diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Class D (http://www.diyaudio.com/forums/class-d/)
-   -   how to decide the triange wave and input signal amplitude.?? (http://www.diyaudio.com/forums/class-d/76538-how-decide-triange-wave-input-signal-amplitude.html)

titanchen68 26th March 2006 04:09 PM

how to decide the triange wave and input signal amplitude.??
 
I saw almost people design the triangle wave is 4Vpp.
If the input signal near 2V , the square wave of comparator output will near 100% duty.
will it damage the output mosfet ??
How to decide the amplitude for triangle and limit of input signal??

Because i fair that if input signal near to triangle wave peak amplitude will damage the output mosfet.
So, I limit the input signal only has 70% of triangle wave peak amplitude.
And Let the Gain of amp become higher to amplifier the signal.
Did the think right??
I probe the output wave , when output amplitude near +/-Rail to saturate.
The wave form has much high-frequency hormonic at high side and low side.
Dose anyone also has the situation??

titanchen68 27th March 2006 01:13 PM

2 Attachment(s)
Because i only use tl084 to generate my triangle waveform.
Its slew rate only 13V/us. I try it only can rise up to about 80KHz.
now , i use tl084 to generate a 83KHz 4Vpp triangle.
Picture above here: T=5us/div V=2v/div

titanchen68 27th March 2006 01:17 PM

2 Attachment(s)
Then i reconnect the mosfet to the pcb.
And add a dead time control befroe ir2110 HI and LO input.
Turn on the power , i saw the current disspation become more.
I probe the output of the mosfet ,,,
Its waveform has some problem.It get a ramp.

(If i remove the L out , the output waveform form looks good.
It is a sqaure waveform , and no other problem.)

Picture above here : T=5us/div V=20v/div

titanchen68 27th March 2006 01:29 PM

2 Attachment(s)
When i turn on power after few second ,,,
I saw the resistor before mosfet's Drain pin starting burned.

( if i didn't add the dead time control , the resistor get bomb immediately)

1. Is that because the modulation frequency too slow ???

2. what is the correct waveform after RC-filter in feedback network ??
someone said the RC-filter freq must equare to modulation freq.
someone said the RC-filter freq must lower than the modulation freq.
I got confuse about it.
I don't know how to designe the RC-filter in feedback network.
If i designe the LPF freq. = modulation freq.
The waveform looks like a square wave also ,but it get smooth at coner.
If i designe the LPF freq. = (modulation freq)*0.8 .
The waveform looks like a triangle wave.
Picture above here. T=5us V=20v/div
which way is right ???

titanchen68 27th March 2006 03:55 PM

I just has a ideal ...
1. Maybe it's because of my gate resistor too large 100ohm.
when i use 200Khz modulation freq , Rg=100ohm is OK .
I will try lower the gate resistor .

2. Maybe the bootstrape cap too small.
Now i use 1uF cap be the bootstrape cap.
But i don't have any ideal to know it's too large or small..
I also will try it tomorrow.

If anyone has any ideal ,,, please tell me your think ... thanks !!

zkaiser 27th March 2006 05:58 PM

if pic in post #3 representing good sqare shape then you suffering from low bootstrap capacitance, in other words speaking try bigger value like 10u or 22u, but use low-ESR cap or smd tantalum, 100ohm might be large and might not, depends, anyway it setting dead time and makes rising edge of gate signal to act slower. . .

Genomerics 27th March 2006 09:18 PM

Have a look at...

http://www.genomerics.org

The first two parts under "Class D Amplifiers" may give you some ideas about how to work out component values for pre-filter feedback.

Sorry, it's not the full story yet, Work In Progress (WIP), but hopefully it's of some use to you.

Cheers

DNA

titanchen68 28th March 2006 02:31 PM

Quote:

Originally posted by zkaiser
if pic in post #3 representing good sqare shape then you suffering from low bootstrap capacitance, in other words speaking try bigger value like 10u or 22u, but use low-ESR cap or smd tantalum, 100ohm might be large and might not, depends, anyway it setting dead time and makes rising edge of gate signal to act slower. . .
I have try to increase the bootstrape cap to 5uF.
Because i only has 1uF tantalum in hand.
So , i parallel 5 to get 5uF. But it seems doesn't improve.
And i try to decrease the gate resistor to 33ohm , 22ohm and 10ohm.
Still do not work too well.
I will think about it more ,,,

Thanks All .....!!

zkaiser 29th March 2006 10:20 AM

2 Attachment(s)
titanchen68, hi !

I've been thinking something, and reason that yours amp doesn't work might be in level shifter for ir2110, and BTW how you supplying triangle generator and lt1016 there are no volt regs. on schematic ?
Use my drawings of voltage levels and proposed level shifter in try to solve out working issues in your amp.

Can you post latest pic of amp board and stuff ?

Best regards
Zeljko Kaiser

zkaiser 29th March 2006 11:25 AM

and please, use current limited power supply if u don't want any burned parts ~!!!!!


All times are GMT. The time now is 04:37 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2