Bridged ClassD 800W in 4 ohms - oscillation problem

[Pafi]

I mean a description of something that reminds us of a ringmodulator,

Wich is not present on output, but inside of the amplifiers.

What sort of effect does it remind you of?

(De-)Modulation, of course. I told it in my first post.

I just did not know that you where allergic to ringmodulatoers, Pafi.

I'm not allergic to them, but here there is not any of them. (What I'm allergic to (for example), are spelling mistakes. They make me very exhausted when I have to decode them.)

It DOES remind us of a ring modulator effect.

1.: Do you really think this was the question?
2.: Who are "us"? I'm not one of them!

Im sorry to say this, but if I change one of the carriers to 96000 Hz, the alien tone goes to 6oooHz... Just as a ringmodulator would do...

I don't know where you've learned distraction. Please calculate it again! This too: "in my case 104000 Hz and 98000 Hz which gives me a measured, de facto, proven and recorded alien TONE of 4 kHZ in the output."

It will affect the carrier amplitude if it is damped by for example a tuned LC network.

Yes, it would affect, if it were damped by a tuned LC. This is not the case. It's not tuned to carrier freq. 1/(2*pi*sqrt(20uH*4.7uF))=16 kHz.

What is your problem really?

I'm allergic to stupidity, ignorance, and people who think much more of themself then they are. But if you want to talk about psychology, you ought to find an other forum for that!

 

[luka]

What is your problem really?

I'm allergic to stupidity, ignorance, and people who think much more of themself then they are. But if you want to talk about psychology, you ought to find an other forum for that! [/B]

That is you problem, keep it to yourself...help people, if you are right do your best to show them that without personal feelings towards others, if you are wrong be open to what others will say and try to find common path.

Thanks

 

[Zoors]

If you add or subtract two different tones you do not get a difference freqency component. However, if a nonlinearity is introduced then there will be a difference freqency. It can be seen as the amplitude envelope getting detected like in an AM-receiver. But without nonlinearity there will be no difference freqency component.

So it must be a nonlinearity here.

I'm allergic to stupidity, ignorance, and people who think much more of themself then they are. But if you want to talk about psychology, you ought to find an other forum for that! [/B



That is you problem, keep it to yourself...help people, if you are right do your best to show them that without personal feelings towards others, if you are wrong be open to what others will say and try to find common path.

Thanks

Thanks.

And thanks for the calculation, Pafi. 16 kHz... I can see that is way to high for this amp, as it only works under 100 Hz. Its a subbass amp. I still think about the Q factor of those LPF's.
And it is really a 2x300W bridge not a 2X400 as I originally said.

I am driving it with 4 ohms passive load (effect resistors) on the bench. Next session I will try to alter the feedback network to see if the carrier can be tuned in anyway. If I check som IR applications there could be a trimmer pot inserted just in the PWM circuits. That I will definately try.
I will also swap the caps to see if theres a change in Q. They work under extreme heavy load, shorting a 100V square power carrier to ground... Leaving like nada.

 

[mag]

Hi Zoors,
if you are driving a 4ohm load with brigded configuration you need to develop two LPF working on 2ohm load.

Your filter 20uH+4.7uF gives you a -3dB cutoff of 20kHz, at 100kHz (carrier frequency) you have an attenuation of around 30dB.

It is a subwoofer amp, so you don't need to amplify the whole audio band. Your carrier frequency is 100kHz so try a filter with around 10kHz cutoff.

A filter with 10kHz cutoff on 2ohm based on the Butterworth prototype is the following:
L=45uH, C=5.6uF. It gives you f0=10kHz, attenuation at 100kHz of 40dB.

Your concerns about Q are related to the Q of the components or the Q of the filter itself?

The Q of the filter is given by the prototype you choose (as Pafi says it is normally between 0.7 and 1, 0.707 for a Butterworth).

At those frequencies don't worry too much about the components Q. A very high quality inductance can be realized using T106-2 or T106-6 cores with a suitable wire gauge.

For the capacitor use MKP types, if you want to reduce the stray inductance use multiple small value capacitors and parallel them togheter.

In any case I don't think the filter itself is the problem. I have made an amplifer similar to yours with two amps not sincronized driving a bridged 4ohm load. Even changing the frequency of one of the two amps I have not noticed any frequency beat. I suspect that the issue is not directly related to the power stage or to the LPF. It can be a layout problem (for example a too weak ground return) that injects part of the signal present on the switching node of one amp and injects it somewhere in the second amp.

ciao

 

[Zoors]

Thanks for that informative post mag.
When I think about it, ground problem is probably the best guess here. And probably easiest way to make it work.

I will let you know the progression.