First, a short history of how I came to have these cards:
-Started with a 2GB 750Ti in my son's first gaming PC 6 years ago.
-A few years ago, upgraded him to a GTX 1060 6GB and moved the 750Ti in my own build.
-in January 2020, I upgraded my 750Ti to a 1650 Super 4GB.
-this month, upgraded his 1060 to an RTX 2060 KO 6GB for $305 shipped.

I tested these 4 cards in my main PC. It's an HP EliteDesk G1 Tower. Bought for $350 as an off-lease refurb a couple years ago and upgraded slightly with existing spare components. Specs:
i5-4590
12GB DDR3
500GB SSD + 2TB WD Green
onboard Realtek sound
320W PSU (proprietary, non-upgradeable, using SATA-to-PCI-E power cables as the PSU does not have PCI-E power cables)
1920x1080 monitor resolution

First, some total system power numbers, measured at the outlet -
-All four GPUs yield about 30W of system power at idle.
-CPU Burner gets the system power up to about 75W
Furmark GPU Stress Test:
RTX 2060: 218W
GTX 1060: 202W
GTX 1650S: 144W
GTX 750Ti: 118W

Now, some synthetic GPU benchmarks:
Userbench GPU Score -
RTX 2060: 83.5
GTX 1650S: 58.1
GTX 1060: 55.3
GTX 750Ti: 19.1

Passmark 2D GPU Score -
GTX 1060: 689
GTX 1650S: 663
RTX 2060: 662
GTX 750Ti: 636

Passmark 3D GPU Score -
RTX 2060: 10647
GTX 1060: 8985
GTX 1650S: 7807
GTX 750Ti: 4682

3DMark Time Spy (DX12) GPU Score -
RTX 2060: 7272
GTX 1650S: 4580
GTX 1060: 3931
GTX 750Ti: 1284

3DMark Firestrike (DX11) GPU Score -
RTX 2060: 18726
GTX 1060: 11955
GTX 1650S: 11571
GTX 750Ti: 4458

Gaming Benchmarks:
Final Fantasy XIV Benchmark, Max Details (Avg. FPS) -
RTX 2060: 96.0
GTX 1650S: 75.6
GTX 1060: 74.4
GTX 750Ti: 31.1

Final Fantasy XV Benchmark, High Details (Avg. FPS) -
RTX 2060: 72.9
GTX 1650S: 50.7
GTX 1060: 46.1
GTX 750Ti: 16.6

Final Fantasy XV Benchmark, Standard Details (Avg. FPS) -
RTX 2060: 90.1
GTX 1650S: 68.2
GTX 1060: 61.6
GTX 750Ti: 24.2

System Shock Alpha Demo 1.22 Ultra Details (Avg. FPS) -
RTX 2060: 103.6
GTX 1650S: 71.5
GTX 1060: 60.3
GTX 750Ti: 22.4

System Shock Alpha Demo 1.22 Medium Details (Avg. FPS) -
RTX 2060: 160.7
GTX 1650S: 147.2
GTX 1060: 142.5
GTX 750Ti: 61.3

DiRT Rally Benchmark Ultra Details, 8X MSAA (Avg. FPS) -
RTX 2060: 78.11
GTX 1060: 77.47
GTX 1650S: 71.79
GTX 750Ti: 28.14

Grand Theft Auto V Custom Max Details (Avg. FPS) -
(4X MSAA, 4X Reflection MSAA, ReflectionQuality, GrassQuality, PostFX set to Ultra, everything else set to Very High)
RTX 2060: 91.2
GTX 1650S: 68.7
GTX 1060: 68.7
GTX 750Ti: N/A (not enough VRAM)

GTA V Custom Max Details (% frames over 60) -
RTX 2060: 98.27
GTX 1650S: 70.32
GTX 1060: 68.67
GTX 750Ti: N/A

GTA V Custom High Details (Avg. FPS) -
(MSAA disabled, FXAA enabled, all details set to Very High)
RTX 2060: 113.9
GTX 1060: 107.0
GTX 1650S: 105.2
GTX 750Ti: 51.9

GTA V Custom High Details (% frames over 60) -
RTX 2060: 99.47
GTX 1650S: 98.99
GTX 1060: 98.54
GTX 750Ti: 4.61

But...does it run Crysis? (all details maxed, integrated GPU benchmark, Avg. FPS) -
RTX 2060: 71.85
GTX 1650S: 66.91
GTX 1060: 66.71
GTX 750Ti: 26.46

Some thoughts/analysis:
-The 1650S and 1060 are very similar in performance. The 1060 does slightly better in DX11, the 1650S does better in DX12
-The 1650S is much more power efficient, easily the best option for my 320W PSU (which is why I picked it to begin with).
-The i5-4590 does appear to bottleneck performance in GTA V and DiRT Rally, especially with the details maxed out.
-The Passmark 3D benchmark has separate DX9, DX10, DX11, and DX12 tests, which are then averaged into a score. This is probably why the 1060 did so much better than the 1650S, since the DX12 score was only 1/4 of the total score.
-The System Shock Alpha demo was benchmarked with MSI Afterburner, running through the same basic script that I manually determined. The results were repeatable enough that I trust them to be accurate for purposes of this comparison.
-The Crysis benchmark is notorious for its poor multi-core support, but was just thrown in for fun.
-The 750Ti is remarkably usable even today if you run on low or medium details.
-The RTX 2060 is significantly faster than the GTX 1060, even on a CPU-bottlenecked system like mine. On my son's i7-4790, the difference is night and day. He can run every game on Ultra/maxed details at well over 60 fps. The 1060 would often dip in the 30-45 fps range with some newer titles like Red Dead Redemption 2 and even games like GTA V with details cranked to the max.
-Quake II RTX is INCREDIBLE!!! It only runs about 40-50 fps but is so amazing to behold and makes an incredibly fun FPS even better. If they can ever get ray tracing in modern games tuned with more efficiency, it's going to make a great card even better.