TPA 3116 is very sensitive to low value bypass caps placed as close to the chip as possible.
Did you put decoupling as close to chippins as possible ? Looks like you added some also too far away from chip, like Original, the long legs inductance and esr make functioning even harder, effective distance is close to "can't do anything" distance already, on YJred, but YJblue too. I read a digital designer saying at 1 inch distance they can not perform any part of task that was intended for them.
That's the point that I am not exactly clear. It is also my plan to use a transformer at the input to adapt single-ended signals to the differential input for all the advantages that you described. The transformer will therefore acts as a balun. Typically, to do that , the single end signal will be connected the one of the primary lead and the other primary lead connects to the circuit ground. In this scenario, the 3 V DC bias at the input should still present even though the cap is removed? However, it seemed that you tried it this way and it worked?
Regards,
Hi irribeo,
I am aware of the importance of distance
I thought of placing 331 cap on the other bypass sms cap nearer to the chip but there is no space with a probably of touching the heatsink.
I was toying with the idea of removing the heatsink that is hardly warm despite after using for an hr or so.
Despite this , this benefit the high frequency.
Maybe I get some more smd cap and scrap away a segment of the ground plane and soldered directly to the pvcc pin .
102 or 101 caps will be ok ?
what is your experience by power supply bypass?
regards kp93300
I am aware of the importance of distance
I thought of placing 331 cap on the other bypass sms cap nearer to the chip but there is no space with a probably of touching the heatsink.
I was toying with the idea of removing the heatsink that is hardly warm despite after using for an hr or so.
Despite this , this benefit the high frequency.
Maybe I get some more smd cap and scrap away a segment of the ground plane and soldered directly to the pvcc pin .
102 or 101 caps will be ok ?
what is your experience by power supply bypass?
regards kp93300
Yes. Sharp31 and dboy first tried input transformers with their TPA3116 Amos, and found that the balanced differential inputs did not have issues with the 3VDC bias. I've since tried it as well and have no problems with DC transformer core saturation either. The sonic benefits are well worth the effort, and I've confirmes this in two other systems besides mine. The big question is if there are any lower priced input transformers that will produce similar results to the CineMags, Jensens, Sowters and Lundahls.
104 is basicly what is needed I think, xr7 will drop value depending on voltage, cog will remain at value chosen. Yellow line in earlier pic I posted is actual path to left side 104 classII decoupling ceramic on YJred, how much will that 100nF do there, path/distance to your Nichicons is actually smaller. End of Arrow is where I think you could put them, but you can get even closer. Sanding a little aluminium off heatink past part that contacts tpachip is very easy.
On dug and some audiobahs I put 10uF classII ceramic next to 100nF, that doesn't sound bad LOL still searching for Rubycon MU to improve boards further:
http://www.rubycon.co.jp/kr/catalog/e_pdfs/pmlcap/e_MU.pdf
On dug and some audiobahs I put 10uF classII ceramic next to 100nF, that doesn't sound bad LOL still searching for Rubycon MU to improve boards further:
http://www.rubycon.co.jp/kr/catalog/e_pdfs/pmlcap/e_MU.pdf
Attachments
Thank for for the information. I will definitely give it a try. In terms of set up, the input caps are removed and the secondaries of the transformers are hooked up to pin 4,5 (right input)and 11, 12 (left input) of the 3116? In addition, I am thinking of increasing the gain of the amp to 32 db with the use of the transformer.
I plan to get a bunch of TPA3116 boards (YJ blue, YBDZ, including mono board that become available recently) during X'mas time. These boards are inexpensive enough to be "disposable" (almost). To stay with the "cheap" theme, I would like to use the Edcor.
Regards,
You think the tiny amount of DC that'll get to them will saturate them?
It could make them sing, and potentially lose thermal efficiency, while still working.
This came from Kevin of K&K Audio, the U.S. seller of Lundahl transformers. I'll take his word for it.
I would not apply DC to either the primary or secondary of a small signal transformer (unless it was designed to cope with DC). I've not measured a significant difference in the bias voltage of the + & - inputs on a TPA3116, so don't have any (significant) DC on the transformer secondary to cause problems. If I was connecting an unbalanced source to the transformer primary, and there was a potential difference between signal and ground, I'd use a coupling cap (or other solution) to prevent DC through the transformer primary.
I'm liking the blue YJ 3116 board and I have the Bootstrap, bootstrap snubber and Panasonic oscons replacement parts on order.
My problem is the lack of bass. I have a raspberry Pi and DAC in the ceiling feeding the amp and then onto some ceiling speakers (please no comments about the ceiling speakers - this is what I'm working with)
The speakers are decent (ish) but I would love to mellow the highs and slightly boost the bass and bottom mids. All I have in the celining is a laptop brick supply but I've been considering a Chinese lm1036 board to provide some tone controls....
I would rather not go down this route and wandered if there are any mods that will slightly increase the bass and bottom mids on the YJJ blue board...
I'm using Volumio otherwise I would provide a small amount of eq in software but I can't,
could the experts guide me here, given that I can't change the speakers...
stu
My problem is the lack of bass. I have a raspberry Pi and DAC in the ceiling feeding the amp and then onto some ceiling speakers (please no comments about the ceiling speakers - this is what I'm working with)
The speakers are decent (ish) but I would love to mellow the highs and slightly boost the bass and bottom mids. All I have in the celining is a laptop brick supply but I've been considering a Chinese lm1036 board to provide some tone controls....
I would rather not go down this route and wandered if there are any mods that will slightly increase the bass and bottom mids on the YJJ blue board...
I'm using Volumio otherwise I would provide a small amount of eq in software but I can't,
could the experts guide me here, given that I can't change the speakers...
stu
Well, I guess it depends on the transformer used. If the DC resistance is low enough, there will be relatively large amount of current flow through the coils and generate relatively large amount induced current/magnetic field.
Have to try it out to see. It might or might not matter.
Regards,
That was my view all along. Just curious, what kind of bias voltage did you observe when you did the measurement?
Regard,
Better to saturate a transformer briefly as you find out your source puts out DC, than to feed DC straight into you amp...
In other words they can act like a safety device but you have to correct any DC problems from whatever source exhibits it.
In other words they can act like a safety device but you have to correct any DC problems from whatever source exhibits it.
Btw my 3116 sounds weak at lower volumes, as well as other inadequacies when I have a volume stepped attenuator with Dale's in it. It's a 10k unit.
The highest input impedance is with 20db gain. It's 60kohm. If it were 100kohm then a 10kohm pot would be ok. As is a buffer or transformer is a better option if sized correctly, than just volume device.
The amp is much better than can be shown with a simple pot. A Bel Canto DAC3 with built in volume would do well.... 🙂
The highest input impedance is with 20db gain. It's 60kohm. If it were 100kohm then a 10kohm pot would be ok. As is a buffer or transformer is a better option if sized correctly, than just volume device.
The amp is much better than can be shown with a simple pot. A Bel Canto DAC3 with built in volume would do well.... 🙂
If I get active or passive audio splitter and connect two Sure TPA amps in parallel, is it feasible to get 100 watt per channel ?
Higher wattage only arises in paralleling if you're running into a lower impedance load (like 2R). Otherwise there's no advantage except that you'll be effectively doubling the supply capacitance in runing two amps (which isn't to be sniffed at).
Edcor
Tonight I bypassed the input caps on the SURE TPA3116 and reconnected the Edcor 10k/10k input transformer. I couldn't really tell if there was an improvement to the overall performance, if connected directly or bypassing the input caps. The improvement could have been there, but it would have taken some alone time at home. So I connected the Edcor, and my brother in law was a 2nd set of ears, who heard this amp for the 1st time. And replayed an HD audio track with several others. With the Edcor and without the Edcor. And we mutually agreed, there was about a 25% improvement over all performance. We tested this in his garage, so im sure if I got alone and had things set up properly at home. Im sure I would bump it up to another 5%, for a total of 30% improvement, using the Edcor. Vocals are more tonally balanced but clearer. Improvement in sound stage, better detail at the highs, bump in the darker lows. IMO for only $33 dollars, its nice value added upgrade, and it bridges the gap to the alternative, the amazing Cinemags, which will run $180 bucks shipped. Or the even more illustrious, world famous, top of the pop Lundahls which will run $325 or so. If anyone else decides to get the Edcor, please post your review here to share.
Tonight I bypassed the input caps on the SURE TPA3116 and reconnected the Edcor 10k/10k input transformer. I couldn't really tell if there was an improvement to the overall performance, if connected directly or bypassing the input caps. The improvement could have been there, but it would have taken some alone time at home. So I connected the Edcor, and my brother in law was a 2nd set of ears, who heard this amp for the 1st time. And replayed an HD audio track with several others. With the Edcor and without the Edcor. And we mutually agreed, there was about a 25% improvement over all performance. We tested this in his garage, so im sure if I got alone and had things set up properly at home. Im sure I would bump it up to another 5%, for a total of 30% improvement, using the Edcor. Vocals are more tonally balanced but clearer. Improvement in sound stage, better detail at the highs, bump in the darker lows. IMO for only $33 dollars, its nice value added upgrade, and it bridges the gap to the alternative, the amazing Cinemags, which will run $180 bucks shipped. Or the even more illustrious, world famous, top of the pop Lundahls which will run $325 or so. If anyone else decides to get the Edcor, please post your review here to share.
Last edited: