Hi Zach, I remember in "Controlled ESR Capacitors" video that increasing ESR of Capacitor of fres=f1 flattens the curve between Z at fres and the peak point, same for the other capacitor, meaning increasing ESR of both capacitors can give a flat response which reduces the supply voltage oscillation in the time domain. I have a question here, doesn't it make no sense to increase ESR of a capacitor for better performance? or since you're talking about "off-chip" capacitors then it's different, meaning, on-chip capacitors implemented in each block in the SoC could require minimum ESR for other reasons? and off-chip could require high ESR for the flat response? could you explain the difference between the two types of capacitors? off-chip capacitors and on-chip for each SoC block modeled as Cdie & ESR?
In that question it was about using a controlled ESR capacitor in a PDN, but that is not really their intended usage. Technically you could use them, but you are correct that it does not make sense to increase resistance if you want to get to a lower PDN impedance. If you used controlled ESR capacitors with higher ESR value to get to a lower PDN impedance then you might need to use a lot of those controlled ESR capacitors in parallel to get the total PDN impedance below target. In terms of on-chip, I will admit I have not looked into the controlled ESR aspect of those chip capacitors because I do not do IC design. I will say that those chips have higher capacitance density AND higher ESR so I think the response at high frequencies would naturally be flatter but I would have to look at data to be sure. There will also be differences between the different types of SoC capacitors (MIM, MOS, MOM, etc.)
Hey Zach!
Hi Zach, in 5:54 equation there is no C.
That's because it's lumped into the natural frequency (omega-0).
Hi Zach, I remember in "Controlled ESR Capacitors" video that increasing ESR of Capacitor of fres=f1 flattens the curve between Z at fres and the peak point, same for the other capacitor, meaning increasing ESR of both capacitors can give a flat response which reduces the supply voltage oscillation in the time domain. I have a question here, doesn't it make no sense to increase ESR of a capacitor for better performance? or since you're talking about "off-chip" capacitors then it's different, meaning, on-chip capacitors implemented in each block in the SoC could require minimum ESR for other reasons? and off-chip could require high ESR for the flat response? could you explain the difference between the two types of capacitors? off-chip capacitors and on-chip for each SoC block modeled as Cdie & ESR?
In that question it was about using a controlled ESR capacitor in a PDN, but that is not really their intended usage. Technically you could use them, but you are correct that it does not make sense to increase resistance if you want to get to a lower PDN impedance. If you used controlled ESR capacitors with higher ESR value to get to a lower PDN impedance then you might need to use a lot of those controlled ESR capacitors in parallel to get the total PDN impedance below target.
In terms of on-chip, I will admit I have not looked into the controlled ESR aspect of those chip capacitors because I do not do IC design. I will say that those chips have higher capacitance density AND higher ESR so I think the response at high frequencies would naturally be flatter but I would have to look at data to be sure. There will also be differences between the different types of SoC capacitors (MIM, MOS, MOM, etc.)
Lee Margaret Perez Daniel Brown Michael
Trantow Lane
Garcia Margaret Perez Jennifer Martinez Anna
Larkin Views