Now, regarding the missed steps, Lewis to be honest I don't know if my system started to lose them right after a Mach software update. I did go from "Mach3_Sherline_V2.04" to R3.043.066 long before I noticed the problem, but I've used my mill rarely these past few months and I don't recall any recent intense direction reversal operations like peck drilling that would have made the problem evident. In my troubleshooting I did try to revert to the older software and this did not resolve the issue.
I've attached some more scope screenshots that show what may be causing the problem in my system. "DirectionalSignal" shows the direction signal coming from the parallel port in green and the direction signal at the input to the microcontroller in the Sherline control box in yellow. The scope is capturing the change in the direction signal from high to low as Mach is trying to reverse the motor direction. You can see how the input signal to the Sherline microcontroller rolls off by about 40 usec compared to the raw output from the computer. Since the next motor pulse occurs about 40usec after the direction change (not shown) I'm guessing that the rolloff in the direction signal confuses the Sherline microcontroller into pulsing the motor in the pre-change direction.
I reduced the rolloff by decreasing the resistor value in the RC filter that separates the Sherline microcontroller input from the raw signal from the parallel port (from 1000 Ohm to 100 Ohm) and the improved signal can be seen in the "DirectionalSignalSharp" image. Again the raw parallel port signal is green and the one at the Sherline microcontroller pin is in yellow. A picture of my lousy resistor soldering job is also attached. The rolloff at the input to the Sherline microcontroller is gone and the system now seems to be working without losing steps. The milling machine is up to ~2000 and counting rapid direction reversals without any step loss.
One last consideration if you are using the Sherline controller box, the power coming out of the box does seem to decrease with time. For instance the hold torque on my Z channel has dropped off quite a bit so I am now using the A channel to run my Z on the mill. From past experience replacing the SLA7044M chip solves this problem. These chips get pretty hot and I'm wondering if there isn't enough cooling going on in the box....
Hope this helps!!
Freddy