Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Planar Transformer Magnetising Inductance Measurement Paradox!!!

Status
Not open for further replies.

Shiladri

Newbie level 4
Joined
Jun 6, 2012
Messages
7
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,350
Hi,

I am facing a peculiar problem in my board-integrated Planar transformer...

It is an 8-layer board where the primary and secondary windings are interleaved(P1 S1 P2 S2 P3 S3 P4 S4)..connections done by vias..

Each primary layer has 2 turns and each secondary layer has 8 turns..thus in total there are 8 primary turns and 32 secondary turns..

The core is a standard E-I core manufactured by Ferroxcube(EI 38/8/25)

I am observing something really strange while measuring the magnetizing inductance(I am measuring this by keeping the secondary open and measuring

the inductance at the primary side using an RLC meter)...The results are as follows

1.Without the core introduced the RLC meter reads 1.5 microH (expected as the reluctance is high because the flux path is through air)

2.With only the E core introduced the meter reads 3.8 microH (again expected as reluctance is slightly reduced..)

3.But the strange thing is that with the 'I' part also introduced(i.e now I have the complete E-I set in place)..the RLC meter shows the terminal

impedance is capacitive(something like 10 nF)..

This really baffles me!!

With the E-I core in place the reluctance is the least and hence the magnetizing inductance should be maximum... how this becomes capacitive is beyond me?

Can anyone plz help..??

Could there be some issue with the board which is causing this problem??But then without the core the board shows 1.5 microH as I said..

Also I have seen that it is working perfectly fine as a 1:4 transformer w/o any core (checked this by applying a sine wave of 1 unit from a function

generator..I get a sine wave of 4 unit at the secondary)...

With the core in place the input signal from the function generator gets distorted and is no longer a sine wave(becomes like flat topped sine wave)...

This problem really beats me!

Any inputs would be greatly appreciated..
 

mtwieg

Advanced Member level 5
Joined
Jan 20, 2011
Messages
3,600
Helped
1,270
Reputation
2,546
Reaction score
1,267
Trophy points
1,393
Activity points
27,098
What kind of RLC meter is it? What is the test frequency?
 

Shiladri

Newbie level 4
Joined
Jun 6, 2012
Messages
7
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,350
What kind of RLC meter is it? What is the test frequency?

Its a FLUKE Make Programmable Automatic RLC meter-The Model is PM 6306

The test frequency is 200 Khz
 

godfreyl

Advanced Member level 5
Joined
Apr 18, 2012
Messages
1,981
Helped
632
Reputation
1,266
Reaction score
624
Trophy points
1,393
Activity points
12,772
There is capacitance between the windings on the PCB. This effectively forms a parallel tuned circuit with the inductance.
 

Shiladri

Newbie level 4
Joined
Jun 6, 2012
Messages
7
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,350
There is capacitance between the windings on the PCB. This effectively forms a parallel tuned circuit with the inductance.

If it is due to the inter-winding capacitance, that capacitance should also be present when the core is not introduced...

Why is it that without the core ..I see some magnetizing inductance...but with the core in place the impedance becomes capacitive??
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,516
Helped
14,271
Reputation
28,805
Reaction score
12,974
Trophy points
1,393
Location
Bochum, Germany
Activity points
280,594
Isn't it obvious? By closing the magnetic path, you get an inductor SRF below measurement frequency, so the total impedance is in fact capacitive. Simply select a suitable low measurement frequency, e.g. 100 or 1000 Hz.
 

godfreyl

Advanced Member level 5
Joined
Apr 18, 2012
Messages
1,981
Helped
632
Reputation
1,266
Reaction score
624
Trophy points
1,393
Activity points
12,772
Why is it that without the core ..I see some magnetizing inductance...but with the core in place the impedance becomes capacitive??
Without the core, the resonant frequency is above your measurement frequency.
With the core in place, the resonance frequency is below your measurement frequency.
 

mtwieg

Advanced Member level 5
Joined
Jan 20, 2011
Messages
3,600
Helped
1,270
Reputation
2,546
Reaction score
1,267
Trophy points
1,393
Activity points
27,098
Yeah self resonance was my guess for the problem. But it's quite surprising that a planar transformer with N=8:32 has such a low SRF, and a whopping 10nF of primary-referred parasitic capacitance. That sounds too big, even for an 8 layer board, but I've never made such a thing myself. You should also check the parasitic capacitance between primary and secondary. It's probably also large, and may be a big problem for common mode noise.
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
48,516
Helped
14,271
Reputation
28,805
Reaction score
12,974
Trophy points
1,393
Location
Bochum, Germany
Activity points
280,594
The said 10 nF parallel capacitance sounds large at first sight. But the order of magnitude seems realistic after transforming it to the secondary (625 pF) and consideing the large windings area and low substrate thickness.
 

mtwieg

Advanced Member level 5
Joined
Jan 20, 2011
Messages
3,600
Helped
1,270
Reputation
2,546
Reaction score
1,267
Trophy points
1,393
Activity points
27,098
Yes, I was envisioning something smaller than the actual core size described in the OP. Even still the capacitance is surprising, even after accounting for transformer action and substrate thickness. I know I've seen planar transformers of near the same time which work in the hundreds of KHz range with no problem, so there must be some kind of trick to how the windings are laid out to reduce parasitic shunt capacitance (probably at the expense of leakage inductance, though). Or maybe they work with a much lower magnetizing inductance than the OP needs.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top