From 63663b9233b3ef24a4caeb4c762f1639e21d70d2 Mon Sep 17 00:00:00 2001 From: EstherLerouzic Date: Fri, 24 May 2024 12:17:08 +0200 Subject: [PATCH] Update release notes of v2.9 Signed-off-by: EstherLerouzic Change-Id: Ib949ff81fa818886a69339117bc66290dc2685b0 --- docs/json.rst | 22 ++++++++++++++++++++++ docs/release-notes.rst | 42 ++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 64 insertions(+) diff --git a/docs/json.rst b/docs/json.rst index ef09cbf35..df3bb9d86 100644 --- a/docs/json.rst +++ b/docs/json.rst @@ -500,6 +500,28 @@ Span configuration is not a list (which may change in later releases) and the us } } +Power sweep functionality is triggered when setting "power_range_db" in SI in the library. This defines a +list of reference powers on which a new design is performed and propagation is triggered +(only gnpy-transmission-example script). + +for example, with one amplifier amp1, and following settings: + + - power_dbm = 0 dBm + - max power amp1 = 20 dBm, + - user_defined_delta_p set by user = 3 dB + - 80 channels, so pch_max = 20 - 10log10(80) = 0.96 dBm + - delta_power_range_db = [-3, 0, 3] + - power_sweep -> power range [-3, 0] dBm + +then the computation of delta_p during design for each power of this power sweep is: + + - with pref = 0 dBm, computed_delta_p = min(pch_max, pref + user_defined_delta_p) - pref = 0.96 ; + user defined delta_p = 3 dB **can not** be applied because of saturation, + - with pref = -3 dBm (power sweep) computed_delta_p = min(pch_max, pref + user_defined_delta_p) - pref = + min(0.96, -3 + 3) - (-3) = 3 ; user defined delta_p = 3 dB **can** be applied. + +so the user defined delta_p is applied as much as possible. + SpectralInformation ~~~~~~~~~~~~~~~~~~~ diff --git a/docs/release-notes.rst b/docs/release-notes.rst index ddd1d92c0..f6bd9dafb 100644 --- a/docs/release-notes.rst +++ b/docs/release-notes.rst @@ -5,6 +5,48 @@ Release change log Each release introduces some changes and new features. +v2.9 +---- + +The revision introduces a major refactor that separates design and propagation. Most of these changes have no impact +on the user experience, expect the following ones: + +**Network design - amplifiers**: amplifier saturation is checked during design in all cases, even if type_variety is +set; amplifier gain is no more computed on the fly but only at design phase. + +Before, the design did not consider amplifier power saturation during design if amplifier type_variety was stated. +With this revision, the saturation is always applied: +If design is made for a per channel power that leads to saturation, the target are properly reduced and the design +is freezed. So that when a new simulation is performed on the same network for lower levels of power per channel +the same gain target is applied. Before these were recomputed, changing the gain targets, so the simulation was +not considering the exact same working points for amplifiers in case of saturation. + +Note that this case (working with saturation settings) is not recommended. + +The gain of amplifiers was estimated on the fly also in case of RamanFiber preceding elements. The refactor now +requires that an estimation of Raman gain of the RamanFiber is done during design to properly computes a gain target. +This results in an increase overal computation time when there are RamanFibers. This will be improved in the future. + +**Network design - ROADMs**: ROADM target power settings are verified during design. + +Design checks that expected power coming from every directions ingress from ROADM are consistent with output power +targets. The checks only considers the adjacent previous hop. If the expected power at the input of this ROADM is +lower than the target power on the out-degree of the ROADM, a warning is displayed, and user is asked to review the +input network to avoid this situation. This does not change the design or propagation behaviour. + +**Propagation**: amplifier gain target is no more recomputed during propagation. It is now possible to freeze +the design and propagate without automatic changes. + +In previous release, gain was recomputed during propagation based on an hypothetical reference noiseless channel +propagation. It was not possible to «freeze» the autodesign, and propagate without recomputing the gain target +of amplifiers. +With this new release, the design is freezed, so that it is possible to compare performances on same basis. + +**Display**: "effective pch (dbm)" is removed. Display contains the target pch which is the target power per channel +in dBm, computed based on reference channel used for design and the amplifier delta_p in dB (and before out VOA +contribution). Note that "actual pch out (dBm)" is the actual propagated total power per channel averaged per spectrum +band definition at the output of the amplifier element, including noises and out VOA contribution. + v2.8 ----