# Flux Density Vs. Switching frequency

Discussion in 'The Projects Forum' started by twenglish1, Apr 9, 2014.

1. ### twenglish1 Thread Starter Member

Feb 25, 2014
38
3
I am working on designing a high frequency power transformer for a half bridge inverter supply, and i am confused

The transformer cores i purchased(EE80 ferrite) state a saturation flux density of 5000 gauss, but from the research i have been doing, and if i understand correctly the max flux density at 40khz should be less than 2000 gauss, why is this? and what happens if i run the transformer at 4500 gauss at the same frequency?

I did some calculations on a high frequency transformer using the same exact core (also running at 40khz) and i have and found that it is being run at about 4500 gauss

Last edited: Apr 9, 2014

Sep 30, 2009
4,095
1,685
3. ### richard.cs Member

Mar 3, 2012
162
31
If you've read somewhere 2000 gauss that will be a guideline for generic cores and some particular idea of acceptable losses. You should go with whatever the core manufacturer reccomends (which will be less than the saturation flux but certainly more than half).

4. ### twenglish1 Thread Starter Member

Feb 25, 2014
38
3
I'll check that site out, and maybe that chart I saw that was comparing flux density to switch frequency, was just generic and showing how frequency effects flux density

Edit: This is where I was reading that about the flux density vs frequency http://www.mag-inc.com/design/design-guides/Transformer-Design-with-Magnetics-Ferrite-Cores

And as I mentioned earlier this welder I have, same input voltage, same core, same frequency, 11 primary turns and I calculated the flux density to be about 4500, but then again the machine can only be run constantly for 3 mins at full power

Last edited: Apr 10, 2014
5. ### RamaD Active Member

Dec 4, 2009
254
33
The core loss goes up with flux density not linearly, but more steeply. Also the hysterisis characteristics is normally linear only upto Bsat/2 approximately. The tendency to saturate will be higher after this. It is better to increase the no. of turns if possible (higher I2R losses) and keep the flux density lower.

6. ### twenglish1 Thread Starter Member

Feb 25, 2014
38
3
This is gonna be a high power, power supply, probably around 7kwatts, the power supply I'm modeling it off of uses the same core and roughly the same power levels, and I calculated its flux density to be about 4500, for a 5000 bsat core, is this a bad design? It's for a welder and does not have to be able to run at full load for long periods of time, the model welder had a 30% duty cycle, meaning 3 mins constant welding, 7 mins standby

7. ### shortbus AAC Fanatic!

Sep 30, 2009
4,095
1,685
With a poor duty cycle like the model welder has, why would you model after that? What I mean is use the basic design, but figure the specs to a better value. Compute a higher duty cycle and get a core that will work, rather than using the one the 'model' has. Many/most of the off brand, cheap welders are only a very bad starter type welder, why not build something worthwhile?

8. ### twenglish1 Thread Starter Member

Feb 25, 2014
38
3
I'm guessing the reason for the low duty cycle is due to problems with heat, some due to the igbt heating up from lack of enough filter capacitors and line inductance(wires between the filter capacitors and igbt instead of a direct connection) and the fact that the power transformer is being driven at almost the max saturation, which if I understand correctly means more losses and more heat build up, correct? And if I understand correctly, using two cores wound together will allow me to run the same 4500 gauss flux density, but the transformer won't saturate until 10,000, since it will be double the bsat, right?

9. ### shortbus AAC Fanatic!

Sep 30, 2009
4,095
1,685
I know nothing about the transformer part of this. I do know from welding most of my life, that this type duty cycle is not something that a REAL welder has.

If you want that same performance as the 'model', it would be way cheaper to just buy another one of those. Like I said before that level of welder is for the guy that dreams about being a welder but can't spend the money to get a good one. But unless you are a good welder, you can't make one of that low level machine do the job.

So why would you try to duplicate that level of performance? Spend a little more time and use the basic design of the 'model', but make it to work correctly. I don't know if they still have it, but both Lincoln and Miller used to have the schematics for their machines on their websites.

If you ask the right questions here there are guys that can and will help you with this.

10. ### twenglish1 Thread Starter Member

Feb 25, 2014
38
3
The more research I have been doing about inverter power supplies and studying the cheap machine I have, the more I realize all the corners that have been cut to build something that just works, and the design for the machine I will build has less and less similarities to the model machine the more research I do, I'm fairly new to inverter power supplies and am still trying to learn about designing and building one, I have been studying the commercial machines to see how its supposed to be done the right way. I believe I have a pretty good design, at least on the power side of it, besides the transformer, I still also need to choose a suitable pwm driver chip and method of current limiting, unless for starters I just use an air gap in the core to limit the flux density which will limit the current if I understand correctly