CAMB (rayleigh): Unexpected feature in 143 GHz cross spectrum

Use of Healpix, camb, CLASS, cosmomc, compilers, etc.
Post Reply
Karia Dibert
Posts: 1
Joined: February 01 2020
Affiliation: University of Chicago

CAMB (rayleigh): Unexpected feature in 143 GHz cross spectrum

Post by Karia Dibert » February 01 2020

Hello,

I’m trying to recreate Fig. 3 from your Rayleigh paper https://arxiv.org/pdf/1307.8148.pdf. I’ve cloned and compiled the CAMB rayleigh branch, and changed only three things in the supplied params_rayleigh.ini file:

output_root=output/run1
get_scalar_cls=T
get_tensor_cls=F

Running this, I find an extra feature in the Rayleigh-primary cross spectrum at l=183 for 143 GHz that doesn’t match your figure (see image at link below). The feature decreases in depth by an order of magnitude if I set accuracy_boost=2, but does not change upon further increasing the accuracy. Any idea what could be causing this?

https://imgur.com/a/yYYEa0y

Thanks,
Karia

Antony Lewis
Posts: 1551
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: CAMB (rayleigh): Unexpected feature in 143 GHz cross spectrum

Post by Antony Lewis » February 03 2020

Not sure, the signal is very small - you many need to increase the other accuracy parameters (L sample boost).

Post Reply