Yes , it's sure to have some resistor load to have the DC feedback closed and stable . This is true not only for D class amps........🙄
To avoid unstable behavior near res. frequency of the output filter , a zobel network is required .
Leads of resistors truncated, 2k2 resistor on output like earlier, new IRF640N, IRS2092, about 450kHz switching frequency, rest almost same as iraudamp7d, added small heatsink, quiescent current at 36V --> 50mA, circuit don't heats too much.
Now with load or no load square waveform before output filter looks exactly the same.
Now with load or no load square waveform before output filter looks exactly the same.
I have connected it to 2x80V SMPS after changing some resistors and dead time to maximum, because quiescent current was more than 200mA! (Now I can't read value on 600mA ammeter, is to small.)
Amp is working, but small inductor heats a bit, maybe 50-60°C, but I think that it shouldn't be dangerous.
Is worth to buy better fets and bigger core for inductor? Arnold Sendust (like now) is good for such applications?
Amp is working, but small inductor heats a bit, maybe 50-60°C, but I think that it shouldn't be dangerous.
Is worth to buy better fets and bigger core for inductor? Arnold Sendust (like now) is good for such applications?
I use t106-2 core with 27uH of enamelled copper wire on it and it barely gets warm.
Last edited:
I used about 20 turns.
I usually just wind it on till I cant get any more on of 18swg enamelled copper wire.
You can test it using a sig gen and scope by making a res circuit wit ha resistor and capacitor. Look for the short circuit across the LC at the res frequency.
Then use:
1
L=----------------------------------------
4 pi pi F F C
20 turns on an Al=13.5nH core only gives 5.5uH!
The simple formula is that L= turns-squared x Al
The simple formula is that L= turns-squared x Al
- Status
- Not open for further replies.