you are welcom katrino. this is a solution but not so simple. the small difference must be used or you get audible distortion.
katrino I wanted your help to understand. I did the test of my cheep class amp d with irfp250n as output. I measure the temperature of the heatsink. Psu + -70v output amplifier loaded by resistive dummy load. 8ohm 56 ° after one hour. 4 ohm 85 ° after half an hour and protection cuts. The test is done with hard rock music at clipping. is it normal. Now I do not have the speaker that support this power to test in real load. Do the speaker require less power than resistive load . Help me please to understand if this temperature is normal. With 6 ohm after one hour 65° measure the temperature of the heatsink. Help me please to understand if this temperature is normal.
Last edited:
Test with dummy or speaker more or less the same.
The big difference if the input is music is less stress compared to continuous plain signal i.e. signal generator.
My experience, the concern of heat is the core losses, not the mosfet. During test a small fan is required just for core cooling. Mosfet is always with proper hestsink and without fan usually less than 50C temp. So if it is.hotter, I believe there is cross conduction OR inductor uH is too high and usually the low side mosfet will be hotter and could die. So check.these two.
thxs katrino. my inductor are 35 uh and a big one. in test the core is at 40° max. if i use irfb4227 heatsink temperature is acceptable 50 -60° full load. but with irfp250n temperature increase very fast.a friend tell me that a speaker current demand is less than simple resistor. resonnance coil speaker at low frequency is high. amplifier see not 4 ohm at this zone but more than 10ohm. and 3/4 of music frequency is in this band. consumption is less than an ordinary resistor. it is true?
i,ve tried but it doesn't help in any way .
i want to try protech approach by changing the 4049 to 74hc00 , since he claimed almost all 4049 are fake. i hope all goes well . i will post results.
the reason i am persistent with sigma delta is that i have mounted up to five channels of class d 200wt rms in one casing without any problem.
i have ever used former and smps . and no noise issue at all or foolish instability where one channel fails for no reason at all . the only challenge is the voltage maximum has to be +/-38vlt dc or the 540 and 9540 fails.
diysmps
. ucd is abit stable but noise issues is hard to deal with, esspecially if you mount 4channel in one casing . also very unstable when you mount more than three channels in one casing.
irs 900 and the likes are a no go zone if mounting more than two channels in one casing at above +/-55vlts dc . if you remove speaker from one channel when playing . the amp fails.
Yes it is with speaker, impedance will vary from time to time. However for average will no much difference.
35uH is too high maybe. To check it just very simple. With standard LPF cap usually 470n or 680n, thr ripple of output at no input (close volume) shall be 2.5 to 3V pp. If less then inductance is too high. If more then inductance is too low.
The problem of BTL for Sigma Delta is the only pwm speed slip. Otherwise your amps not so stable because of parts quality issue.
For UcD again the same problem of noise will occur but in this time noise occur because both channel pwm speed are almost the same speed but not precise. To solve UcD noise then make them synch of make them each at least 20kHz difference pwm speed.
your problem is probably bad grounding i have an experience with that. try to put 100n capcitor between -high voltage rail and the 0v of audio input aop to close as possible to the aop ic. and try to shunt gnd input audio both amplifier together and see result. be careful if you have ground noise issue amplifier can explose if you shunt input gnd together. belive me
I have try some things with a discrete comparator and a IR2011 and lead lag and poles to get low distortions, however the IR2011 model for LTspice give a strange square wave and using dead time give more shoot through then without, strange.
I did put this here because here are some more tryouts and learning is always oke.
regards
I did put this here because here are some more tryouts and learning is always oke.
regards
Attachments
do you have a circuit and pcb your amplifier fullbridge?I want to build it.
I did sim btl version, it give strange outcomes, need a 30 khz higher carrier on one side. better is indeed a full bridge, no need for sync.
Last picture you see the voltage is strange but the current looks oke with very low distortion, not yet dead time implemented tough.
regards
Hello Mr. Kartino. What do I need to change in the circuit to start with a power supply of 65-70 volts?