The Godfather talking
You think I am funny guy huh?
Sonsivri
 
*
Welcome, Guest. Please login or register.
Did you miss your activation email?
March 28, 2024, 11:20:36 23:20


Login with username, password and session length


Pages: [1]
Print
Author Topic: Simulating standard magnetic cores in Micro-Cap  (Read 3310 times)
0 Members and 1 Guest are viewing this topic.
LithiumOverdosE
Senior Member
****
Offline Offline

Posts: 350

Thank You
-Given: 374
-Receive: 568


« on: July 13, 2020, 09:35:59 21:35 »

I'm struggling to properly simulate transformer (on ETD34 former) with 3C94 magnetic cores.
Compared with the real life circuit on the test table the simulated magnetic transformer exhibits much lower magnetization current and way too low current on the secondary side load.
I guess that it may indicate that either Micro-Cap model of 3C94 is wrong or that I'm not setting all the parameters correctly (AREA and PATH).

I assume that Micro-Cap 3C94 core model is correct, so I was wondering if anyone successfully simulated the behaviour of some standard size ferrite transformer (like ETD34/3C94)?
Are there are some guidelines which should be observed?
Logged
optikon
Cracking Team
Hero Member
****
Offline Offline

Posts: 848

Thank You
-Given: 1092
-Receive: 1999


« Reply #1 on: July 14, 2020, 12:04:04 00:04 »

Looks like the default path length is 1cm did you change it to what you need?

Additionally, if you specify a coupling K, say 0.98 does the BH curve look like you expect it to?
Logged

I can explain this to you. I can't comprehend it for you.
LithiumOverdosE
Senior Member
****
Offline Offline

Posts: 350

Thank You
-Given: 374
-Receive: 568


« Reply #2 on: July 15, 2020, 09:36:41 21:36 »

Yup, the first thing I did was to modify default AREA and PATH to fit the ones of ETD34 cores dimensions.
Playing with the coupling coefficient and increasing it from 0.95 to 0.98 does affect the primary current but it is still not correct.

Plotting a BH curve may indicate the accuracy of the ferrite material model and I'll try that to get it off the troubleshooting list.

I'm also attaching the screenshot of the simple test circuit and the values I've modified to fit ETD34 dimensions. 

I am just perplexed that with all those sample circuits provided with Micro-Cap there isn't a single one using non-linear core transformers.
Logged
optikon
Cracking Team
Hero Member
****
Offline Offline

Posts: 848

Thank You
-Given: 1092
-Receive: 1999


« Reply #3 on: July 16, 2020, 12:30:15 00:30 »

Microcap does provide an example circuit called "core" - see my screen shot.

As I understand it, it uses the Jiles Atherton core model. The help topic for it does discuss it a little bit. It is a non linear model.

Now, I am not sure if that vendor 3C94 model core you are using is correct- but in theory, it should be able to be modeled correctly.
The Model program inside MIcrocap lets you build a non linear core model using some data sheet values / curves... so you might want to try that approach also.

Inside the model are a few things I am unfamiliar with like domain wall pinning constant. who knows if thats set right for the vendor model.. how to even figure it out.

It would be good to understand this process.. I would also like to use Microcap for transformer modeling and circuits.


Logged

I can explain this to you. I can't comprehend it for you.
LithiumOverdosE
Senior Member
****
Offline Offline

Posts: 350

Thank You
-Given: 374
-Receive: 568


« Reply #4 on: July 18, 2020, 12:06:51 12:06 »

Microcap does provide an example circuit called "core" - see my screen shot.

As I understand it, it uses the Jiles Atherton core model. The help topic for it does discuss it a little bit. It is a non linear model.


I'm aware of the examples and of the discussion in the manual and it all seems fine but the divergence from the real-life hardware is what makes me cautious to rely on the sims.


Quote
Now, I am not sure if that vendor 3C94 model core you are using is correct- but in theory, it should be able to be modelled correctly.

The manufacturer is Ferroxcube and over the years I've found their datasheets to correspond well with the real-life parts.


Quote
The Model program inside MIcrocap lets you build a non linear core model using some data sheet values / curves... so you might want to try that approach also.

I've tried Micro-Cap's core modelling utility a few years back but found it too cumbersome to work with based on the datasheets of some older cores.
The disadvantage of that approach, at least in my case, is that it takes much more time to make a new model based on datasheet and fit the parameters to correspond with the measured values of the physical part, rather than using the Micro-Cap available core models.
The way I see it, if making the sim behave correctly is taking more time and effort than testing the physical parts on the bench, then its advantages are void. 


Quote
Inside the model are a few things I am unfamiliar with like domain wall pinning constant. who knows if thats set right for the vendor model.. how to even figure it out.

Indeed, that's what's perplexing me as well.
I haven't found the source material that Micro-Cap programmers used as basis for their core models and I haven't find such parameters in the manufacturer's datasheets.
However, they must have based it on something because I suspect they haven't done their own modelling for so many different materials. 
Logged
optikon
Cracking Team
Hero Member
****
Offline Offline

Posts: 848

Thank You
-Given: 1092
-Receive: 1999


« Reply #5 on: July 18, 2020, 02:32:30 14:32 »

can you drive your model and your bench setup with a sinewave? See how those compare. Could be the models frequency behavior was derived that way (?)
How much are the sim measurements off vs the bench?

I agree that if it takes less time to build it on the bench, then the sim is useless.

However, usually up front work is required for the models and then in the future you can spend less time with them --having being validated.

Also you have  a setup to measure B-H curve on the bench? Must be nice! :-)


« Last Edit: July 18, 2020, 05:51:46 17:51 by optikon » Logged

I can explain this to you. I can't comprehend it for you.
LithiumOverdosE
Senior Member
****
Offline Offline

Posts: 350

Thank You
-Given: 374
-Receive: 568


« Reply #6 on: November 19, 2020, 11:04:57 23:04 »

Sorry about the late reply.

In the meantime I did real life measurements, finished that part of development and moved on.
 

I agree that if it takes less time to build it on the bench, then the sim is useless.

However, usually up front work is required for the models and then in the future you can spend less time with them --having being validated.

You are correct on both points.
However, if you're the only one working on some project and you have dead lines then often there is no extra time to trying to decipher insufficiently documented parameters and trying to   fit it to bench measurements. IMHO it seems like something more suited for academic environment or development teams.


Quote
Also you have  a setup to measure B-H curve on the bench? Must be nice! :-)

Nope, but a few of the professors at the local university are friends of mine and if I don't have some particular piece of equipment I can drop by and do necessary measurements on their specialised equipment.
Logged
Pages: [1]
Print
Jump to:  


DISCLAIMER
WE DONT HOST ANY ILLEGAL FILES ON THE SERVER
USE CONTACT US TO REPORT ILLEGAL FILES
ADMINISTRATORS CANNOT BE HELD RESPONSIBLE FOR USERS POSTS AND LINKS

... Copyright © 2003-2999 Sonsivri.to ...
Powered by SMF 1.1.18 | SMF © 2006-2009, Simple Machines LLC | HarzeM Dilber MC