Thoughts on power factor and DC windings

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
After reading Merlin's book on hi-fi preamps, I stumbled upon the fact than many people don't take power factor into account, when designing power transformers. They just add a "safety margin" for current, when choosing wire diameter.

This is how I started too, as some books and websites I read just didn't mention it.

This issue applies to rectified windings, mostly HT DC rail. Using for heaters too, when using DC. Power factor is bad (just a little above 0.5), and we need to use thicker wire, as if the filaments winding drew double the current (because of the RMS value of the current spikes refilling the reservoir capacitor, being higher value than the DC current consumption).
This can be easily simulated on LTSpice.

I suppose it this higher current that has to be used to compute primary current, and it is this final VA power which has to be taken into account when choosing core area, using our prefered formula, and not the real consumed power ( Idc * Vdc ).

Is it how you do it?
Any comments on this topic?
 
Power Factor or PF. is the phasing difference between mains voltage and mains current or drag. It has no bearing on the gauge of wire used for heater supplies or whether you have a rectified DC output or not.In a resistive load the current follows the voltage but a reactive load varies and Smart Meters detect this error and increase the consumption for the electricity companies, costing you more in the end.
In an old style mechanical electricity meter, the meter turns with eddy currents set up in a coil next to an aluminium block that spins with respect to the detected wattage drain.

Take a 20 Watt strip lamp for instance. It has a choke in series with the heaters and across the gas filled pipe. This limits the current but as it is an inductor, the PF is about 0.5 or less so a 2uF capacitor in parallel with the supply, has a phase correction effect. Returning the PF to near 2.0.Without the capacitor, the old style meters would turn at half speed and loses the supply company money.
That is why PF circuits are used in most modern day designs from Television screen to Projectors and large switch mode power supplies. For the benefit of the suppliers, not us.
 
Jon, PF also includes harmonic distortion components (aka rectification spikes). The PF you are referring to is really a subset, and often identified as cos phi PF.

Elerion, I know that some vintage transformer makers specified a max power supply output dc current for the secondary winding rating, and identified a rating for choke filtering and a different level for capacitor filtering (and I think identified the filter capacitance).

PSUD2 can also provide a reasonable stab at winding rms current. Simulation is often the only way to do such a 'measurement', as many meters can't be relied upon for crest factors above 3 (I have to revert to my HP3400A to get up to 10).
 
Power Factor or PF. is the phasing difference between mains voltage and mains current or drag. It has no bearing on the gauge of wire used for heater supplies or whether you have a rectified DC output or not.
http://www.valvewizard.co.uk/heater.html

(at the bottom)

"Another important thing to be aware of is that rectifying AC to DC in this way introduces a power-factor loss of about 0.5, which loads down the transformer more. In other words, to supply a heater with 300mA of DC the transformer actually has to deliver about 600mA of AC, So be careful not to overload the transformer."

http://www.valvewizard.co.uk/psu.html
There's a graph relating Rs/RL with ripple current / dc current ratio.
PF = 1 / ratio

It seems reasonable to me.
 
JonSnell Electronic said:
In an old style mechanical electricity meter, the meter turns with eddy currents set up in a coil next to an aluminium block that spins with respect to the detected wattage drain.
It is my understanding that both old-style meters and modern smart meters measure power - that is real power (for domestic users - industrial users may have different arrangements). Nothing much to do with the topic of this thread.

When designing a transformer for a DC PSU the effect described by Merlin needs to be taken into account. It is taken into account by all competent designers. It is sometimes ignored by newbies, especially when making a DC heater supply.
 

PRR

Member
Joined 2003
Paid Member
It is my understanding that both old-style meters and modern smart meters measure power - that is real power (for domestic users - industrial users may have different arrangements). Nothing much to do with the topic of this thread. ...

Correct. I have the book. It focuses on North American utility power meters. They are carefully compensated to register POWER (real). Large customers which historically reactive (motor) loads also have a PF meter: a high PF for 15 minutes adds an additional charge. Also a Demand meter: a modest average with a few huge peaks (sustained over several minutes) invokes more charges. This is for customers with bills so large it is worth paying people to figure the best deal (capacitors, staggered starting, whatever). Residential customers are simply over-charged for the real power they use (and the meter rental, and government surcharges, and...).

But that is quite different than what this thread is about.
 

PRR

Member
Joined 2003
Paid Member
Here's what we need to know.

A *cap-input* DC power supply makes 1.414 times more DC voltage than the AC voltage we give to it.

There is no free lunch.

So the AC Amperes must be 1.414 times higher than the DC amperes we are sucking.

But a further detail. The cap-input DC power supply draws current in short peak bursts. This is inefficient use of iron and copper. The heating/saggting current on the AC side is more than 1.414 times the DC current.

Depending on whether heat or sag is the limiting factor, the correction may be 1.6 or 1.8. If you do a custom wind you need to do your math. In DIY we shop stock sizes, do not usually find a "just right" AC rating.

My habit is to multiply DC Amperes by "two" and look for that many AC Amperes. At "worst" this is 25% over-size, which is usually a trivial cost difference. And often enough a 2X multiplier just barely works.
 
Exactly. It basically depends on the ratio of load resistance and source resistance (winding secondary + reflected primary).

If out amplifier draws 100 mA at 250 Vdc, it acts as a 2.5 kohm load. If our transformer secondary resistance+reflected is 100 ohm, then the ratio is: 0.04. When ratio is 0.1, the power factor is close to 0.75 (depends on reservoir capacitance), and if ratio is lower (as it is the case in this example), the power factor falls towards 0.5.

It is not a simple little deviation, it is a huge compensation we have to do here. The reservoir capacitor affects power factor too. If bigger, we get less ripple, and less power factor too.
 
first time to see this kind of thread here....
i wonder if there is a close enough calculation to achieve such high power factor...
about the only time i came close to unity power factor was in a 6c33 set amp i built,
the amp is no longer with me so i can not go back and investigate...

is 92% good enough?
 

Attachments

  • IMG_3306.jpg
    IMG_3306.jpg
    91.2 KB · Views: 247
Do you still have the schematic? If so, just use Merlin's chart and calculation to see if the PF matches up to your reading.

that amp is no longer with me so i can not get back to further investigate...
will try to digest merlin's charts and apply them in my other amps..

Tony, that style of mains monitor may not be too accurate with PF - has it got any specs, or do you have a professional power meter?

i realise that, that is a kill-a-watt monitor and in the absence of professional meters, that is good enough for me..

According to Merlin's graphs, reaching such a high PF requires the load resistance to be smaller than the secodary side resistance. Either a low voltage high current comsuption amplifier, or I don't see how right now.

sorry you lost me....thevenin's theorem says for maximum power transfer, internal resistance of the source should be equal to the load resistance, and for maximum power factor, reactive power should be minimised.....how to do this in and amp, i have no clue...i may have guessed, but i truly i have never given it much thought..

for a given size of traffo material, there is a limit to the wire sizes and therefore impedances that you can use...
 

Attachments

  • IMG_4930.JPG
    IMG_4930.JPG
    681.6 KB · Views: 134
sorry you lost me....

All this is easily simulated using spice software, and I get very close results to Merlin's charts. If you look at it, you see that PF increases when source resistance decreases (even when under load resistance value), though it does at a lower rate. This is because the current spikes' rms value decreases, maybe because the conduction angle increases.
 

PRR

Member
Joined 2003
Paid Member
....thevenin's theorem says for maximum power transfer, internal resistance of the source should be equal to the load resistance...

_IF_ you are stuck with a specific power source, then yes: matched gives the most power you can get.

And an equal amount of waste heat in the source.

Most sources will not be happy.

For best *delivered* power, you want source impedance much lower than load impedance.

For example: my house power line is about 0.4 Ohms. My house load runs from 60 Ohms at the moment to 6 Ohms on laundry day (4 to 40 Amps @ 240V). Most houses are wired 0.1r supply and can support 2.4 Ohms (240V 100A) with negligible lamp-dim.

I don't know why we care about Power Factor facing the power company. Unless your home is ALL huge DC supplies, the reactance and peak-drag is not enough to bother the system. (Large factories full of induction motors do pay for Power Factor, and have to consider rectifier effects.)

What we care about is that we MUST buy AC Amps much greater than DC Amps, or suffer sag and heat.
 
All this is easily simulated using spice software, and I get very close results to Merlin's charts. If you look at it, you see that PF increases when source resistance decreases (even when under load resistance value), though it does at a lower rate. This is because the current spikes' rms value decreases, maybe because the conduction angle increases.

thanks, i am not much into simulations....
my idea of high power factor, is when the reactive powers in any amp are minimised....
to me that meant, inductive and capacitive effects in the psu are equalised or cancelled so that only resistive loads exists...

how to do that? that is where the fun is....:cool:
 
The fixes are not the same for low reactive power factor and low harmonic power factor (high harmonic current content). You can't just cancel the latter with a shunt or series reactance. Power factor correctors for capacitor-input filters are actually just a switch mode boost converter. During the part of the input waveform where the voltage drops below what's on the cap, you get no current flow. So it boosts the voltage during that portion of the cycle until it gets current flow. The amount of boost applied is dynamically changed to get the input current to be approximately sinusoidal. I haven't ever seen one used on a standard 50/60 hz transformer supply (usually it is used with switch mode, because the corrector is switched mode) but there is no reason it can't be made to work just the same.

For choke input filters this all goes out the window. Properly loaded, the input current waveform is SQUARE. The choke current is constant, alternately drawing that constant current from either side. The same type correction would not work, since boosting current draw anywhere on the waveform would be counter productive.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.