W hat you see on the scope is a residue of the switching frequency of the amp and is ok as far as I can see .
But you do not seem to understand the mutual dependance of amp and connected load .
Class D amps have a low-pass filter at the output to filter their output which consists of a train of high frequency impulses and convert this into a low frequency signal representing the music . This has nothing to do with any filtering you apply from a DSP coming before the amp .
Class AB amps usually have one or two filters at their output , one to present a minimum load to the amp at high frequencies and another to protect the amp from a load which is too capacitive , both to keep the amp stable under all conditions , and which are not present as such at the output of a class D amp as this works in a different way .
There is one easy to do thing you can try : make a coil of enameled copper wire of 1mm , diameter 20 mm 11 turns tightly wound and connect it between the output of the amp and your speakers . It is a Thiele-cell used in classAB amps to protect the amp from a load with too much capacitance and it might make a difference with your tweeters.
But you do not seem to understand the mutual dependance of amp and connected load .
Class D amps have a low-pass filter at the output to filter their output which consists of a train of high frequency impulses and convert this into a low frequency signal representing the music . This has nothing to do with any filtering you apply from a DSP coming before the amp .
Class AB amps usually have one or two filters at their output , one to present a minimum load to the amp at high frequencies and another to protect the amp from a load which is too capacitive , both to keep the amp stable under all conditions , and which are not present as such at the output of a class D amp as this works in a different way .
There is one easy to do thing you can try : make a coil of enameled copper wire of 1mm , diameter 20 mm 11 turns tightly wound and connect it between the output of the amp and your speakers . It is a Thiele-cell used in classAB amps to protect the amp from a load with too much capacitance and it might make a difference with your tweeters.
Last edited:
Thanks again. I have to decide if adding another passive parts does not lead me to passive crossover anyway.
3 way, 3 modules. I bought resistors and tried the L-pad for the tweeter. It is silent now. I will probably use -10dB (difference in sensitivity + house curve, crossing AMT2 at 5kHz), but there is no need to attenuate so much to beat the noise.
Looks like some sort of interaction amp/tweeter to me . Class D amps have a fixed gain due to the supply-voltage and a feedback which is mostly taken before the low-pass filter they have on their output so the reason for your problem might be found in this .
Their response is related to said output-filter which should be designed for the load they drive otherwise you get variations in response usually at high frequencies
Their response is related to said output-filter which should be designed for the load they drive otherwise you get variations in response usually at high frequencies
I have active three way speakers , filters are Linkwitz-Riley for 24dB slope and flat response with no phase shift at cross-over and one ShuiYuan module for bass/midrange whereas the tweeters are fed by two LM3886 in BTL , giving me the high efficiency of the class D where power is needed and where power requirements are pretty low I use the latter .
I found this to give me the best results. All opamps in filters and ShuiYuan are LM4562. Amps in speakerboxes and the electronics backplate , 30cms x 20cms , is enough heat-sink for the two LM3886
I found this to give me the best results. All opamps in filters and ShuiYuan are LM4562. Amps in speakerboxes and the electronics backplate , 30cms x 20cms , is enough heat-sink for the two LM3886
Last edited:
I'm sorry it's taken me so long to reply to this, but it sounds very elegant! Do you have a picture by any chance? 🙂
Not right now and to make pictures I will have to build two more as I have sold them and they are mounted inside speaker-boxes as well . Give me a week or two and I will see what I can do
I cannot give it a name . It is a PCB from a thread in these forums , optimized through a long process to which a lot of people have contributed in its day . Unfortunately I cannot remember which thread it was but if you give me some time I will try to locate it . Funny thing is I do remember it was post 335 in that thread .
If you send me a pm with your email I can send you the gerberfiles for the pcb
I use two in a BTL-circuit to match the max voltage excursion since the two channels of the ShuiYuan are BTL so the impulse power of the tweeters is not too limited as above all none of the amps must ever go into clipping. Any amp driven to clipping you do not need a high quality amp to begin with🙂
If you send me a pm with your email I can send you the gerberfiles for the pcb
I use two in a BTL-circuit to match the max voltage excursion since the two channels of the ShuiYuan are BTL so the impulse power of the tweeters is not too limited as above all none of the amps must ever go into clipping. Any amp driven to clipping you do not need a high quality amp to begin with🙂
Last edited:
Found the thread and the post : the thread is "An open source layout for LM3886" and the post is 235 . Hope it helps Mike
Hi folks,
I bought (and received) a Shui Yuan TPA3255 module appx 1 months ago.. However I was working on my Peerless LCPF 600 module then could'nt find an opportunity to try it until last week..
So I am listening my "dirty"😀 Shui Yuan for one week at least.
Firstly I noticed that the most of components (caps included) were not fake on the board. Not fake but low quality components were on board.
Then I made some mods;
- Replaced two main reservoir caps with 1800uF/50v Nichicon HE series,
- Replaced the "220uF before XL7015 regulator" with a 470uF Kemet alu polymer solid cap.
- Replaced the "220uF after the choke of XL7015" with a 330uF Panasonic FC,
- Replaced the "half Vcc filter cap (22uF 0805)" with a 180uF Kemet alu polymer solid cap,
- Replaced the audio line 10uF Nichicon FW caps with Elna Silmic II caps,
- Finally replaced the crap NE5532s with OPA2134s...
Am I happy with it?
Absolutely YES,
Is it noisy?
Absolutely YES, but if you are close the tweeters (and mids) more than 10 cms.. Which is enough for music listening I guess.
Conclusion;
For a 30$ amplifier board: A very satisfactory listening impression.
PS: I am using a MW LRS 350-36 PSU adjusted to 34v to not kill my 35v alu polymer caps...
I bought (and received) a Shui Yuan TPA3255 module appx 1 months ago.. However I was working on my Peerless LCPF 600 module then could'nt find an opportunity to try it until last week..
So I am listening my "dirty"😀 Shui Yuan for one week at least.
Firstly I noticed that the most of components (caps included) were not fake on the board. Not fake but low quality components were on board.
Then I made some mods;
- Replaced two main reservoir caps with 1800uF/50v Nichicon HE series,
- Replaced the "220uF before XL7015 regulator" with a 470uF Kemet alu polymer solid cap.
- Replaced the "220uF after the choke of XL7015" with a 330uF Panasonic FC,
- Replaced the "half Vcc filter cap (22uF 0805)" with a 180uF Kemet alu polymer solid cap,
- Replaced the audio line 10uF Nichicon FW caps with Elna Silmic II caps,
- Finally replaced the crap NE5532s with OPA2134s...
Am I happy with it?
Absolutely YES,
Is it noisy?
Absolutely YES, but if you are close the tweeters (and mids) more than 10 cms.. Which is enough for music listening I guess.
Conclusion;
For a 30$ amplifier board: A very satisfactory listening impression.
PS: I am using a MW LRS 350-36 PSU adjusted to 34v to not kill my 35v alu polymer caps...
Last edited:
Dxvideo
Thank you very much for reporting the board modification. I am also not satisfied with the sound quality of this board and will do an am upgrade.
Thank you very much for reporting the board modification. I am also not satisfied with the sound quality of this board and will do an am upgrade.
Thanks UAN,
I must warn you that Shui Yuan uses a very bad quality solder on this two layer board. So its very hard to remove THT components without harming the board!
Be careful!
I must warn you that Shui Yuan uses a very bad quality solder on this two layer board. So its very hard to remove THT components without harming the board!
Be careful!
For 30USD it is very good amp. It sounds good, but connected load-speakers may influence its frequency response. Just a few days ago we compared this amp powered by Meanwell RSP-320-48 (I forgot it is set to 37V) with very expensive class-A Accuphase and Yamaha B-3. Speakers were Infinity Kappa 9.2. Amps were switched immediately, so no guessing, no psychoacoustics. We were listening loud music, so I cannot say anything about distortion on low power, which should be expected according to data sheet. Shui Yuan had absolutely enough power, produced nice bass (even slightly more than competitors with these particular boxes), midrange was attenuated and little bit also high frequencies . If you do not have immediate comparison, you probably do not know. But do not expect it to be a well behaved neutral reference, it is not.
We also compared Yamaha with stock Shui Yuan with NE5532 and then with LM4562NA. Sorry guys, no audible difference.
We also compared Yamaha with stock Shui Yuan with NE5532 and then with LM4562NA. Sorry guys, no audible difference.
I have my own assessment of the quality of digital amplifiers. If there is no quick fatigue when listening to songs, then everything is in order.
Now only amplifiers on the MA12070 chip please my ears.
Yes, the NCore clone is also very good!
Now only amplifiers on the MA12070 chip please my ears.
Yes, the NCore clone is also very good!
Last edited:
I got my basic shui yuan board. i replaced the main capacitor with nichicon kw series. I removed the op amps at the input. i connected the +/- pin of the xlr cable to pin 1 of pin 7 of the opa socket. I replaced the capacitors at the tpa input with elna slimic II. Since the inpedancia of the tpa chip is 20k, so that the previous stage is not loaded, I loaded my tube impedance adapter with an output of 3-400R. Listening this way sounds very good. Now I’m waiting for two weeks to burn in, and then I’ll listen to different music with my friends.