IR Emitter Circuit, PSpice Error

[Mattjd]

I'm working simulating a circuit for a project and am having difficulty running the PSpice simulation.

The circuit in question is driving an IR LED using a power mosfet, and driving the mosfet with gate driver receiving a 38Khz signal from a microcontroller.

(Apologize for the plain blocks, model import wizard wasn't changing those for some reason).




What I have done is find the pspice models library files for

SFH_4715A (IR Emitting Diode)

PSMN1R0-30YLC (N channel power mosefet)

ZXGD3005 (Gate Driver)

I opened Pspice Model Editor Lite->File->Model Import Wizard->Found Model and Hit Next->Finish (Repeat for all 3 models). This created the .olb files

I then created the circuit shown above using the .olb files and performed in Capture CIS

PSpice -> New Simulation -> "enter name" -> Inherit From: "None" -> Create

I created a transient simulation to run for 1000ms, in steps of 1ms, and save data immediately. Then, from within Simulation Settings went to

Configuration Files -> Library -> "added each Pspice Model Library to the design" -> Apply -> OK

I then ran the simulation.

I am getting this output file

https://hastebin.com/oxijiyoquq.md

With the main error being line 178 and 180

ERROR(ORPSIM-16550): Floating point computation failed during matrix solution. Possible solutions: 1)Ensure that all device parameters are in valid range. 2)Try using .options LIMIT

INTERNAL ERROR -- Overflow, Multiply

edit: I just noticed in the picture I have Fall time to be 1E-90, I fixed that but Im still having the same issue.

[Mattjd]

Does my circuit atleast seem correct?

[T3sl4co1l]

Assuming you found the correct models, and hooked them up properly... no.

You didn't note that the ZXGD is a poorly disguised complementary BJT follower, so has no voltage gain -- you'll get about 2.6V Vgs(on) here -- and there's no current limiting resistor in series with the LED, unless you intended it to limit at whatever RS (check the model) plus Rds(on) (1mR I guess from the part number?) runs it at.

Note that SPICE sources are ideal, i.e., your VPULSE can deliver 100kA+ into a low enough resistance, so a gate driver is quite superfluous for simulation purposes; but you won't get the 10V Vgs(on) you presumably wired up, anyway.

Ed: also, you have the PULSE set for 1 nanosecond rise time, 10^-90 (!??) s fall time, 1s pulse width, and 40kHz repeat rate I think??  That seems a little bit odd, too...

You certainly won't get a useful approximation of 1ns rise time, with a 1ms simulation step.  It may slow down to evaluate it better, but that huge of a ratio may simply fault out instead (timestep too small).

I don't know what PSPICE errors mean, if they're more verbose than most (sadly, most simulations are literally two kinds of fault: timestep too small, or singular matrix; absolutely no help whatsoever).

The latter is the linear-algebra equivalent of divide-by-zero, a process which involves very unstable numbers, and could explain the reported overflow.  But I think that's a little unlikely, because (again, assuming the models are connected up right) I don't think you've committed the error that usually results in a singular matrix.  (So, at least there's that. )

Tim

[Mattjd]

Where do you see in the datasheet that the output voltage of driver is only going to be 2.6 given vcc and Vin? I mean that's what spice was showing but it says typical vout is vcc-0.8

[mikerj]

Where do you see in the datasheet that the output voltage of driver is only going to be 2.6 given vcc and Vin? I mean that's what spice was showing but it says typical vout is vcc-0.8

Look closely at the test conditions!  Vout ~ Vcc-0.8 only when Vin=Vcc.  You are only applying 3.3v to Vin.

[Mattjd]

So I need to amplify the pwm signal from micro with a mosfet then, it's a real same we already made the pcb for this.

[Mattjd]

So, I'm used to a circuit that amplifies the PWM signal, and inverts. That voltage coming out of NMOS M1 should be greater than 3.3v, but its not?




[/img]http://

[T3sl4co1l]

Are those the DC averages?

Also, this is a transient analysis, right?  Viewing the waveforms may be more informative.

Tim

[Mattjd]

2.5v for the entire 500ms of the transient analysis