I am building a class d amp circuit with base design from Great Scott's Class D amp which is found here. I am trying to build and simulate the circuit in LTSpice with difference being I won't be using MOSFET Driver IC and will be inverting the signal and feeding it to another half bridge to make a full bridge amplifier.

Here is the Circuit and simulation files:- https://drive.google.com/drive/folders/1_B0tQIh58_niOk3M4CdICHCKTr6wPS08?usp=sharing

I am experiencing some wierd oscillation at the start of the circuit for a few milli-seconds. Can anyone point out the reason and/or a fix. Will the circuit get damaged if I build it as it is?

Here is the output of the simulation:- enter image description here

The oscillation occurs on both of the half bridge after passing through lc filter. The blue line is the load resistor current. (In reality, the current would be half tho. I am using 1 ohm load in design while in reality I have 2 ohm subwoofer)

Also note, the circuit will only be used for sub frequencies, so don't point out that voice will get lower as frequency increases and also, it gets distorted at higher frequencies.

  • \$\begingroup\$ I don't see anything weird - what do you see that is weird? Also what node is V(n009)? \$\endgroup\$
    – Andy aka
    Commented Dec 14, 2020 at 11:49
  • 1
    \$\begingroup\$ If You want to build and use class d amp - really go for driver chip. This will save You a lot of burned transistors and time \$\endgroup\$
    – fifi_22
    Commented Dec 14, 2020 at 13:44
  • \$\begingroup\$ So, should I go for dual IR2110/2113? Also, as suggested in the answer below, do they add a dead-time? \$\endgroup\$ Commented Dec 14, 2020 at 13:49
  • \$\begingroup\$ Adding to the comment from @fifi_22 , no 555 should be used in a class D that's about to be built in real life, unless you want to study the circuit, or play with it. Certainly not for a 90 V (bridge) and a 1 \$\Omega\$ load. 555 or not, that's exaggerated. Not to mention you're simulating with generic BJTs, and there is no feedback, no linearizing. It really looks like you're in for some magic smoke. \$\endgroup\$ Commented Dec 14, 2020 at 15:13
  • \$\begingroup\$ Would a 555 cause problems? Stability? Any alternative? I shoud be able to simulate with any bjt that has HFE > 100 and current capability of 500ma+. I know there's no feedback but I won't really be powering it at 90V (60V in real) also I have mentioned in my question that only sub frequencies will be used so I am not concerned about feedback. and that I will be using 2 ohm load. \$\endgroup\$ Commented Dec 14, 2020 at 15:57

1 Answer 1


Your filters are ringing due to the voltage step at startup. It is worse in the simulation because you are using 'ideal' components with no internal resistance, so the Q is very high.

I added 0.1 Ω resistance to L1 and L2, and 0.01 Ω to C4, C5 and C6 to represent more realistic components, then added a delay of 1 ms before starting the tone burst. Here is the result:-

enter image description here

A more serious problem is you don't have any 'dead-band' bewteen M1 (M4) turning off and M2 (M3) turning on, so there is short period where both upper and lower FETs are turned on at the same time. This causes a 'shoot-through' current of nearly 400 amps!

enter image description here

You need to delay M2's Gate voltage rise (blue line in the plot above) so that it doesn't start until M1's Gate drive (green line) has finished falling.

  • \$\begingroup\$ Thanks for your suggestion regarding the first problem. Also, after I asked this question, I was about to test the shoot through current myself. I have also tested this but I can't seem to think of something to delay the starttup of MOSFET. I think I read somewhere that schmitt triger can be used? Can you please provide some suggestion. I tried increasing the pullup-resistance of the gate but when the shoot through was sufficiently low, my output was not present. Also, I cant rely on simulation for exact tuning of values because of component tolerances or maybe I decide to change my MOSFETs. \$\endgroup\$ Commented Dec 14, 2020 at 13:47
  • \$\begingroup\$ electronics.stackexchange.com/questions/281248/… or the obvious way - use a Gate driver IC! \$\endgroup\$ Commented Dec 14, 2020 at 20:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.