# Transformer Core Temp vs. Winding Temp

#### MGPERF

Joined Jun 8, 2021
77
Hello,

I have a transformer 8 turns primary 240 turns secondary ETD29 core PC44 material running at 30 kHz 14.2 V input I calculated primary turns using the formula/equation (14.2)(10^8)/(4)(30,000)(1800)(5)=1.3 turns primary so I am way over the required number of primary turns however in theory that would prevent the windings from getting hot right? At a steady 4.5 amps the windings get very hot to the point where you can't keep your fingers on them for more than a second or two. Meanwhile the core heats up but significantly less any idea what's going on or how I can solve this problem? I have no air gap currently.

#### ronsimpson

Joined Oct 7, 2019
2,527
I have no air gap currently.
So, this is not a flyback but a forward power supply?
At a steady 4.5 amps the windings get very hot
At very low current is the temperature, OK?

What is the wire like?

#### MGPERF

Joined Jun 8, 2021
77
So, this is not a flyback but a forward power supply?

At very low current is the temperature, OK?

It is a lot lower but still hot to the touch.

What is the wire like?

It is 1mm on the primary and .25 mm on the secondary.
Hi Ron,

It is a push pull step up transformer. I just used the calculator below and I am at the absolute limit for bmax given the frequency. I am thinking maybe I can add a turn on the primary? Or maybe a .1mm air gap?

Maximum Flux Density (Bmax) Calculator. (Formulas & Equations) (electricaltechnology.org)

Joined Jan 15, 2015
7,079
Unless I missed it I am not seeing the wire gauge used?

Ron

#### MGPERF

Joined Jun 8, 2021
77
18 gauge wire thanks

#### DickCappels

Joined Aug 21, 2008
9,512
#18 is probably a bit large for 30 kHz.

#### ronsimpson

Joined Oct 7, 2019
2,527
Usually, I can tell if the core or the wire is heating up first.
You want to have about the same loss in each. If the wire heats first, then we need to think about why.
As DickCappels said #18 will probably have skin effect problems. You seem to know your way around the internet. Some where there is a skin effect verses frequency calculator.

#### MisterBill2

Joined Jan 23, 2018
13,773
Myguess is that there is also a DC component of the current in the wires that is contributing to the heating. What is this for? Is the idea to have a power supply?

#### MGPERF

Joined Jun 8, 2021
77
Appreciate all the help! I did some more digging and yes skin effect is a problem at 30 kHz. I am going to switch to an ETD34, add 1 turn to the primary, and run it intermittently using the shutdown pin and Zener diodes once the capacitor is charged to 300 volts rather than running the ETD29 continuously. With the larger core and extra turn I can go to down to 22 kHz safely.

Thanks,
Jon

#### MisterBill2

Joined Jan 23, 2018
13,773
If the core is becoming saturated magneticly during any portion of the cycle then there will be much more current and heating and that might be part of the problem. So checking that it is not saturating is a good idea.
If the supply voltage is adjustable, see if reducing the supply voltage a bit drops the output voltage. If not, the core is saturated. So there is a simple and easy test to do. Other folks may have additional tests that may be better.

#### MGPERF

Joined Jun 8, 2021
77
That’s a really good idea! I just checked that and volts are going up in a linear fashion.

#### DickCappels

Joined Aug 21, 2008
9,512
Appreciate all the help! I did some more digging and yes skin effect is a problem at 30 kHz. I am going to switch to an ETD34, add 1 turn to the primary, and run it intermittently using the shutdown pin and Zener diodes once the capacitor is charged to 300 volts rather than running the ETD29 continuously. With the larger core and extra turn I can go to down to 22 kHz safely.

Thanks,
Jon
@MGPERF I could not read the data sheet. What is the benefit of switching from an ETD29 to an ETD34. Is core size the only difference?

#### MisterBill2

Joined Jan 23, 2018
13,773
OK, that implies that the core is not saturating, at least that is what I recall from transformer theory class. So the next question is does your application need that much power? If we can presume 90% efficiency, and if you know how much power you need to get from the secondary, you should be able to figure about how much power you need to be delivering to the primary. That might be quite a bit less than you are producing now.

#### MGPERF

Joined Jun 8, 2021
77
Thanks, my peak requirement is 100 watts of power which is likely a bit much for an ETD29 core running at 30khz. It was getting very hot at 65 watts and that’s all I can test it to with my power supply.

#### DickCappels

Joined Aug 21, 2008
9,512
How many watts is the transformer rated for?
Is 30 kHz the frequency this transformer is typically used at?

#### MisterBill2

Joined Jan 23, 2018
13,773
DC is asking the right question. Transformers have definite power limitations. Exceed those and they become quite hot.