UCCNC Plasma - constant velocity output ?!?

[cncdrive]

There is no mistake, it is cased on your faulty understanding of the circuit.
Chinese controllers using cheap hexa output drivers usually which can output only a few tens to a hundred of milliapmps of current per channel.
The AXBB-E using high power mosfet transistors which can output half Ampers per chappens.
So, a very big and expensive DC/DC converter would be needed to make it able to supply that much current.
If the DC/DC power supply would work in the inverse direction from the 24V to the 5V side then it should still handle at least 1Amps which is again a bug DC/DC converter which is not cheap.
So, the easiest is to use separate power supplies.

The AP2401 is an overcurrent protection IC, if you shorting out the ground to the 5V pin on the IDC26 port 3. then it would possibly burn the tracks off the PCB, the same if an internal short circuit happens for any reason.
So, that IC protects agains overcurrent, it shurts the output down in latch if there is a 2.5A or more output current.

Placing a 7805 would not even work as it makes 5Volts and it has an about 2V voltage drop, so then using 5V power would drop it to around 3V and it would be unregulated, because then the output voltage is under the regulation voltage.
And a 7805 does not make an overcurrent regulation, it makes a voltage regulation, if it gets an overcurrent then it gets hot and can burnout.
Also the 7805 does not removes the need for a second powersupply.

The only problem with the 2401 is that it can damage from overvoltage.
In the next revision will planned to place a shottky diode and a fuse before the AP2401 IC which would then cut the fuse in case of overvoltage.
So, then there will be not only overcurrent protection, but also a protection against user error overvoltage on the input.

[stivemaster]

Interestingly, this controller is still on the test table and it does not include any inputs and outputs. I observe its work only on the lighting of the diodes on port 1.
Well, it probably happened to me to be defective, I'll survive it. But I have to keep using Sheetcam to stop and start the THC, because the function you developed works in a very limited range of accelerations on the X and Y axes.
Even without a relay to slow down further, I don't think it can maintain a high speed, adjusting the THC along the Z axis. What I see with the naked eye is that the distances at which there is no constant speed do not change when the percentage changes under which this has to happen. I don't know if this is really the case, I don't have the ability to measure times.
I don't know if you understand what I mean. If you try to work at such a speed and adjust the THC, then the distance from the burner to the material is about 30ms. Something slows down the reaction around corners and in most cases the Z axis jerks down at the wrong time!

https://www.youtube.com/watch?v=txGP-Qbii2E

[cncdrive]

I don't know you, so I have no idea if you are defective or not, but your understanding of the circuit is incorrect for sure.

The THC up/down does not have a limited acceleration range it works basically instantly.
I don't know why it delays at you.

[stivemaster]

I meant the accelerations on X and Y. When they are big the machine is fast and the places where UCCNC stops the output for constant speed and starts it again are very small and obviously do not correspond to the actual speed. after releasing it after the turn, the THC makes the Z axis jump slightly. This means that the output is switched on even before the speed is the set percentage. I put 99%.