Finished Truepath

[BddB]

Hi,
there are some additional info in http://www.tripath.com/downloads/an17.pdf regarding EMI/EMC-suppression in switching amplifiers.
As far as I remember there were schematics of recommended decoupling of the inputs and outputs.
Unfortunately the tripath-site doesn't work anymore, but hopefully Jan has the application note saved somewhere.
/B

[bonalux]

quote:Originally posted by BddB
[br]Hi,
there are some additional info in http://www.tripath.com/downloads/an17.pdf regarding EMI/EMC-suppression in switching amplifiers.
As far as I remember there were schematics of recommended decoupling of the inputs and outputs.
Unfortunately the tripath-site doesn't work anymore, but hopefully Jan has the application note saved somewhere.
/B

Hi Benn,
i've readed it some time ago and probably if i search well, it got to be somewere in my documents folder but, i don't like this approach. Assuming that this kind of class T amp are well engineered (as are the TP and amp1b) and considering that 4 amp1b board that i've assembled are working without (nearly) any EMI emission and with any particular care in layout and grounding, there must be something in my setup that don't work or i am missing. I have to find it and solve it, time it's not a big problem fortunately at the moment, only downside is that is very annoying assemble and disassemble the board to do the tests, and sooner or later something broke and AAAAAARRRRRGGHH!![B)]

Thank you, bye.



Gianluca

[thomaseliot]

Mine is close to a fanless music PC and I have 2 IR, one for volume control in the preamp, the other to command Foobar on the PC: no issues. Even with the portable phone.

I think you must solve first the coils problem. What if only one of the Litz wires is not well soldered at one end? A nice antenna transmitting at (pantented) variable radio frequencies. A good old 16AWG solid wire is all you need.

When coils are excluded, high frequency noise is always due to bad solderings, to not twisted nor shielded cables, big input caps with long straight legs...

Good luck.

[audiorl24]

HI all
About input caps i tried various models,good sound with WIMA MKP but a lot of noise on the ouput.
A resolution OBBLIGATO FILM CAPS with a copper case that they can connect to analog ground , nice sound ,no noise.good price.

[bonalux]

This time got no luck, rewinded the second toroid with solid copper wire but the noise problem it's the same

Problem it's not due to a pick up from the inputs, because with the inputs shorted the noise on output is the same.

Next? Even if i'm sure that SMD components are soldered well, i'll try to redo all of them as a next step, and then check with my multimeter al the SMD capacitor.

Strange thing is that if i grab one of the toroids with fingers, the noise on output drops to an accettable level, but i can't stay all the time in that posture The toroids are the source of the EMI. I've tried to put some shielding material on it but with no effect at all on noise level [?][?]

I'll keep you informed.

See you

Gianluca

[BddB]

Hi Gianluca,
I put some pics of my TP in the pic-gallery.
http://www.41hz.com/forums/photo_alb...asp?sqldtl=427
Hopefully you get some inspiration
/B

[bonalux]

quote:Originally posted by BddB
[br]Hi Gianluca,
I put some pics of my TP in the pic-gallery.
http://www.41hz.com/forums/photo_alb...asp?sqldtl=427
Hopefully you get some inspiration
/B

Thank you again for your support. I can't see nothing wrong with my layout looking at your pictures (as an add i have shielded cable for V5), only thing is that my VDD/GND lines and cable speaker of CH2 intersecates in one point, but considering that VDD and GND are not in the signal path did you think that can cause a problem to the modulator stage? Now that i think to it, having toroids really near to the metal case can be a problem? Toroids emits a strong field in the nearby... maybe another test to do.

Bye



Gianluca

[BddB]

[bonalux]

Hi my friends,
i've tried all the modify that i had listed in the previous post:

1) rewinded the toroid with the (perfectly scraped) litz wire
2) redone all the solderings on SMD components
3) changed the routing of output cables so they now passes away from VDD/VNN cables
4) placed the board and toroids away from the metal case

None of the aforementioned works as solved the emi problem The good new is that the toroids does not overheat anymore.

Since it's clear that the emi comes from the output cables, as a next step i have to try to increase the impedance of the output inductors, and for this reason i've just bought (hope they arrives soon) two MPP toroids with 60u permeability, so i can recycle the litz wire and with less windings i can increase the inductance.
On the output at the moment there are harmonics at least until 100 MHz with 3-4 mV amplitude and on the speakers the noise floor it's too high for my taste.
I hope this modify can put the word "end" to this story

EDIT: anyone as an excel file (or a fast formula) to calculate the cutoff freq of a 2nd order butterworth filter for a class D-T amp? Many thanks.

Bye

Gianluca

[krilli]

quote:Originally posted by bonalux


EDIT: anyone as an excel file (or a fast formula) to calculate the cutoff freq of a 2nd order butterworth filter for a class D-T amp? Many thanks.

Hi, try this:
http://spreadsheets.google.com/pub?k...0aAF-tih_y3bWw

This is a Google spreadsheet for the Amp6 output filter. I don't know for sure if it's the same as in the Truepath.

Edit: wait, that's not editable it looks like. Trying to figure out a way to make it public and editable.

Edit 2: Here's a public invitation to edit a copy of the spreadsheet:
http://tinyurl.com/2m6kzr

You should also be able to save it to an Excel file for convenience.