irs2092 osc freq prob

[philuknow]

hello everyone, i have some probleme with the irs2092.
my oscillating frequency is NOT constant, its as if the triangle wave did not exist, the oscillating frequency is the same frequency im putting in..

i've done the suggested circuit in the datasheet! with some upgrade like zeners and blabla.. the power bridge is working damn well tho but i'll need to adjust some of the OCS setting for low Zout.

any suggestion would be appreciate

[nigelwright7557]

You need to check your circuit again against the datasheet especially around the oscillator.

The 2092 is touchy in regards to decoupling. 100nf b+/gnd b-/gnd and b+/b-.

The other touchy thing is the inductor, I use 2 meters of swg18 on a t106-2 core.
Power inductors just get red hot.

[philuknow]

the probleme is not decoupling i was really careful about it.. i have cap everywhere in my design, the fact is that my oscillation frequency isnt constant, i put a 1khz sinus wave as input, and its the frequency i get on the comp pin.. dunno why

[nigelwright7557]

Originally Posted by philuknow

the probleme is not decoupling i was really careful about it.. i have cap everywhere in my design, the fact is that my oscillation frequency isnt constant, i put a 1khz sinus wave as input, and its the frequency i get on the comp pin.. dunno why

You can check the osc frequency on the output as there is a residual on the output signal. My 2092 circuit runs at around 200KHz.

You can also check pre inductor for the osc frequency.

[Eva]

In general, in phase-shift self oscillating circuits, when playing 1khz at full output, both the 1Khz component and the carrier (oscillation) component should appear at the input of the comparator with equal amplitude. That's how modulation takes place.

[philuknow]

Hi, i didnt know it was a phased-shift oscillator, in fact i never heard of these before.. i just looked them out a bit as you say it.. i tought it was generating a triangle wave whom was comparated with my input signal, it explains some things!

in attachement i have put the comp pin with no input (the triangle) and with sinus input. is it normal, then, that i have a triangle wave of 1MHz ? is it the anti-glicht system ? cause it does not appears every time...

Also, the high side OC set up if i put a cap to the ground as output filter.. do you have any idea ? i have tried from 10uH to 250uH and it doesnt seem to change anything

and i almost forgot.. the supply power seem to have much variation when i put a load, even if i put many cap.. nothing heat tho

this chip is more fun then i thought :P

[nigelwright7557]

Did you resolve your problems ?

[philuknow]

YES, its all good i have a sweet pwm output, but If i adjust my oscillator frequency, it goes bad! i think its because im on breadboard, in fact im almost sure its for that. tho i have adjusted, with a pot, the oscillator resistance to get something OK and put a buffer before the amp input to adjust impedances.

[luka]

heh, good thing that it even works on breadboard

[nigelwright7557]

Originally Posted by luka

heh, good thing that it even works on breadboard

Does it work into hard clipping without resetting ?