What should the stepgen signal look like?

I'm still trying to get setup, and all I've gotten so far is closed-loop stepper motors shaking like a back massager, with one occasionally moving in the commanded direction (while still vibrating).

Putting an o-scope on the pulse signal going into my CL86T driver, I was expecting to see a nice 50% duty-cycle square wave.  Instead, what I'm seeing is 5 volts, with pins pushing down to ground on an evenly spaced basis.  

Is this expected, or do I have some timings screwed up?  I had the drivers wired up for a differential signal, but was getting nowhere.  I switched to a common-anode configuration when I saw that the differential setup wasn't in version 4 of the CL86T documentation.  That is when I got some motor movement in the right direction.

Stepgen signals are pulses of the programmed width (steplen)
(in mode 0 = step/dir)

They are only square waves in mode 2 (quadrature)

Is steplen the same as "Step Space" or "Step On-Time" in the "Stepper Info" box of the motor configuration page?
'Cause, I have both of those maxed out.

pncconf calls it Step On-Time
( Though the actual HAL parameter name for Mesa hardware is steplen)

Normally this should be set to the drives minimum step length *1.5 or so.
Setting it longer than necessary limits the maximum step rate and has no benefit

Step Space is the minimum time between step pulses

The only numbers I can get from the documentation is a max Pulse Input Frequency of 200kHz, and that the pulse width should be set to 2.5 micro-seconds or longer.

I think I could make a nice contribution to PncConf by just researching and making sure that there are units on all the inputs, and maybe doing some pop-up helps dialogs or maybe even entry cross-checks.