Fuel economy: Poodle vs Reliant (traveling to moons)

[JoeSchmuckatelli]

I've forgotten so much, and my google-fu is failing me.  Trying for my first Minimus trip of my first career since returning to the game.  At current tech, Poodle and Reliant look like my best Kerbin orbit to Minimus Orbit engines.  Poodle is 1.75 tons with 250vac Kn.  Reliant is 1.25 tons with 240vac Kn.  These are pretty close... so - from a fuel-economy standpoint is there a real difference between the two?  

Does it make sense to use the lower thrust, lower weight rocket as my final stage, or am I gaining something by going bigger?

EDIT: Detailed Question Below - based on current answers (thanks)!

Edited May 5, 2020 by JoeSchmuckatelli

[Bingleberry]

The vacuum isp on the poodle (350)  is better than the reliant (310)

[Guest]

For a regular Minmus trip I use a Terrier and/or Spark. A Poodle is great but you can shift an entire space station with that, it's way overkill if you just want to swing by Minmus.

[JoeSchmuckatelli]

1 hour ago, Bingleberry said:

The vacuum isp on the poodle (350)  is better than the reliant (310)

 

1 hour ago, Brikoleur said:

For a regular Minmus trip I use a Terrier and/or Spark. A Poodle is great but you can shift an entire space station with that, it's way overkill if you just want to swing by Minmus.

Hmmm - Okay; help me put all that together, please.  - Worried I have a fundamental misunderstanding of what's going on.

For the purpose of the question, I'm using the in-game vac numbers on the engines themselves; they are as I show in the OP question, as are the weights.  Do I calculate the TWR by dividing the thrust in kn by the weight of my total craft in tons?  Such that with an 11 ton craft, I get TWRs of ~19.61 for the Poodle and 19.59 for the Reliant? 

• Poodle: 11 ton craft + 1.75 engine = 12.75 / 250 vac thrust = 19.61 TWR
• Reliant: 11 ton craft + 1.25 engine = 12.25 / 240 vac thrust =  19.59 TWR

And if those numbers are correct;

• Is there any functional difference between the two?
• Does one get 'better fuel economy' because it's raw thrust number is better?

Finally (regarding using Terrier): I thought that using a more powerful engine offered better use of fuel because it got the craft 'up to speed faster'.  Using a Reliant or Poodle = a shorter burn time.  Doesn't using a Terrier with 60kn of vac thrust = ~ 4x the burn time?

Am I just totally lost on how this all works?

Thanks for helping a duffer out!

Edited May 5, 2020 by JoeSchmuckatelli

[AHHans]

3 minutes ago, JoeSchmuckatelli said:

• Is there any functional difference between the two?
• Does one get 'better fuel economy' because it's raw thrust number is better?

Yes, there is a significant difference which is that - for maneuvers in vacuum - the poodle gets significantly better fuel economy. But that has nothing to do with the amount of thrust that the engines generate, but with the specific impulse - the Isp values - of the two engines. Higher Isp means that the engine uses the fuel more efficiently. But the Isp value of an engine depends on the pressure of surrounding atmosphere. The Poodle as a high vacuum Isp (350s, one of the highest in the game) but its Isp at sea-level on Kerbin is a lousy 90s. So the Poodle is good for maneuvers in vacuum but really bad for lifting anything from Kerbin. The Reliant has a decent Isp of 265s at sea-level on Kerbin but only a mediocre vacuum Isp, making it a good booster engine to lift stuff from Kerbin, but not the first choice for vacuum maneuvers.

17 minutes ago, JoeSchmuckatelli said:

Finally (regarding using Terrier): I thought that using a more powerful engine offered better use of fuel because it got the craft 'up to speed faster'.  Using a Reliant or Poodle = a shorter burn time.  Doesn't using a Terrier with 60kn of vac thrust = ~ 4x the burn time?

Yes, with a Terrier you have less thrust and thus a longer burn time. But once you are in orbit it doesn't matter much how long a burn takes (and if it takes too long then you can usually split the maneuver into two burns(*)). What does matter is how high your Delta-V (the rocket-man's equivalent of how far your fuel will take you) is, and this depends on Isp.

You are probably confusing this with the situation of launching something from Kerbin - or any other planetoid - into orbit. There the TWR does matter! (Which becomes obvious when you think about what happens if the TWR is smaller than 1.)

P.S. (*) Yes, there is a limit when a low TWR leads to efficiency losses also of orbital maneuvers. But in KSP the practical limit is usually the patience of the player when burns become just way too long.

[Guest]

TWR only really matters for lifters and landers, and even there you'll only need about 1.4 fully fueled (from Kerbin). In space you can just burn longer, the efficiency loss is minimal; if you're chasing efficiency, find the lightest, highest Isp engine you have and use that. If you have a really low TWR you might have to split it into two burns. 

The real trade-off is that it's tedious to sit through a 10-minute burn, even with physics warp, and routinely doing multiple-burn ejections is even more tedious. A basic Terrier-powered Minmus mission shouldn't take much more than 2 min to eject from LKO so that's not a problem. 

To give you an idea, here's an extremely early-career Minmus lander/return module. I clearly didn't have the Spark yet because the Terrier is way overpowered for this job. I probably had a transfer stage or aa grunty lifter to get her part of the way to the Minmus transfer, but there's certainly enough dV here to do the Minmus insertion, landing, and transfer back to Kerbin atmosphere; the heat shield on the command module will handle the hot re-entry.

And here's one from a little bit later -- a basic space station with a Spark-powered lander and return module. As you can see I used a Poodle to get the thing there in the first place.

[jimmymcgoochie]

As others have said before, Reliant is a good launch engine as it can produce most of its total thrust at sea level, but it isn't particularly efficient in space plus it has no engine gimbal so has a lot less control. Poodle is big and fairly weak for its size, but it has very high efficiency in vacuum- no stock engine can match it (DLCs and NERV excluded). For moving large/heavy things around in space, the Poodle is probably your best choice short of taking the nuclear option and using the NERV.

However, for going to Minmus and back it's also completely overkill, a much smaller engine like a Terrier or even a Spark will provide more than enough thrust as Minmus has puny gravity and a typical lander isn't all that heavy. If you're using a Mk1 pod and a single crew, a Spark will probably be enough; for a larger Mk3 pod, use the Terrier. If you were going to the Mun then the Poodle might just be justifiable on a 2.5m heavy lander, and for going to Duna the Poodle is a great choice as while Duna's gravity is a lot more than the Mun, its atmosphere is so thin the Poodle loses almost no ISP from it.

[Empiro]

On 5/5/2020 at 11:13 AM, JoeSchmuckatelli said:

 

Hmmm - Okay; help me put all that together, please.  - Worried I have a fundamental misunderstanding of what's going on.

For the purpose of the question, I'm using the in-game vac numbers on the engines themselves; they are as I show in the OP question, as are the weights.  Do I calculate the TWR by dividing the thrust in kn by the weight of my total craft in tons?  Such that with an 11 ton craft, I get TWRs of ~19.61 for the Poodle and 19.59 for the Reliant? 

• Poodle: 11 ton craft + 1.75 engine = 12.75 / 250 vac thrust = 19.61 TWR
• Reliant: 11 ton craft + 1.25 engine = 12.25 / 240 vac thrust =  19.59 TWR

Here you're calculating acceleration by dividing the thrust by mass, so your crafts will accelerate at 19.61 and 19.59 m/s/s. The TWR is a Ratio between the acceleration due to the Thrust, divided by the acceleration due to gravity (the Weight). Your TWR changes depending on your reference body and how far you are from that body, though if omitted, it's typically relative to surface of Earth or Kerbin (both 9.8 m/s/s). To get TWR, you'd divide your answers by 9.8 (or do what I do, and just divide by 10) to get a TWR of around 2.

Like many folks have said, TWR matters when lifting off. Under 1.0 and you won't be able to lift off at all. In orbit, TWR matters very little unless it's extremely low (<0.1g).

Quote

 

And if those numbers are correct;

• Is there any functional difference between the two?
• Does one get 'better fuel economy' because it's raw thrust number is better?

Finally (regarding using Terrier): I thought that using a more powerful engine offered better use of fuel because it got the craft 'up to speed faster'.  Using a Reliant or Poodle = a shorter burn time.  Doesn't using a Terrier with 60kn of vac thrust = ~ 4x the burn time?

Am I just totally lost on how this all works?

Thanks for helping a duffer out!

 

I think you're on the right track. I think your misunderstanding is coming from thinking that more thrust = more efficient, without taking fuel consumption into consideration. The Terrier will take 4x longer to burn, but compared to the Reliant, it will be able to use the same fuel and burn for more than 4x longer. The ISP measures efficiency. The higher the better. If you're interested in the math, it's calculated by taking the Thrust and dividing by fuel consumption per second, and then dividing by 9.8/s/s (it is by convention to measure ISP in seconds). KSP gives fuel consumption in weird units (not kg/s) so it's good that it provides you the ISP information in seconds.

However, the ISP isn't the full story, because the mass of the engine matters too. The Terrier is often chosen because it both has good ISP and low mass. It is almost always chosen over the Reliant in space. Poodles have only slightly better ISP but far more mass, so its good for larger crafts.

When comparing the Poodle with the Reliant, it's a bit trickier, since the Reliant is lighter, but the ISP is lower. KSP shows you the delta-V in the VAB, but by default shows you stats for in-atmosphere, which is really bad for engines like the Poodle. In the VAB, try out the two different engines, making sure to set the Delta-V mode to vacuum. 

If you're interested in learning the general rule about engine selection and how to judge an engine, I can help with that. It helps you determine what engines you should and shouldn't be looking at when deciding what to use.

[JoeSchmuckatelli]

4 minutes ago, Empiro said:

 

If you're interested in learning the general rule about engine selection and how to judge an engine, I can help with that. It helps you determine what engines you should and shouldn't be looking at when deciding what to use.

Oh my lord, yes please! 

 

I've just basically cave-manned rocket selection by what didn't work and by sticking a wet finger in the air - and then sticking with what worked, until it no longer does.

[Cavscout74]

My usual engine choice for crewed pods for trips to either of the moons:

1.25m pod = Pug (from Restock, 1/2 thrust of Terrier).  The stock Spark works too, but since I have Restock installed anyway, the Pug is a 1.25m engine, so no size changes to deal with. 

1.875m pod = Terrier

2.5m pod = Cheetah

I usually end up with a ~0.5 - 0.75 TWR at full fuel, which is plenty for orbital operations.  Basically, I pick the vacuum engine one size down for crew pods and it works pretty well - it does require engine plates or creative use of aerodynamic fairings to cover the different sized engines at launch

Edited May 8, 2020 by Cavscout74

[JoeSchmuckatelli]

On 5/5/2020 at 3:20 PM, Brikoleur said:

.

While they look cool - is there any advantage or disadvantage to the RU looking lander pods - aside from (if I read correctly) no reaction wheels?

[Empiro]

This is just a general guideline on how engines in KSP compare, so you'll want to do some exploration and studying on your own (the KSP wiki is helpful).

The two most important stats of an engine are the TWR and ISP. The other stat to look at is mass, but it's much less useful.

ISP is important because as we've discussed above, it's the basic efficiency of an engine, taking both thrust and fuel consumption into account. Vacuum ISP is typically used since you spend most of your time in a vacuum (as soon as you're 20+km high, the ISP is very close to vacuum). If you compare all the engines, you'll notice that engines with really high ISP in vacuum tend to have very poor ISP in atmosphere (the lower ISP comes from having lower thrust, while still consuming the same amount of fuel).

KSP doesn't provide the TWR, but you calculate it by dividing the thrust by just the mass of the engine and nothing else (so for the Twin Boar, make sure to subtract away the mass of its 36-ton tank). An engine with great TWR on its own will also provide great TWR attached to a rocket. TWR is important when taking off and landing, and you don't want it too low in space (<0.1 in space is quite problematic. I generally have more than 0.5 to avoid long burns. If you compare all the engines, you'll notice that engines with really high TWR in vacuum tend to lose less ISP and thrust in atmosphere, so high-TWR engines have good TWR nearly everywhere. The VAB also has a TWR display that shows you atmospheric and vacuum TWR of your craft in different environments.

Note that raw thrust is not something I look at on its own. This is because TWR is a more useful stat that encompasses both thrust and mass/weight. If an engine has a great TWR but not enough thrust on its own, you can easily double or triple-up an engine to get the oomph you need.

Finally, mass matters, but ONLY because you can't have half an engine. TWR already takes into account mass. Engines like the Spark get used on small crafts because you can't have a quarter-Terrier with 1/4 of the thrust and 1/4 of the mass and same ISP.

All this provides us a basic way to compare engines and summarize the important points:

• You want engines with both high TWR and high ISP, but those don't exist -- engines with great TWR tend to have relatively lower vacuum ISP and vice versa.
• Low TWR engines have really low ISP in the atmosphere, which means their TWR becomes abysmal, and are worse in every way compared to high-TWR engines in atmosphere (as it has both lower ISP and TWR).
• More massive / late-tech engines tend to be better in both TWR and ISP and break away from the trend a bit. For example, the Poodle is heavier than the Terrier, but actually has better ISP and TWR in vacuum. The Aerospike is pretty good in every situation (though never the best). These engines are great if you have the tech and big enough craft. You'll find that for small crafts, these engines provide more thrust than necessary and add too much to the mass. (The engines in KSP are pretty balanced these days, but in the early Beta versions, some engines clearly were outside of these trends)

Putting together these stats, you can roughly group engines into three classes:

• Lifter engines with high-TWR and relatively poor vacuum efficiency. Mainsails, Reliants and SRBs fall into this category. They're engines you use off the launch pad in Kerbin and Eve, where there's high gravity and atmosphere.
• Vacuum engines with really great ISP but poor TWR. The Terrier is the most accessible example of this. Vacuum engines are great to use as soon as you're high up in the atmosphere. Even though the TWR is poor, they're still useful on landers because most places you land on have no atmosphere and low gravity compared to Kerbin (TWR of only 1 on Kerbin is 6 on the Mun). Duna's atmosphere is thin enough for most vacuum engines.
• Sustainer engines that are a middle ground. I find the Skipper a really useful example in this category. You can use them off the launchpad if you augment them with SRBs, or use them as a second stage that puts you into orbit. They're also really useful to keep using in Low Kerbin Orbit.

Hopefully this is helpful to you. Let me know if you have more questions.

Edited May 8, 2020 by Empiro

[RoninFrog]

1 hour ago, JoeSchmuckatelli said:

While they look cool - is there any advantage or disadvantage to the RU looking lander pods - aside from (if I read correctly) no reaction wheels?

Pros:

• Built in heatshield, fairly heat tolerant
• Built in decoupler
• Pods are all the same size for 1, 2, and 3 kerbal pods

Cons:

• Very draggy
• No reaction wheels
• Decoupler decouples at an angle
• They can roll if you land on a slope

Edited May 8, 2020 by RoninFrog

[archiebald]

29 minutes ago, RoninFrog said:

• They can roll if you land on a slope

....and roll and roll and roll....

[Superfluous J]

1 hour ago, RoninFrog said:

Cons:

• Very draggy

That's a big pro when you're coming home. They slow down fast and almost guarantee you'll come down with the pod at the top. But I won't use them until I can put them under a fairing at launch.

Another pro, they're light for how many Kerbals they seat, esp. if you ditch the ablator.

Edited May 8, 2020 by Superfluous J

[Guest]

6 hours ago, JoeSchmuckatelli said:

While they look cool - is there any advantage or disadvantage to the RU looking lander pods - aside from (if I read correctly) no reaction wheels?

They fit 2 or 3 crew.

[jimmymcgoochie]

Is nobody going to mention that those MH crew pods are named after Soviet WW2 tanks? Surely that's a pro?

However, despite their built-in ablator they can overheat on re-entry with the re-entry heating turned up as the heat shield only covers a small area on the bottom and the sides are still exposed. I cooked one on re-entry once, and since then I haven't really bothered to use them as I'd need to put another heat shield on the bottom and at that point I'd be better off with a Mk1-3 pod instead.

[Guest]

19 minutes ago, jimmymcgoochie said:

However, despite their built-in ablator they can overheat on re-entry with the re-entry heating turned up as the heat shield only covers a small area on the bottom and the sides are still exposed. I cooked one on re-entry once, and since then I haven't really bothered to use them as I'd need to put another heat shield on the bottom and at that point I'd be better off with a Mk1-3 pod instead.

That's never happened to me and I use them a lot. Returns from Mun or Minmus orbit no problem. Did you have something on it that unbalanced it perhaps?

[JoeSchmuckatelli]

14 hours ago, Empiro said:

This is just a general guideline on how engines in KSP compare...

… Great Stuff...

Hopefully this is helpful to you. Let me know if you have more questions.

Apologies for the long post

Okay - so let me see if I'm getting this (I just discovered the Dv tools button, so bear with me):

Presume for a moment that I'm averse (in a sense) to 'do it right' by sending a bunch of smaller payload missions - and I want to Kerbal it up by sending to the Mun a lander attached to a polar orbit resource scanning probe to maximize science and stuff to do in a single trip.  Also sending it with an engineer passenger and an OCTO as the brain / pilot - because the stranded Kerbal is a scientist, and together they can set up a bunch of experiments before blasting back off to rendezvous with the return pod. Lander is a Munar Excursion Module using a RoveMate as the brain (because it looks cool and fits better) which can lift two non-pilot kerbals back to space. 

My end stage has a lot of junk, is heavy and is long.  Lander is powered by a Terrier (best I can do at current tech).  It has

• A Kerbin TWR of 0.29 - which doesn't really matter because
  • it's an end stage,
  • by the time we use it, we're in space and all it has to do is
    • switch a Munar orbit from equilateral to polar...
    • and then land...
    • and then orbit and finally
    • link up with another orbiter
  •  -- and it's actual vacuum TWR is 1.16, which is more than enough, and honestly doesn't matter, because:
• despite 85 seconds of ISP (Sea Level) it has 345s of ISP in vacuum - which should give me plenty of puff time around the Mun -- and ISP > TWR in vacuum
• My OP thoughts on fuel economy were boogered because no matter what, the engine will, at full throttle, burn all it's fuel in 96 seconds (whether at sea level or in vacuum); what changes is where I use the Terrier matters more because that determines how much work it can do in the specific environment during that burn time... which in turn determines the efficiency.

To get that to the Mun, I will need an Orbital stage to help me stabilize my orbit around Kerbin and get me started on to my Mun Encounter... Thus, I need an engine that performs well at altitude and vacuum and can handle its own weight as well as that of my end stage.  So I should look at something like the Poodle or Bobcat here.

Poodle has

• a 15km ISP of 332s (with the above load and its own fuel can), and a vac ISP of 350s

• a 15km TWR of 0.98

  • Which is a problem, but might be mitigated by already having speed from a lower stage

• a vac TWR of 1.04 which is okay but not great

Bobcat has

• a 15km ISP of 309s and a vac ISP of 310s

• a 15km TWR of 1.63 -- much better than Poodle's

• a vac TWR of 1.64 - again better than Poodle's.

So if I was only focusing on TWR, the clear choice would be the Bobcat... but that would be the wrong choice because I ignore the better ISP of the Poodle both in the upper atmosphere and in orbit.  So Poodle is the better choice for a mid-stage.

Now, having decided on a mid-stage, I need to get the whole contraption up and through the atmosphere.  Because it's top-heavy, I need weight at the bottom and a big engine.  I will need a 'sustainer' engine as well as something to get me off the pad.  Choosing the Skipper (again, forced by tech) I discover it ain't gonna work.  At least, not by itself.

• With everything above and a bunch of its own fuel, it has a sea-level TWR of 0.79, and an ISP of 280s - which matters less b/c sea level needs greater TWR (or at least positive). 
• Thus to improve TWR, I need to add motors.  I can
  • Add a pair of Bobcats to either side of the Skipper, with additional fuel - that gives me a sea level TWR of 1.28 and an ISP of 296s.
  • Add a pair of Kickbacks - that gives a sea level TWR of 1.47 and an ISP of 216, or
  • Add a pair of Polluxs - which gives a sea level TWR of 1.66 and an ISP of 212
• So, despite the better sea-level ISP of the Bobcats, I need more TWR at the pad... and thus, I should choose the Pollux as the side stages of the sustainer

Thus - a Skipper sustainer with Pollux help should get me moving fast enough... we hope.

And now we've built a rocket!

 

Please feel free to point out and criticize all errors!  (and thanks!)

 

 

[Empiro]

You generally don't need to calculate exact ISP and TWR values at specific altitudes -- for example at anything 20km and above, I simply use the vacuum values (even if you start burning at 15km, you'll gain altitude very quickly). The ISP and other guidelines are there to help you whittle down the engines you should be looking at different stages, but the actual TWR of your rocket and delta-V determine the construction.

If both the Poodle and Bobcat have enough delta-V, then I would go with the Bobcat -- the better TWR means less burn time, and a more forgiving flight trajectory (I don't need to worry about falling back down). If I have to go with the Poodle to get enough delta-V, then the low TWR also helps inform me of my choices for the 1st stage -- I know I would need a rocket with better TWR and fly a slightly more vertical trajectory.

For your first stage, your choices of engines are all pretty good, and I think each design could be viable. My final choice will depend on my 2nd stage choice (if Poodle, then higher TWR). Real-life rockets like the Saturn V had an on-pad TWR of only 1.15 or so. In KSP, a 1.15 TWR rocket requires roughly 3800 m/s to make it into orbit, compared to a 1.6 TWR rocket that only requires 3500 m/s.

[JoeSchmuckatelli]

2 hours ago, Empiro said:

You generally don't need to calculate exact ISP and TWR values at specific altitudes -- for example at anything 20km and above, I simply use the vacuum values (even if you start burning at 15km, you'll gain altitude very quickly). The ISP and other guidelines are there to help you whittle down the engines you should be looking at different stages, but the actual TWR of your rocket and delta-V determine the construction.

If both the Poodle and Bobcat have enough delta-V, then I would go with the Bobcat -- the better TWR means less burn time, and a more forgiving flight trajectory (I don't need to worry about falling back down). If I have to go with the Poodle to get enough delta-V, then the low TWR also helps inform me of my choices for the 1st stage -- I know I would need a rocket with better TWR and fly a slightly more vertical trajectory.

For your first stage, your choices of engines are all pretty good, and I think each design could be viable. My final choice will depend on my 2nd stage choice (if Poodle, then higher TWR). Real-life rockets like the Saturn V had an on-pad TWR of only 1.15 or so. In KSP, a 1.15 TWR rocket requires roughly 3800 m/s to make it into orbit, compared to a 1.6 TWR rocket that only requires 3500 m/s.

Hmmm... then I am clearly missing something.

 

So - going back over what I wrote for the Bobcat / Poodle issue: is it because the vac ISP of both engines are relatively close (350 vs 310) but the TWR of 1.64 being so much better than the Poodle's 1.04 that drives that?

What if the difference were

• an engine with 1.64 TWR with 221 vac ISP vs
• an engine with  1.45 TWR with 350 vac ISP?

(I flipped the ratios)  -- What engine do you choose for the middle stage?

[Rhomphaia]

11 hours ago, JoeSchmuckatelli said:

Hmmm... then I am clearly missing something.

 

So - going back over what I wrote for the Bobcat / Poodle issue: is it because the vac ISP of both engines are relatively close (350 vs 310) but the TWR of 1.64 being so much better than the Poodle's 1.04 that drives that?

What if the difference were

• an engine with 1.64 TWR with 221 vac ISP vs
• an engine with  1.45 TWR with 350 vac ISP?

(I flipped the ratios)  -- What engine do you choose for the middle stage?

What you are missing is the rest of the rocket the engine is attached to.  How much of it is dead weight for the rocket to lift and how much is Propellant for the rocket to burn.

This, coupled with the ISP of the engine is what determines the DeltaV.

For an upper stage (handles final orbital insertion, and then can go in to perform orbital manuvers) you want the option that gives the most DeltaV, but for a pure  ascent stage you want the option that gives enough deltaV while still providing enough thrust. SInce you are using this engine to send you to the Mun it falls more into the upper stage category rather than what I would consider a middle stage from a lifter perspective

For a middle stage of a lifer this will depend on the performance of the first stage too.  but in KSP I rarely bother with middle stages on my lifters unless I am trying to make a historical replica. 

Roughly building based on the numbers you gave in the earlier post, either engine will be more than enough for a Mun trip, but the poodle would be my choice out of those two, especially if you plan to reuse this design to go further.  From the design of the first stage the upper stage doesn't have to do any actual lifting (the skipper/pollux combo makes orbit without it). It is cheaper than the Bobcat also.

 

Edited May 9, 2020 by Rhomphaia

[Empiro]

On 5/8/2020 at 4:03 PM, JoeSchmuckatelli said:

Hmmm... then I am clearly missing something.

 

So - going back over what I wrote for the Bobcat / Poodle issue: is it because the vac ISP of both engines are relatively close (350 vs 310) but the TWR of 1.64 being so much better than the Poodle's 1.04 that drives that?

What if the difference were

• an engine with 1.64 TWR with 221 vac ISP vs
• an engine with  1.45 TWR with 350 vac ISP?

(I flipped the ratios)  -- What engine do you choose for the middle stage?

Yes and no -- it comes down to whether both engines have enough delta-V to get you where you want to go. I don't know what your entire rocket looks like and I don't know where you're planning to go with it, so I can't tell if both engines will work or not. However, you're right that 310 vs 350 isn't *that* huge of a difference, so in many cases, both would work.

If the ISP difference were 221 vs 350, then it's highly likely that either: the 221 engine wouldn't have enough delta-V, OR the 350 engine has so much excess delta-V that I can remove fuel and get the same TWR.

A slightly more advanced tip is that efficiency becomes more and more important the later your stage is. The design of a stage not only affects that stage, but also the performance of every stage before it. It's critical that your upper stages are efficient and light, while low ISP solid rocket engines are totally OK for the 1st stage. For a 2nd stage, it matters some, but in my experience you have some flexibility in your selection.