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.
Hello,

Real power that turns into heat is read as real watts at the utility power meter and shows up on the bill the beginning of next month.

Claiming that heat is caused by increased Power Factor is wrong. The heat is lost as lost efficiency; do not blame Power Factor for the loss.

If it gets hot build a bigger transformer with a larger derating factor.

DT
 
Last edited:
Transfo specify that, heat is not a problem only as far as insulation is a possibly melting varnish and glue.

You don't even pay for PF loss in residential consumption.

PF is only the power the electrical company needs to produce at the power plant to generate your watts, transformer lose nothing.

Only major concern is mega factories with MWatts of induction motors (electrical motors)

With new switching power supplies and CFL's this is now an issue in commercial buildings too.

Most of the problems originate from all the wires and wire-less routers which can get a polluted signals, so filters are installed in big buildings. A real concern for construction electricians, I worked as a cost estimator / proposal writer / and I don't know of any concern for an amplifier.
 
They really hate it in commercial buildings because they typically use 3 phase power. Large 3rd harmonics can wreak havoc, even overloading (undersized but code compliant) neutrals in feeder circuits. Lighting ballasts are often PFC, but that doesn't do anything about all the office equipment.
 
DF96 said:
It has already been made clear in this thread that people who know what they are doing take account of this, and people who do not know what they are doing do not take account of this. It is unclear to me what else there is to discuss.

I agree.

That is the part that is missing from the conversation.
What needs to be quantified is how much is lost as heat and how much is lost as Apparent Power (PF)?

Well, AFAIK this is simple. Leaving PF aside, power is lost because of DC resistance loss, which depends mainly on the windings wire (function of area and length), and magnetic core losses.

Anymore has to be taken into account?
 
@Elerion, are you more confident now than when you started this thread?

in any traffo, there is such a thing as copper losses and core losses that both lead to elevated heating...

our goal when designing and building trafo is to quantify both losses and dimension traffos with those in mind..
 
Ex-Moderator
Joined 2011

Attachments

  • 20140828071707_89991.jpg
    20140828071707_89991.jpg
    79.1 KB · Views: 63
  • 20140828071733_72350.jpg
    20140828071733_72350.jpg
    91.9 KB · Views: 61
Care is usually required when reading a peaky current waveform (whether with current range or the voltage across a sense resistor), as per a secondary winding current that is rectified. Even the amazing Keysight 6++ digit lab meters can only stay within accuracy spec for 3:1 crest factor if near full range - typically have to choose a range where signal is well below full scale, and even then the spec goes just up to 10:1.
 
I guess cheap comes with a disadvantage - a lack of proper specs. I can see there is an all encompassing accuracy.

Using that accuracy spec, the 227.4V would be +/- 1.6V assuming 600V range. The 0.054A would be at circa +/- 0.02A, depending on range. So power accuracy is circa 3.8% if the meter is within spec. But there is no definition of crest factor, and those waveforms are pretty benign.

I must admit, the gang on EEVblog just jump on the latest cheap meters and give them a good going over with whatever you beaut lab meter they happen to have on the bench.
 
ST has an app note on secondary side PF correction, so the trafo can be used up to it's AC ratings. you only need to select a lower transformer voltage as the boost PFC always needs a higher output voltage than the peak ac voltage from the rectifier. you may consider turning off the PFC circuit at low loads and idle.
 
There are two problems with that approach, of course. First is that it turns a simple core-coil power supply into a switch mode. Yeah, you have an SMPS with a big 60 Hz trafo. The other is that you lose the comfortable (at least 10X) short term overload capacity, unlesss you size the PFC to handle ridiculous currents. PFCs can't correct properly over too large of a range of output current. Of course if you're willing to live within the rules it is a solution.
 
The other is that you lose the comfortable (at least 10X) short term overload capacity, unlesss you size the PFC to handle ridiculous currents. PFCs can't correct properly over too large of a range of output current.
Valve amps typically have constrained limits on supply rail currents, unless there is a fault. For a B+ example, the worst typical scenario would be class AB operating substantially in B, so the 'normal' supply rail swing could exceed 2:1 from idle.

Yes you want the PFC to operate 'normally' in the required range, and to be able to gracefully manage start-up and fault limits in a current limited manner (and preferably with a fold-back mechanism for a faulted output, as HT fusing may have no protection outcome).

Was there a wider load range that you had in mind that would be a problem?
 
Typical "high current" class AB amps that can put out 40 to 60 amps or more peak into an ohm or two. This is where old school transformer amps (can) shine compared to lightweights with switching supplies. The supply will go into limit in 20 milliseconds or so, where the old school supply will happily put it out till the breaker trips or the heat sink on the amp gets too hot. Asking any switcher, PFC or not, to deal with those kind of peak currents all the way down to idling is an intractable design problem.

Large class A amps, where the supply current will swing at most 2:1 or tube amps which tend to be high biased and have overhead like heater currents to deal with stand a much better chance working with a PFC. If you can stand the switching artifacts or are really good at getting rid of them.
 
PFC is the least of my concern....while nice to look at in the meters,
does not really matter much to me....

when i design my linear psu, my goal always was least losses in both core and copper, and huge rectifiers, 70 amps and 150 amps, and humongous filter caps...>100kufd...

this of course makes soft starting mandatory...no two ways about it...

even for my 16kt88 power amp, i used a 16amp bridge rectifier in the plate supply,
and why shouldn't i? they come very cheap from discarded boards...
 
16 KT88's? A 16 amp rectifier might not be considered overkill there.

For a given technology/device family, a higher current rectifier will always have a higher leakage current and a longer reverse recovery time compared to a lower current version. A higher voltage diode will always a have a higher forward voltage drop and a longer reverse recovery time compared to its lower voltage counterpart. The extra rating isn't exactly "free" even if you didn't care about $4 vs. 6. You do have to keep an eye on secondary parameters and make sure something unexpected doesn't drive power dissipation, regulator dropout, noise, etc..
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.