Spice oddity. Is it me or a glitch ?

[Mooly]

It comes up after about 240ms (I got rid of your 16 for Ncycles of V5) but it is odd that that LTspice is still thinking that your servo is there until you remove the op-amp (you can remove the voltage V1 and replace it with a wire and it will still start immediately).
You also get the delay if you replace the op-amp with another device,

I never thought of running it for longer. Weird isn't it !

 

[Andrew Eckhardt]

I don't think you'll want to keep the diode. I don't really know what's going on here though I notice if you load the current source to ground or supply with less than about 175k, that'll get it to run too. I was just poking around trying to figure out why it wants to reverse bias Q1 right out of the gate.

I don't think LTspice gets reminiscent..

 

[Mooly]

Thanks Andrew.

Seems it really is one of LTspices' little (of many 🙂) oddities I guess. All good fun.

 

[godfrey]

Hi Mooly

On an unrelated note, aren't M3 and M4 living dangerously? If the output voltage exceeds +-15V, their reverse protection diodes will start conducting and short the output to the +-15V rails, or am I missing something?

 

[Mooly]

Hi Godfrey,

Excellent excellent point and tbh I hadn't even thought of that at this stage... thanks. So no, it's me that's missing something on that one.

It's all just a doodle at the moment for helping me learn a bit about LTspice and trying weird ideas. In simulation the HEXFET's take over quite nicely.

Hmmm... I guess a couple of diodes 'aint going to cut it for this one.

 

[godfrey]

A couple of high speed rectifier diodes in series with the drains of M3 and M4 should work, assuming you want rail-switching.

 

[cbdb]

Welcome to the simulation zone. (you might want to change your signature)😀

 

[Mooly]

Welcome to the simulation zone. (you might want to change your signature)😀

😛

A simulation experimental zone, simulate it, shout at it, burn a few virtual components, shout a bit more... then go build it for real 🙂