Transformer Weirdness

[CJ]

So I am working on reversing a particular transformer that shall remain aynonomous at this juncture.
I will say it is a pretty old transformer.
I can't get a good null on the Gen Rad impedance brdige. Mushy need action. Then I read something over at K and K about transformer "break in".
Being ever so the skeptic, (or is it septic!), I put transformer "X" on a signal generator for a week to "break it in"
.
.
Well, guess what! Now the thing has a real sharp nulling action!
Please don't take this as anything but audiophoolory at this point, but I am definately seeing a difference.
Perhaps this transformer has not been used in quite a while.
.
.
Sooo, I do not know why I started this thread.
(but if any of you run into this same thing....)

You might want to give your new iron a while to "settle in" before passing judgement.

Or this might be a bunch of bull, but you have to explore all the options.
At this pint, I do not know what to make of this, so I will try the same thing with some other iron and report back.

And please don't tell PRR!
I don't want to get laughed out of Dodge!
(or Ford, in his case!)


cj

 

[CJ]

I am thinking more along the lines of the Weissections (groups of molecules that switch back in forth with the signal) getting looser in the core steel.
There isn't really any heat in a line level transformer unless there is dc applied.
But I am open to all ideas at this point, so you might have something there.

Just a note on saturation, the Weissections bend to a certain angle, back and forth in the core when ac is applied. When they can't bend anymore, (180 degrees is the thereoritical limit although most core steel isn't that good) this is when you get saturation.

Keep those ideas coming!

 

[PRR]

A transformer that has been put on an ohm-meter can be magnetized.

The GenRad bridge probably won't de-mag it.

High level from a signal generator, turned down slow, will.

I think using the GenRad bridge to look at audio transformers is using a too-sharp tool. What is this "null" that causes you such trouble? We don't null nothing, and any sharp null is likely nulling-out the broad parameters that we hear. And if this is a fixed-frequency bridge, it gives you tunnel-vision.

I would plot impedance magnitude and phase from a known (roughly "matching") impedance. If you can take it down near 1Hz and up to 100KHz you can know all the impedance parameters that are likely to matter.

 

[owel]

You gotta train those electrons. Sometimes, they become slackers and forget what they're supposed to do. :grin:

 

[CJ]

If I can't get a good null, I can't get an acurate inductance measurement.
The Gen Rad works at 1 k hz.
I can inject my own frequency from an external oscilator, which I have been ploting.
Very hard to null below 10 hz, which is why you don't see many inductance curves from manufatures of vintage x-formers below this range.
Anything above 1 k hz is hard to get also.

I am still working on a Crowhurst test bench, but I need two Y inputs plus the x input.
So my Tektronics dual trace scope won't do the Crowhust setup unless I build an electronic switch to change between the two Y inputs.
But at what frequency do I switch at?
Anybody know how to make an electronic switch for scope signals?
I need about a 100 times gain on one of the Y channels also.
This will allow me to plot phase shift more accurately.
There was some other text involved with the Crowhurst article describing in more detail the values needed for the pots, etc. so I am fishing in the dark.
I have done a X-Y input on a 10 ohm series restor for one input, and the winding for the other input, which gives me a horizontal line at no phase shift, a circle for a lot of shift, and a vertical line for 180 degrees shift, but Crowhurst says this is not as accurate as the three signal input that he used.


I have a Lundahl that has a sloppy null also.
It is a brand new 7903.
I am going to put it on the signal generator and see if it follows the same pattern as the other transformer I tested.

 

[thermionic]

Sorry for the thread-creep and general trivia, but they had a documentary on Norman Crowhurst last night on tv. He ran an electronics company, and faked a one man round the world voyage, check this link for details:

http://www.jimloy.com/books/crowhurs.htm

I wonder if he is related to the transformer-icon Crowhurst? :?:

Justin

edit: After a bit of Googling I reckon the two are unrelated, bit of a coincidence that the sailing Crowhurst ran an electronics firm though.

 

[adrianh]

wrong Crowhurst