Skip to content

Instantly share code, notes, and snippets.

@darconeous
Last active February 21, 2024 00:37
  • Star 97 You must be signed in to star a gist
  • Fork 10 You must be signed in to fork a gist
Star You must be signed in to star a gist
Save darconeous/8c7899c4d2f849b881d6c43be55066ee to your computer and use it in GitHub Desktop.
Hacking the Rectangular Starlink Dishy Cable
@jbowler
Copy link

jbowler commented Oct 6, 2023

Resistance between the two independent ground spikes is fairly low.

I'm very interested in knowing what that resistance is :-)

@darconeous
Copy link
Author

I'm very interested in knowing what that resistance is :-)

If I remember correctly, it was a few ohms. I don't remember what it was exactly, I just remember thinking "wow, that's lower than I expected". I have an outlet nearby with a ground that is connected to the house ground (not the ground that the antenna is connected to) so one day I got curious and pulled out the multimeter. But a low resistance makes sense, we have a lot of minerals in our clay-type soil that also has considerable moisture content. And I really buried that rod quite well.

I'm curious now, we just had a lot of rain. I can check again tomorrow.

@crdiaz324
Copy link

Is it possible to use 2 of these making 2 custom Y cable to use only the 2 powered pairs of each injector?

Yes but you have to know exactly what you are doing and it's pointless since the Tycon, as documented above, works just fine so long as correct wire pairs are swapped. If you don't want to do that the one I posted links to on Amazon doesn't require the wire swap. Seriously though, the cable has to be cut with any of these solutions (except the two most recently posted) so attach the RJ45 modular plus and it really is no more difficult to crimp those on with the correct wire swap than it is to crimp them on without a wire swap. (Both are error prone of course.)

A single APOE03G costs the same as a single Tycon PoE (USD15 from Amazon, USD10+postage direct from Tycon).

I've successfully used the techcharmer.com kit with my modified dishy without having to cut or swap any wires. The process involved connecting the provided cable from the kit to the plug in the back of the dishy, and subsequently to the POE's power output. Then I connected the POE to my router utilizing a standard ethernet cable I acquired from Amazon. This setup has proven to be incredibly reliable, running seamlessly for approximately 4 to 5 months. The kit was roughly $30. Hard to beat that IMO.

@JustOneGuyHere
Copy link

JustOneGuyHere commented Oct 20, 2023 via email

@jbowler
Copy link

jbowler commented Oct 20, 2023

@JustOneGuyHere; @crdiaz324 was commenting on his previous post from July:

https://gist.github.com/darconeous/8c7899c4d2f849b881d6c43be55066ee?permalink_comment_id=4645398#gistcomment-4645398

It's worth noting that TechCharmer now sell that Tycon lookalike PSU with what appears to be YAOSHENG adapter I posted a link to (two posts down from the above). Here: https://techcharmer.com/collections/our-products/products/starlink-dc-adapter-kit The Amazon link I posted no longer works but it's here:

https://www.amazon.com/YAOSHENG-Rectangular-Adapter-Connect-Injector/dp/B0BYJTHX4P

It's a nice move by TechCharmer; they've undercut the YAOSHENG very expensive PSU ($81) with an adequate and well proven power supply (assuming it is one of the Tycons), but they certainly seem to be selling the YAOSHENG adapter as part of it. Everyone benefits. The only downside to the TechCharmer solution is that there is no surge protection that I can see, but there is none on the official StarLink router (aka incredibly expensive PSU) so far as I know.

@WIMMPYIII
Copy link

Has anyone tried the new gen3 kickstand dishy.
Wondering if power layout is the same and if we are good to use the same POE PCBs that we use on the gen2?
i know the power requirement is higher and closer to gen1 / beta unit power demand.

@jbowler
Copy link

jbowler commented Dec 14, 2023

The cable seems to be an RJ45 and there is a separate PSU (which powers the router which in turn powers the antenna):

https://www.starlinkhardware.com/starlink-gen-3-dish-launched/

Here is the StarLink "specification sheet":

https://api.starlink.com/public-files/specification_sheet_standard.pdf

So that's apparently RJ45 all the way which means StarLink are likely to either be using modern standards or the Gen1 design. The new router offers Gen1 compatibility but still uses the Gen1 PSU, so the Gen3 dish is most likely using the bt standard. Minimally the RJ45 connection to the dish has to be protected; think what happens if (when) someone plugs their device RJ45 into the StarLink port or plugs the Gen3 dish into a regular PoE.

The dish goes up in stated wattage, "75-100W" from the spec sheet, despite losing the motors. Losing the motors may well help with the connector arcing problem and, anyway, RJ45s are much easier to replace :-)

There's a picture of the power brick here:

https://www.starlinkhardware.com/starlink-gen-3-router-review/

and, zooming in, it's only faceplated at 60W (30V DC 2A) and it is a UL listed LPS, so unless Elon has built in one of those perpetual motion devices that's the maximum the dish can take on average. I'm guessing that the "spec sheet" is hooey and that the "100W" just corresponds to a peak current for a fraction of a second (as in the Gen2 measurements). The pictures on that page also show the router face plate and it lists "Input 1: 30V 1A", so I'm guessing that is just the router box, not the pass through to the dish.

One really weird thing is the compatibility of the router with the Gen2 dish; scroll down the last link above and you will see a picture of the setup. The new router is connected to what looks like the standard Gen2 ethernet dongle, read the text and you will see it is the configuration we all know and hate; the Gen2 (dish) router becomes a separate power brick for the Gen2 dish. Just goes to show what a dogs breakfast the Gen2 dish cabling is.

Even though the Gen3 dish is using what appears to be an RJ45 connector (it has an extra rubber boot) that doesn't mean that it is using a standard PoE solution; I'm just suggesting that given the widespread available of very efficient electronics (FETs instead of diodes for protection etc) it only really makes sense to use one of the standards. The standard implementations can deliver way more power than the power brick.

If the dish has "diode" style protection then it might still work with the swapped connectors so long as it is fine with 48V (not the 30V from the PSU) and so long as it doesn't do active negotiation. That said I wouldn't try it; connecting a passive PoE to an arbitrary device not explicitly documented as requiring the given voltage and wiring strikes me as dangerous.

It may be that the Gen3 dish can be connected safely to an active PoE; it may not work but the active PoEs are designed not to deliver more than a few volts unless the connected device does the right resistor dance. Still, absolutely no guarantees; given the wacky Gen2 dish wiring I regard any such non-approved connection as an experiment.

@WIMMPYIII
Copy link

I find it very odd that the power is said to be higher but the power supply is 30v 2a 60w vs 48v 2a 96w. gen1 brick is even a little higher at 56v.
I am disappointed that they dropped voltage to 30v. this is a really ugly in-between number and not as capable extending distance as 48-56v hardware. It could be alot more finicky adjusting for voltage and amp drop.
30v would not be a bt or any typical POE standard.
We need a brave soul with a gen3 to plug in a cable with isolated exposed pairs to do some pick testing while the dish is running. figure out what pairs are doing what and at what voltage.

@Shangrila385
Copy link

Shangrila385 commented Dec 15, 2023 via email

@WIMMPYIII
Copy link

I want to see a brick and cable combo that will carry a signal and voltage a long distance. I develop condo highrise projects. I see the need to put antennas on the roof and drop cable down 20 floors to individual units.. gen 2 will not work without major cable mods..  Sent from Yahoo Mail for iPhone On Friday, December 15, 2023, 1:09 AM, WIMMPYIII @.> wrote: Re: @. commented on this gist. I find it very odd that the power is said to be higher but the power supply is 30v 2a 60w vs 48v 2a 96w. gen1 brick is even a little higher at 56v. I am disappointed that they dropped voltage to 30v. this is a really ugly in-between number and not as capable extending distance as 48-56v hardware. It could be alot more finicky adjusting for voltage and amp drop. 30v would not be a bt or any typical POE standard. We need a brave soul with a gen3 to plug in a cable with isolated exposed pairs to do some pick testing while the dish is running. figure out what pairs are doing what and at what voltage. — Reply to this email directly, view it on GitHub or unsubscribe. You are receiving this email because you commented on the thread. Triage notifications on the go with GitHub Mobile for iOS or Android.

On a building you could put an enclosure location with ac for the power brick and then 200ft or so down from the brick for a total of 330ft or less. This is a hard issue for a tower or tree installation though.

@jbowler
Copy link

jbowler commented Dec 15, 2023

I find it very odd that the power is said to be higher but the power supply is 30v 2a 60w vs 48v 2a 96w. gen1 brick is even a little higher at 56v.

As you say for PoE standards compliance the minimum voltage at the power injector is 50V, going up to 52V for 802.3bt "type 4" (the 100W variant). All I could find was the data for the power brick and the input of the the latest router (Gen 4?). The maximum voltage in the standards is still 57V, so if the new (Gen 3) dish is PoE 802.3 compliant there is a voltage doubler in the router; these are cheap to implement and very efficient in some of the charge pump implementations.

The PSU brick I found a picture of is apparently not the version distributed in the package with the Gen3 dish. The picture shows a two prong NA-only wall-wart (the prongs can't be changed) but the StarLink and other pages all show an AC cable in the full package and an IEC-320-C7 ("figure-8") socket on the PSU. The wall-wart is more than adequate for the new stand-alone router when used with Gen 1 and Gen2 because they both had separate power adapters for the dish itself.

It's even possible that the "real" PSU has multiple outlets and different voltages, but then I would expect the router to list multiple inputs.

We need a brave soul with a gen3 to plug in a cable with isolated exposed pairs to do some pick testing while the dish is running. figure out what pairs are doing what and at what voltage.

The Gen3 dishes still seem to be "invitation only", but the corresponding router is meant to be available for $199 on its own (possibly with a different PSU). Testing the router should not be a big problem, for one thing there won't be any voltage on an RJ45 breakout plugged in to the dish port after a fraction of a second if it's an 802.3 PoE.

@Shangrila385
Copy link

Shangrila385 commented Dec 15, 2023 via email

@jbowler
Copy link

jbowler commented Dec 15, 2023

So tell me what I need to know in plain English if you would. I think I would need rj45 cable, but what gauge and cable type is best?

You're at the limit of copper (twisted pair) Ethernet tech. You need to span a 200ft rise plus however many feet there are to the dish location and, at the other end, the router in someone's apartment (if I understand what you are trying to do). Twisted pair ethernet is limited to 100m (~328ft) max for all cable combinations. For StarLink you could drop from gigabit ethernet to 100Mbit but users can tell if they know how to check.

Your best bet is probably the Gen1 kit at present (since we don't know what Gen3 is yet...) Gen1 has a power injector and uses RJ45, I believe that allows the injector to be remote from the router. Alongside the 100m of Ethernet cable you run 100m of local code approved power cable (what this is varies enormously with region) correctly sized for the run length and local code requirements. In the US IRC it's something like no more than 10 or 20% power loss. This is discussed in this gist (the original stuff, right at the top!) In fact at 240V you can probably ignore this; 300ft of 14AWG power cable is about 5Ω (there and back), 100W at 240V is 1/2.4A, so that's under 1W loss in the cable if I got the arithmetic right.

With Gen3 if it really is possible just to use off-the-shelf RJ45 in place of the StarLink (150ft) cable then you need CAT5e (as used with Gen2) or CAT6 (including CAT6A). CAT6 normally uses 23AWG in the varieties of cable you would use but quadruple check that it really is (ideally measure the actual resistance!) Normally look for "pure" copper cable (not CCA - copper clad aluminum) but if you are buying a pre-fabricated cable (easiest, normally the best connections) check the actual resistance; send it back if it is too high.

Resistance figures vary, this site https://www.kbe-elektrotechnik.com/en/service/awg-table/ has 70.1Ω/km, so that 7Ω/100m. All 8 conductors are used in Gen2; we certainly don't know about Gen3 yet, so 100m has 7/4Ω there and back, 3.5Ω round-trip. So look at the original gist; the original assertion is that Gen2 becomes unstable at 2.5Ω. If that is correct you are limited to 70m for a CAT6 cable. Doing the same calcs for CAT5e at 24AWG, 89.2Ω/km, and I end up with a max run of 56m which is close to the longest, 150ft, official StarLink cable (CAT5e).

You could try a longer cable and check for stability of the system with the original router and, for Gen2, one or other of the ways of adapting the connectors at each end to RJ45 (minimizing the probably 24AWG CAT5e that is left.)

If you are actually building the structure then make sure there is building power on the roof, make sure the apartment (etc) owners/lessors pay for the shared power in their lease agreement, stick a suitable (SFP) router up there and run fibre to each floor, or maybe even each unit. Get some techie in to set up a commercial internet link, possibly a commercial StarLink one and do the maintenance (which will be required.) It is possible to do individual StarLink antennas with that arrangement but you put the routers in bypass mode on the roof and run the data down using VPN tech; techie required (a good one.)

@Shangrila385
Copy link

Shangrila385 commented Dec 15, 2023 via email

@jbowler
Copy link

jbowler commented Dec 15, 2023

On a general note WiFi-6 has a stated range of 45m; curiously close, well, the same as, the length of the longest StarLink Gen 3 cable. This is "line of sight" - i.e. nothing between the sender and receiver other than thin (we hope) air. WiFi-5 comes in at 25m or so, about 80ft, but I run WiFi (not WiFi 6 yet) between my own buildings which are about 100ft apart.

So a Gen3 with power and the standard (package) setup can, in fact, do a WiFi link at way more than the StarLink max speed to something 150ft away; WiFi-6 maxes out at 9.6Gbps. I manage that on my CAT5e Cu twisted pair network but only in a carefully tested setup; less than 100ft (maybe 50). In fact WiFi-5 at a theoretical 3.5Gbps is still way more than StarLink (residential) and my own copper network.

TP-Link (who I regard as reputable) have a range of "WiFi extenders" starting (based on Amazon prices) from $20 for a WiFi-5 one. So far as I can see it should be possible to use one of these to set up what is effectively a point-to-point solution; no need for copper or fibre. I've looked at p2p in the past but they have been somewhat more expensive ($100's rather than $10's) and often lower bandwidth (10Mbps).

The solution does require power at both ends but I find that fairly easy (at low power) even without solar.

@SailorBruce
Copy link

I have a CradlePoint router w/ a PoE++ (802.3bt) port that will push up to 60W. I don't live in a snowy area, so I plan to keep the snow melt off. My thought was that I could swap the pinouts to make the antenna work directly from the router as a PSE (power source equipment) and eliminate a whole bunch of custom cables, and a dedicated PoE injector. However, it doesn't seem to be working....

Is this possible? If so, how should I build the termination that comes out of the antenna and connects to the PoE router port?

@jbowler
Copy link

jbowler commented Feb 19, 2024

Is this possible?

No. The StarLink antenna doesn't support any of the PoE protocols so a PSE which uses them simply won't send out any power. The latest 802.3 standards have plenty of power but it would require a 48V 802.3bt PoE tap (a splitter) to get the 48V out , e.g:

https://www.amazon.com/StarTech-com-Industrial-Gigabit-PoE-Splitter/dp/B08KYHX1QS

Then it would be necessary to take the 48V output and wire it back to a passive injector... In other words pretty much twice as much work (you just save power cabling from the 802.3bt PSE to the PD, the antenna.)

@jbowler
Copy link

jbowler commented Feb 19, 2024

Off topic but still relevant to understanding the Gen 2 antenna here is a good, clear, description of how the 802.3 PoE standards work:

https://www.skyworksinc.com/-/media/SkyWorks/SL/documents/public/white-papers/understanding-the-ieee-8023bt-poe-standard.pdf

The circuit breakdown of the Gen2 router suggests to me that it lacks the electronics to do that, although it does seem to have the ability to measure current draw so there might be some communication capability there. The mixed up wiring of the ethernet pairs, however, prevents 802.3bt working; IRC the antenna looks like a shorted-out PD.

@WIMMPYIII
Copy link

I have a CradlePoint router w/ a PoE++ (802.3bt) port that will push up to 60W. I don't live in a snowy area, so I plan to keep the snow melt off. My thought was that I could swap the pinouts to make the antenna work directly from the router as a PSE (power source equipment) and eliminate a whole bunch of custom cables, and a dedicated PoE injector. However, it doesn't seem to be working....

Is this possible? If so, how should I build the termination that comes out of the antenna and connects to the PoE router port?

60w isnt going to cut it.

@SailorBruce
Copy link

SailorBruce commented Feb 19, 2024 via email

@WIMMPYIII
Copy link

WIMMPYIII commented Feb 20, 2024

Thanks for the info. I have a much better understanding of the issue now that I've failed a bit and read the SkyWorks doc. I now have working power and signal to the rectangular (gen2) dish. It makes sense that you have to flip the conductors to put power on the wires that Dishy wants, then swap the wires back so that the signal is back to where it is supposed to be. I'm still having problems though. About every 40-50 seconds, the PoE injector reboots. I get the same behavior with different cables, and when I plug dishy into my laptop or the router. This is from the router:
@.***: /]$ tcpdump -n -i starlinkvlan tcpdump: verbose output suppressed, use -v[v]... for full protocol decode listening on starlinkvlan, link-type EN10MB (Ethernet), snapshot length
262144 bytes
13:19:32.638287 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request
from 3a:30:44:66:d7:83, length 303
13:19:36.638274 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request
from 3a:30:44:66:d7:83, length 303
13:19:38.181021 IP6 fe80::2412:acff:fe1a:8001 > ff02::2: ICMP6, router
solicitation, length 16
13:19:39.698323 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request
from 3a:30:44:66:d7:83, length 303
13:19:42.788281 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request
from 3a:30:44:66:d7:83, length 303
13:19:46.244886 IP6 fe80::2412:acff:fe1a:8001 > ff02::2: ICMP6, router
solicitation, length 16
13:19:48.458300 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request
from 3a:30:44:66:d7:83, length 303
13:19:51.538347 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request
from 3a:30:44:66:d7:83, length 303
13:19:51.540257 ARP, Request who-has 192.168.100.100 tell 192.168.100.1,
length 46
13:19:53.556754 IP 192.168.100.1.67 > 255.255.255.255.68: BOOTP/DHCP,
Reply, length 300
13:19:53.578276 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request
from 3a:30:44:66:d7:83, length 315
13:19:53.588961 IP 192.168.100.1.67 > 255.255.255.255.68: BOOTP/DHCP,
Reply, length 300
If you wait 40 seconds, it happens again. When I am plugged into the laptop, I can get 18 (sometimes 17, but never 20!) ping responses before the interface goes down and the laptop reports no path to host. It is as if Dishy is expecting something from the router and when it doesn't get it it reboots.... To make sure I hadn't broken something, I plugged everything back into the Starlink router. I used a different (and tested) ethernet cable, but the same RJ45 keystone connector that is wired into Dishy. Everything worked swimmingly, so I didn't break anything when I cut the back off of Dishy for the mount (www.starmount.com, no affiliation). That was a relief, but still not helpful. The only thing that I haven't eliminated from the problem is the PoE injector. However, it is brand new and works fine when I power my Raspberry Pi and move lots of data through it. Does anyone have documentation on the exact meaning of the LED status lights on the Yaosheng Model: YSNEAPL12001A ? Is there an extra step that I have missed in configuring the network that Dishy creates?

How well have you read over this thread. 60w is not enough to run a SL gen2. You realistically need near 100w power supply run stable. Gen2 peaks at well over 60w. The stock power supply is 96w and is pushed to its absolute limit at 150ft.

@SailorBruce
Copy link

SailorBruce commented Feb 20, 2024 via email

@Missoulajeff
Copy link

Missoulajeff commented Feb 20, 2024 via email

@jbowler
Copy link

jbowler commented Feb 21, 2024

The PoE injector will support up to 120W. Once I have it more juice from the bench power supply, it all got stable. The 5A current limit was starving Dishy.

Ah! Yes; this is a reported problem with boost converters. It seems from my measurements that there is a sudden very brief amp draw at pretty much exactly the time you reported the PoE/dish reboot. Here's the data:

https://gist.github.com/darconeous/8c7899c4d2f849b881d6c43be55066ee?permalink_comment_id=4466245#gistcomment-4466245

So there is a large inrush (maybe) just after 40s; around 3.4A (see the comments at the end of my post). Buck converters seem to handle this ok while some (at least) boost converters do not. The Yaosheng is apparently designed for a battery supply (12V or 24V) and that would give it a large surge capability. One contributor above uses a super-capacitor on the output side but I don't like that because, while it would handle an inrush, it could fry the connectors or the ethernet magnetics; better to put it on the supply side of the boost converter.

Here's an earlier, and more conservative, measurement of the power requirements, the amperage graphs above are likely to be more accurate:

https://gist.github.com/darconeous/8c7899c4d2f849b881d6c43be55066ee?permalink_comment_id=4463407#gistcomment-4463407

I couldn't find Yaosheng documentation either. It seems likely that they have done the measurements properly (i.e. using better equipment than I have) to get robust designs but that doesn't mean they designed the 120W version to work off anything other than a battery...

On average the current draw at 48V without snow-melt or pre-heat is under 1A. During boot it goes up to maybe 1.4A but this corresponds to the effects (on the average) of the large inrush at 45s (3.4A). Adding snow melt (etc) seems to add maybe 0.25A. Solar systems with a 48V battery should be fine but they need to provide at least 1A continuous (over 24 hours) to keep the Dish up and running.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment