Quote Originally Posted by fasteddy View Post
As I have got you here, what is the voltage rating for the 3Nm nema 23 motors you sold me? Has been suggested I may need to upgrade my
24v PSU for them...
I run mine with the same PM752 drivers on 70V, so you can easily increase yours which will make a huge difference to the speeds you get. You can use up to 50V for the motors on the PM542... 48V is common and convenient. Easiest option is probably to get a big toroidial transformer to run both on 48V, but it would be better (faster) if you had two transformers to get 48V and 70V. There used to be someone selling lots cheap power supplies on eBay which were just under that voltage and ideal for this application. They've all suddenly gone now but you can buy them from Zapp for several times the price.

I did have a similar problem to yours at one point...
Since you're saying that X and Y both go at highest speed when moving at 45° you must have their feeds set the same. On my router I started with X a little slower than Y, so to get both to move at highest speed (which makes the actual speed the vector sum of X and Y maximum velocities) I had to move it at a different angle. At that angle or close to it the machine would often stall. The first thing I tried was to reduce the feedrate further on X from about 10m/min to 8m/min then 7m/min which each helped (stalled less often), but that was getting a bit slow...I had a spare transformer, so added that in to power the Y (and A) axis and left both the X's and Z on the bigger transformer. That makes the two axis independent, so from the driver's and power supplies point of view it's no different if one is drawing a lot of current or both. The voltage wont drop.
This didn't make a lot of difference, so I analysed both set-ups with oscilloscope. I can't find where I saved the graphs, otherwise I'd post them ... but it did show there was some ripple (couple of volts I think) with both setups and the ripple was slightly less with them split. Still evidently it was not enough to make much difference anyway.
Since that implies the power supplies were only a small part of the problem I decided to finally get round to swapping the computer to a better one I had (1.4Ghz AMD to a 3Ghz something or other), which made a massive difference. Feedrate on X went up from barely reliable at 7-8m/min to fine at 15m/min! It almost doubled the feedrate ... purely due to the computer's evidently better parallel port as I kept everything else the same. Since then I have increased the kernel frequency to 35kHz as that's sufficient to use 1600step/rev. You only need to use the minimum kernel speed to get the feedrate you require at the chosen microstepping value, so having it at 45kHz versus 25kHz is pointless and unnecessarily unstable if the former is sufficient.

Putting a bigger capacitor on the power supply may help, depending on what the value is currently?
Definately use the driver test in mach3 to check the parallel port. Perhaps try another computer?