Electronic – AD620 in Proteus: timestep too small

I've made this circuit in Proteus, as a part of a practice report:

In real life it works, as I built and tested it. But in Proteus, when I want to run the simulation, I get these errors:

[SPICE] transient GMIN stepping at time=3.41026e-006
[SPICE] TRAN:  Timestep too small; timestep = 1.25e-019: trouble with instance d:u1:3.

I realized the error disappears when the gain of the amplifier is 1: that is, when I remove the resistor R1 (letting it be "infinite"). Otherwise, it won't work.

Why do I have this error, and how can I solve it?

UPDATE # 1

As @Andyaka pointed out, I tried with very small (as Proteus complained about 0 Ohm) resistance values. The circuit I got is this one:

But when I run the simulation, I get other errors 🙁

[SPICE] Gmin step [45 of 120] failed: GMIN=4.21697e-007
[SPICE] Gmin stepping failed
[SPICE] Source step [12 of 120] failed: source factor =  10.0000
[SPICE] Too many iterations without convergence.

UPDATE # 2

I tried the procedure @berto said in the answers section. I connected a load resistor of \$10k\Omega\$ to \$V_{\text{out}}\$ and the ground, and then connected the voltmeter in parallel with it, but I got these errors:

[SPICE] transient GMIN stepping at time=2.14996e-007
[SPICE] TRAN:  Timestep too small; timestep = 1.25e-019: trouble with node #b:u1:e3#branch.

Two interesting facts when I played with the circuit at this stage:

1. When I disconnect the voltmeter and put a load resistor of \$0 \Omega\$ (thus, a connected wire from \$V_{\text{out}}\$ to the ground), the simulation starts and continues with no errors.

2. When I disconnect the voltmeter but leave the proposed load resistor (\$10k\Omega\$) from \$V_{\text{out}}\$ to the ground, the simulation stops with the same timestep too small error.