sine becomes triangle at higher frequencies

[rmallik]

hi all, i made a class d amplifier with HIP4080 and 4 x STP55NF06, aiming at 150W into 4 ohms. The schematic is basically AN9525.2. The error amplifier used is LF353, Fsw is 200kHz. Aiming for a B/W of about 5kHz only. The circuit is working, drain ringing is about 10% of the supply (+40v) at full output. The clipping is symmetrical, and clean. Byt the signal starts becoming triangular instead of sinusoidal from about 400Hz and up. Also, the -3dB bandwidth wont go above 2.5kHz, no matter what I do with the feedback loop. What seems to be the problem? Tri wave is 3V Pk-Pk.
bye,
rmallik

[Baldin]

Could you post your schematics, even though it resembles the AN?

Is it only at full power that the output becomes triangular?

Could you post a scope picture ore something?

I'm building on af similar project, so I hope I can help you.

[rmallik]

Hi...thanks for replying. It starts becoming triangular from about 35watts level or so. With no load connected load, at higher output level say from about 12-13v rms level, it becomes triangular. Is the error op amp slewing? Also what about the bandwidth? It will not go above 2.5kHz... I will try to post the schematics soon.

[janneman]

What power supply do you use? It may be, that at the higher power levels, the power supply is limited in current, and therefore the output slews at that max current.

Jan Didden

[EC8010]

Yes, as Jan says, you are seeing slew rate limiting due to insufficient current to charge a capacitance somewhere and all the feedback in the world won't help you. You need to identify the capacitance that is causing the problem and either reduce it or increase the current driving ability of the stage feeding it.

[rmallik]

OK but then it is happening without any load also. The power supply is overrated. Filter caps are 2 x 6800uF. And why is the bandwidth limited to 2.5k....The loop cutoff is much higher. I tried reducing the Tri wave amplitude( increasing open loop gain actually) but that too did not help. Even with components exactly as in AN9525, only LF353 instead of CA5470, the same problem is observed.
Thanks all...for replying.

[Gorilla]

Looking at Fig. 2, if there is any DC in the feedback loop you will saturate that capcitor in the error integrator feedback.

Try adding R8. That will limit DC gain.

Also make sure your amp isn't oscillating..

BTW: Is it suitable to use an electrolytic capacitor in such a feedback loop as the schematic depicts?

[Eva]

Your output filter is probably out of tune by an order of magnitude or so... Component selection is critical there.


You can check what signal is being actually fed by the switching stage to the output filter by employing a couple of auxiliary RC low-pass filters.