CNC power supply - is there a better way?

[erikg]

Hi, all;


  I'm reworking the electronics on a CNC system.  The original system used small stepper motors but was obviously designed for a range of motion hardware - I suspect the "high end" model would have used servos instead of steppers.  So, I'm refitting it with servos since it doesn't work right now anyway.  I'm using slightly oversize (ok, a lot oversize) servos so I can directly drive the ball screws... that way I get no backlash from belts, gear drives, etc.  So, oversize servos and servo drives to match... each axis servo (X, Y and Z) needs about 3 amps continuous at 230VDC, peak is about 10A each.  I'm going to (approximately) size the power supply to handle normal continuous load on all three axes.  With the draw from the servo drives' electronics added, call it 10 amps at 230VDC, or about 2.3KVA.  I've done this before on a milling machine and helped other folks with CNC machines using similar power supplies - basically, a linear supply with a large step down transformer feeding a rectifier and enough capacitors to smooth out ripple.  I feed this to a DC bus with a clamping regulator on it to avoid overvolt conditions.  Linear supplies have been recommended to me a lot for CNC use because they tolerate the braking voltages and noise associated with driving servos.

  So, where I'm at right now is building another supply.  I have all the parts except the transformer in my stockpile.  The problem is coming up with the proper voltage - 230VDC means I need a transformer with a 240V single phase primary and a 160 volt secondary.  Not a terribly common item, especially sized for 2.5KVA.

  Here's my question - I'm getting set to disassemble and rewind a transformer to produce the voltage I need for rectification, and before I get started with all that work I wanted to see if anyone could think of a better way to build this power supply.  For example, does someone out there sell a CNC supply this size for a decent price?  Maybe 3x supplies of 750VA instead? 

  Does anyone know of an easier way to get a solid 230VDC from 245VAC mains current @10 amps?



 

[jmelson]

Well, if you use a 2:1 step-down transformer from 245 VAC, you get 122.5 VAC, then through a full-wave rectifier and capacitor-input filter, x 1.414 = 173 V DC, which might be close enough.

Otherwise, you are pretty much having to find or make a custom transformer.

Jon

[alpher]

2 buck transformers connected in series should give you ~180VAC at their combined output.
You should pick 2 x 120V-16/32V transformers that have their primaries combined with the series connected secondaries as close as possible to 120V, join 2 of them in series and you should  get primaryVAC - 64VAC easyli, multi KVA are cheap and of the shelf at most electrical suppliers.
Marcus or Hammond brand are most popular here.
Hope this helps.

[erikg]

Awesome... never thought of doing that

[erikg]

Just a quick update to let people know how this came out.

Looking through the various options, I found a transformer pulled from an old piece of equipment that was labeled on a sticker as producing (two taps) 84VAC and 14VAC.  I noticed that the H side had jumpers on it for HV/LV and they were set to LV   So a little rewiring and a test later and I have a transformer that accepts 245VAC in and produces 168VAC out.  Running that through a rectifier gave me 236VDC out, which is close enough to what I needed.  I am going to keep the above posted suggestion in mind for the next power supply I build, though

Putting a capacitor bank on this sized for 3 servo axes on the output of the rectifier (17600uf) gave me a real inrush problem.  I could have pursued an NTC thermistor solution, but that would require a rather large thermistor and I'd need to switch it out of the circuit after start up to avoid lots of loss anyway.  So I settled on a simple resistor solution - I installed a large resistor to limit the inrush (57R/200W) and a contactor to remove it from the circuit.  I use a packaged time delay relay with a small step down transformer attached to the mains feed in parallel with the large transformer and set it for 1 second to toggle the contactor.  Circuit startup now works without blowing breakers, and I end up with my desired DC voltage on the cap bank.

Next steps will be mounting the whole thing in a steel enclosure with appropriate fusing, and indicators, a clear plastic cover for the cap bank, and warning stickers.  Then I can finally move on to configuring the servo drives

[metrologist]

If you have a chance to do this again, you might consider AC drives and skip the power supply altogether. When I was rebuilding my CNC, Leadshine drives were mentioned, along with another name I've forgotten.

[langwadt]

Just a quick update to let people know how this came out.

Looking through the various options, I found a transformer pulled from an old piece of equipment that was labeled on a sticker as producing (two taps) 84VAC and 14VAC.  I noticed that the H side had jumpers on it for HV/LV and they were set to LV   So a little rewiring and a test later and I have a transformer that accepts 245VAC in and produces 168VAC out.  Running that through a rectifier gave me 236VDC out, which is close enough to what I needed.  I am going to keep the above posted suggestion in mind for the next power supply I build, though

Putting a capacitor bank on this sized for 3 servo axes on the output of the rectifier (17600uf) gave me a real inrush problem.  I could have pursued an NTC thermistor solution, but that would require a rather large thermistor and I'd need to switch it out of the circuit after start up to avoid lots of loss anyway.  So I settled on a simple resistor solution - I installed a large resistor to limit the inrush (57R/200W) and a contactor to remove it from the circuit.  I use a packaged time delay relay with a small step down transformer attached to the mains feed in parallel with the large transformer and set it for 1 second to toggle the contactor.  Circuit startup now works without blowing breakers, and I end up with my desired DC voltage on the cap bank.

Next steps will be mounting the whole thing in a steel enclosure with appropriate fusing, and indicators, a clear plastic cover for the cap bank, and warning stickers.  Then I can finally move on to configuring the servo drives

don't forget a bleeder resistor and maybe a warning light, getting zapped by that thing is not going to be fun

[erikg]

Thanks... yes, I have a bleeder resistor picked out, and I'm planning on three warning lights for the power supply at the moment - "Switch On", "Power Good" and "Capacitor Voltage Present".  I may actually replace the third one with one of those cheap Ebay voltage meters.

Plus I'm planning a warning light on the gantry for when the control is active and the thing can start moving on its own.