[morzh]

In order not to reinvent the tex I will put what I wrote on ARF.

So I went and looked at the transfer of the remaining good xfmr.
I fed it signal from 1kHz to 6kHz directly from my HP oscillator, while looking at the input signal directly across the Primary and at the output directly across the secondary. 

This is the setup.

The HP200D oscillator I havn't fired in over 10 years (restored way back).
   

The scope I fixed this summer with two channels. The channels are connected right across both windings, with the sig gen also right across the primary. So what we see is what is across the windings.
   


Now the inputs are in order: 1kHz through 6kHz.
Every photo will have the secondary to primary volts ratio (which is, more or less, should be the turn ratio, right? The claimed one is about 3:1)

1kHz - 0.65
   

2kHz - 0.7
   

3kHz - 0.825
   

4kHz - 1
   

6kHz - 1.66
   


So, so much for the promised 3.

I guess, the core type, the parasitics and the losses (what losses? 10M scope probe).

I feel justified replacing them.


PS. I was told that this is not the way to check an interstage.
Well, what I will do is, I have just ordered the P-T156 from AES, they will come in a week or so (free shipping USPS, you know) and I will repeat the experiment with something that has the modern core.