I’m an Italian PhD student and a Rubin in-kind collaborator, at the moment I’m working on the estimation of etendue and total system throughput in the context of the technote SITCOMTN-151. For our simulations we need to use the filters limiting wavelength and we referred to the LSE-30 document, but we get a bit confused about upper/lower limits with percentage 97 and 103. Can you help me understanding it better and calify what do they refer to? Overplotting the upper/lower red/blue values on the filter curve looks like this and we were wondering which of the two combinations is the most correct for our purposes.
Are 97% and 103% related to the subtended area under the filter curve or simething else?
Although not my area, I was curious to see if I can help a bit, and then if needed find a Rubin staff member with the right expertise here.
To start, it sounds like you were referred to LSE-30. Since that is a requirements document and doesn’t (necessarily) represent the as-built system, it might not be what you need for SITCOMTN-151? I say that because based on your linked plot, it looks like you’re using the measured transmittance as in Fig 5 of SITCOMTN-151, and not, e.g., the baseline throughputs.
Next - and maybe you already know this, but just to establish what’s in LSE-30 for readers - the section “u-band Response Envelope” in LSE-30 sets the specification that “The area weighted mean u-band filter response normalized to the in-band average (as measured between u_inBandBlue and u_inBandRed) shall lie between the upper and lower envelopes defined in the tables below.” And that’s where the u_upperBlue(1.03), u_lowerBlue(0.97) etc. values are defined.
On page 176 of LSE-30, Document-16295 is cited for the definition of the area weighted mean response function, and that document starts off with a caveat that “The algorithm described here is primarily meant for the purpose of evaluating the filter witness samples … [and] is not intended to be used for acceptance testing of the final filter” which makes sense in the context of LSE-30 as providing the specifications for the system.
So maybe that helps to clarify, maybe not – but for the question of what might be most correct for your purposes, it might help to hear more specifically about the purposes?
Hi @MelissaGraham , Thanks a lot for finding the time to reply to this post.
Following what you said I understand that the two percentage refer to the area subtended by the transmission curve normalized, so obviously the 103% limits are wider than the 97% limits. I’m fine and agree with that up to the moment I look to the 0% limits that are even wider than the 103% ones. At this point it make no sense to consider that limits as the ones at which the normalized area is the 0% of the one calculated using the u_inBandRed and u_inBandBlue, but it seems that the % is referred to the response value of the curve. The 0% limits are so wide that the filter response has already dropped to 0 at that wavelengths, but how do we justify seeing the filter having a 103% response in that scenario? My concern is just about understanding the reference of the % and see if it is really what I’m looking for.
Probably to solve ambiguity I can use the as built filter curves as in SITCOM-151 and compute by my self the wavelength at which the subtended area is the same as the expected one, but for doing this I wuold need to make the computations listed in Document-16295 by myself and so be exposed to possible errors. I would prefer to use tabulated values (even requirements and not as built data) rather than number I made up by myself to be more error proof.