TPA3250 somebody is listening?

Jump to Latest
#782
chermann said:
from post 775

so the tpa 3250 has a gain of 26db so with this opamps config i get about 33db so its 7db more gain


is that correct?

so the volume knob- potentiometer is a voltage divider for the input and the inverted opamp has a gain about 7db....correct?


which values should i find an the pcb to get this? can some fresh brain and expert help me?


thanks chris
Click to expand...
On - forum.vegalab.ru they are engaged in the Hypex UcD amplifier and write:
«Completely excluded OU as well as WG. The signal is fed directly to the modulator. Gain is enough for comfortable listening directly from the source. The input impedance is not enough, but the output buffer of the source pumps. This option is more like.» They have an input impedance of the amplifier (modulator) of 10 kom, the TPA3250 has a resistance of 24 k per channel. Or 12kom with two channels connected in parallel.
 
#784
forestsgump said:
thks fr the suggest on buck, will chk them out. I was thinking of low noise LDO LT1963 to rep. LM317.

Chris btw, you done tons of work on Fx502spro ~ cld be most well-versed abt it among many of us. Did you spot any fake component used?

~ i wld hope replace any of these components, bef they cause trouble & kill the amp!
Click to expand...


Hi Forestsgump


Sorry but friday checked the caps near the op amps. this little 100µF electrolytic caps are bad..after suing many hours the are burned in - so no excuse....and have an ESR of 0,2R🙄🙁
i changed it to nichicon fine gold 1000µF/25V..this are the only one which i had. i my setup its working because my lm317 is set to 20V..
if you change these caps and keep the lm317 origin you have to be in minimum 35V


therefore the sound check give me better timing and darkness.
as always 4 xchecks ...this time with master album by melody gardot + stefano bollani...😉



chris
 

Attachments

  • fx502spro coil compare4.JPG
    fx502spro coil compare4.JPG
    87.9 KB · Views: 569
Last edited:
#788
Hi noasds


as i wrote ....i do not agree at this measurements because its origin amp...they miss some modification😉.....now its much better..


i did some measurements...maybe on the very beginning of my journey...please read it...
now i have the latest frequency response: post 735 https://www.diyaudio.com/forums/class-d/315681-tpa3250-listening-74.html#post5648131


at this graph :from 10khz to 20kz minus 1,5 dB is not super dramatic..
the dis balance of left and right...hmm.. normal you have to adjust with your complete system the dis balance between left and right...so no problem

XLO Reference Recordings Test & Burn-In HDCD-Elusive Disc
track 1 +track 2 +track 3 is a good start😉

chris
 
#789
This frequency response drop at 20kHz is typical when measuring class-d-amp with LC output filter loaded by some low inductance dummy load. Loading output with some inductive load like a real speaker is a different story and may turn into a slight peak at the upper audio frequency corner.
Having said this frequency response measurements with dummy loads should be taken with a grain of salt😉
 
#790
Thanks chermann. Your mod is great.
Of course, I do not mean an dis balance, but a fall after 10K on the chart.
I noticed improvements in sound quality after replace default (probably fake) opamps with LME49720NA. And sound really very good. But after reading that review I think may be to improve anything else 🙂.
 
Last edited:
#791
voltwide said:
This frequency response drop at 20kHz is typical when measuring class-d-amp with LC output filter loaded by some low inductance dummy load. Loading output with some inductive load like a real speaker is a different story and may turn into a slight peak at the upper audio frequency corner.
Having said this frequency response measurements with dummy loads should be taken with a grain of salt😉
Click to expand...


thank you voltwide...this is correct🙂


noasds:
thank you for the roses....that´s fine..😀


with the TI Filter designer you can see the simulated frequency response of the output filter you changed at the amp.
http://www.ti.com/tool/LCFILTER-CALC-TOOL


chris
 
#792
voltwide Ok, I realized that with 8 Ohm speakers the frequency response should be flat enough. Need to tell to amirm to make changes to his measurements of class D amplifiers. There is rather an excess of high frequencies and he probably makes the wrong conclusions.
 
#793
chermann said:
Hi noasds


as i wrote ....i do not agree at this measurements because its origin amp...they miss some modification😉.....now its much better..


i did some measurements...maybe on the very beginning of my journey...please read it...
now i have the latest frequency response: post 735 https://www.diyaudio.com/forums/class-d/315681-tpa3250-listening-74.html#post5648131


at this graph :from 10khz to 20kz minus 1,5 dB is not super dramatic..
the dis balance of left and right...hmm.. normal you have to adjust with your complete system the dis balance between left and right...so no problem

XLO Reference Recordings Test & Burn-In HDCD-Elusive Disc
track 1 +track 2 +track 3 is a good start😉

chris
Click to expand...




sorry at the post 735 at V5 board its not 1µF output filter cap...its 0,68µF
 
#794
Knowing about this issue of pure pre-filter feedback class-d amps you can measure frequency-response both with low inductive dummy load and open circuit. This imho gives the worst case range of possible frequency responses under all different load conditions.
I did this with my personal TPA3118 ref design using STEPS & USB-soundcard producing this result with 10R load / open circuit, respectively. Sadly my NOS-ears do not honor this frequency response anymore.
 

Attachments

  • fq_response_tpa3118_2x100pf_1k6_10r_no_load.png
    fq_response_tpa3118_2x100pf_1k6_10r_no_load.png
    41.8 KB · Views: 479
Last edited:
#796
Without load is to see how the amp performs undamped. Some implementations tend to be unstable without load, pushing the magic smoke out of your output filter (caps and or snubber resistors) due to i.e. excessive peaking. An open-load safe amplifier is some kind "safety feature" in case of broken/unconnected wires/voicecoils.
 
#797
doctormord said:
Without load is to see how the amp performs undamped. Some implementations tend to be unstable without load, pushing the magic smoke out of your output filter (caps and or snubber resistors) due to i.e. excessive peaking. An open-load safe amplifier is some kind "safety feature" in case of broken/unconnected wires/voicecoils.
Click to expand...


AAAAh...it ´s good to have here 2 experts....:yes:
Is there a rule of dump how long you can handle without load and be quick in measureing before the amp is smoking 😀

thx chris
 
#800
There is a rule of thumb. Excessive peaking requires hi-level input with high frequency content. So, do monitor filtered power output. Choose input voltage for a non clipping output level, i.e. well within power rail boundaries. That's it.
 
You must log in or register to reply here.
Top Bottom