Re: [CAD_CAM_EDM_DRO] PSU choose

g bon wrote:

> I'm not really shure of how I should correctly choose my servo power supply.
> Can someone tell me if I'm right or not ?
>
> My 3 motors are rated at 72V - 38A (in peak) - 8.5A (continuous)
>
> I don't knows how are labeled power supplies.
> For example, www.kelinginc.net sell 72V - 20A PSU.
> Is the transformer 50V and output equal 50x1.41 = 72 V
> Or, is the transformer 72V and the output is 72 x 1.414 = 102 V
>
> If I choose a transformer of 50V - 20A, my output will be at 50 x 1.414 =
> 72V
> Is this ok for 72V motors ?
> Or transformer should be less than motor maximum voltage ?
>
> Then, capacitor should be 100V and C=(80000*20/50) ?
> Should I replace 50V in this formula with 72V ?

Hoo, boy, you don't ask the easy ones.

If you have three motors rated at 8.5 Amperes continuous, your power
supply had better be able to handle 3*8.5 or 25.5 Amperes continuously
or the magic smoke will escape.

Power supplies are rated by output voltage. A 50 Volt power supply
should have 50 Volts coming out of it (well sort of, there are things
like regulation and ripple which come into this.)

Depending on the design, the transformer could be 52 Volts or 102 Volts
CT. Huh? Using a full wave bridge rectifier you have 2 diode drops to
deal with, so the 52 Volt transformer only puts out about 50 Volts after
subtracting the diode drops. Using a diode pair and a center tapped
transformer, you will see only one diode drop, so 1/2 of 102 is 51 Volts
and after a diode drop you see 50 Volts.

If you have a choke input supply, all this changes!

If you have a PFC supply, all this changes.

Your formula may work, but you need to realize that you will be able to
pull 38 * 3 or 114 Amperes at 72 Volts peak. You will need diodes
capable of handling over 100 Amperes EACH! The transformer will droop,
but thermal lags on the diodes are small. Now, if your output swings 22
Volts on each cycle at 114 Amperes, you need C=I*dT/dV, just for round
numbers I=114 dV=22 dT=1/120 so you will need 114/120/22 = 0.043 Farads
or 43,000 uF at 100 Volts. Smaller voltage swings will need bigger
capacitors, but the capacitor is supplying 114 Amperes when the output
of the transformer is below peak, so the capacitor ESR becomes
critical. Let's say the capacitor has an Effective Series Resistance
(ESR) of 25/1000 of an Ohm. 114*25/1,000 = 2.85 Volts of additional
droop, you will see 25 Volts of ripple! Hmmm, time to look for half a
million microfarads and an ESR of a milliohm or two. Then the supply
won't cause /as many/ servo faults during normal operation when you have
X, Y, and Z rapids at the same time.

Of course if you are designing a system from scratch, you can not allow
full acceleration on all axes simultaneously to save power supply loading.

BTW, how are you feeding the supply? Even with 100% efficiency, you
will need 8.2 KW input at times, that is 37 Amperes at 220, better
figure 40 Amperes at 220 peak. Now, if you can go 3 phase, you can get
away with a lot less capacitor, but you need 3 phase power and more
diodes.

The realities are that this is a big, expensive supply and it will be a
bit daunting for the average hobbyist. It will also have power levels
which can kill you. Build safety interlocks and /*do not */bypass them
unless you know /*exactly*/ what you are doing and why.

--
David G. LeVine
Nashua, NH 03060