The built in revolutions-per-minute unit is unclear to me.
The description of the "rpm" unit is "Revolutions per minute", but the definition is in "Radians per minute".
As a result, when converting an operating frequency from "rpm" to "Hz', the calculated result is "radians per second" instead of "revolutions/cycles per second".
For example, if I have a component that operates at 900 rpm, it completes 900 cycles per minute, or 15 cycles per second = 15 Hz. That is the result I expect.
Instead, I get 94.248 Hz. The value of 94.248 is Radians per second, not cycles per second.
On the other hand, the description of the "radpm" unit is "Radians per minute", but the definition is in "per second". So my 900 revolutions-per-minute component is completing 900 x 2 x pi = 5655 Radians per minute. I expect 5655 radpm to equal 15 Hz, but instead I get 900 radpm = 15 Hz.
Is this a bug in SMath or in my understanding of units?

There is nothing wrong with SMath regarding unit definitions.
Sorry but your unit understanding is wrong.
In SI, 1 revolution (rev) = 2*π, not 1, hence you make this fault.
900 rpm = 15 rev/s = 15 rev*Hz, not 15 Hz.
This has been debated 2 years ago.
https://en.smath.com/forum/yaf_postst21432_rpm-to-rad-s.aspx

Regards

See page 22 in https://doi.org/10.25933/OPUS4-2947

The text is in German but translation should be a simple task for any decent AI.

I'd not say that it is your fault, because SMath is a bit sloppy with telling frequency from angular velocity.

If you use unit rpm then you want to define a frequency (events per time, with the event being a full rotation). If you use rad/s then you want to define an angular velocity.

In this case the default conversion to Hz by SMath is wrong.

I recommend to use min^-1 instead of rpm. Direct specification as angular velocity saves you later conversion but you become inconsistent with textbooks which always have explicit conversion between frequency and angular velocity by the factor 2 pi.

Near all CAS software has this conversion.
It is not a bug and not specific to SMath.
Mathcad has same units.

Regards

https://support.ptc.com/help/mathcad/r7.0/en/index.html#page/PTC_Mathcad_Help/frequency_and_angular_velocity_units.html

Thanks for the replies. I'll read the other thread carefully.
I just checked Mathcad and it indeed behaves the same way.

Wrote

Near all CAS software has this conversion.
It is not a bug and not specific to SMath.
Mathcad has same units.

SMath shows Hz as default unit for results with dimension 1/time. This is correct, if the quantity is a frequency but it is wrong when the quantity is an angular velocity (or a strain rate)
As SMath has no way to know what the actual meaning of the quantity is, the only correct default unit for all cases is 1/s but definitely not Hz. So I still think it is a bug.

Also, it would be helpful if the units dialog would not say that radpm is a unit of frequency instead of angular velocity.