matt_garman somebody reported something like you did earlier in this tread,not sure if it was solved.
-close to defective electrolytics Jamicons had such an effect on YJblue I was told
- in greenconnector you can insert wire a lot deeper then with YJblue on pcb connectors, speaker is +-+- from left
-do both speakers lack treble ? an accident with switched outputcapacitor SMD's for inputcapacitor values will filter treble nicely I think.
no smd caps output maybe, don have pcb here😀
-close to defective electrolytics Jamicons had such an effect on YJblue I was told
- in greenconnector you can insert wire a lot deeper then with YJblue on pcb connectors, speaker is +-+- from left
-do both speakers lack treble ? an accident with switched outputcapacitor SMD's for inputcapacitor values will filter treble nicely I think.
no smd caps output maybe, don have pcb here😀
Last edited:
Even at idle it is the same warmth - mildly warm that you can feel. So is the TPA3116D2 heatsink is about the same +5 maybe +10 deg C above ambient. I checked the gain resistors they are correct at 100k and 20k. I checked the bootstrap caps = 0.22uF, LC cap is 560 nF (spec says should be 680 nF), snubber caps are 10nF and 1nF per spec. Output snubber resistor is off at 5.7ohm vs 3.3ohm.
Sounds good so far though so I won't make big deal of it.
So at idle if there is warmth that must mean there is some AC oscillations. Since this is a BTL load, zero out at speakers occurs because finite signals are 180 deg out phase but nonetheless current is flowing through inductors. This means quiescent current flow is high and will not be best candidate for battery power compared to Ybdz Wiener which is absolutely cold.
I do remember reading about that, I think it was on the Audio Circle thread. My Ybdz Wiener board doesn't have Jamicons. I can't remember the brand now (don't have the board on hand at the moment), but it wasn't Jamicon (something I'd never heard of). But, point taken, they could be (semi-)defective regardless of manufacturer. At any rate, it's a safe bet they are not premium high-end caps.
On my Ybdz board, speaker output is +--+. At least that's how it's labeled. I do remember reading in this thread that there was a board on which something was mislabeled. I can't remember which board it was, or what was not labeled correctly.
I would like to go through this thread again in it's entirety (and add more content to the wiki), but unfortunately don't see myself having the time for that.
SMSL TPA3118D2 seems like a good buy for $55?
This amp sure looks like it uses quality components. I did a search and it is only $55 shipping included from Amazon. Irrebeo, does it sound good?
The 3118 is same chip but via heatsinked so rated for 25 watts/ch. If you want a good amp out of the box, this may be it. The price includes a nice case and volume control. Oh, and let's not forget it comes with bootstrap snubber as standard. It also looks like it has Panasonic power caps. Is that OSCON 330uF stock or did you put that in Irrebeo?
Amazon.com: SMSL SA-36A Pro 20WPC TPA3118D2 Digital Amplifier AMP + 24V Power Supply Golden: Electronics
Resistor in snubber already on YJblue is 10 ohm, TI example has 3.3R, that might differ too. SMSL is only one I have seen that uses prefilter snubber:
This amp sure looks like it uses quality components. I did a search and it is only $55 shipping included from Amazon. Irrebeo, does it sound good?
The 3118 is same chip but via heatsinked so rated for 25 watts/ch. If you want a good amp out of the box, this may be it. The price includes a nice case and volume control. Oh, and let's not forget it comes with bootstrap snubber as standard. It also looks like it has Panasonic power caps. Is that OSCON 330uF stock or did you put that in Irrebeo?
Amazon.com: SMSL SA-36A Pro 20WPC TPA3118D2 Digital Amplifier AMP + 24V Power Supply Golden: Electronics
I think it is +- +- starting from power connector (which is -+) and moving to front of amp. You may just have phase backwards on one? It will not be +- -+ which means it is crossed relative to layout on chip and they will not do this as it requires multilayer and vias to do a route crossover.
Ahh, then it may be indeed be simple "operator error"! Because I definitely wired it +--+. And now that you said why the outputs must not be the way I wired it, it makes total sense (should have deduced that on my own, but half the reason I'm here is to learn).
So I read up on what out of phase speakers sound like... now I'm not sure if my original impression of the problem (and my description above) is accurate, i.e. saying the highs were filtered out. The music definitely sounded "wrong", but that's a vague term (though the effect was very obvious). It was very thin and weak. And clearly if the speakers were out-of-phase, some musical information would be cancelled out, and presumably create such a thin/weak effect. Though it wouldn't necessarily only affect the high frequencies, it would I assume mostly affect sounds that weren't panned (i.e. mixed equally to both channels, or "mono" parts of the mix).
Anyway, sorry to ramble, and thank you for the insight. I definitely wired the speakers wrong, will re-test soon. Good learning experience.
Anyone have any thoughts on the "rising resistance" value of the gain setting resistor? One suggestion I got was that there could be a cold solder joint creating a capacitor. Any other theories on that?
Thanks again guys!
Measuring resistance on board connected to internal cap inside chip will have varying voltage and as result varying R as DMM charges up the cap with its sensing current.
...and my Ybdz Wiener board is good! Thank you irribeo and xrk971 for the tip!
When I said my outputs were labeled +--+, that was a bit inaccurate. It's simply labeled "ROUT" on one side of the terminal block and "LOUT" on the other. I interpreted (wrongly!) that to be +--+, but it actually is +-+-.
And I have to say, this sounds great now! This is for my desktop rig, and I have two amps, my modded tpa3116 YJblue (sloppy soldering, with the big Bourns 2300 coils, see pics a few posts back) and a Sure tpa3110. The 3116 case is easier to work with, so I swapped the modded YJBlue board with the Ybdz Wiener. While I did that I listened to the 3110.
I noticed an immediate improvement with the Wiener. I've always struggled to pick a winner between the YJBlue 3116 and the Sure 3110. But this stock Ybdz is definitely the "wiener" of my lot. (And this is on my "backup" speakers, as I'm finishing the cabs on my new ones right now.)
At $14, definitely worth a go. Now for the missing heatsink... running without one right now, though the chip is barely warm to the touch (13.8V PSU, 8 ohm load). I might contact the seller and see if I order a second they'll throw in two heatsinks.
It looks like the consensus is that the YJBlue is a "must mod", but the Ybdz is definitely a step-up in terms of out-of-the-box quality. Not sure where the absolute ceiling is with these chips, but wherever it may be, I think the Ybdz starts out much closer to it.
When I said my outputs were labeled +--+, that was a bit inaccurate. It's simply labeled "ROUT" on one side of the terminal block and "LOUT" on the other. I interpreted (wrongly!) that to be +--+, but it actually is +-+-.
And I have to say, this sounds great now! This is for my desktop rig, and I have two amps, my modded tpa3116 YJblue (sloppy soldering, with the big Bourns 2300 coils, see pics a few posts back) and a Sure tpa3110. The 3116 case is easier to work with, so I swapped the modded YJBlue board with the Ybdz Wiener. While I did that I listened to the 3110.
I noticed an immediate improvement with the Wiener. I've always struggled to pick a winner between the YJBlue 3116 and the Sure 3110. But this stock Ybdz is definitely the "wiener" of my lot. (And this is on my "backup" speakers, as I'm finishing the cabs on my new ones right now.)
At $14, definitely worth a go. Now for the missing heatsink... running without one right now, though the chip is barely warm to the touch (13.8V PSU, 8 ohm load). I might contact the seller and see if I order a second they'll throw in two heatsinks.
It looks like the consensus is that the YJBlue is a "must mod", but the Ybdz is definitely a step-up in terms of out-of-the-box quality. Not sure where the absolute ceiling is with these chips, but wherever it may be, I think the Ybdz starts out much closer to it.
Took the plunge on the SMSL TPA3118 via amazon, thanks to your last mention of its features, xrk. 
http://amzn.com/B00JXCEDGS
This one is slightly cheaper, and Prime. 😀
Heck of a bargain.
http://amzn.com/B00JXCEDGS
This one is slightly cheaper, and Prime. 😀
Heck of a bargain.
Von Ah,
That SMSL unit does look sweet - I think it may be the only boxed 2.0 available? It is very "Topping" like but I think better. Let us know how it sounds when you get it. It is indeed a great deal as it includes the power brick too.
That SMSL unit does look sweet - I think it may be the only boxed 2.0 available? It is very "Topping" like but I think better. Let us know how it sounds when you get it. It is indeed a great deal as it includes the power brick too.
I think picture I just googled might be Rhings SMSL. Oscons are no in there standard, 47uF's are. Just had best view on pre inductor filter
I just tested the black and blue amp with a ~100dB sensitive homebrew 4-way speaker (Beyma TPL-150, Saba alnico 8" fullrange, Meyer Milo 120 12" + 600W active sub below 80Hz).
It sounded okay, somewhat harsh everywhere. Very loud! I turned the volume pot to 5-10% for normal listening levels. The 50 watts is there in full.
I then quickly changed the input DC decoupling caps to four 3.3uf Mundorf M-caps and the difference was significant. More creamy mid bass and mids, but still a sometimes harsh treble. I have been using a Muse T-amp before. It is weird, because the TPA3116D2 has the same sound signature of the T-amp, but a lot more power and detail. I used the same power supply to both of them - a crummy 12V 6A brick for LED lights.
I noticed how the Meyer milo 120 12" midbass woofer (80-700Hz) came alive with a more powerful amp. It now plays quite a lot of bass - as much as with a 800W class A/B Bittner Audio amp I used before, but with more detail.
The upper mids and treble lacks the magic of my 2A3 tube amp. I will continue modding it, but I don't expect it to rival 2A3 + 6SN7 tubes anytime soon. But good luck using a 5W tube amp for midbass! Here the TPA3116D2 will really shine.
I have ordered the big Coilcraft inductors (Bourns were better for smooth treble?) and I expect even more detail from these.
Great experiment, and I will continue using it for a while. I needed several cheap amps for midbass duty in a modest Meyer Milo 8x12" woofer array and this is perfect.
It sounded okay, somewhat harsh everywhere. Very loud! I turned the volume pot to 5-10% for normal listening levels. The 50 watts is there in full.
I then quickly changed the input DC decoupling caps to four 3.3uf Mundorf M-caps and the difference was significant. More creamy mid bass and mids, but still a sometimes harsh treble. I have been using a Muse T-amp before. It is weird, because the TPA3116D2 has the same sound signature of the T-amp, but a lot more power and detail. I used the same power supply to both of them - a crummy 12V 6A brick for LED lights.
I noticed how the Meyer milo 120 12" midbass woofer (80-700Hz) came alive with a more powerful amp. It now plays quite a lot of bass - as much as with a 800W class A/B Bittner Audio amp I used before, but with more detail.
The upper mids and treble lacks the magic of my 2A3 tube amp. I will continue modding it, but I don't expect it to rival 2A3 + 6SN7 tubes anytime soon. But good luck using a 5W tube amp for midbass! Here the TPA3116D2 will really shine.
I have ordered the big Coilcraft inductors (Bourns were better for smooth treble?) and I expect even more detail from these.
Great experiment, and I will continue using it for a while. I needed several cheap amps for midbass duty in a modest Meyer Milo 8x12" woofer array and this is perfect.
Last edited:
I do hear some more "hiss" in the background than with the Muse T-amp. Not sure if it is picking up more static with the big input dc coupling caps. I think I heard it before as well. The sensitivity of the speakers is around 100 dB so this may be a reason. People with 89dB speakers would probably not hear it. I use a SMSL DAC that is very quiet usually. No Minidsp that is adding static either, only 1st order passive XO.
Last edited:
DC power caps and inductors are more important regarding 'pops and cracks' during ON/OFF. That mod should be the first step to do!
Also, you need a toggle switch in between PSU and the power input on the board. Perhaps even better psu and up to 24V, then you will have better defined treble and stronger bass, and most importantly - no more 'thumps' in speakers! 😉
Input DC coupling caps were the only thing I could shop locally here in Oslo. Had to scavange the potentiometer from the T-amp I used before. Next time I write I will have more mods in place.
I am awaiting four Coilcraft SER2915L 10uh inductors. I read that someone liked the Bourns better for smoother treble. Can anyone confirm this?
I am awaiting four Coilcraft SER2915L 10uh inductors. I read that someone liked the Bourns better for smoother treble. Can anyone confirm this?
You need to implement the Giancarlo anti thump mod. A resistor a cap and a diode is all it takes.
Took the plunge on the SMSL TPA3118 via amazon, thanks to your last mention of its features, xrk.
Amazon.com: SMSL SA-36A PRO 2* 20W TPA3118D2DAP Amp Stereo Digital Amplifier + Power Adapter, by Gemini Doctor: Electronics
This one is slightly cheaper, and Prime. 😀
Heck of a bargain.
Hey Von
I think the pictures show a TA-2020 powered unit Not the TPA3118. Only 12 volts input. Not the same insides as the other unit,,
Amazon.com: SMSL SA-36A Pro 20WPC TPA3118D2 Digital Amplifier AMP + 24V Power Supply Golden: Electronics
But who knows,,? Maybe you will get the TPA3118 powered one.
Last edited: