Why is not working with Pseudo GND?

[SpoNNw]

Hi, I’m working on a small project to capture EMG signals, and I plan to read them using a microcontroller’s ADC. In a very general overview my setup includes an instrumentation amplifier (AD620ANZ), a second-order bandpass filter to capture frequencies from 20 Hz to 500 Hz, full wave rectifier amp op, and last a peak detector. However, I need a dual-rail power supply (+12V and -12V) to power my ampOps, the ideal solution would probably be using a dual power supply, which I don't have. After doing some quick research, I found that using a voltage divider and a buffer configuration might be sufficient to generate the dual-rail voltage. The issue I'm encountering is that my differential amplifier works perfectly fine with V+ and V- supplied from dual sources of -12V and +12V, along with their respective ground. However, when I apply a voltage divider and a buffer configuration and generate a PseudoGND, something strange occurs in my simulation circuit (LTspice); I might be doing something wrong! I expected to have +12V and -12V connected to the supply terminals, instead it, I'm having -24V at V+ and -48V at V- ! Can someone pls help?

https://prnt.sc/i4rFAIj2ZGAT Dual Supply (Working fine)

https://prnt.sc/loR_ikVnlpnk GND configuration

[ledtester]

It would help if you could post the .asc file and the relevant component models.

What happens when you just use the UniversalOpAmp model instead of a LM741 for the rail splitter?

I have a feeling that SPICE doesn't handle this case very well -- attaching GND to the output of the LM741. I would just attach ground to a more conventional location and subtract the value of PseudoGND from whatever voltage node you want to plot.

[Zero999]

Yes. It's a problem with the models. The circuit will work.

UniversalOpamp doesn't model the supply volage pins and currents. Use UniversalOpamp2, which does.

To change it so it approximates the LM741, right click on the symbol and replace the default peramaters with the following:

Avol=200k GBW=1Meg Slew=500k
Ilimit=25m Rail=2 Vos=0.8m
En=0 Enk=0 In=0 Ink=0 Rin=2Meg

[SpoNNw]

Yup, I can confirm, it's definitely a problem with the models. It's working now, thanks!

https://prnt.sc/lJ2ykn0gpF9R Simulation