee.org/document/6847793/?reload=true&arnumber=6847793 https://nepp.nasa.gov/DocUploads/C288941E-C4DF-486A-9ADD317D00A26BC3/07-118%20Irom_JPL%20Guideline%20for%20Ground%20Rad%20test.pdf

Regards.

Feb 5, 17 / Pis 08, 01 23:03 UTC

Unlikely, Avionics typically favour realtime, and ulitmately demand reliability. Cache helps both these, commonly. And if it's unrequired, then that adds extra weight... These are not comonly built with COTS components, like http://www.xilinx.com/esp/mil_aero/collateral/presentations/radiation_effects.pdf, and likely feature sheilding suitable for operating the sensitive cache.

I find in incredibly unlikely that an on-chip cache be disabled due to radiation, if you can't trust the data stored in the cache(commonly also stored with checksum, to spot "errors") then you shouldn't be able to trust the caclulation of the processor itself, realistically, unless there's only one stray particle and it just happened to miss the rest of the processor and hit just that one cache "cell". Looking at the JPL link it suggests some SRAM caches are more sensitive than others, likely a function of the package than it's contents(Possibly a clue) - but if you actually pay attention not only are these the sort of processors you'd find in a museum(or my old PC stack, I have a few of those processors listed, and some older) and therefore irrelevant for study as we're unlikely to be deploying PIII or PowerPC chips but several clear for operation with cache's enabled.

You may find interesting further reading in: http://ams.aeroflex.com/pagesproduct/appnotes/ApNote_MRAM_SEFI_Work-Arounds.pdf http://ieeexplore.ieee.org/document/5722061/ http://klabs.org/mapld05/presento/138_czajkowski_bof-m.pdf http://radhome.gsfc.nasa.gov/radhome/papers/radecs05_SC.pdf

Feb 7, 17 / Pis 10, 01 07:18 UTC

Yes, it is a great subject, many techniques can be Applied. I already worked in Xilinx virtex QPro FPGA line in a lab for electronic design, they have great products. Yes some protection radiation techniques limit the hardware performance. As the link I posted, that turn off the memory cache in some situation. It doesn't mean that cannot never use the memory cache, the link, they point the technique as you can read is not fully, it is partial (the limitation can be done with the L2, using L1 with great reliability). However, all those ways impact in the performance or the manufacture limitation, and is hard to believe applying in a Lake Crest that has a huge dependence of HBM2, for example. Therefore, it is import to think in a shielding for the environment as was cited and NASA did.

Another point, that I was thinking about this subject. It was discussed about the insecurity of has private companies as our partner, but it will be done in some level, because we are here talking in how to produce technologies in space that it not entirely ours. What would be the level of these partnership?

Feb 7, 17 / Pis 10, 01 19:22 UTC

About the shielding the environment, I was reading some article today, and the big deal is to create a force field like the natural one that earth produces around itself. Of course, this idea faces a power consumption problem, but it is a way.

May I am running away from the topic's subject, but some links:

NASA http://www.sciencealert.com/nasa-plans-magnetic-force-fields-and-deep-sleep-chambers-for-mars-astronauts CERN http://cds.cern.ch/journal/CERNBulletin/2015/32/News%20Articles/2038160?ln=en

Regards.

Feb 7, 17 / Pis 10, 01 20:12 UTC

For me at least the intent was to develop technology that is ours. Using open source hardware projects wherever applicable to springboard development. Ultimately open sourcing anything we come up with.

That's not to say we can't work in concert with other companies, I don't really see what would be the advantage to them. Especially if they're already balls deep with various initatives like putting on-chip 3G leading to direct access of the core that controls all the other cores, and isn't made accessible to the operator of the hardware. I don't particularly see why they'd be eager to work with us to remove all the backdoors and implants into their products after they've put so much effort into them being there.

At least in the early phases, not everything would need producing "up there" - but the more of it that can be will reduce the required lifting mass for construction of additional equipments. With things like a seed factory, most of the bulk will be panelwork, bars/tubes, cogs/belts/chains etc and easily built from scrap floating nearby but to clone itself fully it'd need the entire electronics control package again. Things like motors also be required but it might be feasible to build some steppers etc. Reducing the lift to just the electronics in order to complete the facility is a massive improvement(and some could potentially be lifted with the initial seed, it'd be clever to has spares) but ultimately produced up there is the better option and long term essential. Can't be waiting really for things to lift from surface then throw out past Mars to the belt in order to pop off new facilities or repair/replace existing over there and harvest faster, ideally it'd require production on that end. Which would be easier if first developed in local orbit. If developed on the ground before that, the initial seed will go up powered by our tech.

I'd considered to investigate the automated tool crapple use to take apart old iphones with view of building something similar - there's likely a lot of usable components in the scrap floating about, that could reduce lift further.

Ofc, once we have orbital facilities for production and supermassive supplies of material it'd make sense to encorage others production to occur through our facilities - the more industrial processes removed from the surface of the Earth the better, IMHO. Plus it's a way to get rid of the exponentially increasing stash of harvested materials.

Feb 7, 17 / Pis 10, 01 20:27 UTC

There's no real details regarding the radiation shielding in that sciencealert link, It's possibly not anything noteworthy - the CERN one for superconductors shows more promise. Not entirely sure how this immense magnetic field isn't going to attract object, or be prevented from interfering with other systems.

I did occur some rumours involving "pyramid energy" suggesting the field emitted by a rotating pyramid is able to "wavebubble" waveform radiations - but lacking any particular experiments in this I'm unsure how much faith is wise to place in such. It is a muchly contested topic.

Feb 8, 17 / Pis 11, 01 00:19 UTC

Yes, I agree. Open Source project is the way that we (Asgardian) already can start to searching and Integrating into. For the future, when it will be necessary we already have some north. It can avoid the dependence from some malicious companies, and keep on of one concepts of Asgardia (technological one).

Yes, may, for now, it's hard to find something that companies can "get" from us. I believe in the early partnership, they will need to "buy" the Asgardia's concept like we did. Of course, probably, they will give more than receive. However, I believe it will unavoidable, because in Asgardian concept it's already open for interested companies. Maybe, it's a good point that the main "heads" on Asgardian can share if companies in this subject already approached us. I don't see many representatives giving their opinion about the vision that we are discussing daily on the Asgardian Topics (my fault, I'm starting on the Asgardia's Forum).

I will search, on next weekend, for open-source hardware projects that can be useful for us, not just about produce electronic, but also communication projects as small satellites that I believe can be a start for the Engineering/Technological division from Ministry of Science.

Let me know if already know a project. Regards.

  Last edited by:  Felipe De Freitas Bertao (Asgardian)  on Feb 8, 17 / Pis 11, 01 00:21 UTC, Total number of edits: 1 time
Reason: Typo

Feb 8, 17 / Pis 11, 01 04:12 UTC

With regards to give more than recieve - not overall, long term. Sensible firms recognise this, and they also recognise the current way of operating is simply unsustainable, especially with how it's about to scale. Look at things like the open source car for example. You'd really not expect any major players in the auto industry would consider such a venture, if you didn't know that things like self-drive and subscription based fleets deployed by companies like google and uber - ignoring things like uber's ex-NASA engineer claiming they'll likely have VTOL in less than three years, Zee Aero's reportedly working prototype, Honda's Fuzo system they've been sitting on whilst everyone else catches up and then it's worth developing a system for the cars to talk to each other - then car ownership is highly likely to be a niche market in a environement of already dwindling new car sales - sensible firms recognise the money will be in the parts, which is why they're working together to make sure they slot together like lego.

Considering the way the IT/PC industry taking a dive it's quite likely to follow a similar path but I'd not actually count on this.

Sure, in the beginning they may be teaching us tricks - but once you reach down and bring someone up to your level then when they exceed you, you can learn tricks.

As for open source projects, you'd really need to be more specific - It's really a case of: where to start. As this thread specifically topics Microprocessor then the likes of OpenPiton should definitely deserve a mention - open source multicore CPU (OpenSPARC). The Bridgman–Stockbarger technique isn't entirely a well kept secret, and I assume in the worst case senario a centrafuge can emulate gravity - not 100% on crystaline formation in microgravity(do know it's been tested). http://cuervo.eecs.berkeley.edu/Lava/simulators.htm and http://www.cerc.utexas.edu/utda/publications/tsm_peng_elias can possibly also be considered relevant. Everything from there forwards can be considered "standard robotics", I think. See a open source GPU on hackaday.

With regards to comsat specifically, I'd say start with the likes of the HAM radio buffs, they're already aligned in that general direction. I've seen posts for in here somewhere, you can find some. They'll help with signals and equipments/technologies I'd wager, whilst things like http://librecube.net/, http://mstl.atl.calpoly.edu/~bklofas/Presentations/SummerWorkshop2012/Fitzsimmons_Open_Source_Software.pdf, https://2014.spaceappschallenge.org/project/cogs-project---cubesat-open-source-ground-station-project/, https://github.com/else-sa/xU are likely to be considered of use in actually building a satallite.

  Updated  on Feb 8, 17 / Pis 11, 01 04:16 UTC, Total number of edits: 2 times
Reason: fixing link

Feb 19, 19 / Pis 22, 03 23:52 UTC

Could we lease a few pods in the Von Braun Space Station that the Gateway Foundation is building? We could control our own manufacturing processes and quality and 3D printing of our own robots to run on our own OS. Once we have enough supplies and product to start assembling the Kingdom Space Station we can cancel our lease. Then we can start building the Solar Shields to protect our Kingdom and the Earth. What I would like to see our attention focused on is the long term effects of micro gravity / LEO gravity on our human bodies. Scott Kelly has outlined many of his difficulties back on Earth in his book Endurance.