作者kyokanasaki (念个书吧~哞!)
看板Electronics
标题[问题] Circuit optimization by Hspice
时间Sun Jul 5 23:13:02 2009
Hello~everyone
Here I have some question about circuit optimization by Hspice
Could I get these best W/L of each transistor by for loop method?
There is no answer through Google and I came up with some idea to solve it.
The idea is that I write some statements like that below:
Step 1:
....
.atler
.param w1=200, l1=5, w2=100, l2=3.6 .......brabra ...Vout is saved to file
.end
If I want to try many w/l(ex:w=200 to 300, step=10,keep the same l1), then
changing some values of some transistors and repeat statements
Step 2:
....
load txt file to Excel and use macros to get the best W/Ls to meet output.
....
I think the work mentioned above is okay(I haven't tried it yet).But it's
exhausting. Do u anyone have any better idea? Thank you all!
--
※ 发信站: 批踢踢实业坊(ptt.cc)
◆ From: 122.123.134.39
※ 编辑: kyokanasaki 来自: 122.123.134.39 (07/05 23:20)
※ 编辑: kyokanasaki 来自: 122.123.128.189 (07/06 19:49)
1F：推 juc:usually there're dozen of transistors in a analog ckt 07/09 23:46
2F：→ juc:tweak one of them could not optimize overall performance 07/09 23:50
3F：→ juc:I think you can consult some traditional design flow first 07/09 23:55
4F：→ juc:once you are familiar with it, you would know which 07/09 23:58
5F：→ juc:transistors should be tweaked larger or small, you 07/10 00:00
6F：推 juc:can do this by iteration. you should know what parameter 07/10 00:09
7F：→ juc:you want to be optimum(power, swing, linearity...) 07/10 00:12
8F：→ juc:different parameters result in differdnt difficulty on 07/10 00:16
9F：→ juc:optimum. there are no one rule can optimize all of them 07/10 00:17