Re: [CAD_CAM_EDM_DRO] Re: Sorting out the options for a CNC mill system

The G540 can only handle 50v. So a 48v power supply is maxing it out.

Most of what wildhorse is saying is gobbledygook to be honest. Lemme
clarify that.

In designing switching power supplies, the feedback to adjust the
current's switching without overshooting or undershooting IS a
difficult problem. This is why you won't be able to design a switching
supply yourself. However, in any professionally designed supply, it HAS
been done. The supply voltage will be very very consistent across any
current switching that device can do. And of COURSE they have an output
capacitor! Its internal capacitor does not have to be large because
it's refreshed at at least 100,000 times a sec rather than the 120 times
a sec of a 60Hz supply, and has tuned inductances around it. Adding an
external capacitor is neither necessary nor usually even desirable.

These supplies are designed to run digital equipment which switches
current on and off quite rapidly. Also they're not designed for one
specific type of load. So they're designed to be able to take just
about anything and still provide a tightly controlled voltage.

I'll scope mine out later tonight just to say I did. If my G540 causes
the power supply to have more than like a volt or two of variation when
running, fine, I'll eat my words (I don't expect to see more than a
quarter volt honestly). If there's no spikes in the regulation, well,
this IS the exact configuration I recommended, a very common type that
is applicable to most people's situations, which you're saying could
have problems. If it doesn't have problems, then, well, there are no
problems with this configuration.

On the other hand:
The big 60Hz supplies suffer from 2 major problems. One, the
transformer does not compensate for varying line voltage, which may vary
from 125V or 110V. So any attempt to get 48V out gets 42v-48v, not even
including power drops and spikes when the house's AC kicks in or whatnot.

Second is that cap DO have ripple. They must. The way to combat it is
higher capacitance, but cap sizes become unreasonable both in terms of
capacitance and ESR requirements, and then this configuration starts
causing significant bad harmonics on the power line. The way ripple
works, you will get a significant voltage difference when the supply's
loaded and unloaded.

Due the significant variations in possible voltage, you'd either have to
use a voltage way under what your controller could take to ensure that
its peak voltage won't damage it, OR use a linear regulator. No
off-the-shelf linear reg will handle the voltage and current required,
and the heat generated burning up the voltage difference is tremendous.
Even still, there's a big concern that the ripple under load, when line
voltage is low, will bring down the voltage enough to make the output
voltage fall.

These home-designed supplies are not virtuous like people think.
They're huge, expensive, big on heat, bad on the power line's harmonics
and power factor, and typically have significant design flaws. Most
significant is not getting enough ripple rating in the capacitor to
perform the job reliably. Many amps of ripple rating is again very
expensive and large. Even the big computer caps need to be checked as
far as what they're rated at vs what they're actually going through.

Danny

The 60Hz transformers (

wildhorsesoftware wrote:

> Douglas,
>
> I'm going to differ with Danny on the matter of power supplies.
>
> First, you should use the highest voltage that your controller will
> accept. Second, unregulated, linear (like the one described at the
> hobbycnc.com site) are better for CNC operations.
>
> Here is the reasoning behind my statements.
>
> Virtually all controllers on the market today are "choppers". They
> work like this:
>
> The output of the controller is set, in amps, to the value desired
> for the stepper motor being driven. (Controllers that drive multiple
> motors have an adjustment for each motor.) Notice that the setting
> is in amps, not volts. You can have a stepper motor that is rated at
> 8 vdc controlled by a controller that is fed with a 50vdc power
> supply.
>
> The controller has a feedback circuit that can "sense" the amount of
> current (amps) that is being fed to the motor. Now the next part
> gets a little complex. When a motor needs to step, the controller
> turns on the current to that motor (feeding current into the motor so
> that it moves the correct direction). But the current does not rise
> immediately to the desired value. The slight delay (or ramping up)
> of the current is a function of a complex relationship between the
> power supply voltage, the capacitance of the circuit and the
> impedance of the motor winding. You don't need to understand what
> all of that means, just understand that the varying values of the
> various components go into controlling how fast the current rises.
> This is where the "chopping" action of the controller comes into
> play. The feedback circuitry of the controller senses the amount of
> current flowing through the motor at any given moment. When the
> current reaches the desired value, the current to the motor is cut
> off.
>
> The controller goes through this current on, current sense, current
> off sequence 1000's of times a second (usually 20,000 times or more).
>
> Just as the current does not rise instantly, it does not fall
> instantly when cut off. Again various factors (capacitance,
> impedance, etc.) determine how fast the current falls (decays). The
> delays in the rising and falling of the current cause a "smoothing"
> effect on the current. Thus instead of "seeing" chopped current, the
> motor "sees" a more or less constant current.
>
> Power supply voltage and type (linear or switching) is a significant
> factor in determining how much power a motor produces. The higher
> the voltage, the more "pressure" there is to ramp up the current
> quickly. (Voltage can be compared to the pressure in a water hose.
> The higher the pressure, the more water comes out.) From our
> understanding of how a chopping controller works, we can see that
> voltage regulation (holding the voltage to a close tolerance) is not
> particularly important. If a power supply rated at 60vdc actually
> delivers only 58vdc, it is not going to impact motor performance very
> much.
>
> What is needed is a "reserve" of power. That reserve exists in
> linear power supplies in the form of the large (10,000 mf or greater)
> filter capacitor(s). These capacitors are "reservoirs" of power that
> are immediately available when the current to a motor is switched
> on. And during any "off" time, the capacitor replenishes itself so
> that power is available on the next cycle.
>
> Now why are switching power supplies not well suited to CNC
> applications? A switching power supply operates in much the same
> fashion as a chopping controller. The difference is that it controls
> voltage, not current. When the CNC controller is "idling", no
> reserve of power is built up. Only when the motor is switched on
> does the switching power supply start to react. The demand for power
> causes the switching power supply to "go into action" releasing more
> power to the controller board. But because the switching power
> supply has a feedback circuit to monitor the output voltage, the is a
> slight delay in delivering the current to the controller. Now we
> have two places where delays occur in getting the power to the
> motor. Not a good situation.
>
> Switching power supplies work well where voltage regulation is
> important and the demand on the power supply is fairly constant.
> They are generally cheaper than a linear power supply of a similar
> rating. This is because the transformers and large filter capacitors
> used in linear supplies are expensive items, especially as the power
> demands increase, They are also bulkier, so they require a bigger,
> more expensive housing.
>
> One way (the only way as far as I know) that switching power supplies
> can effectively be used for CNC operations, is to put a large (again
> 10,000 mf or larger) filter capacitor on it's output. The capacitor
> again acts as a reservoir of energy that is immediately available
> when the motor is switched on, giving to switching power supply time
> to react to the increased demand for power.
>
> This explanation could be 10 times longer and still not do justice to
> the subject matter. I've done my best to condense it while still
> giving enough information for a beginner to understand why I believe
> that linear power supplies are better for CNC operations than
> switching power supplies. In doing so, I have probably introduced
> some vagueness and information that can be easily misinterpreted. I
> apologize in advance for these short comings.
>
> And again, I don't have a particular fondness for the Gecko drivers.
> I just think that you don't get the extra value for the money. There
> are thousands of hobbycnc controllers out there that drive the Taig
> mill just fine.
>
> And if anyone wishes to take exception to my opinions, feel free to
> critique my work. But please don't engage in character
> assassination. I'm just one guy out here, trying help my fellow
> CNC'ers. I don't pretend to set myself above anyone.
>
> Regards
> Gary
>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, Danny Miller <dannym@...>
> wrote:
>
>
>> Get a 48v regulated power supply with the current greater than... I
>> think it was like 70% of the sum of motor currents.
>>
>
>