My 1st Class-D Amp -TPA3255 with PFFB

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
J

JohnnyInBris

Member
Joined 2016
I hadn't looked at Class-D for a couple of decades and I was surprised at the performance levels now being achieved. This prompted me to have a go at building my own. After much reading I decided on the TPA3255 with post filter feedback in stereo BTL configuration.

The power supply is a 400W 48V LED supply out of china turned down to 45V. I chose this one because it has a thermostatically controlled fan which doesn't kick in at normal listening levels. The minimal heatsink gets warm to the touch at ear shattering levels from my Klipsch floor standers.

I'm still looking for a suitable enclosure but in the meantime I've breadboarded it on some acrylic.

The inputs are balanced but my DAC isn't so I'm using balanced to unbalanced cables.

The chokes I wound myself using 15 turns of 2 x 1mm stands on sendust cores. with AL=51nH/T. These barely get warm.

I'm still waiting on some 48V to 15V DC-DC modules so in the mean time I've fitted a 7815 with a couple of 100R/1W droppers.

How does it sound? Well I don't have golden ears but it's on par with a tube hybrid that I have except that the bass seems tighter and it has a very neutral and uncoloured sound. Despite the single power supply, imaging is also very good. And it is so quiet.

Gerbers for the PCB are available if anybody wants them. How did I load this board? rather crudely! I apply solder to all the SMT pads using my iron and then dip each component in rework flux before reflowing them into place using a hot air tool.
 

Attachments

  • TPA3255_PFFB.png
    TPA3255_PFFB.png
    135.3 KB · Views: 3,027
  • IMG_20170225_175934.jpg
    IMG_20170225_175934.jpg
    881.6 KB · Views: 3,005
I

impuls60

Member
Joined 2011
Wow! I want to try out this chip so bad!! A couple of years ago I started on a huge amp build for 2x tpa5630 cards only to find the sound quality lacking in those. So I never bothered to finish it. So now I am waiting for the 3255 to go mainstream, so I can finish my dream amp. I don't understand why this chip hasn't generated more of a buzz on this forum! Everyone that has built one says it sounds really good, and getting a usable 48V psu doesn't cost that much either. I've built and bought several tpa3116 amps but this is the next big thing!! Let'ss jump on the 3255 wagon instead :)
 
J

JohnnyInBris

Member
Joined 2016
Yes, I am planning to do some THD measurements. My oscilloscope doesn't have the resolution needed but I will soon be getting a USB audio interface with the following spec for line input:

Dynamic Range 106 dB (A-weighted)
Frequency Response 20 Hz to 20 kHz, ±0.1 dB
THD+N <0.003% (minimum gain, -1dBFS input with 22 Hz/22 kHz bandpass filter)
Maximum Input Level+22 dBu
Gain Range 50 dB
 
M

mantabernd

Member
Joined 2012
Nice work!
I also collected some experiences with TPA3255 in PFF configuration.

I posted my info here: http://www.diyaudio.com/forums/clas...th-x-amplifier-made-mobile-application-6.html

Now I wanted to ask you... Do you have problems with your PFF snubber (1uF + 3.0 Ohms) when there is no load connected?

In my design this snubber dissipates a lot of power and sometimes even dies if amp
clips without load. I need to replace the R then... But I'm actually using 1.33uF + 3.3 Ohms so corner frequency is a bit lower in my design.

Did you try this?

Thanks and regards
Bernhard
 
doctormord

doctormord

Member
Joined 2007
360customs.de
mantabernd said:
Nice work!
In my design this snubber dissipates a lot of power and sometimes even dies if amp
clips without load. I need to replace the R then... But I'm actually using 1.33uF + 3.3 Ohms so corner frequency is a bit lower in my design.
Click to expand...

What do you expect, do a simulation and see for yourself. It's totally normal without any bw-limit on the inputs. With values from application note, the snubbers resonance frequency is at about 47kHz.

Been there, done that:

http://www.diyaudio.com/forums/clas...ng-them-everything-comes-182.html#post4987732

30A current in the snubber. ;)
 
Last edited:
M

mantabernd

Member
Joined 2012
JohnnyInBris said:
With no load and with audio signal levels just below clipping there is no heating of the snubber resistors. I would suggest that there may be a HF instability in your design or that you are testing with unrealistic signal levels in the 10s of KHz region
Click to expand...

Thanks for checking this in your design.

It only happens when the amp clips. Shortly before clipping it's all ok.
And no, my test signal was 1kHz sinus.

But actually I could solve this problem already by compensating the post filter feedback loop. It's stable now even with no load.

My phase shift is now between 15° and 45°.
 
E

earsofgold

Member
Joined 2018
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom