New TPA3255 dual mono in PBTL

Member
Joined 2004
I personally would flip all the THT parts and connectors on the other side of the board! :)
(maybe even some caps)

That way you have a killer plate-amp that can also be mounted in just a regular amplifier chassis as well!
 
  • Like
Reactions: 1 user
I personally would flip all the THT parts and connectors on the other side of the board! :)
(maybe even some caps)

That way you have a killer plate-amp that can also be mounted in just a regular amplifier chassis as well!
I was never a fan of plate amps, or any other electronic inside a subwoofer - there is a lot of shaking so you have to flood the amp with insane amounts of glue to prevent rattle, and then you have a cooling issue. Klipsch subs are known for dying early because of that. Testing is next.
 
Uncalibrated, spdif into DAC then Scarlet 2i2. So cumulative THD for the full sound chain.
50w-4ohm.png
 
  • Like
Reactions: 1 user
Member
Joined 2004
I was never a fan of plate amps, or any other electronic inside a subwoofer - there is a lot of shaking so you have to flood the amp with insane amounts of glue to prevent rattle, and then you have a cooling issue. Klipsch subs are known for dying early because of that. Testing is next.
Just putting a wooden board between the main volume and volume for the amplifier is enough.

I have been working with those over 20 years and I definitely don't recognize your issues.

Anyway, it wouldn't really change design, yet it would be an awful lot more interesting for other people.
 
Just putting a wooden board between the main volume and volume for the amplifier is enough.

I have been working with those over 20 years and I definitely don't recognize your issues.

Anyway, it wouldn't really change design, yet it would be an awful lot more interesting for other people.
I don't see how the separate volume would isolate the board from wall vibrations. Either way, one could easily just mount all THT components on the other side using the same pcb.
 
Member
Joined 2008
have no problems with amp practicaly just behind 15" woofer in same volume, but it is "encased" in heatsink as all smd are on bottom heatsink side and there is thermal pad for full pcb size except cutout where tpa chip is located, tht parts are dense so small bit of glue holds it all together, and there is no microphonics issue with ceramics. it is bit of nervous breakdown to fit it as there is little room for error and tht parts pins need to be trimmed almost flush , but it is very solid when mounted.
on topic, with that bodge wire to reset from dvdd and without mcu for startup beware of weird problems at turn on/off, even bit of dc is not uncommon, if speakers are expensive please at least use relay on outputs.
also note this, cp from datasheet:
In BTL modes, to accommodate bootstrap charging prior to switching start, asserting the reset input low enables
weak pulldown of the half-bridge outputs.

.... A rising-edge transition on reset input allows the device to resume operation after a fault. To ensure
thermal reliability, the rising edge of reset must occur no sooner than 4 ms after the falling edge of FAULT.
 
Member
Joined 2018
I have never been able to hear differences between caps, even when I was using 1206 X7R as dc blockers.
.So I'll pass...; because it's impossible for me to approach the notion of ugrade without discriminating and qualifying capacitors.
 
Member
Joined 2018
I don't think there is any audible effect from the cap, as long as you use some film type, not something questionable.
.../...
For me there is a very audible effect even from different film caps (even between polypropylene caps).
FKP3, MKP4, PHE450, PH426, R71, etc... all sound different.
 
I have a surplus of 5 boards that I can let go for 5 eur each. But the heatsink holes are 0.8mm closer than what they should be in order to install a standard heatsink. If you are planning on using your own heatsink then the hole distance does not matter, otherwise you would need to expand m3 holes to m4 and then standard sink will fit too just fine.
 
I have a small concern with this dual-mono approach: THD signatures for two channels are not identical. I'm not sure why this is; it might be just a manufacturing deviation between IC batches, or I messed up something during assembly process. The PCB is routed manually with as much symmetry as possible between two channels.

I'm afraid that this might lead to less than ideal imaging. I haven't listened to the amp yet, but having both channels on the same IC might have it's own advantages after all...
 
Member
Joined 2015
I have a surplus of 5 boards that I can let go for 5 eur each. But the heatsink holes are 0.8mm closer than what they should be in order to install a standard heatsink. If you are planning on using your own heatsink then the hole distance does not matter, otherwise you would need to expand m3 holes to m4 and then standard sink will fit too just fine.
Board without components I assume?
 
Member
Joined 2021
For me there is a very audible effect even from different film caps (even between polypropylene caps).
FKP3, MKP4, PHE450, PH426, R71, etc... all sound different.
I believe you think you hear differences, I don’t doubt that. But can you, in a blind test say which one is which? Otherwise it’s really likely to be just mind games.
 
Member
Joined 2020
90Wrms into 4ohm, 48V PVDD. View attachment 1306562
Considering your measurements I see a slight decrease at 20khz.
looks like a non calibrated setup.
And THD figures are worse than TI specs without PFFB.
For better results I recommend the REW self-calibration procedure in a loopback configuration.
A sample frequency of 192kHz and an extended measuring bandwidth of 90khz.
And please show the result of loopback calibration.
Next step measure frequency response with and without load in the critical range of output filter resonant frequency.
Btw your PFFB does not make sense to me.
Could you elaborate on this?
 
Last edited:
Member
Joined 2020
THD-figures are dominated by the output inductors, rising with output current, i.e. lower load impedance.
The ones you choosed are the best available on the market for my knowledge.
Nothing you can improve here.
The LM5017 is an excellent choice as well, but being a 600mA regulator some kind of overkill here.
Keep in mind that for save operation the saturating current of its inductor has to be bigger than 600mA.
For this reason I choose the LM5019 150mA version with an inductor of 220~470uH.
Plus some circuitry for extra low noise output.
 
Last edited:
Considering your measurements I see a slight decrease at 20khz.
looks like a non calibrated setup.
And THD figures are worse than TI specs without PFFB.
For better results I recommend the REW self-calibration procedure in a loopback configuration.
A sample frequency of 192kHz and an extended measuring bandwidth of 90khz.
And please show the result of loopback calibration.
Next step measure frequency response with and without load in the critical range of output filter resonant frequency.
Btw your PFFB does not make sense to me.
Could you elaborate on this?
PFFB is just copied from TI's app note. Which problem do you see with it?

https://www.ti.com/lit/an/slaa788a/slaa788a.pdf
 
Top