That was my first inclination as well. How is this "50A per phase" being measured? Cheap meters cannot always accurately interpret high harmonic loads. Even when not running, the drive is still rectifying into the DC bus and bleeding off some of that power to run the electronics themselves. If the meter is only seeing peaks, however short, it might be falsely showing something that is not really happening.
I like ozmosis' story too though. Using the analog signal as the run command was a common "trick of the trade" for automated pumping systems that had no automation intelligence, primarily agricultural pump systems where the users are notoriously cheap. In other words, the designer (typically a pump supplier) was too cheap to put IN a "Start" command device if there was some way he could get away without it! When people discovered they COULD do it with just the analog signal only, they jumped on it. But on occasion, it did create problems like what ozmosis witnessed; 0 speed is not really "Off". Now days, many drive mfrs have fixed the problems by adding what is referred to as "Sleep Mode" which monitors that exact situation and takes it to the next step of actually turning the VFD off if the input analog signal is below a threshold for a programmed amount of time. That was not always the case, it was developed by a pump supplier here in California for Saftronics (Yaskawa) drives about 10 years ago and Saftronics had an exclusive on it for years. That didn't stop the major players from developing their own versions however once they knew the need was there. But even now not all VFDs have that, I think the marginal players can't afford the R&D or have never chased that niche market so they are unaware of it. So if the system designer was PLANNING on using that feature, but the job went to the lowest bidder who used a drive that doesn't have it, that scenario is very plausible.
"Will work for (the memory of) salami"
