I have designed a class d amplifier with 2 amplifiers on the same pcb.
I have kept them separate except for +/-45 volts and ground.
I have star grounded the two amplifier inputs so hopefully I don't get crosstalk.
I have a phase splitter on one amp so I can bridge the amplifiers or use stereo which is decided by a link on the pcb.
Is there anything wrong with this approach ?
The amps are irs2092 based.
I have kept them separate except for +/-45 volts and ground.
I have star grounded the two amplifier inputs so hopefully I don't get crosstalk.
I have a phase splitter on one amp so I can bridge the amplifiers or use stereo which is decided by a link on the pcb.
Is there anything wrong with this approach ?
The amps are irs2092 based.
Thanks John for the AN numbers . Read them again and again , and one thing not mentioned here , what is the advantage of synchronizing the clocks ! I read hours and hours on this , and one says the IM distortion goes up , and the other says it's only to prevent HF carrier garbage on your speaker wires .
I have a IRS2092 5 channel amp ( as you probably all know by now on this forum
) and I'm VERY happy with the performance in 5 channel mode , but in Stereo , there's something wrong . The female voices are "walking over the podium" in one breath .The rest of my stuff is pretty good , and I have other amps ( class A and AB) that don't give the placement error .So it must be some sync error ?In direct mode of my Emotiva preamp I use 2 channels of the D amps , one is clocking at 412 kHz and the other at 425 kHz with no signal . I switched channels that have ( slightly) different carrier frequencies , and it seems it sounds different ( no AB test possible here ) .
I'm pretty sure I can get the clocks tuned to a couple of Hz's or synchronize them , but I like to understand what I'm doing , so please explain !
Cheers ,
Rens
412Khz and 425Khz is a pretty bad combo. That means an IM of a very audible 13Khz. Ideally they should either be perfectly synchronized, preferably 180 degrees out of phase with eachother (or 120, 90, 60 and so on depending on the number of sync'ed modules). Or they should have clock frequencies that are so far apart that the IM is well outside the audible frequency band. But not so far that the IMPs moves into audible band either.
Last edited:
The L + R channels are phase inverted . to prevent buspumping , but what is "Audible" I don't hear any distortion or any noise , not even at very low levels , and should that noise be around 13 kHz ? The only complaint I have is an unstable placement in the center of the sound field , mostly with female voices . It still sounds VERY good , no distortion . If I play DTS music , Eric Clapton Unplugged , or The Eagles or whatever , the sound stage is unbelievable , the problem only exists in stereo mode .
Cheers ,
Rens
It's not noise or distortion that would be audible it's the roughly 1ms continuous shift in timing between the 2 channels that would be audible. In effect, one channel constantly moves from sampling the input and producing the output slightly before and after the other channel. The closer those frequencies are to eachother, they more noticeable it is until they are perfectly sync'ed. In DTS mode I will assume it's masked because there's so much more deliberate phase shifting going on.
Last edited:
Your answer makes more sense then the hundreds of PDF pages I read so far .
Thanks ! I 'll start with some multi turn's to get the clocks as close as possible ( or more then 20kHz apart). Or sync them if that doesn't work . The DTS mode is using the center channel , so that's probably masking the shifting problem .
I'll take the system apart after the Christmas carols or have war here at home !
Let you know how it works !
Cheers ,
Rens
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.