I've used days instead of weeks since the outset. In order to get a goodly number of sectors between the edge of the empire/federation and its capital, I use the arbitrary figure of one year travel time in one direction max. To me it feels roughly analogous to the Age of Sail in the 1700s-ish. Any farther than that and the span of control simply becomes too great.
If you look at Limzano Sector in the rule book and calculate travel times you'll find it takes months to years to cross it. As an example, Ulmo to Oshan, requiring only 8 jumps, takes 47 weeks to accomplish; Ulmo to Magha, the longest trip from South to North as it were, is 82 weeks, almost two years. And keep in mind that's doing nothing but travelling. I think logically what you'd find are clumps of polities strung along the jump routes. (Which is fun in its own way but it's not congruent with the intent of the game).
I have zero experience with SF games unless you count Gamma World, not out of any prejudice so much as habit. So I may be making suggestions that are already standard in interstellar games. But I agree with Greg that the Age of Sail is a good place to look for inspiration. I mean, Gene Roddenberry did. And you already have variable times of travel, much like ships dependent on currents, wind speed and wind direction, so it seems like a good fit. (I don't know if you already have an in-universe reason for the variability, but if you don't, I have thoughts.)
It also depends on what role travel plays in your game. If communications can outrun travel, you can know something is happening, but your response is delayed. If travel outruns communications (which is very Age of Sail), then you aren't even really able to call for help; letters travel faster than radio. (Of course, if that is the case, it raises the question of why you aren't sending messages by "carrier pigeon" FTL drones, which is something you need to consider.)
But the meta question really is, how long do you want it to take to send for help, or go get help? If travel and communication is slow enough, then the players are always on their own; there is no utility at that point for making it longer, other than aesthetics. If you want help to be a possibility, either because the players ask for it, or because you want the players to have a shot at responding in time, then you need to tailor travel and communication speeds to allow those scenarios.
So I guess the question is, how long in-universe does it take for a scenario to play out? And how much prep can be done by the PCs while in transit? Will there be delays to resupply or obtain special equipment before they have to leave? Is any of that even a part of your game? Keeping in mind that you can always manufacture reasons for delays in communications, but if you make communications or travel too slow, you remove the option of creating scenarios where there is time to respond. So perhaps err on the side of too fast, rather than too slow?
If travel itself is risky, then the system should probably be a bit more complex, because subjecting players to risk without choice isn't really all that fair, or interesting. So I think if travel itself is risky, there should be ways to manage that risk. Maybe there are routes that are fast but but riskier, or slow but safer.
Keep in mind that the visual evidence of where a star is just shows you where the star used to be years/decades/centuries ago -- even before you consider how the gravity of seen and unseen bodies between you and the star might have bent the light. Even if someone has been there and back, since everything is in motion, still travelling away from the big bang, all that person can tell you is where the star used to be, its speed and its bearing. So it wouldn't be a leap to suggest that going directly through deep space, instead of taking an indirect route by system-hopping, could carry a greater chance of getting lost.
System hopping between friendly polities is safer and means you have access to supplies and repairs, but it may be very indirect. Hopping to unfriendly systems, or unpopulated/patrolled systems, might save time, but it carries the risk of encounters with unfriendly navies or pirates. Deep space travel is very direct but carries a greater risk of getting lost, or travelling into uncharted hazards, plus who knows what is out there?
At the risk of stating the obvious why not just use the Traveller "standard" of essentially two weeks (seven days actually spent in jump space, plus two or three days transit time to the jump point in the departure system and two or three transit time from the jump point in the arrival system)? This "standard" is well known, easy to use, makes some physical sense, and it supports the kind of "imperial" SF that you're trying to present. This is one area where I don't think you really have to reinvent the wheel. As for comms, FTL communications completely destroys all SF credibility to me.
There are two reasons I don't simply want to use Traveller's system. The first is, I suppose, vanity. I don't want Thousand Suns to be distinct and Traveller's system, while great, is too closely associated with that game. The second is that I want the transit time to be variable. I'd some sense of pseudo-distance in which some star systems are "farther" from others via D-Drive than others and a consistent transit time like Traveller doesn't achieve that.
I tend to use historical wet navies as my goto. I mean my general SF spaceship rules are pretty much based on non-carrier WW2 vessels (space fighters do not exist, but space gunboats do). Similarly the stutterwarp drive from Traveller 2300 generally emulates the WW1 era restrictions on speed and refuelling [coal] range. An earlier game used Age of Exploration sailing ships as a measure (and actually emulated a lot of history of the period, which is mostly ignored in Australian schools).
That said my current "Traveller" game is pretty much based on massive alien starliners (so I can use the first edition jump route table for ship availability). Since a lot of interaction takes place on board these ships. So spending a week on board provides plenty of opportunity. Travel is not an intermission in the campaign, it is fundamental. [There are no private or even government ships.] If you are skipping over space travel in actual play then why not make it instantaneous?
If there is a heavy interstellar war focus than you want strategic room to maneuver. How soon do you want friendly forces to learn what happened. [A Starfire game with communications delivered in envelopes at the appropriate time is ... fun.] Or to listen in vain to the massacre of your friends and colleagues, helpless to interfere.
If you want an exploration game then you want a frontier that is out of contact. A lot of SF fans tend to consider ship types to be the size of the ship rather than the role of the ship. The TOS Enterprise is a cruiser because it is expected to "cruise" the frontier, independent of core worlds support, for five years. That's why it had a big secondary hull filled with supplies. [I hold one of the reasons that the Federation had a large territory that it claimed was that the Prime Directive forced it to ignore many planets that the other races didn't/couldn't. So, for example, the Klingon's couldn't pass a planet with an available labour [slave] force to exploit it.]
Distances are an interesting question. In most of my games a lot of space is filled with unusable stuff (economic limits). As H Beam Piper said in Space Viking, it's easier to go a little further for something better [paraphrased] so marginal environments tend to be ignored. [It's trivial to set up a hidden base somewhere - supplying it is where it is difficult and the main way of it being discovered.] This is also a problem I have with a lot of non-jump-point FTL. Most RPGs tend to think of FTL as intersystem and not insystem transit.
Then again I did use actual physics in my High Frontier "Traveller" game (although the drives were much more efficient than they should have been). So Hohmann orbits, transit windows, and a lot of low berths to save supplies. Getting places cost time and dedication; which was interesting since it was a large multiplayer game with players scattered all over the inner solar system (although one did leave the campaign by taking the multiyear trip to Farpoint Station in the inner Oort). Thanks to a false clue another player left. Fun.
I think from a purely "plays at the table" point of view, having travel between systems in days allows adventures that span multiple systems, while travel in weeks makes each system more of a singular adventure. By which I mean, having to move to other systems to track down clues, or make contacts, or pursue a fleeing enemy feels "tighter" if it takes a few days, not a few weeks.
I've used days instead of weeks since the outset. In order to get a goodly number of sectors between the edge of the empire/federation and its capital, I use the arbitrary figure of one year travel time in one direction max. To me it feels roughly analogous to the Age of Sail in the 1700s-ish. Any farther than that and the span of control simply becomes too great.
If you look at Limzano Sector in the rule book and calculate travel times you'll find it takes months to years to cross it. As an example, Ulmo to Oshan, requiring only 8 jumps, takes 47 weeks to accomplish; Ulmo to Magha, the longest trip from South to North as it were, is 82 weeks, almost two years. And keep in mind that's doing nothing but travelling. I think logically what you'd find are clumps of polities strung along the jump routes. (Which is fun in its own way but it's not congruent with the intent of the game).
Just my opinion.
I have zero experience with SF games unless you count Gamma World, not out of any prejudice so much as habit. So I may be making suggestions that are already standard in interstellar games. But I agree with Greg that the Age of Sail is a good place to look for inspiration. I mean, Gene Roddenberry did. And you already have variable times of travel, much like ships dependent on currents, wind speed and wind direction, so it seems like a good fit. (I don't know if you already have an in-universe reason for the variability, but if you don't, I have thoughts.)
It also depends on what role travel plays in your game. If communications can outrun travel, you can know something is happening, but your response is delayed. If travel outruns communications (which is very Age of Sail), then you aren't even really able to call for help; letters travel faster than radio. (Of course, if that is the case, it raises the question of why you aren't sending messages by "carrier pigeon" FTL drones, which is something you need to consider.)
But the meta question really is, how long do you want it to take to send for help, or go get help? If travel and communication is slow enough, then the players are always on their own; there is no utility at that point for making it longer, other than aesthetics. If you want help to be a possibility, either because the players ask for it, or because you want the players to have a shot at responding in time, then you need to tailor travel and communication speeds to allow those scenarios.
So I guess the question is, how long in-universe does it take for a scenario to play out? And how much prep can be done by the PCs while in transit? Will there be delays to resupply or obtain special equipment before they have to leave? Is any of that even a part of your game? Keeping in mind that you can always manufacture reasons for delays in communications, but if you make communications or travel too slow, you remove the option of creating scenarios where there is time to respond. So perhaps err on the side of too fast, rather than too slow?
If travel itself is risky, then the system should probably be a bit more complex, because subjecting players to risk without choice isn't really all that fair, or interesting. So I think if travel itself is risky, there should be ways to manage that risk. Maybe there are routes that are fast but but riskier, or slow but safer.
Keep in mind that the visual evidence of where a star is just shows you where the star used to be years/decades/centuries ago -- even before you consider how the gravity of seen and unseen bodies between you and the star might have bent the light. Even if someone has been there and back, since everything is in motion, still travelling away from the big bang, all that person can tell you is where the star used to be, its speed and its bearing. So it wouldn't be a leap to suggest that going directly through deep space, instead of taking an indirect route by system-hopping, could carry a greater chance of getting lost.
System hopping between friendly polities is safer and means you have access to supplies and repairs, but it may be very indirect. Hopping to unfriendly systems, or unpopulated/patrolled systems, might save time, but it carries the risk of encounters with unfriendly navies or pirates. Deep space travel is very direct but carries a greater risk of getting lost, or travelling into uncharted hazards, plus who knows what is out there?
Just my $0.02.
At the risk of stating the obvious why not just use the Traveller "standard" of essentially two weeks (seven days actually spent in jump space, plus two or three days transit time to the jump point in the departure system and two or three transit time from the jump point in the arrival system)? This "standard" is well known, easy to use, makes some physical sense, and it supports the kind of "imperial" SF that you're trying to present. This is one area where I don't think you really have to reinvent the wheel. As for comms, FTL communications completely destroys all SF credibility to me.
There are two reasons I don't simply want to use Traveller's system. The first is, I suppose, vanity. I don't want Thousand Suns to be distinct and Traveller's system, while great, is too closely associated with that game. The second is that I want the transit time to be variable. I'd some sense of pseudo-distance in which some star systems are "farther" from others via D-Drive than others and a consistent transit time like Traveller doesn't achieve that.
Roger that. So, maybe 1D6 days per jump period?
Probably 1D12, since that's the dice the game uses.
I tend to use historical wet navies as my goto. I mean my general SF spaceship rules are pretty much based on non-carrier WW2 vessels (space fighters do not exist, but space gunboats do). Similarly the stutterwarp drive from Traveller 2300 generally emulates the WW1 era restrictions on speed and refuelling [coal] range. An earlier game used Age of Exploration sailing ships as a measure (and actually emulated a lot of history of the period, which is mostly ignored in Australian schools).
That said my current "Traveller" game is pretty much based on massive alien starliners (so I can use the first edition jump route table for ship availability). Since a lot of interaction takes place on board these ships. So spending a week on board provides plenty of opportunity. Travel is not an intermission in the campaign, it is fundamental. [There are no private or even government ships.] If you are skipping over space travel in actual play then why not make it instantaneous?
If there is a heavy interstellar war focus than you want strategic room to maneuver. How soon do you want friendly forces to learn what happened. [A Starfire game with communications delivered in envelopes at the appropriate time is ... fun.] Or to listen in vain to the massacre of your friends and colleagues, helpless to interfere.
If you want an exploration game then you want a frontier that is out of contact. A lot of SF fans tend to consider ship types to be the size of the ship rather than the role of the ship. The TOS Enterprise is a cruiser because it is expected to "cruise" the frontier, independent of core worlds support, for five years. That's why it had a big secondary hull filled with supplies. [I hold one of the reasons that the Federation had a large territory that it claimed was that the Prime Directive forced it to ignore many planets that the other races didn't/couldn't. So, for example, the Klingon's couldn't pass a planet with an available labour [slave] force to exploit it.]
Distances are an interesting question. In most of my games a lot of space is filled with unusable stuff (economic limits). As H Beam Piper said in Space Viking, it's easier to go a little further for something better [paraphrased] so marginal environments tend to be ignored. [It's trivial to set up a hidden base somewhere - supplying it is where it is difficult and the main way of it being discovered.] This is also a problem I have with a lot of non-jump-point FTL. Most RPGs tend to think of FTL as intersystem and not insystem transit.
Then again I did use actual physics in my High Frontier "Traveller" game (although the drives were much more efficient than they should have been). So Hohmann orbits, transit windows, and a lot of low berths to save supplies. Getting places cost time and dedication; which was interesting since it was a large multiplayer game with players scattered all over the inner solar system (although one did leave the campaign by taking the multiyear trip to Farpoint Station in the inner Oort). Thanks to a false clue another player left. Fun.
Sorry for the length.
I think from a purely "plays at the table" point of view, having travel between systems in days allows adventures that span multiple systems, while travel in weeks makes each system more of a singular adventure. By which I mean, having to move to other systems to track down clues, or make contacts, or pursue a fleeing enemy feels "tighter" if it takes a few days, not a few weeks.
Journeys of weeks (or longer!) allow for more scope for shipboard events, if that's what a table is interested in.