When another value than 0° is set as inclination in Ascent Guidance, then some correction occurs during ascent.
If it's set to 0°, the first try/attempt of correction occurs on circularization ...

@lamont-granquist Thank you for the reply.

Btw I have a "simple" question:
Are you actual builds only destined to be used when RO/RSS is installed or can they also be used when "only" RSS is installed and no RealFuels, too?
Like "Shuttle PEG Release #17 KSP 1.4.x"
I'm not quite sure about that ...

PEG builds can even be used in stock.

The problem is that PEG starts by assuming a flat earth / linearized gravity approximation which is broken by the fact that Kerbin is only 600km in diameter and not ~6000km and has a density some 10,000x that of Earth (so has a much less flat gravitational field), and typical launches require a coast phase, but PEG can't optimize the coast phase. So while they "work" you're stuck beating on an algorithm that wasn't quite designed for the use-case, and I tried it and couldn't really figure out any kinds of satisfactory heuristics on how to figure out how to launch a vessel other than trying and failing and trying again with a bit more or less loft and more or less coast, with no good idea how close to optimal I was actually getting...

It should work fine on any real-scale Earth-like launch, so RSS + RO/SMURFF or 10x Kerbol + RO/SMURFF. It doesn't know anything about RF (there is the ullage stuff in MJ but those are under utilities and should auto-disappear if RF is not installed, nothing to do with PEG ascents).

note though that I didn't ever merge PEG to core, because there is that small angle approximation and its very difficult to explain to game players when you hit it what its doing and how to avoid it -- you can even encounter that on launches in RSS.

If you see:

tgo 41.0s
phi 45.00

Then you're not going to space. The phi value MUST start counting down from 45 degrees before tgo is 40 seconds. What is additionally confusing is that once you're below 40 seconds then phi always counts down to zero. So the trick is to look for when phi starts counting down from 45, and if its > 40 seconds you're okay. If not, you won't wind up where you planned, in a very confusing way.

I actually ping'd the author of a better implementation which is what the shuttle uses, which gets rid of that nonsense, but haven't heard back from CSDL as to if they can release it publicly.

If you're talking about launches to inclinations below your launch latitude using the classic ascent algorithm, then this isn't a bug -- or at least not a bug that will ever get fixed. Those kinds of launches are very expensive in terms of dV and should mostly be avoided. If you need to do that you should launch to the inclination that matches your current latitude, and then execute a plane change at the next AN/DN. There is not a lot of cost savings in trying to launch directly to that orbit compared to doing a launch and a plane change. Most of those trajectories (that aren't too small to matter) will wind up looking like that anyway and will require an optimized coast to the equator to do a plane change.

@lamont-granquist
I always wondered why so many mods/software users out there in the world suffer from a lack of documentation ...

You know the deal:

• Be a "coder" - if earning money, hire a documenter.
• Create a raw wiki when releasing the first alpha, never finish it.
• When asked for a documentation by a noobish user, yell "I'm a coder, not a documenter"

;)

Yeah, in this case I keep trying to explain it to people on slack/discord/irc and failing, so if I can't explain it, I can't really document it. It boils down to "NASA doesn't have this problem because they have real trajectory optimizers, and you're much worse at rocket trajectories than you think you are -- get gud and stop asking the algorithm to do things it wasn't designed for..." which results in "but i want magicsauce button to press and not have to think...."

As a result though I am working on a real trajectory optimizer though, but it is not ready for prime time....