I received this in the mail today, TPA3116 50W x 2 Dual Channel Class D Digital Audio Amplifier Board 100W 8V 25V | eBay.
It was a bit better than my other YJ boards but it had the same hint of ringing\harsness as the other boards did. I replaced the input caps with my trusted 1uF smds and put them right up to the chip legs and the ringing went away!(I had to file a bit of the cooling fin). Even with wrong 22uH inductors on my 5-6ohms speakers it doesnt sound harsh.. I notice that many people switches inductors here on this tread.. I've done that too but it hasn't helped nearly as much doing the input cap mod! I've changed the output coils to suit my speakers but it doesnt effect how warm they get or the ringing. If anyone one of an amp were the input caps are smds and uber close to the chip please let me know!
It was a bit better than my other YJ boards but it had the same hint of ringing\harsness as the other boards did. I replaced the input caps with my trusted 1uF smds and put them right up to the chip legs and the ringing went away!(I had to file a bit of the cooling fin). Even with wrong 22uH inductors on my 5-6ohms speakers it doesnt sound harsh.. I notice that many people switches inductors here on this tread.. I've done that too but it hasn't helped nearly as much doing the input cap mod! I've changed the output coils to suit my speakers but it doesnt effect how warm they get or the ringing. If anyone one of an amp were the input caps are smds and uber close to the chip please let me know!
I use the emty holes from the two large blue film capacitors. The hole closest to the power caps on each side. Ground is original input pin, but I'm sure there is a ground closer to the chip..
No I removed all the 4 caps that were in series with the input. 2 on + side and 2 on - side. I have written more indepth about it a couple of pages ago🙂
I basically rewired the input side with smd 1uF caps and placed them within 2-4mm from the tpa3116 chip. I scraped off the paint above the traces and ofcourse made 4 small cuts and then placed the miniature smd cap on top! Its a pita to make the solderings! From there it traces goes normaly to the X2's input caps first leg(counting from the chip side).
I basically rewired the input side with smd 1uF caps and placed them within 2-4mm from the tpa3116 chip. I scraped off the paint above the traces and ofcourse made 4 small cuts and then placed the miniature smd cap on top! Its a pita to make the solderings! From there it traces goes normaly to the X2's input caps first leg(counting from the chip side).
Last edited:
beauty of solderings isn't more important then sound here 🙂
Sorry I missed earlier post. I was hoping the funny antennashaped pcbtrace was out now, but that was not the difference. You left signalshield at old inputconnector. Did you connect SMD for negative to ground somewhere close to chip, or just trace back to old capacitor?
(I replaced all inputcaps and main filtercap and sound only got more brillant, so I put back Original caps. Replacing bootstrap added the missing detail but also added brillance again. Somehow YJblue always seems to add brillance and it is brillant already, so interested when something seems to tame that a little).
Sorry I missed earlier post. I was hoping the funny antennashaped pcbtrace was out now, but that was not the difference. You left signalshield at old inputconnector. Did you connect SMD for negative to ground somewhere close to chip, or just trace back to old capacitor?
(I replaced all inputcaps and main filtercap and sound only got more brillant, so I put back Original caps. Replacing bootstrap added the missing detail but also added brillance again. Somehow YJblue always seems to add brillance and it is brillant already, so interested when something seems to tame that a little).
Having reread the sloa119a pdf I agree that for BD modulation as with the 3116 that the approach they state is to effectively 1/2 the speaker impedance when performing calculations, the values derived are unchanged.
I plugged some values into the calculator "Calculate the transfer function for RLC low-pass filter with R, L and C values" at RLC Low-Pass Filter Design Tool . You can state the target frequency and aim for optimum Q or damping factor, I supplied an inductor value as a start point. An example below:
R = 5.6Ω
C = 0.33uF
L = 22uH
Cut-off frequency
fc = 59067.9394874[Hz]
Quality factor
Q = 0.685857127979
Damping ratio
ζ = 0.7290148044
R is half the actual value so this equates to a speaker impedance of 11.2ohm which is good for the MA 10.3 I am modelling for.
I would appreciate it if you could verify this is a valid approach.
Also is 60khz to high a cutoff frequency?
Last edited:
I dont think its noise in the traditional sense that causes the ringing. The signal wires are pretty long underneath (and far apart) and still no noise in this setup. Could it be something to do with the way the chip is constructed/+3V applied to the inputs?? I have tried other input caps on the yj red board, but the ringing were there always. At first I thought is was the power caps, then inductors, then bootstrap but no🙁 Now I firmly belive its the input caps that are too far from the chip. I'v modded 3 different boards now with this mod, and the ringing goes completely away with my latest discovery. Now there is zero harsness which I didn't think possible with my speakers/class D.
I noticed that the larger smd 250V bootstaps gives less distortion when the amp is pushed. But overall I think its the smd input caps and the closeness to the chip that is the solution to making these amp sing! If anyone has an unmodded board please try this mod to confirm my suspicion.
I noticed that the larger smd 250V bootstaps gives less distortion when the amp is pushed. But overall I think its the smd input caps and the closeness to the chip that is the solution to making these amp sing! If anyone has an unmodded board please try this mod to confirm my suspicion.
For MA alpair10 the peak created at 60khz isn't likely to matter. Certainly not when noisefloor is only thing amplified 20dB or so. If 60khz is (un)filtered switchingfrequency smps PSU you use maybe amplifier can put out a real loud 60khz??? and that being constant take away power from amplifying musicsignals???could be wine and beer and to little knowledge talking🙂
Filtervalues you mentioned look fine. Bottom graph okawa shows impuls behavior, if you enter lowest impedance alpair10 there is hardly dampening, lowering impuls of those frequencies. Limited peaking as in SLOA119 till a little above 20ohm.
Dumb question here but I see a lot of posts about ringing. Can you tell me what it is and what the sonic effects are?
Thanks for that. Using a linear power supply anyway (not a fan of SMPS in critical applications).
Regarding bootstrap caps, I was wondering if anyone has try instead of the recommended X7R, the ECHU film caps line from Panasonic. They seem to be superior than ceramic X7R.
ECHU Series Product Descriptions | Surface Mounted Film Capacitor | Passive and Electromechnical Components | Panasonic Industrial Solutions Global
ECHU Series Product Descriptions | Surface Mounted Film Capacitor | Passive and Electromechnical Components | Panasonic Industrial Solutions Global
Noise with a 400khz source nearby isn't like noise in a classAB amp. NXP bias is a little lower, but that might be more sensative, don't know.
Looked at YJblue here, how did you connect the two negative inputsmd's to ground? I still would expect smd's to sound brighter then the X2 used, like the Wima's.
Stability after bootstrap replacement might be better too, but cd I mentioned really had no lowlevel on YJblue with epcos bootstraps, it was worse then gettoblaster wih cdplayer, that did play some background lowlevel on that classical cd. I guess the epcos have higher inductance like some filmcaps do, dissipation of most films is way lower then ceramics, so esr is lower, only inductance added by filmcap, not leadlenght inductance, TH ceramics have same leadlenght, but the capacitive part inside could make a difference, if the epcos/siemens are genuine. The part costs close to nothing, really strange.
Those ECHU caps looks awesome! How expensive are they?
ChrissMmm, the ringing is there even with a toroid and huge caps. The best I can describe it is an over emphasis on sss sounds. For example: Delain (Breathe on me) every time she sings breathe the sound does not sound natural, the s sound in breathe distorts.
From What Is Sibilance? - Definition in Stereos :
Definition: Sibilance is a sound characterized by pronouncing words with the letter 's'. In audio reproduction the letter 's' should sound clear and distinct, not smeared or distorted as in 'sh'. The letter 's' should sound like a hissing sound.
Irribeo, I shorted out the leads between the holes for the old little grounding caps underneath the board.
ChrissMmm, the ringing is there even with a toroid and huge caps. The best I can describe it is an over emphasis on sss sounds. For example: Delain (Breathe on me) every time she sings breathe the sound does not sound natural, the s sound in breathe distorts.
From What Is Sibilance? - Definition in Stereos :
Definition: Sibilance is a sound characterized by pronouncing words with the letter 's'. In audio reproduction the letter 's' should sound clear and distinct, not smeared or distorted as in 'sh'. The letter 's' should sound like a hissing sound.
Irribeo, I shorted out the leads between the holes for the old little grounding caps underneath the board.
Last edited:
When I first got the YJ blue/black I compared it with my Ybdz with listening and measurements using a mic in the same setup and it did feel a little brighter. Not much but you can see here in the response at listening position. This is stock on both amps.
If the brightness bothers you, a mild -1.5dB EQ (Q=0.5) high shelf from 6kHz should do the trick. Or tlt more off axis.
Last edited:
I think this is the feature separating that board from all the rest:
"5. Open writers impact protection design, improve the switch machine blown, demanding please note."
(Where's the "WTF" emoticon?)
Hi,
I have changed the power caps from 1000uF to 470uF then 330uF then 220uF and 20uF still rings, the boot strap caps were changed to 220nF and input caps were changed to MKP10 2.2uF, still rings and the high still harsh, will try your method later this week on the boot strap cap and input caps, need to get some smd caps. Again what value of SMD cap you replaced the power cap with ?
When I first got the YJ blue/black I compared it with my Ybdz with listening and measurements using a mic in the same setup and it did feel a little brighter. Not much but you can see here in the response at listening position. This is stock on both amps.
If the brightness bothers you, a mild -1.5dB EQ (Q=0.5) high shelf from 6kHz should do the trick. Or tlt more off axis.
All 1uF's on YJblue and the 560nF were replaced with Wima 1uF same as on my Hiamp boards, difference in brillance was gigagigantic, not able to listen to for long bright, so YJblue got all original parts back then. Jamicons were already replaced, gain set to 20dB, nothing else was done on YJblue. It could be the warm inductors are actually 5uH when amp needs to deliver 5watt into my speakers, that would make difference smaller on speakers needing 0.5 watt to deliver same SPL. (#watts just as example)
Keep following this thread, and more I read about this - so called - 'ringing' more is confusing, and my brain starts 'ringing' as well 😀
I have YJ blacky side by side to 2x6N3 preamplifier board in the same box. YJ is modified, as I did post earlier - new Coilcraft's inductors (22uH) and 2 Elko's DC caps (1000uf 35V) and that's about it. No hissing, no noise, and NO 'ringing' - no matter what source is IN; turntable, CD or PC via DAC, and as I'm just listening my favorite 'hissing' record 'Tango in the night' by FM have to admit more transparency, highs are perhaps more emphasized than through the tube amplifier, but that's what I was expecting from a D-class amp and a little chip. I remember those 'ssshhheeshh' in music before (voices or harsh cymbals,..) even via tube amps, but once have replaced polyesters with polystyrene's or polypropylene's and better electrolytic caps the story has changed
However, I do not say there is no room for more improvement and better (or more expensive) components to try with this lilliputian but I'm done with this one. Still have to do similar mod to the blue ChengZhe and finally try Audiobah's board that I hope does not need any modification at all?!
I have YJ blacky side by side to 2x6N3 preamplifier board in the same box. YJ is modified, as I did post earlier - new Coilcraft's inductors (22uH) and 2 Elko's DC caps (1000uf 35V) and that's about it. No hissing, no noise, and NO 'ringing' - no matter what source is IN; turntable, CD or PC via DAC, and as I'm just listening my favorite 'hissing' record 'Tango in the night' by FM have to admit more transparency, highs are perhaps more emphasized than through the tube amplifier, but that's what I was expecting from a D-class amp and a little chip. I remember those 'ssshhheeshh' in music before (voices or harsh cymbals,..) even via tube amps, but once have replaced polyesters with polystyrene's or polypropylene's and better electrolytic caps the story has changed
However, I do not say there is no room for more improvement and better (or more expensive) components to try with this lilliputian but I'm done with this one. Still have to do similar mod to the blue ChengZhe and finally try Audiobah's board that I hope does not need any modification at all?!