This is incredible. Thank you all so so much! The first pass of the ISS over Phoenix is coming up shortly so hopefully we will hear something as well!
5 Likes
QARMAN is also received and we have some decoded frames from it:
https://network.satnogs.org/observations/1724779/
3 Likes
It looks like a QARMAN message too at:
https://network.satnogs.org/observations/1724748/
(Phoenix observation)
2 Likes
First of all: congratulations to all successful deployments!
From a first look at the data, ASU-Phoenix has an issue with the proper encoding of the AX.25 adress field: the source and destination callsigns are encoded in ASCII without shifting left. This might pretend some decoders to work!
2 Likes
QARMAN seems to do a proper encoding of the AX.25 header, indeed!
1 Like
A quick update:
Several frames from QARMAN and Phoenix. Nothing heard until now from the rest of the satellites.
Thank you for this note, Patrick. Our health beacon was encoded in ascii to make it easier for us to decode on the ground. Since we were entirely unversed in other ways the amateur community decodes packets, we didnāt consider this. I will make a note to record this as a lesson learned from a first cubesat project.
Sarah, it is not the ASCII encoded info field! It is the address fields in the ax.25 header! āYouā did not implement the protocol specs properly. So if a decoder tries to read the address fields it gets wrong results in the lowest bitā¦
To be more precise: source and destination are 7 bit ASCII, shifted left by one bit.
1 Like
That is a very weird spec. Not something I would expect. I guess that is why they look like gibberish when just received with AFSK and not decoded.
1 Like
This is a spec from the 80s. Nothing weirdā¦
1 Like
A quick update until now:
QARMAN and Phoenix still transmitting, difficult to say which one is in each observation though. Also their signal started to show doppler shift due to not accurate TLE.
Nothing heard from RadSat-U and Argus-2.
From AztechSat there are possibly some signals, however not demodulated, also there are the tweets bellow that say it also managed to receive something:
2 Likes
@fredy So how did you assign an āarbitraryā TLE data set to use with these birds ?
Iām puzzling over the logistics of the Network/Schedule system, not the orbital mechanics, 'cause using ISS to start with makes absolute sense of courseā¦ until they start to drift and their velocities impart a slightly different Doppler curve than whatās expected for ISS, (as weāre seeing now), at which point one might want to start ātuningā TLE states a bit to settle things until official elements are releasedā¦ (and maybe for some time after as well)
Iām just curious as to how a non-NORAD-supplied TLE could be assigned to the bird in this systemās DBā¦ like say, if someone were performing their own independent OD, and maybe even calculating ābetterā state estimates. 
(all for cubesat/amateur/public hardware people !! donāt nobody get bent out of shape or nothingā¦ not proposing anything that makes some orbit folks a little squeamishā¦)
but I would like to better understand the proper channels, actions, and approval processes if ācustomā TLEās are ever going to be allowedā¦ (even if itās just to feed into my own station currently in testing mode), and feel free to yank this particular conversation onto its own thread. I just asked here because of the current events for these birds demonstrates that it can be done when needed. Iām just curious as to how.
Cheers,
-Lucky
2 Likes
The ISS TLE is usually accurate enough until CSpOC releases TLEs for the new objects. For non-ISS deployments we usually generate an estimated TLE based on public information about the orbit. In some cases weāve had a few stations operate STRF and improve the estimated TLE with actual data and use that.
5 Likes
@cgbsat Thanks for the clarification on āsourceā of the TLE, but what about the process for applying the TLE to the DB ? Who does that & How ?
Like say, CSpOC releases a ābadā TLE for Phoenix or QARMAN ? either a poor fit to the orbit, or a complete mistag with something else.
In either case the āstaleā fit of the ISS TLE might be better, so how can someone ārequestā a particular TLE (or slightly old one) be used when scheduling Obs ?
I see from a cursory scan of the Gitlab repo that this concept has been brought up here for satnogs-db, issue #338.
ā¦and there appear to be other related concerns as well, particularly this one for satnogs-network, so I know itās not a clean-cut kind of deal.
Iāll be moving my line of questions on this stuff elsewhere so as to not pollute this thread, but I do have a question for the folks who are running Phoenix and QARMAN, so Iāll ask here:
@sarah_srogers and @amandine_on4eya Do either of your teams plan to track & study the evolution of the TLEās issued by CSpOC/NORAD/Celestrak over the respective lifetimes of your birds ?
Like for example, monitoring newly issued TLEās against the last set, to see if large discrepancies in orbit shape, inclination, etcā¦ occur.
i.e. Take a recent TLE, propagate it backwards in time, and see how things āline upā with an earlier TLE propagated forwards ?
Nothing will match up exactly of course (thatās just the nature of SGP4 after all), but you really start to notice the outliar TLEās when they get issued, and from that you also tend to notice the trends for when and how often that happens.
I know T.S. Kelso has done just this kind of sanity check to validate where and when a potential mistag (or plain bad OD fit) has screwed up planned observations for parties outside of SATNOGS, but heās just one guy, and canāt be everywhere at once, even though he does kind of fit the role of Keplerian Santa Claus for TLEās 
The reason I ask: everything that gets done, and assessed ādownstreamā, such as good/bad observations, health & condition of the bird, etcā¦ hinges on the quality of the TLE sets, and if youāre not keeping a running log on their quality and evolution over time, all your other derived assessments are just guesses with un-attributable errors.
Iāve heard through the grapevine of a few Cubesat projects that did this and subsequently identified a need, for later classes to undertake & implement, to start performing their own OD work as part of running a bird & collecting the science.
There is some obvious trace-ability in the SATNOGS Observations of which TLE was used when conducting the Observation in question, but I donāt think thereās an equivalent ranking system to post-qualify the TLE itself as āgood/bad/suspectā, and thatās something Iām going to bring up on my other thread (link to which will be added here, after I get around to creating it)
ohā¦and because I neglected to say this before:
Congrats to all on a successful lift, and subsequent deployment !! Best of Luck on your respective missions !
Cheers,
-Lucky
3 Likes
Here are some quick answers, I owe you a deeper one in the thread about orbit determination.
Currently except from the initial calculated, as @cgbsat described, TLE or preliminary ones that satellite teams provide, we use TLEs from celestrak, spacetrack, amsat and calpoly lists.
In DB are automatically updated if a valid NORAD ID has been assigned to the satellite, if not, like in cases of temporary NORAD IDs, then we manually added them.
In Network, and soon this functionality will be moved to DB, we do the same as DB except that we have the ability to set for each satellite with temporary NORAD ID to follow a valid NORAD ID and use its TLE.
Manual TLE currently are added by admins/moderators in DB and Network. There are some ideas and discussions about TLE use and generation but Iāll analyze them in my deeper post.
For now the only way is to ask in a new thread here in community or better in an issue in SatNOGS Operations repo where we track operation tasks. We have tools to check if the suggested TLEs are good to use and proceed on try them. Itās not perfect but is a workaround for now.
From the start of SatNOGS and due to several factors, we have NORAD ID as identifier for a satellite, however this is not true. Satellites are not always identified and sometimes misidentified, also there are satellites that have changed their NORAD IDs and these examples are against on using NORAD ID as identifier of a satellite. The solution is to generate a random identifier as we do for transmitters, and use this instead. More on pros and cons on this idea will be on my deeper post.
Thatās for now, sorry for the sort answer. I promise to come back with post with more details as most of the subjects you set for discussion have been discussed, face to face online and offline, and there is a direction we head to, but due to limited human resources is not as fast as we would like to be.
2 Likes
Hi, Lucky,
We were not planning to study the evolution of the TLE data. Though if it would help improve anything, I would be happy to look into it.
~Sarah
2 Likes
Hey @sarah_srogers !
Just curious.
No need to go adding to your list of To-Doās or tack on additional requirements on the Phoenix program for my sake. I already have a fairly good idea of the track-record for āgoodnessā and ābadnessā for TLEās associated with multiple objects being dropped off at or around the same place and time in orbit.
Where I think it could help you though, is in your own self-assessment of your system down the line. After the vehicle performs its mission, and yāall go back and look at all the "weird sht" that happened along its life, ex. failed to check in as expected, or realized we tracked the wrong object for a full 24hrs, etcā¦ youāll have some corroborating evidence for why that happened.
Thereās not much you can do about those occurrences, after the fact, but youāll at least understand why and with what regularity they occurred, and THAT helps in future design decisions for follow-onās to Phoenix, etcā¦
also doesnāt hurt to include in formal reports to the folks āfundingā the project, LoL 
(i.e. not our fault we lost some scienceā¦ TLEās were crpā¦ 
)
Yes, you can go looking for all those archived TLEās after the fact, but itās laborious, and youāre at the mercy of someone else having kept that archive clean and accurate.
Easier to keep a running log of each and every one yourself. Kind of like keeping all the receipts for the maintenance done on your car.
Furthermore, keeping track while the mission is unfolding would simply give you a little more insight, and possibly sooner, when Observations donāt happen as expected.
i.e. it improves your own SSA for your bird, while youāre trying to work it.
For example, say you schedule an observation on a GS that has a great track record, with all the bells & whistles, like properly polarized, high-gain patterns, with PreAmpās, bandpass filters, etcā¦ (i.e. the Bentley of ground stations) and you see nothing. If that happens to coincide at/around a TLE change, and you werenāt checking what those were telling the SGP4 propagator, youād be left with a few too many questions:
- Was it the TLE inaccuracy ?
- Did we glitch on orbit ?
- Was it the Ground Station ?
- Is this serious ?
But if you knew the TLE was suspect, you might consider scheduling an observation differently, requesting different mixes of GS equipment. A GS with an AzEl rotator is dandy when you know the pointing is right, but tight beam patterns and bad TLEās just means you were looking and listening intently at the wrong place in the sky. Maybe you catch the RF in a side-lobe, maybe you donātā¦ So you might want to next schedule an omni pass or two, just to confirm the timing in the ephemeris predictionsā¦ and so forth. It all helps in refining your scheduling plan when you get a chance to task this Network for help.
In the end, if youāre just set up to take {anyoneās} TLE (ex. CSpOC, amsat, ESA, some sketchy dude in a dark alley, etcā¦) and use it, thatās all you got, but itās as much of a pedigree of the vehicleās life as its science data history.
Cheers & Best of Luck with the mission !
-Lucky
QARMAN initial dashboard: https://dashboard.satnogs.org/d/98UPKYQWk/qarman?orgId=1&refresh=1m
@amandine_on4eya: please check the current conversion formulae and make suggestions how to improve this dashboard 
@sarah_srogers: ASU-Pheonix decoder coming soonā¦
3 Likes
Hi everyone!
Iām sorry we are a little late to the party, but my name is Sergio Bernabeu and Iām the Ground Operations Lead for Argus-02. Iāve been using SatNOGS for almost a year now but certainly still have ~many~ questions regarding spacecraft ownership within SatNOGS. Is there any edition/control I can have over Argus-02ās DB? We both broadcast telemetry every 45s and beacon ax.25 every 30s in between broadcasts. Is there a generic decoder for ax.25 available? And overall for GFSK9k6? As well, see attached an image of Argus-02, as I noticed there was none in its profile. At the moment we have not been able to establish contact with the spacecraft but are hoping to attain so tomorrow. We would use some of the past passes recorded in SatNOGS but itās been complicated to differentiate our data bursts with AzTechSat-1ās, who transmits on 437.30MHz, just 10kHz above usā¦
Please let me know any comments or direct me to any forum where I can learn more about the capabilities of SatNOGS for spacecraft owners!
Lastly, has anyone had trouble tracking any of these spacecraft with ISS TLEs so far?
Best,
Sergio
3 Likes