MC34063 with Very Bad eff. ?? !!

[ahmed osama]

i am trying to make a dc-dc converter

from 9-12 Volt to 18 volt 4 amp

my first trial was using a discerect comp. using 300 Khz but without feedback just constant load ( i was about to begin to make the feedback cir. )

the mosfet peak curr was about 10A and coil peak curr was about :10A

now i said to use MC34063 since desing the feedback maybe it will so diff. , i faced a problem that the IC could only support switch curr less than 1.5A and using MC34063 datasheet design equ. my peak will reach 16 Amp. so i have to use an ext. transistor and on freq, 50Khz

but i shocked i found
the mosfet peak curr became 48A and coil peak curr was about :48A

i used the same mosfet & the same diode in both design so i guess changing them will not solve the problem

so any tips to improve the eff.

thanks
bye

 

[danciurea]

MC34063 controller was designed to drive bipolar transistor as an external switch. From your post I understood that you used a MOS transistor. It is possible to use a mos but with special cautions. Have you took care of that?
Simply replacing the bipolar with a mos causes turn off problems which can give you those high currents. Have you designed you inductor properly so it will not saturate?

 

[ahmed osama]

MC34063 controller was designed to drive bipolar transistor as an external switch. From your post I understood that you used a MOS transistor. It is possible to use a mos but with special cautions. Have you took care of that?
Simply replacing the bipolar with a mos causes turn off problems which can give you those high currents. Have you designed you inductor properly so it will not saturate?

well i tried alot alot to use the bipolar using 2n3055 but i failed so much i don't know why indeed
so what such special cautions about using mos
about the coil i will use air coil so no saturation prob

thanks
bye

 

[danciurea]

Since you used an air coil there is no problem with coil saturation.

The first step would be to insert a push-pull stage made out of two bipolar transistors between the MC34063 output and the mos-fet gate. This will greatly improve your mosfet switching characteristic.
Once I saw a car laptop charger schematic with a mosfet and MC34063. Now I only remember that the mosfet was cleverly connected to the controller circuit but I cannot remember any details. I will try to find it and post it here for you.

 

[FvM]

I think, MC34063 can be operated with a MOSFET, but it can't work correctly without a current sense resistor. Please consult the application notes.

 

[ahmed osama]

I think, MC34063 can be operated with a MOSFET, but it can't work correctly without a current sense resistor. Please consult the application notes.

in datasheet it says Rsense min is 0.2 ohm and in my case it is about 0.07 so i removed , and if i put 0.2 it didn't work well and 0.07 so diff to be done in real life ??

 

[FvM]

The point is to understand that the MC34063 control principle is based on a current sense that ends the on-time period. The 0.2 ohm minimal value is meant for operation without external transistors, to keep the maximum current ratings. With external transistors, safe operation according to their specification has to be achieved.

I want to place a general comment in addition. MC34063 is one of the cheapest SMPS controllers on the market. It shouldn't be expected to give optimal efficiency with high power and external transistors, although it can be used in this field, too.

In my opiinion, it's much simpler to setup a SMPS with a classic controller IC, e.g. 3524 (voltage mode) or 384x (current mode) or one of the more recent types.

 

[ahmed osama]

"........ 3524 (voltage mode) or 384x ............"

ummmmmmmm well some how complicated and very bad doc. and much less resorces and LAST THING NO SPICE MODEL !!! so nice

 

[FvM]

Then you should use MC34063 according to it's specification. AN920 has e.g examples with external transistors (and respective low current sense resistors).

 

[ragav777]

Thank you..helpful