transformer parameters to decrease heating / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

transformer parameters to decrease heating

1 2 next

Move Thread

LAN_403

IamSmooth

Sun May 23 2010, 10:02PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

I have an iron powdered toroidal core with a cross sectional area of 1.5cm^2. I have four of them with 17 turns of #10 wire. When I put 300vac across @ 100khz I calculated that he flux density is 2600 gauss per toroid. I got this using the formula

Bmax = Ex10^8/(4.44xAxNxF)

The supplier says that I should not exceed 5000 gauss. With four stacked together, does this roughly decrease my density by a factor of 4?

At the power levels I am operating at the toroids/wire start to get warm. Would a lower permeability be advantagous? Should I stack more toroids together?

GeordieBoy

Sun May 23 2010, 10:35PM

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

Run a few different design scenarios through Micrometals iron-powder magnetics design software. That's what it's there for and it is superb at telling you how hot each different core material will get whilst operating in your application.

If peak flux swing and operating frequency are low and you need lots of inductance for a low price then Type-26 is probably the best material.

If this runs slightly too hot then look at the more expensive material like Type-52 for lower core losses with the same permeability.

Type 34 or 35 will give you lower core losses again but with a decrease in permeability so you'll need more turns to get the same inductance. Hence you'll trade some core loss for copper loss. It's also more expensive.

Then there are material types 6 and 2. These have lower permeability again and much lower core loss. Type-2 for instance is very hard to saturate, and has the lowest core losses of all even at very high frequencies of operation and large AC flux swing. The downside is that the permeability is low though, so its best for high frequency applications where you don't need much inductance.

I would tend to avoid stacking cores unless you have to. The reason is that it is usually best to use either a single larger core, or a lower loss material. Putting twice as much of the wrong material in the AC field isn't really going to arrive at as significant improvement as changing to a low loss material or a better core shape. There are some instances where core stacking can help though.

As a final note, look at the thermal aging results from the Micrometals design software! Iron-powder cores use an organic binding material that is susceptible to thermal aging. If you operate an iron powder core well below its critical temperature it will have a long and happy life. However, if you operate it above this temperature, the material degrades, core losses increase, and the material gets even hotter. Evetually thermal runaway takes place and the thing will burn up. I've seen many switch-mode power supplies that used to have yellow-what type-26 toroids that ended up with crispy brown ones because there was insufficient air-flow over the iron-powder core! Something worth bearing in mind if you want it to have a long and drift-free life.

I hope this helps,

-Richie

IamSmooth

Sun May 23 2010, 11:21PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

That's a nice link Richie. I've already started playing with it.

If I go with a lower permeability the inductance goes down. If this is the case, and this coil is inverter's load, then wouldn't the current draw also go up? I would then need more turns to compensate, but this will decrease the voltage on the secondary.

Would I then have to add more turns on the secondary?

GeordieBoy

Mon May 24 2010, 02:22PM

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

> Would I then have to add more turns on the secondary?

Yep, that's how it works. You are trading off core losses for copper losses. Usually you want to split the overall losses about 40:60 between the core and the windings because it's harder for heat to be removed from the core than the windings.

It seems a bit strange that you are using powdered-iron for a transformer core at 100kHz. I would have thought that a Mn/Zn power grade ferrite would have run cooler, given tighter magnetic coupling (reduced leakage inductance) and a much higher magnetising inductance too. A ferrite material like 3C90 or 3F3 would have been my first choice for a 100kHz power transformer.

-Richie,

IamSmooth

Mon May 24 2010, 02:51PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Any suggestions where I can buy some 3c90 ferrite toroids?

kell

Tue May 25 2010, 11:45PM

Registered Member #142 Joined: Sat Feb 11 2006, 01:19PM
Location:
Posts: 102

Amidon.
Now I'm thinking back to the last order I placed with them. The girl on the phone had an extremely heavy Japanese accent. Odd.
But Amidon is a good outfit to order from, especially for very small orders. They have a wide selection of iron powder and ferrite cores, magnet wire etc. and have always catered to the hobbyist/ham market.
If you order from them, make sure they send you their info sheet on iron powder and ferrites. It has lots of good info, including that equation you wrote in your first post, and lots more besides.

IamSmooth

Mon May 31 2010, 04:15AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

Well, I got some 3C90 ferrite cores and they work very well. The cores are relatively cool. However, the wire (#10) got very hot and melted the coating. Please note, I am running about 35-45A through the transformer at 100khz with 18 turns.

I took 16g magnet wire and braided 8 pieces into a cable. I tried this. The wire gets very hot. I looked up ampacity charts and I can see that I am not utilizing all of the wire due to lack of skin depth penetration. Also, depending on the chart, I am still a little short on proper current handling capacity, but there is little room for bigger wire.

My next thought for cooling is either forced air or a water-cooling jacket around the transformer. I don't want to try mineral oil as I would need some type of metal radiator-type container to remove the heat. I don't want metal near this section of the project.

I am hoping someone out here might have some suggestions based on experience with cooling some high current component like this.

The pictures show the burned wire. I removed the fused pieces and took the second picture.

Dalus

Mon May 31 2010, 09:03AM

Registered Member #639 Joined: Wed Apr 11 2007, 09:09PM
Location: The Netherlands, Herkenbosch
Posts: 512

High temp litz wire and forced air cooling?

Steve Conner

Mon May 31 2010, 04:18PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I've used copper pipe in fibreglass sleeving for winding high-current inductors before.

You have water cooling, so you could do this and pump water through the pipe too.

Also you could try aircraft grade wire, it's rated to withstand very high temperatures.

GeordieBoy

Tue Jun 01 2010, 05:04PM

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

Silver plated copper tube, flat copper strap or parallel bundles of Litz wire are the best options for high-current high-frequency windings.

I'm fond of the LCLR work-coil arrangement for high-power induction heating myself because it keeps the large AC work coil current out of the isolating transformer.

-Richie,

1 2 next

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