12 vdc to 18 vdc isolated P-P

[jamesrnz]

i have been trying to get this going but i can only get about 50% efficiency out of it.

right now the current loop is not connected

and the voltage loop is stable.

it is 11-14 vdc in and 18 volts out at 23 amps

i am shooting for minimum 80% efficiency

it is running at 110 khz

i have to charge a 14 volt battery.

aaahhhhhhhhhhhhhhhhhhhhhhh

[mflorin]

I can undesrtand, in some situations (probably high power apps), the decision of using two separate transformers. However, for your 414W output power @ 110kHz I think that a single transformer will keep things simple and under control.
Assuming that both transformers are using the same core type (which can not transfer all the necessary output power), you can put the cores in parallel.
If you're using "off the shelve" transformers, then I suggest you to remove all copper and try making your own transformer.

You can also use for reference the voltage at pin 16 of MC33025. This way you can skip the 7809.

For charging purposes, the average current mode control is well suited.

[jamesrnz]

i wanted to use them in series to keep the step up ratio low, but what ever works will happen.

i tried a single EE set but could had too much loss and inductance, ultimately i wanted 54 amps. i used litz wire foil etc.

i have much good help on various projects from the forum. the low current ones work well.

the primary is each side wound with 3t ct 3t, 30 wires of #36 (2 sets each side of pri) then 80 wires of #36 then the pri again.

the cores are pot OP43622UG.

good point on the reference pin..........


so you are saying just parallel the transformers. change the sec winding ratio a bit.

i wanted to keep it around 1000-1200g for continuous duty.

i NEED to make this work..

thanks

[mflorin]

Unfortunately, pot cores can not be put in parallel (or I don't know how).
An even number of E cores (read it "halves") can be arranged together to build a larger core (EI, U also work). Toroid cores can be stacked for the same purpose.
For the other types you just have to use the right core which has a volume determined by the application specific conditions (output power, switching frequency, topology, core and copper losses, ...).
Look for Erickson's book "Fundamental of power electronics". An optimized method for choosing the right core is presented there.
Yo can also try to see how much power can be transferred through your pot cores @ 110kHz by keeping the core losses to an acceptable level (let's say 100kW/m^3).

Assuming a theoretical 50% duty cycle for each switch, your primary to secondary transformer turns ratio should be around 0.61 (11/18). So, for 3 turns on the primary side you need 5 turns on secondary. I would say that is a low enough "step up ratio".

Knowing the transformer turns ratio, the primary number of turns for a given core can be calculated by using Faraday's law.

[mflorin]

check this link:
http://www.mag-inc.com/pdf/2006_Ferr...nformation.pdf

[jamesrnz]

mag-inc sent me some samples of 0P43622UG that they suggested for my requirements.

[N-Channel]

Hi James,

Not to disrespect m-florin's suggestion about paralleling secondaries, but I like the idea of series-connecting the two secondaries. Seriesing the secondaries also guarantees equal current sharing, and as you said, reduces the number turns needed for the secondaries. So, even if there is the slightest difference in the secondaries between the two xfmrs, series operation doesn't care.

Here is a schematic of the power section of a 2.5kW inverter I have that uses multiple transformers with secondaries seriesed. Obviously, this is for high voltage (+165VDC) for the 60Hz output, but the same principle applies for your circuit. If I can find the pic, I will also post it. HTH,

Steve

[N-Channel]

Here is a pic of the inside.

[jamesrnz]

cool. i am not far off. i think i got this idea from pressman.

i am going to replace my current shunts with some precision ones to see if my Pin and Pout calculations are off. which of course would effect the efficiency, but i don't think they are that far off.

i HAVE to make this work

thanks steve

jim

[jamesrnz]

well the currents shunts made no difference, except they are more precise and have better temperature stability.

so i need to figure out they efficiency problem