An absolute value would mean nothing: you have to compare it to something, the magnetizing inductance or the magnitude of the load impedance converted into reactance for example.
For a forward or symetrical converter, a leakage value of 0.1% of the magnetizing inductance is the normal order of magnitude (+/- many 10's or 100's of %: depends on the permeability of the core, etc).
If you convert the leakage inductance into Ω at the base frequency, you can also estimate the drop it will cause with a given load, but it will be just that: an estimation.
With purely linear elements in a sinusoidal regime, you would draw the phasor diagram and compute the vector losses accordingly, and with DC you would treat all impedances as purely scalar and apply straight Ohm's law, but the reality with a converter will be something in-between: it just allows you to put limits on what you should find, no more.
Ok, so we wounded the transformer 2 times today. This was the two ways we did it:
1st:
- Half primary
- Complete secondary
- Half primary
2nd:
- Half secondary
- Complete primary
- Half secondary
We got the same results with both: 47V at about 2A, 42V at about 4.5A and 28.5V with a 4 ohm load (before we got 32V here). I don't know what else to change, maybe adding more capacitance at the 220V or maybe that's all the power I can get but I was expecting much more from an EE42 core at 100KHz
Switching PSUs usually spec low ESR caps. I built one, and lacking the required low ESR cap, I substituted a "normal" one and got lots of ripple and the capacitor also got hot.
More parallel wires? The DC resistance is very low. I have no area left in the transformer, the window area is completely filled :/
I still think the core can take more copper around it.
I do not know what core you are using exactly, but for example if it was a Epcos E42/21/20-N97, it should easily take 200mT at 100kHz.
Therefore you might put much less than 33 turns around it.
0.5mm wire is good up to 100kHz without too much losses in to skindepth.
If your transformer is to be connected to the wall outlet, I would wind the primary first, even though it is not optimal. Just for safety reasons that is.
5mm insulation on each side of the winding.
26mm window width - 10mm for the insulation means 16mm window width for the primary.
Lets put 16 turns (did no calculations! Took it from my sleeve) around the core with the 0.5mm wire and make that with 4 parallel wires.
Wind it bifilar, do not twist them in to a Litz.
You will get only 2 layers and have used only 1mm from the available window height.
Then put the grounded shield in there next.
Now make the secondary winding, and use the whole window width for it.
And, make the secondary with FOIL, not with wire.
Note, I have no info about your core material. It may not be stressed to this level, but: Even if you doubled the turns for the primary, you should be able to fitt all the copper to the window. Doubling the primary means 2mm used from the window height (= nothing).
Ok! We calculated the primary for a Bmax if 100mT so with 200mT, 16 turns should be ok. Using FOIL you mean aluminum paper (not English speaker
)? Interesting never saw that in the book. We will do this and post the results! Thank you!
Ok! We calculated the primary for a Bmax if 100mT so with 200mT, 16 turns should be ok. Using FOIL you mean aluminum paper (not English speaker
Copper foil. Not aluminum.
Available from the local "hobby shop" as sheets, cut to needed width. Thickness 0,1 - 0,2mm...
Or should also be available coiled from Ebay in 25mm widths... Note, NOT TAPE.
Found a German manufacturer once, but cant remember anylonger what/where...
Sold ridiculously pure 99.98% copper foil for transformer use...
But any foil will be pure enough. 99.8% still OK...
Check from www.ebay.de for ex...
"kupferfolie"
Last edited:
Ok! Will try to get it, we can't buy from eBay :/ In case we can't get it, is any other alternative? Using bifilar too maybe? Why litz wire is not good here?
Great! As I have 16 turns on the primary side instead of 32, should I change secondary to have 6+6 turns or keep it with 12+12 turns?
6+6
Wind the two foils both at the same time, one above the other, both foils at once that is. Use Kapton tape as insulator between the foils and turns.
Bend the foil to 90deg angle to get out from the core. You can fold that bend over once or twice to make it nice and narrow enough. Cut some bobbin pins to make room for the foil to get out between them...
Wind the two foils both at the same time, one above the other, both foils at once that is. Use Kapton tape as insulator between the foils and turns.
Bend the foil to 90deg angle to get out from the core. You can fold that bend over once or twice to make it nice and narrow enough. Cut some bobbin pins to make room for the foil to get out between them...
The core is EPCOS B66329 (E 42/21/20) N87 material so I think it will be ok!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.