do I need PFC ?

[andy55]

hi electronics gurus,

I'm planning a servo system based on the surpluscenter sanyo denki servos. will use three 400W servos for axes and one 1kW servo for spindle.
i.e. 2.2 kW of power together (in addition I suspect that the quoted wattage of the servos is the mechanical output and not the electrical input, so the power requirement is even higher...)

A simple "transformer+diode bridge+filter caps" would be my first thought, but what about PFC ?
The psu will be hooked up to 4x PWM servodrives that drive the inductive motor coils.

If I use the simple design without pfc, by how much would it be reasonable to oversize the psu ?
I probably need to buy two isolation transformers for this, is it OK to wire the secondaries in parallell (identical transformers)?

what about EMI or disturbing the local electricity network in general. do I need RF filters on the primary lines ?

what's a reasonable maximum power to hook up to normal single phase electricity ? (I think I have 16A fuses on the 230VAC mains here...)

thanks for any answers,

[Al_The_Man]

Originally Posted by andy55 hi electronics gurus, A simple "transformer+diode bridge+filter caps" would be my first thought, but what about PFC ? The psu will be hooked up to 4x PWM servodrives that drive the inductive motor coils. If I use the simple design without pfc, by how much would it be reasonable to oversize the psu ? I probably need to buy two isolation transformers for this, is it OK to wire the secondaries in parallel (identical transformers)? what about EMI or disturbing the local electricity network in general. do I need RF filters on the primary lines ? what's a reasonable maximum power to hook up to normal single phase electricity ? (I think I have 16A fuses on the 230VAC mains here...) thanks for any answers,

What makes you to think that you would need PFC? A 'Simple' ps is all that is needed, Transformer bridge and capacitors.
Your power supplier is only going to see the transformer load, there should not be a any need for PFC.
The secondary is feeding a capacitor/inductor load and which is isolated by means of the transformer, power transformers have inherently good high frequency reflected suppresion, also there is a large capacitor bank between your load and the TXR secondary..
It is possible to wire transformers in parallel, providing a they are a close match, they must be phase correctly, however.
You can usually hook up any load that is within your capacity of your supply .
You could check into the possibility of larger fusing capability for your panel.
Al.

[andy55]

Originally Posted by Al_The_Man What makes you to think that you would need PFC? A 'Simple' ps is all that is needed, Transformer bridge and capacitors. Your power supplier is only going to see the transformer load, there should not be a any need for PFC. The secondary is feeding a capacitor/inductor load and which is isolated by means of the transformer, power transformers have inherently good high frequency reflected suppresion, also there is a large capacitor bank between your load and the TXR secondary.. It is possible to wire transformers in parallel, providing a they are a close match, they must be phase correctly, however. You can usually hook up any load that is within your capacity of your supply . You could check into the possibility of larger fusing capability for your panel. Al.

thanks for these answers.
A PF of less than one means that I have to oversize the transformer by 1/PF which will cost more...
Also, with ever more stringent regulations if I put some homemade 2kW device on the network I might get the power company guys knocking on my door real soon...

I found cheap AC2DC PFC units for sale on ebay but they all produce around 380VDC or so (some kind of industry standard?) and not the 150-200VDC I would need.

[gar]

060104-1539 EST USA

andy55:

Power factor is basically defined for a sunusoidal voltage and current with no harmonic content. How do you, or your power company, define power factor when the voltage is sinusoidal and the load current is not?

Now lets consider a transformer with a bridge rectifier and capacitor input filter on the secondary.

With the secondary disconnected the primary current will have a substantial inductive component and this lightly loaded condition on the power company line will have a low power factor, but not much current relative to full load.

Next connect the secondary to the bridge and put a load on the dc output to fully load the transformer. What does the primary current look like? It still has the small magnetizing component of the transformer, but added to this is a peaked waveform from charging the filter capacitor. The peaks are in phase with the transformer primary voltage.

How do these peaked waveforms fit into your definition of power factor?

These peaks should be equally objectional to the power company, because they increase the RMS heating of distribution wires and transformers, as compared to a sine wave current load.

So how upset the power company will be is a function how they define power factor, or other criteria.

When transformers are designed for capacitor input filters some assumption is made about the current waveform and its RMS heating of the windings.

A criteria that Signal Transformer uses for a full-wave bridge rectifier capacitor input filter is:

Maximum DC output volt-amperes is equal to AC volt-ampere rating of the transformer. This is obviously ambient temperature dependent.

.

[JRoque]

Hello. You're probably thinking about switched power supplies and not linear like the one you plan to use. As Al mentioned, the transformer windings, diodes and caps will likely damp any glitches resulting from the chopper drive. You can always add a filter network on the mains side if you're concerned with noise.

An efficient linear transformer (ie: toroidal) put very little load on the AC line by themselves. Also, even when firing all motors, you'll notice that total load will be far less than the full rated printed on the motor - thanks to the PWM drives.

JR

[gar]

060104-1743 EST USA

JRoque:

What do you mean by "load" is it power or current?

.

[JRoque]

Hello Gar, I mean Watts consumed. Same as you used it in "...this lightly loaded condition". BTW, my comments were for Andy55's post - I didn't see your post before but it certainly explains things further.

JR

[gar]

060104-1940 EST USA

JRoque:

Consider a motor running under steady state conditions, meaning constant speed and torque, over a time interval of concern, Then averaged over this interval :

Mechanical power ouput of the motor will be less than the electrical power input to the motor.

Electrical input power to the motor will be less than the electrical input power to the chopper circuitry.

Electrical output power from the DC power supply will be less than the input power to the DC supply.

All these derive from the conservation of energy. Dominately all the losses as you go down thru this chain of elements are thermal. There may be some minor radiation losses.

The chopper circuit average input current will be less than the motor average current because you essentially have a step down circuit as most of these drivers are designed. But the average of the instantaneous input v * i will be greater than the average of the instantaneous output v * i .

.

[unterhaus]

Anytime current and voltage are not sinusoidal and in phase, the power factor is less than one. Linear power supplies also have non-sinusoidal current due to the filter capacitors. Power factor meters often have two readings, one is Displacement Power Factor, which is cos(phi) or the phase between the current and voltage, and Total Power Factor, which is the ratio of the total power to the RMS power. If you have a linear system as a load, these will be the same thing, but if there is a rectifier feeding filter capacitors, they never will be. Basically, the power company is going to hook one of these meters up to the line and say Total Power Factor is not one, fix it.

I really wonder how much the electric company is going to be able to notice these harmonics. I can't imagine my electric company caring too much about it. Do they really go around with power factor meters?