Hi, i im layouting new pcb for lateral mosfet and found this schematic of OPS.
My question is about decoupling caps on DRAIN of mosfets. On some schematic i see 220nF between DRAIN and GND, and on some 470nF or 100nF value.
Does this mean that i need to add to each lateral mosfet DRAIN pad as close as possible 470nF?
What is recommended value for DRAIN capacitor for each mosfet (i will be using 2 pairs)?
I will be using EXICON lateral mosfets 2pairs in TO264 case.
And i read that MKT 2,2uF at DRAIN improves THD below 10kHz.
And some members suggests adding 1000uF/63V electrolytic cap to DRAIN and GND and solder on legs of this electrolytic cap 100nF MKP1837 capacitor.
I im not sure if this large capacitor will not give trouble to drain of mosfets when mosfet turns on/off as this large capacitor charges/discharges.This idea have description that it improves large currents/demands from fast slewrate audio amplifier and it is close to DRAIN to give accumulated energ to DRAIN for fast slewrate amplifier.
So what is recommended way?
My question is about decoupling caps on DRAIN of mosfets. On some schematic i see 220nF between DRAIN and GND, and on some 470nF or 100nF value.
Does this mean that i need to add to each lateral mosfet DRAIN pad as close as possible 470nF?
What is recommended value for DRAIN capacitor for each mosfet (i will be using 2 pairs)?
I will be using EXICON lateral mosfets 2pairs in TO264 case.
And i read that MKT 2,2uF at DRAIN improves THD below 10kHz.
And some members suggests adding 1000uF/63V electrolytic cap to DRAIN and GND and solder on legs of this electrolytic cap 100nF MKP1837 capacitor.
I im not sure if this large capacitor will not give trouble to drain of mosfets when mosfet turns on/off as this large capacitor charges/discharges.This idea have description that it improves large currents/demands from fast slewrate audio amplifier and it is close to DRAIN to give accumulated energ to DRAIN for fast slewrate amplifier.
So what is recommended way?
Hi!
As far as I can see, the 470uF and the 0.1uF are decoupling the DC power rails, where the Mosfet drains are connected to.
This is very common and recomended to be as close to the transistors as possible, since they are the active components that drains current from the power supply. And if you add 1000uF there is no problem at all. The capacitor will not discharge, it will be connected in parallel with the power supply.
The idea to add a film capacitor to the electrolytic is due to fact that the electrolytic is not good at high frequencies (has inductance and ESR).
In addition you decouple any parasitc inductance the PCB and wiring may add.
Any film capacitor will do the job - no need of special type.
As far as I can see, the 470uF and the 0.1uF are decoupling the DC power rails, where the Mosfet drains are connected to.
This is very common and recomended to be as close to the transistors as possible, since they are the active components that drains current from the power supply. And if you add 1000uF there is no problem at all. The capacitor will not discharge, it will be connected in parallel with the power supply.
The idea to add a film capacitor to the electrolytic is due to fact that the electrolytic is not good at high frequencies (has inductance and ESR).
In addition you decouple any parasitc inductance the PCB and wiring may add.
Any film capacitor will do the job - no need of special type.
Don’t put anything over 100uF on the amp boards. Big caps will shunt power rail noise to the local ground. That is the most common mechanism that you get 120 buzz noise.
Last edited:
I always use at least 470uF to the ground. The lead between the capcitors shall be very short and have a connection to power that is very near the V+ and V- leads in order to minimise the distorted magnetic fields from V+ and V- leads. And the power supply is only grounded via that channel amp pcb.
That means that you will have two separate power supplys for a stereo amp.
It is possible to use one 2 winding transformer and use only 1 winding for each channel but then you need to double the big capacitors. I do that for amplifiers under about 50w.
That means that you will have two separate power supplys for a stereo amp.
It is possible to use one 2 winding transformer and use only 1 winding for each channel but then you need to double the big capacitors. I do that for amplifiers under about 50w.
That is not a given. It depends on the PCB layout. I connect PSU ground-return lines as close to the on-board large-value caps as possible so voltage drops due to ripple current don't get too far.
I designed, laid out and built a JLH-like HA using that principle, and recently tested it using REW. I expected to see some peaks at 60 and 120 Hz but....nothing.
Thanks for detail explenation.
I will use from V+ to GND 1000uF // 100nF mkp1839 direct mounted to every drain of lateral mosfet. Gnd will be going from minus of 1000uF capacitor with fastcon connector on pcb to connect to star ground on main psu gnd.
And in the middle on pcb i will add 1000uF // 220nF and from V+ 1R/2W+470nF mkp1839 to GNd for OPS
For ips+vas uses vreg+ 470uF+100nF mkp1839
Why mkp1839? They are axial type will be verical mounted to pcb to use as little space as possible as mkp1837 in mouser are only black case and best is in blue case.
I will use from V+ to GND 1000uF // 100nF mkp1839 direct mounted to every drain of lateral mosfet. Gnd will be going from minus of 1000uF capacitor with fastcon connector on pcb to connect to star ground on main psu gnd.
And in the middle on pcb i will add 1000uF // 220nF and from V+ 1R/2W+470nF mkp1839 to GNd for OPS
For ips+vas uses vreg+ 470uF+100nF mkp1839
Why mkp1839? They are axial type will be verical mounted to pcb to use as little space as possible as mkp1837 in mouser are only black case and best is in blue case.
A 2cm lead in series with 100 nF will give a resonance in the region of 5MHz.
I dont think it will be serious because there is another resonance with the 470uF electrolytic.
Anyway the best is to consider this with short leads and power line to capacitor and out from the other side of the solder pad to filter out HF as much as possible.
I dont think it will be serious because there is another resonance with the 470uF electrolytic.
Anyway the best is to consider this with short leads and power line to capacitor and out from the other side of the solder pad to filter out HF as much as possible.