Forced air cooling of transformers / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Forced air cooling of transformers

Move Thread

LAN_403

Dr. Dark Current

Thu Feb 26 2009, 05:26PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

I was wondering, how much can forced air cooling increase the continuous power available from a line frequency transformer?
Can a small PC PSU fan do anything useful?

GeordieBoy

Thu Feb 26 2009, 07:50PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Maximum power rating of a line frequency transformer is a function of the copper losses in the windings. These are basic resistive losses for a line frequency transformer so the power dissipated is equal to the winding's current I squared times each winding resistance R.

Contrary to popular belief the transformer wont saturate if you draw excessive current, provided that it is being driven with the correct voltage, frequency and wave shape that it was designed for.

Dissipation, and therefore temperature rise, is likely to increase with the square of the current though! So if you run a 100W transformer continuously at 200W it is likely to be dissipating 4 times as much power in copper losses than it was designed to do. This will result in a correspondingly higher temperature rise above ambient temperature.

Forced air cooling always helps, and you'd be suprised how even a small draft over a heatsink or power component can give significant benefit. That's why all serious power electronics kit has a fan - Even a small fan really does reduce junction temperatures and improve reliability.

You can quantify the benefit yourself by performing a test. First measure the ambient temperature. Then do a test running your transformer at it's rated power (or slightly more) with no forced air cooling for an hour and measure the winding temperature. Then run the same test for an hour with forced air cooling and repeat the temperature measurement.

You can measure the winding temperature using a thermocouple meter, but the best way to do it is by the resistance change method. Right after you switch off the power you measure the winding resistance every 30 seconds, plot it and extraploate back to the moment when you shut off the power. There is an equation that allows you to calculate the temperature rise of the winding from the resistance increase.

This method of measuring the temperature of the winding is much better than using a thermocouple. The reason is because hot-spots usually exist deep within the winding where heat cannot be dissipated by convection to the air. In transformers wound with PVC coated wire inner layers can even melt and short together!

If you pick an upper limit for the winding temperature and a typical ambient operating temperature, then you can quantify how much the transformer's power handling ability increases with a given airflow.

-Richie,

KLH

Thu Feb 26 2009, 08:00PM

Registered Member #1819 Joined: Thu Nov 20 2008, 04:05PM
Location:
Posts: 137

GeordieBoy wrote ...
Maximum power rating of a line frequency transformer is a function of the copper losses in the windings. These are basic resistive losses for a line frequency transformer so the power dissipated is equal to the winding's current I squared times each winding resistance R.

There are also eddy current, hysteresis, and core losses that will affect the power dissipation. Core losses will increase with an increase of AC flux density, which increases with power output. Eddy current and hysteresis losses will likely not change much, but the given losses from these factors in the beginning can affect how much extra power you will be able to get.

Cooling efficiency can be greatly improved by mounting the transformer on a vertical surface, where a lot of the heat is removed through the "chimney" effect. Putting a fan under it will greatly increase the cooling effectiveness.

As for a "small" fan, from what it sounds like you want to do, will not work. You should pick a fan size in relation to your transformer; I would recommend a fan diameter larger than the width of your transformer (unless it is one of those really low profile designs) to get some airflow around the sides of the transformer.

Proud Mary

Fri Feb 27 2009, 02:37AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

A modified vacuum cleaner can be made into a very good forced air cooling unit, especially useful for big finned valves (USA: "tubes")

Dr. Dark Current

Fri Feb 27 2009, 10:57AM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

killah573 wrote ...

There are also eddy current, hysteresis, and core losses that will affect the power dissipation. Core losses will increase with an increase of AC flux density, which increases with power output.

I think the AC flux actually decreases a bit when drawing power, how could it increase?

Anyway, can anyone guess what's the ratio of core and winding loss in a typical line frequency xfmr?

GeordieBoy

Fri Feb 27 2009, 02:16PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

> Anyway, can anyone guess what's the ratio of core and winding loss in a typical line frequency xfmr?

It depends entirely on the load. Under light load, winding losses are dwarfed by core losses. Think of an MOT running with its secondaries open circuit.

A very rough general statement might be that core and winding losses will likely be around the same order of magnitude when a power transformer is loaded to its rated VA.

If core losses dominate winding losses, then a given transformer should probably have more turns of thinner wire on its windings. Conversely if winding losses dominate core losses, then it should have less turns of thicker wire filling the winding window. If the winding windows aren't almost full, then this should set alarm bells ringing about the transformer design!

-Richie,

Sulaiman

Fri Feb 27 2009, 08:41PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Don't know, haven't googled yet !
I guess 20% core 80% copper due to cost of copper?

GeordieBoy

Fri Feb 27 2009, 11:05PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

If you have a wattmeter it's quite easy to measure the approximate core loss and copper loss of a transformer:

First short circuit the secondarys and increase the primary voltage until the rated current is drawn. Feeding the transformer from a wattmeter will then show the power being dissipated in the winding resistances.

Then open circuit the windings and apply full rated voltage to the primary. Now the wattmeter will show the core loss due to in-phase primary current.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob