schematic of my future D amp for the pb12

Status
Not open for further replies.
Jump to Latest
#107
OK, Ikadar.
Your measurements are perfectly reasonable. The simulation shows 6.14Vpp of ripple with your values. It increases to your measured 6.4Vpp as soon as the real value of the cap is 450nF instead of 470, or there is a small deviation in supply voltage, coil inductance, freq, etc.

So if you can, you should increase sw. freq. up to at least 280-300KHz. What mosfets and drivers are you using?

Best regards,
Pierre.
 
#108
Hi!

I don't know exactly :bawling:... Here is the schematic. Could you explain? The largest difference is that I use IR2110 instead of IR2011 because I couldn't buy it here 🙁.
Ohh... I've an idea... if the output of the opamp looks like a trapezoid wave, could it be problem 😕??? Should it be look like a triangle wave?

Regards, lkadar
 
#110
Pierre said:
OK, Ikadar.
Your measurements are perfectly reasonable. The simulation shows 6.14Vpp of ripple with your values. It increases to your measured 6.4Vpp as soon as the real value of the cap is 450nF instead of 470, or there is a small deviation in supply voltage, coil inductance, freq, etc.

So if you can, you should increase sw. freq. up to at least 280-300KHz. What mosfets and drivers are you using?

Best regards,
Pierre.
Click to expand...

Hello!

Thanks again... 🙂. In my previous post you can find the schematic... I'll try to rise the freq. as somebody can help me to find the problem.

Regards, lkadar.
 
#111
But once you know that your ripple is just as high as it should be for the sw. freq. you have, what's the problem?
If you want to increase the frequency, you only have to play with R6, C6 or C2 until you get 300-400khz, for example. This way you will reduce your ripple to about 3Vpp, which is reasonable.
Have you got some other problems? (do the mosfet get hot without load, or it doesn't soon very good?)

About the output signal, it can be an almost perfect sine or similar to a triangle, it doesn't give so much info. You must look at the PWM output. Is it perfectly square? Does it have over/undershoot? What are the rise/fall times?

Best regards,
Pierre
 
#112
Hello again,

Well... I replaced the pot. in the integrator with a 5KOhm one and now the whole thing resonates at 440 kHz with or without load at no input. The output riple gets much much lower (now about 1.75V)... (the resonation freq. can be rised but I don't want to. It's satisfying)
Next week I'm going to connect it to my DIY speakers and let's rock... 😉

Thanks again everybody!

Regards, lkadar.
 
#113
Have in mind that the higher the sw. freq, the higher the losses and the hotter your mosfet will become. Having a 3Vpp output ripple is not bad (that's 150mW on 4ohms at an unaudible freq. 😉

BTW: Your mosfets have an VdsMax=100V, just at the limit! You should use at least 150V mosfets for +/-50V rails for reliability.

Best regards,
Pierre
 
#114
Pierre,

You're so fast 😀 ...

Thanks the suggestion... Then I'll use a lower switching freq...
MOSFETs... For the first time IRF540 is good enough, because they're cheap... so if they burn... it doesn't hurt so much 🙂.
Next week I'll try to capture the waves on the scope somehow. The PWM output just before the coil is almost perfect square. It has about 3V to 4V overshoot at the end of the rising and falling edges. I didn't measured the rise and fall times yet.

Best regards, lkadar.
 
#115
My experience is that overshoot gets worse as output power increases. Be sure that you measure it with full load or you can end with avalanche death of the mosfet's.
Read the "mosfet reliability in Class-D amplifiers" thread. There is a lot to learn there.

Best regards,
Pierre
 
#116
Hi,

Pierre is of course correct but, no need to be too paranoid about Fs being too fast. You want it fast enough to reduce THD and also as input and output signals increase with amplitude there is some frequency modulation so it will slow it down which is OK but you want it to remain fast enough under all conditions for optimal sound (filtering, ripple. THD..)

It's not like it shoots up to 1Mhz and beyond, so you're switching losses aren't intolerable. ~450kHz should be just fine.

Let your finger be the judge while idle, it's switching fastest then so your greatest losses due to switching will be with no input signal, but I would be aiming for ~450kHz myself, and I wouldn't sweat 500 or even 550kHz either.

You should of course have some added thermal mass though.

It's good to see you're making progress.

Regards,
Chris
 
#117
I would be interested to know if this solution sounds as good as the Tripath solution. Tripath really beats on PWM in there advertisements, in all honesty I don’t see how there class-T solution differs. But this thread would seem to sagest that PWM could be used full range and not just for low frequency reproduction.

The parts are certainly easier to get. I see everything is available from digi-key. The cost would probably be slightly (trivially) smaller.

From reading this thread I don’t totally understand what was wrong with the original component values, apart from the need for the muting relay? Is the updated final schematic available?

Leve
 
#118
What function in the original IRF schematic did the TC7WH08FU (U5) perform and why was it eliminated. It looks like it stops signals from being passed to the gate driver when SD2 is pulled low. I can see it is hook up to what looks like the DC protection circuit, is that its only function?
 
#119
Hi, the U5 is part of the protection system into the original design, it is used to disable both outputs in the event that it stop to socillate and then lock-up at +vcc or -vcc (or both :xeye: ) so there is no more current draw from the ouputs.

hth

Pat Allen
 
Status
Not open for further replies.
Top Bottom