Elegant users forum

Dear Michael,

I have a question regarding the longitudinal behavior I observe in ELEGANT simulations.

I compared two versions of the same lattice:

Case 1: Four dipoles active (first dipole with a 5° bend) → R56 ≈ 1.9×10⁻⁶, bunch length ≈ 149 µm.

Case 2: All dipole bending angles set to 0°, but the septum and kicker remain active → R56 ≈ 9.8×10⁻³, bunch length ≈ 96 µm.

The energy spread (Δp/p = 10⁻³) is identical in both cases, so I would expect a longer bunch for the larger R56 (three orders of magnitude higher). However, the opposite occurs. Could you please help me understand why this happens?

In addition, I noticed that even when varying the beam energy from 90 MeV up to 1 GeV, the final bunch length remains almost unchanged. It seems that ELEGANT may be assuming β ≈ 1 (particle velocity equal to the speed of light).

Is this β≈1 assumption still valid at 90 MeV, or could it introduce a noticeable error in the longitudinal dynamics, such as in the bunch length or R56 effects?

Thank you very much in advance for your time and clarification.

Best regards,
Samira Fatehi

Samira,

It is hard to comment without having the input particle distribution. If it has a time-momentum correlation of the right sign, then having larger R56 could easily result in a shorter bunch.

Also, be careful to look at St rather than Ss. As explained here, Ss is the spread in distance traveled, while St is the spread in arrival time.

Elegant does not assume beta=1. However, even for 90 MeV, beta=0.99998..., which is close enough to 1 that you shouldn't see much effect.

--Michael

Dear Michael,

Thank you very much for your reply, and my apologies for the very late reply. I had trouble accessing the forum several times.

I should mention that no initial 𝑧−𝛿 correlation was defined in my input distribution, please see the attached file.

Also, I repeated the simulations with sigma_dp=0, and the results were unchanged.
This confirms that the bunch-length evolution is not driven by energy-dependent effects. ​

Instead, it seems that geometric path-length differences arising from transverse betatron amplitudes in the deflecting elements play the dominant role.

Thanks and best regards,
Samira

Samira,

The bunch lengthening results from second order effects in the path length. E.g., in a drift space of length L, one has T522=T544=L^2/2.

You can see this by plotting the T5?? elements from your matrix.sdds file. Also, if you set the emittances to zero, you'll see that the bunch length increase largely disappears. (There's still some tiny increase due to the energy spread.)

--Michael

Thanks a lot for the explanation,