IRS27951 / IRS27952

[nigelwright7557]

Originally Posted by lorylaci

Rewind the transformer primary for 13 turns. For 50V secondary you need 4 turns. Then measure the leakage inductance and the magnetizing inductance.
Then we can move forward.

I have rewound primary to 13 turns and secondaries to 4 turns.
This gives (with no gap) 200uH and 96uH

What next ?

[lorylaci]

Originally Posted by nigelwright7557

I have rewound primary to 13 turns and secondaries to 4 turns.
This gives (with no gap) 200uH and 96uH

What next ?

According to the ferrite datasheet, your total primary inductance should be 752uH with 13 turns and no gap. According to my approximations your leakage inductance should be about 15-30uH.
So there must be some problem with your winding or your measurement. Could you make a close picture about your transformer winded?

[nigelwright7557]

Originally Posted by lorylaci

According to the ferrite datasheet, your total primary inductance should be 752uH with 13 turns and no gap. According to my approximations your leakage inductance should be about 15-30uH.
So there must be some problem with your winding or your measurement. Could you make a close picture about your transformer winded?

The core material is 3c90.

I checked my LC meter with some known inductors and it reads OK.

[nigelwright7557]

Originally Posted by lorylaci

According to the ferrite datasheet, your total primary inductance should be 752uH with 13 turns and no gap. According to my approximations your leakage inductance should be about 15-30uH.
So there must be some problem with your winding or your measurement. Could you make a close picture about your transformer winded?

Bu that would give a k factor of 25 when the recommended k factor is 3 to 10.

[nigelwright7557]

Originally Posted by lorylaci

According to the ferrite datasheet, your total primary inductance should be 752uH with 13 turns and no gap. According to my approximations your leakage inductance should be about 15-30uH.
So there must be some problem with your winding or your measurement. Could you make a close picture about your transformer winded?

I was getting about 700uH with 36 turns so I wont get it with 13 turns.
In fact scaling the turns down does give what my meter is saying.
And the secondaries are the same number of turns so no change there which again sounds correct.

[nigelwright7557]

I used the transformer as you suggested.
I worked out Cr as 100nf.

I powered up the SMPS and it outputs +/- 50 volts DC.

I need to do some more testing on loads to make sure it works ok.

[lorylaci]

Originally Posted by nigelwright7557

I used the transformer as you suggested.
I worked out Cr as 100nf.

I powered up the SMPS and it outputs +/- 50 volts DC.

I need to do some more testing on loads to make sure it works ok.

Nigel, your transformer is seriously not ready.
It wont work OK, it can even destroy the FETs and the controller.

Ferroxcube ETD54 made of 3C90 has an Al of 5000nH. So your 13 turns should have a total inductance of 845uH without a gap.

Your transformer is very ugly wound, and the distance between your primary and scondary is almost 1 cm. Just try to move the secondary a bit closer (aobut 2mm) to the primary, your leakage inductance will decrease rapidly.

We arae aiming a leakage inductance of 15-30uH. After we get something like this, we apply a gap to set tha magnetizing inductance, tog et a proper k value. Setting the magnetizing inductance is the last (since it is easy).
However if your meter reads a different value than we calculate from the datasheet of the ferrite core and the turns, we must think on "Where is the error?", because it can be caused by the following:
- ferrite half cores ar not grinded
- you dont push them together when measureing (it is just a measurement problem, it will not cause us harm, however we must check it to rule out other errors)
- there is a magnetic pathway short
- there is some problem with the wire insulation, its leaking

So better check out.

Also you are using thick wire with only one strand. Unfortunately I do not have a picture of my transofrmer without the yellow insulating tape, however I can show picture about a nice wound transformer.

As you can see it uses multiple strand of thin wire to mach currents. Look at how nicely it is wound.

[nigelwright7557]

My transformer looks fine, the cores meet up ok and are tight.
The inductance is 100uh and 200uh which I cant make bigger without adding extra turns.
Adding a gap makes the situation even worse.

[nigelwright7557]

I found an old etd34 core and wound 13 turns on that to see what I get.
I get 20uH !
I checked my meter and that is fine on some inductors I had lying around.

I took out the cores on my original transformer and turned them 180 degrees.
I now get 250uH and 100uH.

I rewound the transformer with thinner wire but in pairs.
I still get 240uH.

[nigelwright7557]

I also found another etd54 set of cores and tried those.
I get 240uH and 100uH.

I took apart an old fly-back SMPS I had with an etd54 transformer and measured the primary inductance with 40 turns and it was about 1000uH.
So 13 turns at 250uH isn't miles away from what is expected.