thank your answer, in my schematic, it just happens with one sheet, I have 3 sheet in project and 2 other still working I will try your solution.![]()
↧
RE: loss of components when open schematic on another PC
↧
RE: loss of components when open schematic on another PC
Well, if it's just one page out of three, the problem isn't global settings that I was trying to address. Sorry about that. But, it's weird that it isn't happening across all machines. But, you might get lucky and it goes away with a new Capture.ini file.![]()
↧
↧
Import Schematic from Mentor Tanner to Cadence
Hi, I've recently migrated to Cadence from Mentor. Could you please let me know if there is a way to import my Mentor S-edit files into Cadence? I would really appreciate the help, since this would save me hours of schematic redrawing. Thanks!![]()
↧
RE: DC-DC Converter/ Feedback/ Verilog-A
hi dear pyroblast, do you have the dcdc veriloga model? can you share whit us? i try to write the dcdc power state model with veriloga, but i fail.![]()
↧
什么是COM/JCOM信道合规技术
在当今这个数以十计/两位数Gbps的数据时代里， 工程师的工作越来越不容易，正确地设计并表征系统以符合不断刷新的业内标准搞得大家焦头烂额，不仅要对高速串行链路及其所有损耗进行仿真，还得通过合规测试。 雪上加霜的是，大家熟悉的依靠比较不同频域参数（例如插入损耗、回波损耗、串扰等）从而发现设计缺陷的传统合规性检查方法已经过时 了！由于没有考虑到不同参数的相互依存性和权衡裕量 问题，传统方法容易导致过度设计和成本过高问题。例如，如果一个设计满足插入损耗和串扰指标且有很大裕量，而仅有回波损耗一点点不达标时，那么即使它实际上能够完美运行也一样会被视为设计上的不合规。那怎么办？ Channel Operating Margin（COM）——通道裕量，是IEEE 802.3bj-2014 100 Gb/s背板操作和铜缆以太网标准中介绍的合规性检查方法。 作为时域规范，其定义如下： COM = 20 log 10 (As/Ani) 其中，As和Ani 是指接收器输出端的信号和噪声幅值，并 考虑整个端对端通道特性（包括发射器（ TX ）、接收器（ RX ）和通道损耗）的影响。 As 由通道中的 s 参数计算出来，包括近端和远端的串扰路径、 TX 和 RX 器件封装模型、 TX FFE 和 RX CTLE 滤波器、 TX 和 RX 输出 / 输入阻抗模型、以及 RX DFE 模型。相对于来源于发射器、均衡后的残余 ISI 、抖动（转换为幅度噪声）、峰值串扰的噪声与 COM 指定的 RX 噪声滤波器级联的 CTLE 噪声，由于 FFE 、 CTLE 和 DFE 的均衡抽头都实现了优化， As 也相应实现了最大化。噪声幅度 Ani 是在假设为高斯噪声分布的情况下，根据传播到接收器输出 端的所有噪声源的组合分布函数来计算的，并通过将累积分布函数等同于检测器误差率(DER0)来求解，以使得到的噪声电平对应于该统计误差率。当COM值超过某个阈值，如IEEE 802.3bj标准中的3dB，即满足设计合规性要求。值得一提的是， COM是由许多相互关联的信号变量组成的，因而工程师才能在满足BER或DER 0 要求的前提下进行设计权衡。 JCOM是针对JEDEC JESD204C C类物理层规范的合规性方法，用于将数据转换器（ADC和DAC）连接到逻辑器件（如ASIC 、 FPGA等）的高速串行链路（数据速率范围从6.375 Gbps到32 Gbps）。 JCOM = 20 log 10 (As/Ani) 与COM一样，JCOM也是在接收器输出端计算信噪比（SNR）的时域测量值，并允许设计人员灵活管理链路中TX 、 RX和通道损耗之间的权衡。 它包含对COM的几种改进，比如定制化器件封装模型以及频率相关的TX/RX 输出 / 输入阻抗模型。当 JCOM 值大于阈值 2 dB 时，设计即满足合规性要求。 那么，使用COM或JCOM 将对我们的工作 有何益处？ 主要有以下三点： 在通道损耗、串扰、发射器和接收器指标之间进行设计权衡，以避免过度设计和成本增加 一旦获得考虑串扰因素的通道S 参数，可直接、迅速地确认设计是否满足合规性要求 可使用COM / JCOM来帮助确定系统架构设计中的决策问题，例如FFE、CTLE、DFE配置和均衡抽头数 COM / JCOM 可帮助我们以更快速高效的方式通过合规性检查 并减少过度设计，这对于今后的高速串行链路设计至关重要。 Cadence的 Sigrity SystemSI 现已包含COM通道合规性功能，可帮助工程师快速便捷地判断系统设计是否满足要求。 JCOM通道合规性功能将于不久后问世，敬请期待。如果您正在被通道合规性仿真问题所困扰，请联系我们，我们将运用SystemSI 的 COM 通道 合规性技术快速解决您的烦恼。 有关COM计算的详细信息，请参阅IEEE802.3bj-2014 标准 的附录93A ； JCOM 信息，请参阅 JEDEC JESD204C标准的5.12节和附录E 。 *原创内容，转载请注明：https://community.cadence.com space 欢迎订阅“PCB、IC封装：设计与仿真”博客专栏， 或扫描二维码关注“CadencePCB和封装设计”微信公众号，更多精彩内容期待您的参与！ space 联系我们：spb_china@cadence.com![]()
↧
↧
Error with Calibre PEX using a certain PDK in Cadence Virtuoso
Hello, This question is related to Calibre tools using in Cadence, and I have already asked Mentor this question and the PDK vendor. I did not know where to post this question, so I am posting it in this forum. But I had questions relevant to the Cadence Virtuoso version I am using with Calibre and the PDK. I am now using Cadence Virtuoso (1.6.7.64b.78) and an updated version of Calibre (2015.2). I am getting errors with the PEX Rules file compilation but not with with the LVS or DRC rules file compilations. Calibre will not compile the PEX rules file and gives me the following error (attached). The vendor of the PDK tested this with Cadence IC 1.6.7.500 and Calibre 2015.4. A colleague of mine can run the same PDK with Cadence 1.6.7.500 and Calibre 2014, not Calibre 2015 - in other words an older version of Calibre, a newer version of Cadence Virtuoso - and an older version of Calibre than the PDK vendor sugests. What reason, then could it be that I get a compilation error when my Calibre version is newer ? Is it the older version of Cadence 1.6.7 ? Can an older version of Cadence affect the working of Calibre ? Can this error be fixed by updating to a newer version of Cadence Virtuoso ? What can I do ? Thank you.![]()
↧
RE: Error with Calibre PEX using a certain PDK in Cadence Virtuoso
This isn't really a relevant question for this forum, because a compilation error in a Mentor tool must be the responsibility of that Mentor tool. I cannot see how the Virtuoso could have an influence. It's quite possible that there's an issue in a particular range of Calibre versions - you're still using a version that is older than the PDK vendor suggests (they said Calibre 2015.4 and you're using Calibre 2015.2). Maybe that issue didn't exist in Calibre 2014, was introduced in or before the subversion you're using, and was resolved by 2015.4. This however is complete guesswork. BTW, I assume you meant IC6.1.7.xxx rather than "Cadence" 1.6.7.xxx (there is no release stream called "Cadence" and I think you have the version numbers mixed up). Andrew.![]()
↧
RE: Import Schematic from Mentor Tanner to Cadence
You've posted this in the wrong forum - this is a forum for feedback, suggestions, and questions about the forum itself, not for technical questions. I can move it into the right forum and (possibly) answer it, but first you're going to have to say which Cadence tool you're using. There is no tool called "Cadence", so you might be talking about "Capture" for the PCB world, or "Virtuoso Schematic Editor" for the IC world. Please clarify...![]()
↧
RE: is it possible to extract the subcircuit names/models and the associated pins from input.scs ?
Hi Andrew, Thanks a lot, it worked, with the note taht teh name of the file is actually subckts.info.subckts . I have next question: for some subcircuits there is no info about teh parameters and pins (they are empty strings). My guess is that this is happening for the subcircuits which are not used in my design, but are defined in the model files. Is that true? Also, is there any documentation about the format of this file, or a skill API to interrogate this database? Best Regards, Marcel![]()
↧
↧
RE: Annotations are just way too overpowering
Glad you got it all working - that looks good! If you don't want to go via the cdsParam() approach, you could simply use OP(strcat(geGetInstHier() "/" instName) "gm") - the main downside of that is that you then won't get the ability to change what is annotated via the menus in ADE, or using the annotation setup form in the schematic - up to you. Regards, Andrew.![]()
↧
RE: is it possible to extract the subcircuit names/models and the associated pins from input.scs ?
Hi Marcel, The subckts.info.subckts is a "PSF" database that is saved into the psf directory if you have where=rawfile specified - there's an entry on the Outputs->Save All form which saves that by default. If you want to process the PSF results, you can simply do: selectResult('subckts) ; to avoid having to add ?results 'subckts on the end of the outputs() or pv() functions below: foreach(subckt outputs() terminals=pv(subckt "subckt_terminals") params=pv(subckt "subckt_params") unless(blankstrp(terminals) printf("%s : %s : %s\n" subckt terminals params) ) ) You're right - unused subckts (e.g. from model files) are output with an empty terminal list. Regards, Andrew.![]()
↧
RE: constraints file
ok sir. thank you.![]()
↧
RE: is it possible to extract the subcircuit names/models and the associated pins from input.scs ?
Hi Andrew, Thanks again. You are very kind, not too much work left for me: I just had to copy->paste :) Best Regards, Marcel![]()
↧
↧
RE: ams simulator vs spectre: empty schematic view
Thank you Andrew. Now it is clear. Rosario![]()
↧
RE: Calibre fails to export schematic netlist
Hi when i am doing calibre lvs. i was facing the same issue as mentioned by Mohammad. So after changing " setenv CDS_Netlisting_Mode Analog" i am facing ERROR (OSSHNL-514). Can You help me regarding this issue? Regards, Amit![]()
↧
PCB Editor Database Attachment xml form.
Hello Everybody, Can you please help me? I want to use the SKILL command axlCreateAttachment() to store XML file directly to the database as attachment. I can sucessfully use the axlCreateAttachment() in the 'string mode, however, I have issues with the 'file mode execution. Can somebody explain to mee how to use the 'file mode? In the documentation there is example only for the 'string mode. I've supposed that I'll need to open the XML file with infile(), then readout the conntent with read() close the XML and then somehow send the data to the attachement, but this doesn't work for me. Any simple example will help. Thanks, Erik![]()
↧
RE: Calibre fails to export schematic netlist
Hi Amit, Please read Guidelines for the Custom IC Design Forum . There you will see that you should post a fresh topic thread for your question rather than posting on the end of a 5 year old thread. Please also provide the complete error message (not just the code) and the IC version you're using (as given by Help->About in the CIW). Regards, Andrew.![]()
↧
↧
Ronto and Quecto Are Not Cheeses
The International Bureau of Weights and Measures (its initials are BIPM because it is in France, the Bureau International des Poids et Mesures) recently redefined the kilogram. Meera wrote about it late last year in Size (of the Kilogram) Matters . They didn't actually change the kilogram, they switched the official definition from being the mass of a particular block sitting in Paris to a definition in terms of Planck's constant. There are a number of problems with a block of platinum and iridium sitting under a bell jar on the other side of the world, the most obvious of which is how you use that block to do anything since it is far too important to let people touch it. Plus, in a problem familiar to people who work with semiconductor fabrication, you have to deal with particles. In manufacturing, they mess up the lithography and we end up with non-functional chips. In the case of the kilogram, the mass changes slightly. Prefixes Well, here's another change that involves the BIPM and semiconductors. There is a recent paper in Measurement (although it is already available, that edition of the journal is dated April 2019). The paper is On the nature of SI prefixes and the requirements for extending the available range . The SI prefixes are things like nano and pico on the small end, and giga and exa at the high end. I don't know how many industries use both the very small measurements, and the very large, but we certainly do, with dimensions in nm, delays in ps, but also data volumes in exabytes (EB) and frequencies in gigahertz (GHz) and up. You've probably noticed that the SI units, like common usage in English and other languages, go up in units of three zeros. Thousands, millions, billions. You might think this is universal, but at least a couple of billion people will tell you that's not so. If you learn how to count (beyond the small numbers) in Chinese, you discover they count in units of four zeros. So 10,000 is 1 wan. So you have to remember things like a million is 100 wan. If you need larger numbers, you get to yi (100Ms). When I had a development group in India in the 1990s, I had to discover that Indians count with FIVE zeros, lachs, which are 100,000 of something, usually rupees. But the next unit up, the crore, is just 100 lachs, so 10M. It is surprisingly difficult to think in wan (China) or lachs (India) when you are completely used to numbers being split up into groups of three. I guess Chinese and Indians have that problem the other way around. Gate oxide thickness is often measured in Ångstrom units (Å), which is 10 -10 meters, and about twice the diameter of a hydrogen atom, or the diameter of a chlorine atom. But that's not an official SI unit (they only have multiples of three). One thing I learned working for Virtutech, where our engineering was in Sweden, was the Å is a character in the Swedish alphabet. It is not an A with an accent on it, like the way the French alphabet consists only of the unaccented characters, and the accents are added later. The alphabet runs from A to Z but then has 3 more characters, Å, Ä, and Ö. At the other end of the scale, another unofficial unit is the Googol. Apparently the founders of Google named their company after it (because they were dealing with a lot of data), but didn't get the spelling right. A googol is 10 100 . The term was invented by the 9-year-old niece of mathematician Edward Casner, who used it in his famous 1940 book Mathematics and the Imagination . Somewhere in one of the boxes in my garage, I have a copy. But to be honest, in the Amazon era, it is probably quicker for me just to pony up $18 and get a new one in a couple of days, rather than searching the garage. Yes, despite the 1940 publication date, the book is still in print almost 80 years later. Even available on Kindle. The book also introduced the name googolplex, which is 10 google , not to be confused with the Googleplex, which is in Mountain View. The Proposal Anyway, back to that paper. The author is Richard Brown, who is the chief metrologist for the National Physical Laboratory in Teddington in the UK. That's in the West of London. Trivia fact of the day, nearby Teddington Lock, which is actually a weir that can be bypassed by three locks (for different sized boats) is the tidal limit of the River Thames. Richard is worried that if we don't come up with new units to extend the scale that we have, then informal words that are starting to come into use will become ingrained. So a proposal has been made to the aforementioned International Bureau of Weights and Measures, BIPM, in Paris. As Emilio Prieto (who works at the Spanish Metrology Center) told Science magazine: Once people start using the wrong prefix names it is impossible to go back. Brontobyte and geobyte are apparently already becoming popular. So what are the newly proposed prefixes that Richard came up with? There are four of them, two big ones and two small ones. They preserve the very valuable ability to abbreviate them with a single letter (like we say GB and nm), which is a challenge. X, W, and V are apparently out, and U is out since we commonly use um (instead of the official µm) for microns. I'm not sure why S is not acceptable, maybe because 1ss looks silly for a tiny unit of time. But carrying on from zetta and yotta (abbreviated as Z and Y) we get to R and Q as the next available letters. So the current high prefixes are mega, giga, tera, peta, exa, zetta, yotta. The proposal is to use ronna for 10 27 and quecca for 10 30 . The current low prefixes are micro, nano, pico, femto, atto, zepto, yocto. The proposal is to use ronto for 10 -27 and quecto for 10 -30 (with r and q as their abbreviation). If the proposal is accepted, the new units would (probably) be introduced in 2022. The previous big and small prefixes were added almost 30 years ago in 1991. After that, there is only one available letter pair left and that is B/b. So, in the future, if a new prefix gets added, that would be the way to bet. The names to get used to again: ronna and quecca for really big numbers (next after yotta) ronto and quecto for really small numbers (next after yocto) XKCD There's an XKCD for everything, but not yet for these new prefixes. But there is one for the redefinition of the kilogram. Sign up for Sunday Brunch, the weekly Breakfast Bytes email.![]()
↧
RE: Annotations are just way too overpowering
Thanks Andrew - I'm actually also very interested in your other approach. I did try a search for "op" and "op(" to read more but their are too many keywords containing or ending in "op" that contaminate my search results. Can you please point me to the documentation for this.![]()
↧
RE: Import Schematic from Mentor Tanner to Cadence
Given that the storage format for S-Edit is EDIF (or a variant of EDIF), the best route is to use the EDIF export from the S-Edit tools and then import that into Virtuoso. I've had good success with that, although I have had to do some pre- and post-processing to clean things up - sometimes due to PDK incompatibilities, sometimes due to limitations/bugs in both EDIF interfaces (in the source and destination tool). You may want to try this and if you have trouble, please contact customer support - although it might need a service if the work needed is extensive. Regards, Andrew.![]()
↧