Author Topic: Simulate Mosfet Ringing in LTSpice  (Read 2715 times)

0 Members and 1 Guest are viewing this topic.

Offline TomsikTopic starter

  • Contributor
  • Posts: 35
  • Country: cz
Simulate Mosfet Ringing in LTSpice
« on: June 26, 2019, 08:18:55 pm »
Hello, I'm trying simulate ringing on mosfet transistor by LTspice, but I'm not able to do...Please help me :)
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Simulate Mosfet Ringing in LTSpice
« Reply #1 on: June 26, 2019, 10:07:09 pm »
There's actual several possible reasons why you can't see ringing. First is that the rise and fall times for the pulse are both set to 1u - that's way too long to excite ringing between the MOSFET capacitances and the stray inductances; try setting that to 20n.

Second is that you are measuring the voltages outside of the resonant networks; measure the actual drain and gate voltages, instead.

Third is to set a minimum timestep in the simulation command to something appropriate for the frequencies you expect to see, though note that this greatly increases simulation time; try 100n first, then 10n if you still don't see ringing.

EDIT - forgot one, which is the most elementary of them all: the strays are overdamped by the resistances, so no ringing will occur. Though you'll likely to some overshoot on the source terminal.

« Last Edit: June 26, 2019, 10:16:28 pm by MagicSmoker »
 

Offline TomsikTopic starter

  • Contributor
  • Posts: 35
  • Country: cz
Re: Simulate Mosfet Ringing in LTSpice
« Reply #2 on: June 27, 2019, 08:46:58 am »
I've set rise and fall edge to 20n and set the simulation timestep to 100n a then 10n. Still not ringing.
Sorry, but i don't understand your last sentence...
 

Offline MagicSmoker

  • Super Contributor
  • ***
  • Posts: 1408
  • Country: us
Re: Simulate Mosfet Ringing in LTSpice
« Reply #3 on: June 27, 2019, 09:50:48 am »
I've set rise and fall edge to 20n and set the simulation timestep to 100n a then 10n. Still not ringing.
Sorry, but i don't understand your last sentence...

Each inductor you added forms a series resonant network with the relevant stray capacitance of the MOSFET, but also in series with those resonant networks are resistors which dampen the ringing. That is to say, as the damping resistance in a series resonant network increases, the amplitude of the ringing, and the time it takes to decay completely, decreases*. In the case of the BSP89 MOSFET you picked, the .model directive shows it has a gate-to-source capacitance (Cgs) of 70pF and a drain-to-source capacitance (confusing referred to as Cjo) of 10pF. The characteristic impedance is determined from the formula (L/C)^0.5 which comes out to around 12 Ohms for the gate LC network, but the series resistance is 50 Ohms, so this LC network is definitely overdamped (in fact, with such small LC components I doubt you'll see ringing even if the gate resistor was reduced to 1 Ohm).

I'll leave determining the drain and source node characteristic impedances to you as an exercise, but note that ringing at these nodes really only occurs during turn-off, not during turn-on (for another exercise, try to figure out why).


* - and for a parallel resonant network the damping resistance is in parallel, with ringing amplitude decreasing as resistance is decreased, so the inverse of the series LC network.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf