DCDC BUCK的效率随频率先升高,再下降
我用3V CMOS 0.35um工艺制作一个DC/DC BUCK, 使用TYPE3类型补偿,
是最简单的PWM电压型,PMOS和NMOS开启电阻都是1欧姆左右,
工作都是稳定的,
工作频率有四个档,分别是:2MHz,4MHz,8MHz,10MHz,,
输入电压3.6V,输出1.8V,输出电流60mA情况下,
效率分别为:82%,88%,89%,88%,
就是效率随频率先升高,然后下降,
8MHz到10MHz效率下降可以理解,是电容损耗开始占主导,,
2MHz,4MHz,8MHz随频率升高而升高,就不理解了,
电感=4.7uH,电容=4.7uF,电容是陶瓷电容,低ESR,
开始考虑是电感电流可能会出现负电流情况,分析没有,
实验用10uF电容替换4.7uA电容后一样的结果,
并且测试点输出电流10mA,20mA,30mA,40mA,50mA,60mA,70mA,80mA,
都是这样的情况,
求教各位,哪位仁兄能指点一下。
频率高了,电流纹波小了,导通损耗里面的交流损耗小了,所以增大频率可能会提高效率的。我是这么理解的
Since your inductor is fairly large, the current ripple is small, so the cross point is about 8MHz, and conduction loss is dominant below 8MHz, that's why you see efficiency goes up with frequency below 8Mhz.
谢谢回复,我的疑问主要在于仿真和测试的差距,
仿真结果这2,4,8MHz,3个频率工作效率都是在91%左右,没有明显差别,
10MHz 开始效率下降到90%,
所以2,4,8MHz 工作的仿真和测试差距巨大,
尤其是2MHz 效率从仿真的91%降低到82%,
差距太大,不能理解,其他还好,降低些说明仿真时候一些寄生电阻和电容导致的损耗没有考虑进去,
仿真时候用的还都是比较理想的器件。
谢谢各位,谁有什么高见,再说说。
For a good IC designer, you can not just buy in what simulation results tell you, device model sometimes is deceptive. Do you check the connection b/w the inductor and controller?
I just designed a buck converter with max 1A load, working at 20MHz, inductor 60nH integrated, peak efficiency~85%, it's not just prototype, it's product now.
For a good IC designer, you can not just buy in what simulation results tell you, device model sometimes is deceptive. Do you check the connection b/w the inductor and controller?
I just designed a buck converter with max 1A load, working at 20MHz, inductor 60nH integrated, peak efficiency~85%, it's not just prototype, it's product now. I can calculate all the efficiency using Excel spreadsheet
VERY GOOD
非常感谢您的答复,您说的b/w 是指什么呢?还望解释一下。
between
Buck 的频率需要这么高吗?
power loss 分两个大方面。
一个是内部的switching 方面的loss。当然是switching freq越高loss 越大。除conduction loss外,你的freq升高,效率升高,是什么原因,不好说。
另外就是 外部连接的部分。从你的pad到电感,从电感到电容,esr是一方面,连接是另一方面。
电容损耗是什么意思
niubility.. wha is the part number.