DIYHiFi.org

For the sake of audio
It is currently Tue Nov 21, 2017 1:31 pm

All times are UTC




Post new topic Reply to topic  [ 8 posts ] 
Author Message
PostPosted: Sat Sep 01, 2007 7:40 am 
Offline
Hen
User avatar

Joined: Sat May 19, 2007 7:52 pm
Posts: 5
Location: Copenhagen
In all of my designs so far I have had some problems with a high pitched sound when two channels are mounted together. I think this is normally called the beating frequency, and is caused by the difference in swhitching frequency, and coupling between the two amplifier modules.

Does anyone have any good way around this?

One way is to place the switching frequency far enough apart e.g. 25 kHz difference, so that the resulting difference freq is outside the audio band.
... But in an self oscillating desing, the switching freq wil of course change with the music, and the beating freq will therefore not be eliminated, but wil maybe just be another form of distortion!

Another way is to syncronize the freq, but for a self oscillating design, thets not realy a good option, I think!

Where can the difference tone enter?
- coupling between the two output filters .... better shielding?
- through the supply lines ...... better filtering, better PSRR?
- Through ground ?
- .... other ways??

Is this a problem with commersial products like UcD ... and if not, how do they get around it?

Best regards Baldin :wave:

_________________
Visit SensibleAudio, click on the website below.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 01, 2007 10:36 am 
Offline
Account disabled on owner's request

Joined: Wed Dec 29, 2004 8:31 pm
Posts: 899
Look up heterodyning.

Look up EMI.

Look up PCB layout techniques for high power/frequency / low noise.

That's how Hypex does it (about the only one), not letting EMI be a problem to begin with, consider anything else a bandaid used to mask a serious problem that's better eliminated at the source.

You don't stand a chance of making a good power switching product without a firm grip on that, and I think that's where most fall short of being impressive (newclassd, all the zap pulses in relation to the former owner, nuforce, etc, etc, etc).

As a slightly more direct response:

"- coupling between the two output filters .... better shielding?
- through the supply lines ...... better filtering, better PSRR?
- Through ground ?
- .... other ways??"
yes, yes, yes, and yes.


"Is this a problem with commersial products" YES "like UcD" NO

"how do they get around it?"
By a in depth understanding of the nature of the beast... see the above.

Cheers


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 01, 2007 11:58 am 
Offline
Cow

Joined: Wed Jan 05, 2005 7:18 am
Posts: 118
The problem is coupling of the switching transient of one channel to the pwm comparator of another channel. You have to minimize this coupling, either by cleaner switching of output transistors or by any other means already mentioned. Just think how switching transient propagates to the other channel.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 08, 2007 9:20 am 
Offline
Hen
User avatar

Joined: Sat May 19, 2007 7:52 pm
Posts: 5
Location: Copenhagen
Ok. This is what I'm thinking:

1. Control the switching spikes/ringing on the rails by using snubbers directly over the mosfets. This should reduce the noise induced on the rails.

2. RC filter for all low voltage supply lines (something like 100 ohm/100uF) before the regulators. Have tryed to simulate some T-filters (L-C-L) instead, but it dosen't seem to be very effective, and you have to know the impedance quite closely, not to make matters worse!

3. Chokes in the the supply lines from PSU to amp board. Do I have to use same core material as for the output coil?? .... could I use a normal 1mH noise supression coil?? ..... I'm a bit afraid of putting a coil into the rail, as my intuition says it might actually make things worse ... any experience in this area??

4. Separate ground planes for input/PWM and output circuit, connected as close to the main PSU GND as possible.


You do not have the required permissions to view the files attached to this post.

_________________
Visit SensibleAudio, click on the website below.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 08, 2007 3:19 pm 
Offline
Cow

Joined: Sun Jun 11, 2006 3:36 am
Posts: 112
...hm, if it is the beating, then it could theoretically be also the unavoidable beating of the two acoustic signals in the air.. Do you have the high pitched sound also if you run both amps, but connect only one speaker?
If no, then I would guess you will only be able to tackle it by reducing the ripple at the output, means going to a higher order filter...

If yes, then I would also guess that you are in trouble with switching noise that feeds into the control.
In this case I would expect that you can find with your scope one or multiple signals, which show undesired noise. And if this signal(s) is(are) looking systematically better when you run only one amp, then you probably found the relevant signal(s).
Next step would be systematically searching the dominant noise path(s)...

Normal noise suppression coil: What is a normal noise suppression coil? You will find as much answers as much people as you ask.
Your proposal of a 1mH choke is pointing to a choke, which is probably not suitfull to filter switching noise. Switching noise has usually the harmful signal content in the harmonics above 10MHz. So you need a choke which has a high impedance there, means a choke with low parasitic capcitances. A 1mH choke usually has quite some parasitic capacitances and a resonance frequency below 1MHz. In the range of 10MHz and above it will act mostly like capacitor. Furtheron in this frequency range the layout is really some sort of higher art, because the geometries of you copper tracks and GND areas do form tons of small capacitors, transformers and antennas. Up to now I did not dig into depth for this (just some very basic surface know how), but I should learn more about that before I make the layout for my Gen2. So if you find good links, please let me know. I remember one short and simplyfied information paper from Bruno... -will tell you when I find it again.

_________________
Best overall performance is derived by the right trade off.


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 08, 2007 3:34 pm 
Offline
Cow

Joined: Sun Jun 11, 2006 3:36 am
Posts: 112
http://www.hypex.nl/applications/emi.htm

_________________
Best overall performance is derived by the right trade off.


Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 09, 2007 1:08 am 
Offline
Cow

Joined: Sun Jun 11, 2006 3:36 am
Posts: 112
...when you are going to examine small signals about noise you cannot connect the GND of your scope somewhere somehow to your circuit. You must reference GND as close as possible to the concerned signal.
I.e. like in the attached picture. Even then there might remain some noise error in the measurement, but usually not troublesome. (If you want to know how much, then simply check with the probe signal needle how much noise will remain when you touch the needle
to the GND connection point.)


You do not have the required permissions to view the files attached to this post.

_________________
Best overall performance is derived by the right trade off.


Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 09, 2007 5:47 pm 
Offline
Hen
User avatar

Joined: Sat May 19, 2007 7:52 pm
Posts: 5
Location: Copenhagen
Hey that's a neat little trick ... I'll try it :grin:

_________________
Visit SensibleAudio, click on the website below.


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 8 posts ] 

All times are UTC


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group