完善资料让更多小伙伴认识你,还能领取20积分哦, 立即完善>
嗨,我是ADS的新用户,只知道基础知识;
我必须创建一个具有8个混频器的电路,并将其IF基带信号上变频为8个频率,并使用两个4x1分路器将结果信号组合在同一条线路上,该分路器与一个2x1分路器相结合; 已经使用来自minicircuits.com的下载数据的数据项获得了分离器。 (我附上了一些图像以显示原理图设计)我一直在模拟结构,直到它有6个频率和任何分离器工作。 它还使用4个频率和两个分离器。 问题来自于我试图用分路器模拟8个信号(后来,我还想添加放大器和噪声模拟):它以“异常模拟终止”结束; 显示的原因通常是“内存不足”或“没有收敛”(我尝试将求解器更改为Krylov,降低8个基频的顺序,尝试使用瞬态求解器辅助)。 我的内存是4GB,我的处理器是intel centrino 2; 我在具有26GB可用空间的windows7 32位操作系统上使用ADS 2009(我需要32位系统才能看到布局)。 事情应该足够了。 有什么东西我不见了吗? BMixerswithcombiners.jpg56.8 KBCHBsettingswithsimulation.jpg281.0 KB 以上来自于谷歌翻译 以下为原文 Hi, I'm a new user of ADS and know only the basics; I have to create a circuit with 8 mixers in parallel upconverting their IF baseband signal to 8 frequencies and combining the the resulting signals on the same line using two 4x1 splitters which combines with one 2x1 splitter; the splitters have been obtained using data items with downloaded data from minicircuits.com . (I've attached some images to show the schematic design) I have been simulating the structure piece by piece, and until it had 6 frequencies and any splitter it worked. It also worked with 4 frequencies with two splitters. The problem came when I tried to simulate the 8 signals togheter with the splitters (later, I will also want to add an amplifier and noise simulation): it ends with "abnormal simulation terminated"; the reasons displayed were typically "insufficient memory" or "no convergence" (I have tried changing solver to Krylov, decreasing the order of the 8 fundamental frequencies, trying to use transient solver assisting). My ram is 4GB, and my processor is intel centrino 2; I use ADS 2009 on a windows7 32bit operating system with 26GB free space (I need 32 bit system to be able to see the layout). Things should be enough. Is there something that I'm missing? 附件
|
|
相关推荐
4个回答
|
|
我可能已经找到了问题的根源:我有太多基本频率,订单太大;
我试图提取一个基本频率(现在有七个),并将订单数量设置为2或3,现在它在15分钟内工作而不会因为需要太大的计算而中止; 我删除的频率可以从其他频率的组合中找到。 现在的问题是只用2或3次测量不能给我一个很好的解决方案,而且我希望将来还包括其他频率; 有没有其他方法可以模拟这个电路,或者更有效地使用谐波平衡? 我给你所需频率的值,以MHz为单位:130,250,410,615,820,940,1189,1400; 我使用了以下频率(订单数量):130(4),410(2),615(2),820(3),940(2),1189(2),1400(2)我没有' 使用了250,但仍然发现它,可能是因为它是其他频率的组合。 你能否告诉我这个低顺序是否足以导致错误? 我已经看过手动订单> 15甚至100 ....感谢任何回复的人,问候,J.L。编辑:John Lack于2013年8月3日上午3:29 以上来自于谷歌翻译 以下为原文 I've probably found out the source of the problem: I have too many fundamental frequencies, with too big orders; I've tried to remve one fundamental frequency (now there are seven), and put the number of orders to 2 or 3, and now it works in 15 minutes without aborting because of too big computation required; the frequency I've removed can be found from a combination of the others. The problem now is that measuring with only 2 or 3 order doesn't give me a good solution, and also I would like to include other frequencies in the future; is there any other way I can simulate this circuit, or for a more efficient use of Harmonic Balance? I'm giving you the values of the required frequencies in MHz: 130,250,410,615,820,940,1189,1400; I have used the following frequencies (with number of orders): 130(4),410(2),615(2),820(3),940(2),1189(2),1400(2) I haven't used 250, but still found it as a result, probably because it's a combination of other frequencies. Can you tell me also if this low order is enough for the misurements? I've seen on the manual orders>15 or even 100.... Thank you to anyone who responds, Regards, J.L. Edited by: John Lack on Aug 3, 2013 3:29 AM |
|
|
|
林宇宣55 发表于 2018-12-4 18:19 频率加上谐波和互调音调的数量肯定是任何谐波平衡模拟复杂性的主要因素。 减少这些肯定会减少内存需求。 您可以做的另一件事是利用任何与谐波相关的频率。 例如,在您的情况下,820 MHz是双410 MHz,因此您不需要同时包含这两者。 通过对这两个输入重复使用相同的410 MHz设置,使用值本身为1和2x为另一个,可以“简化”模拟。 以上来自于谷歌翻译 以下为原文 The number of frequencies plus harmonics and intermodulation tones is definitely a major contributor to the complexity of any harmonic balance simulation. Reducing these will certainly ease memory requirements. Another thing that you can do is to leverage any harmonically related frequencies used. For example in you case 820 MHz is double 410 MHz so you do not need to include both. The simulation can be 'simplified' by using reusing the same 410 MHz setting for these two inputs using the value itself for one and 2x for the other. |
|
|
|
|
|
|
|
非常感谢,它减少了频率数量(已删除820 MHz)和订单数量。
最大顺序是关键,但我没想到计算时间会有这么大的差异。 此外,抱歉迟到的回复,但我有帐户问题。 最诚挚的问候,拉克兰 以上来自于谷歌翻译 以下为原文 Thank you very much, it worked reducing the number of frequencies (820 MHz has been removed) and the number of the orders. The maximum order was the key, but I didn't expect such a high difference in computation time. Also, sorry for being late for the response, but I had problems with the account. Best regards, Lackland |
|
|
|
只有小组成员才能发言,加入小组>>
1240 浏览 0 评论
2352 浏览 1 评论
2164 浏览 1 评论
2030 浏览 5 评论
2914 浏览 3 评论
983浏览 1评论
关于Keysight x1149 Boundary Scan Analyzer
714浏览 0评论
N5230C用“CALC:MARK:BWID?”获取Bwid,Cent,Q,Loss失败,请问大佬们怎么解决呀
813浏览 0评论
1241浏览 0评论
小黑屋| 手机版| Archiver| 电子发烧友 ( 湘ICP备2023018690号 )
GMT+8, 2024-11-30 20:24 , Processed in 1.448196 second(s), Total 85, Slave 67 queries .
Powered by 电子发烧友网
© 2015 bbs.elecfans.com
关注我们的微信
下载发烧友APP
电子发烧友观察
版权所有 © 湖南华秋数字科技有限公司
电子发烧友 (电路图) 湘公网安备 43011202000918 号 电信与信息服务业务经营许可证:合字B2-20210191 工商网监 湘ICP备2023018690号