TL072 simulation gives incorrect gain curve

I’ve been getting to grips with the NGspice simulator and I simulated a basic TL072 op amp circuit. I used the Tl071.301 file from Texas Instruments, but when I simulate the -3dB point of the roll-off is somewhere around 50kHz, when the unity gain bandwidth of the TL071 is given as 3MHz on the datasheet. What is the reason the difference? am I doing something wrong?

0.0mdBV is quite close do zero.
So it does not work at all.
Always watch those scales!

From the plot you show, 50kHz is where you have the -3 mdB point, not -3 dB (note the “m” prefix).

Your vertical scale is very zoomed in. I think you need to increase the upper frequency limit of your sweep to something higher, say 10 MHz or 100 MHz.

ah! of course, i missed that. But how can I scale the axis? Must I increase the X axis range or can I just set it somehow?

From playing around for a minute, it doesn’t look like you can zoom out arbitrarily far. Once you get to a zoom level that shows you all of the simulated data, you can’t zoom out further.

So unless I missed something, I think you’ll need to widen your sweep frequency (X) in order to get an expanded Y scale in this case.

I am suspicious of the scale of dbV instead of dB.

One simple way of expanding your “Y” scale is to run the simulation to a few MHz.

What is this:
Those little square boxes hint at open connections.
To veryfiy if KiCad accepts a connection, you can drag a schematic symbol. Another symbol should move with it, or the wire gets extended.
Press [Esc] to cancel the drag, so the schematic symbol snaps back to it’s previous location.

U2B is not connected at all. ngSpice does not like that.

I am suspicious of the scale of dbV instead of dB.

dBV is what I get on the Y axis when I run an ac sim, so I don’t think that’s an issue