Adding inductance to transformer primary

[wonka]

2020 48V 15KW PowerJack.  

I have some ferrite e-cores to play with, and I'm trying to get the idle current to drop a bit.  Currently, there is one OEM choke on the primary, and from that choke it splits to connect both mainboards. Would it be better to install 1 additional choke for each mainboard (adding 2 more chokes)? Or both boards combined in a single additional choke, in the fashion of the OEM PJ choke?

I wonder if each board having it's own additional choke in series would provide for some isolation between the two?  any benefit here? Is there any concern of causing the 2 mainboards to be (more) imbalanced?

 

Option 1:                                       PJ A1-----------(new L)-------+====(PJ L)====xfmr

                                                      PJ B1-----------(new L)-------+

 

Option 2:                                       PJ A1-----------+===(new L)===(PJ L)=====xfmr

                                                      PJ B1-----------+

 

Forgive my ASCII art, I cannot get an image to upload/attach to this post. 

 

[Sid Genetry Solar]

As long as the ferrite combinations are balanced between the 2 mainboards, you should be just fine.  Definitely would not recommend imbalanced ferrites between the 2 mainboards.  Both mainboards are driven with exactly the same drive signals (as in, same connection points, not separate outputs).

The stock PJ ferrite is pretty good...IF it has 2 full turns of wire around it.  In my experimentation, it's roughly the same as an E-core (have experimented with those as well).

The ferrite is basically absorbing/filtering the 24KHz SPWM carrier from the FETs.  The main transformer will have an extremely low impedance at 24KHz, causing extremely high no-load current due to the transformer deadshorting the FETs.  The more of this "carrier" that is absorbed by the ferrites, the better the no-load current will be.  The lower limit of the no-load current is the main transformer's magnetizing current + switching losses / control circuitry power requirements; it is not possible to reduce the no-load current farther than this with ferrites.

[wonka]

Thanks, Sid, that's what I was thinking. I tried the E-Cores in both configurations; both produced favorable results, but the idle consumption was a tad bit lower with 2 separate chokes.  I was able to get it down to 46W at idle (down from ~110W originally).