Site maintenance in progress (expect intermittent wonkiness). Thanks for your patience!


B-H Curves
« on: August 12, 2004, 10:18:43 PM »
I have been researching transformer testing techniques.
Here is some dope on B-H curve tracing.
This curve diaplays the magnetic properties of the transformer core.
It is basically a visual display of how a core beomes magnitized when a signal is applied, and how it de-magnitizes after the signal is removed. There is a little lag in the magnetics, and it varies depending on the core material. It is also called a hysteresis curve. The curves can be traced at different frequencies, different voltages and different transformer loads, so building up a file on one transformer can take a while if you include phase shift, frequency response, DC resistance, inductance, distortion, leakage, capacitance......
If anybody has any additional info on testing, feel free!
I have the Crowhust article alrerady.
Dang it! I think Scenaria's new toy will do all this automatically!
I know it will do phase shift.

Here are some links that have some B-H circuits.
Some are simple, some are complicated:
If I can't fix it, I can fix it so nobody else can!


B-H Curves
« Reply #1 on: August 13, 2004, 12:17:40 AM »

Scen's new toy will do a lot of things.  Phase is definitely one of them as I had mentioned in a thread before.  Measuring these other things is possible with an AP, but not automatic!

There are some schemes for measuring impedance (with phase), but there are some limits since the AP's generator can only drive to 13.3Vrms max with a sine wave, less with a square.  You could use an external amp to get more drive though. But using this method can give you a lot of info.

The guys at the transformer companies seem to know the most about this 'voodoo that you do'.  I can only tell you from watching two guys that I have worked with over the last ten years, that designing a trafo is about 70% calculation and about 30% 'get out some cores and wind some wire'.  These designs have been for class D amp output filters and for high freq (~30kHz) power trafos.

The AP is a very powerful tool, but it is basically a (sweepable) signal generator, a phase meter, an RMS voltmeter, a distortion analyzer, all with data logging and automatic graphing.  This is an oversimplification, but the point is that you have to have at least some knowledge of how to use it, in order to get what you want out of the machine.

"The sow would rather have her ear than a purse." - PRR


B-H Curves
« Reply #2 on: August 13, 2004, 02:52:44 PM »
You are right , Charlie!
The engineer I worked with at a transformer company had 20 years experience, and even so, he did a lot of hoping and praying!
If I can't fix it, I can fix it so nobody else can!


B-H Curves
« Reply #3 on: August 24, 2004, 09:27:00 PM »
I posted a BH curve question at Magnequest and got some more info.
Here is the response and the link:

In Reply to: Plotting B-H curves posted by cjenrick on August 12, 2004 at 15:56:40:
Cheap B-H loops.
You can use a Variac and an isolation transformer to drive the part at 60 Hz. Use an RC set to 0.6 Hz or lower to make the B curve. The capacitor must be a polypropylene (not mylar, not aluminum). Use a resistor in the ground leg at less than 10% of the winding resistance to measure current and then look at the curve on a scope set to X-Y.
Measure the voltage across the coil with an average (not true RMS) volt meter to get the saturation voltage of the B-H loop point you are testing. You need an average reading meter because saturation is average based issue, not true RMS based issue.
Another interesting test is to drive the low voltage side from a 600 ohm signal generator and look at the primary voltage at different frequencies (beware, The primary side is a serious shock hazard.) This gives you an idea of what the magnetization current is doing to your signal.

Play safe and play longer! Don't be an "OUCH!" casualty.
Unplug it, discharge it and measure it (twice) before you touch it.
. . .Oh!. . .Remember: Modifying things voids their warrantee.
If I can't fix it, I can fix it so nobody else can!


B-H Curves
« Reply #4 on: August 24, 2004, 11:38:11 PM »
Finally getting some good curves!
I wish they would have told me to switch to square wave when tracing BH curves!
Finally in the 5th article I read, they mention that little important detail that everybody left out.
will take some digi pics tomorrow.
sometimes you just have to keep on pluggin til you get the results you want.
If I can't fix it, I can fix it so nobody else can!


B-H Curves
« Reply #5 on: September 03, 2004, 10:45:14 PM »
Here is a way to tell when your transfomer is saturating.
You put a resistor in the leg of the primary and put a scope across the resistor. You will be looking at the magnetizing current waveform across the resistor.
When your peak exciting current, B,  is twice the average exciting current, A,  saturation occurs.
This chart is for a square wave.

I also measured the input signal with an average reading voltmeter, not true rms, as recomended by Mike over at Magnequest.
Here is a transformer hitting the wall at 40 hertz with a sine wave.
The lower wave is the mag curent, upper is input to primary.
14.8 volt peak to peak saturates the x-former. Avg reading was 6.22 vac.

Here it is at 20 hertz.
Notice that the saturation voltage decreased to 7.40 volts peak to peak.
This is due to the lower frequency. The transformer saturates easier at the lower frequency. Avg. reading 2.77 vac.

Here is the same transformer getting smacked by a 40 hertz square wave.
The scope reading is wrong, there is actually a 11.4 volt peak to peak square wave being input.. Avg. reading was 4.42 vac.
The square wave saturates the transformer a little more easily.

Here is the square wave at 20 hertz:
The input voltage required to saturate the x-former dropped from 11.4 to 7.20 volts peak to peak, avg was 2.32 vac.

Just some more tidbits to file under your "useless information" directory!
Party on, Garth!

cj :guinness:
If I can't fix it, I can fix it so nobody else can!


B-H Curves
« Reply #6 on: September 03, 2004, 11:23:00 PM »
Another useless tip:
Use brass hardware when bolting together transformers.
Brass is non ferrous and therefor won't mess up your flux.

Use it when mounting transformers also, like the API output.

OK, here are a few pics of some BH curves I traced.
I am guessing 80 permalloy, what you say?

If I can't fix it, I can fix it so nobody else can!


B-H Curves
« Reply #7 on: September 04, 2004, 11:07:08 AM »
Chris I did an experiment I installed a small 10:1 microphone transformer in the square microphone.  It had a an amophous core.  It sounded great at speech level, it also sounded better with  female voice.  When it was driven harder a "fizz" was heard in the output.

  What was interesting is that the sound was good to a fairly well defined level/ fequency.  I found that interesting it was like a switch.

  Out of habit I tend to like big cores with microphones.

  What got me thinking about this is the size of the u87 transformer.  It seemed small to me but it did not sound bad.  It appers that neumann controls the gain of the circuit as to not overload the core.

  Maybe all one wants/needs is just enought core for the biggest level the transformer will "see".


B-H Curves
« Reply #8 on: September 06, 2004, 12:46:26 PM »

That's all GOOD stuff!  I think you are really getting the hang for that stuff.:thumb:

One observation:  Have you thought about getting a different scope?  We have seen a lot of things on an analog scope that those LCD Teks will "miss".  HTH!

"The sow would rather have her ear than a purse." - PRR

B-H Curves
« Reply #9 on: September 06, 2004, 07:10:57 PM »
Quote from: "SonsOfThunder"
70% calculation and about 30% 'get out some cores and wind some wire'.  Charlie

Most of transformer design is based on model measurements.
You wind some transformer, measure it and if you wind different
(with the some winding), use similarity to compute parameters
(winding capacity, leak/shunt inductance ratio and so on)

Is using models woodoo? I mean, that not.
Water turbine design is based on other similarity and they are
engineered very fine.

Mathematics is model also. But this model is not solved by measurement,
but by computing.



Related Topics

  Subject / Started by Replies Last post
12 Replies
Last post March 25, 2006, 04:21:24 PM
by Gus
10 Replies
Last post July 03, 2008, 02:08:52 AM
by PRR
1 Replies
Last post May 28, 2009, 11:45:35 AM
by emrr
0 Replies
Last post August 09, 2010, 10:21:00 AM
by listen202
7 Replies
Last post March 03, 2011, 11:49:21 AM
by ruffrecords