I did using a transistor and a green LED but the transistor had a low Rbe which helps so the limiting factor was the HFE, two of the circuit was used and the ref voltage the green LED was common giving me a random noise and a 3dB cut. There are 1000s of these circuit in Magnum's working well, at the time LM741 was the class act op-amp and NE5532 series was to be honest shit.
The worst source of noise was the PSU, but since that is my design skill set and I had to work on PSU with 10nV noise floor at lees than 1Hz no real problem. And some at 55V 100,000Amps (battery charger not for mobile phones) using transistors and thyristors and transformers at 7 English tons, tiny 1mA PSU's are easy.
Fun days and big bits no real SMD and pebble glasses to wear, now I am 65 they seem to be getting thicker every month, or are they cheating and making them smaller ?
Oh I used spice in this case only to act a simply drawing package, in real life there are 2 NPN and 2 PNP transistors in each stereo pair of constant current source on the same substrate and the design has been working for 40yrs, oh when I took it to Rotterdam I had the response by the Prof this can't work, oh dear I took the prototype out of my briefcase and guess what it works.
So Spice was not used and at the time there was no Spice only in Curry. he he