Losing Z steps

It has been mentioned that 24 volts is low ( I would say too low). I had similar issues when I first started playing with steppers and all of my problems went away when I changed from 24 volts to a 48v power supply. Most systems run on 48v-80 volts. What is the maximum voltage the gecko can handle? If you are not up near that number, you are leaving torque on the table.

I know there is a lot of theory in calculating the voltage to use but its not rocket science. Just change your power supply before you do anything else. I never did the maths and I never have had a problem at 48 volts. Also somewhere on Gecko's site, they say if you sum the total rated amps on the motors, you can exceed the power rating of the power supply by 40%. Probably a 7 amp power supply and increasing the current limiting to be equal to the maximum the motors are rated for will work out pretty nicely. I've got 4 steppers and 2 of them are NEMA 34's so I upgraded to a 10 amp supply.

The gecko are 17 to 80 volt for most of them

No its not. The Gecko 540 is rated for 3.5 amps at 50 volts.
www.geckodrive.com/g540-4-axis-digital-stepper-drive.html
Perfect for a 48 volt power supply. I know thats what the Aussie Gecko distributor sells and recommends for it.

Thank you all for the very quick reply.
I am running 48V with a 7.5A supply. I calculated the voltage for the steppers an it cames back to 52V of so.
The current limit resistor is a 3.5kOhm which is the current limit of the steppers. Could the additional wire resistance play a role and cut into the voltage overhead? Maybe I am not even getting the 3.5 A on the steppers.
By time do you mean the period (high and low timing) for the communication FPGA to G540 ? If not can you elaborate on that?

Yes the step time is the active time of the step pulse. There will be an .ini file entry
with this parameter for each axis, I suggest doubling this value