They are same size as the 6.8uF inputcapacitors on the Wiener, all are 6.8uF on mine. Actually it is a tiny bit shorter then 6.8uF inputcaps, just looked, it measures 6.4uF, input to ground beside it measures 7.5uF, input to ground furthest right measures 6.4uF too, input near connecto measure 7.5uF and 7.4uF
Last edited:
Yes, ideally I would have made one change at a time (as suggested earlier in the thread) but unfortunately I do not have the patience for that. I was in "git 'er dun" mode.
An externally hosted image should be here but it was not working when we last tested it.
As a friend of mine used to say...
"Anything that can't be done right away is probably not worth doing." 🙂
So are you pretty happy with the sound that this amp is providing now. We can all see the importance of good quality capacitors on amps. How does it compare to your premium tube system?
Thats the case - MULTICOMP - G750
The input caps are Wima 2.2u 63V. Power supply outputs 19.5-20 volts. Filter is set by 6 x 3300uF 35V caps and one 330nF. The transformer is 100W (14V 7.14A).
The yellow caps at the output are the original.
The 3116 sounds very, very good now. I am quite happy with it. In fact I was up till nearly 1:30am listening to it.
I did not do any back-and-forth comparison with my tube amp or my STA508 last night, but from memory I can say that the 3116 is definitely in the same league at this point, with respect to clarity, presence, imaging, etc. I will be doing much more listening and comparisons over the next few days and I don't want to jump to conclusions, but last night's listening session had me thinking I like the 3116 better than the 508. Something about the overall coherence of the 3116...
listening now to YJblue, first impression: lowlevel indeed has appeared, bass seems louder too, but that is from memory, maybe I only focussed on lowlevel with the Silmics last time I listened to this YJblue. need more time to hear if lowlevel now is equal to hiampmini for example, but it is a relief to hear what is in that Adagio recorded by Teldec😀 (Now ofcourse it is a distraction from music again😀) And would the solder on pcbtraces matter, it looks kind of retro, scraping away all black on traces would be fun, nahh?
Attachments
Yay! Cool. 🙂
I feel vindicated because I felt partially responsible for steering you down this path just to be disappointed. Really glad that your ears confirm my feelings about this amp all along.
Nah, there was no need to feel that way. In fact, you steered me into buying this specific version due to my intent to modify, and you were right, so thanks for that.
I didn't really expect it to sound fantastic in stock form. I just didn't expect it to sound quite as bad as it did, based on reading much of this thread.
I really like the way it sounds now, but the tweaker in me is already thinking about those coils...
Did you experience any ring like Implus60 did with the Slimics ? I just picked up 4 Slimics II 470uF 50V, they were out of 330uF and not sure if I should use them or order the OSCON.
I have four of these Bourns capable of 5 amps that I want to put in:
http://www.digikey.com/product-detail/en/SRR1208-100ML/SRR1208-100MLCT-ND/2353016
http://www.bourns.com/data/global/pdfs/SRR1208.pdf
Some folks like the Coilcraft but they are huge and need creative mounting methods. The coils may be the next high value mod.
I also hear that using batteries is also very good as it kills all that SMPS noise (unless you are already using a fat linear supply with toroid and cap bank).
Have just put away YJ black and testing ChengZhi blue board and BINGO - no more cracks on powering on/off! Sound is excellent via 1x12AU7 tube Dynavox tpr-2 preamp, and certainly no more issues on higher volume here.
Yet, I'm noticing chip heatsink and inductors getting hot. PSU is 70W 24V 3A.
With 12V 1A YJ black had also warm to hot all of these?!
Yet, I'm noticing chip heatsink and inductors getting hot. PSU is 70W 24V 3A.
With 12V 1A YJ black had also warm to hot all of these?!
I've considered ordering some of the Bourns, along with some other caps and stuff I want, but the $20 shipping fee at Mouser always stops me from plugging in my CC number. My cart right now adds up to $53 with the shipping fee. For a $20 amp? Seems silly. And the amp sounds fantastic now, so...
Re: batteries - I've never tried running any of these amps off of batteries, but one of my future projects is to build an audiophile-grade sound system into my travel trailer (it has a real POS in it from the factory) so eventually I will get to enjoy one of these class D amps running on batteries. 🙂
Boostrap Capacitors and Snubber Circuit?
The topic of what bootstrap caps work or don't work on the TPA3116D2 seems to be the hot topic. I am not sure if this has been previously asked but I just realized that when I built my TPA3118D2 "Dead Bug Point-to-Point" amp, I followed the EVM circuit schematic. If you look at the EVM datasheet (http://www.ti.com/lit/ug/slou336b/slou336b.pdf), Fig. 2 shows a 330pF 50V COG ceramic cap connected in series with a 10ohm resistor to ground tied to the 220nF bootstrap cap:
Here is a closeup of the EVM snubber shown above:
Is this to "snub" the ringing that so many folks hear and complain about? When I listen to my dead bug amp, it sounds cleaner - less sibilance and I wonder if this is the case? Should we all be seriously trying this mod next? I think it might be a simple case of butt-soldered 330pF-10ohm connected to the 220pF bootstrap cap for SMT mods on Ybdz for example, and in the case of the YJ Blue/Black a simple 330pF axial and 10ohm 1/4-watt resistor can be added easily.
For comparison, the basic circuit that all of the Chinese boards out there use (except the one EVM copy) seems to be missing and may be a big part of the sound quality. The basic datasheet is here: http://www.ti.com/lit/ds/symlink/tpa3116d2.pdf Here is a closeup of the standard filter and bootstrap circuit:
What do you all think? Probably worth trying... certainly it is very inexpensive to add.
The topic of what bootstrap caps work or don't work on the TPA3116D2 seems to be the hot topic. I am not sure if this has been previously asked but I just realized that when I built my TPA3118D2 "Dead Bug Point-to-Point" amp, I followed the EVM circuit schematic. If you look at the EVM datasheet (http://www.ti.com/lit/ug/slou336b/slou336b.pdf), Fig. 2 shows a 330pF 50V COG ceramic cap connected in series with a 10ohm resistor to ground tied to the 220nF bootstrap cap:
Here is a closeup of the EVM snubber shown above:
Is this to "snub" the ringing that so many folks hear and complain about? When I listen to my dead bug amp, it sounds cleaner - less sibilance and I wonder if this is the case? Should we all be seriously trying this mod next? I think it might be a simple case of butt-soldered 330pF-10ohm connected to the 220pF bootstrap cap for SMT mods on Ybdz for example, and in the case of the YJ Blue/Black a simple 330pF axial and 10ohm 1/4-watt resistor can be added easily.
For comparison, the basic circuit that all of the Chinese boards out there use (except the one EVM copy) seems to be missing and may be a big part of the sound quality. The basic datasheet is here: http://www.ti.com/lit/ds/symlink/tpa3116d2.pdf Here is a closeup of the standard filter and bootstrap circuit:
What do you all think? Probably worth trying... certainly it is very inexpensive to add.
Attachments
Last edited:
RFS1412-223KE Coilcraft are not huge, taller yes but fit in the place of those originals. That's my plan with black board, as got some free samples from generous folks at Coilcraft!
A question regarding DC power caps - Datasheet suggested value: 220uF, so could Elna Silmic 220uf/25V do the job well?
You might want to check out all the inductors that DUG has spec'd for the group buy - there is a long list and Wurth or Coilcraft seems to be his favorites.
http://www.diyaudio.com/forums/group-buys/257996-gb-tpa3116-8-pbtl-bare-pcb.html
470uF is not supposed to be HF decoupling, that far from chip🙂 removing them doesn't only impact ringing I fear, but removes a nearby powersupply for chip. YJblue here now has 470uF silmicII and 0.1uF closer to chippins. That doesn't remove brightness/clarity but does remove harshness add-on (to good recordings) that was there sometimes, and not only that, being a very close powersupply it does more. Like the bootstrap does more then just allowing some specific lowlevel information in a recording to be heard. It is just a coincidence we found a recording that had backgroundnoises completely filtered by Original bootstraps on YJblue, and then you listen a dozen times concentrating if it really is gone to conclude it is. Then inputcaps were replaced and same track was listened too, no change in lowlevel, then Original back on pcb and yellow filtercap replaced, no influence on lowlevel, original back on pcb, and noticing original yellows had a somewhat warmer glow then Wima's tried there, the Wima's sounded more plastic but that is huge exaggerated ofcourse. Panasonic Hybrids that replaced very bad Jamicons were pulled of pcb again and replaced by Elna SilmicII, lowlevel didn't change. And so on with YJblue 😀
Now it seems for first time YJblue is a good amp, electrolytics/0.1uF decoupling/bootstraps changed, nothing else yet. Worries or wonders are about why inductors warm up without music and warm up quite a lot when power is left on for 24 hours, when music is playing they stay cooler, I felt it too this evening (europe). With the same PSU's, same pot's, same wallsocket the other boards can be left powered unused for days without feeling anything alarming. But you can remove power and amp sounds good now. Didn't have time yet to compare to other amps.
Thanks for the link.
I did some digging through my electronics bits and bobs while home for lunch and to my surprise I found 6 pieces of 10uH ferrite-core toroidal inductors. I had ordered them a long time ago for another project (TA2020) and then never ended up using them. They are of questionable origin and quality, so I do not have high expectations. They are the same as these 6pcs Toroid Inductor 10UH 6A Toroidal Ferrite Core Choke Coil TC22 T5E | eBay
I'll install them this evening and post my thoughts later.
The low price of these boards tends to encourage the use of low price (often poor quality) power supplies, which I believe can be more detrimental to sound quality than many other elements.
I'd encourage anyone new to these amplifiers to invest in a decent PSU before doing any other mods. So far I've had good results with a medical grade open frame 100W 20-24V SMPS (switching >100KHz) as well as a 40VA EI core transformer with heavy duty rectifier and 3x 3300uF Rubycon ZLH caps. I've had noticeably worse results using one of the cheap 19V SMPS laptop chargers bought for the Amp Camp Amp (that amp can have a CRC supply so audible PSU noise is less of an issue, although better PSUs were an improvement there too).
After that I'd consider input coupling (single ended cap coupling noticeably inferior to a balanced input transformer), output filters (low ESL caps, coils that won't saturate and values optimised for the loudspeakers in use) and local decoupling (polymer/electrolytics on amp PCB plus low ESL bypasses as close to chip as possible).
This is my experience of course. The point I wanted to make was the order to address these issues (how each one is handled is a separate issue) - no point trying 10 different amp board electrolytic decoupling caps while you're using a £5 laptop charger as PSU.
I'd encourage anyone new to these amplifiers to invest in a decent PSU before doing any other mods. So far I've had good results with a medical grade open frame 100W 20-24V SMPS (switching >100KHz) as well as a 40VA EI core transformer with heavy duty rectifier and 3x 3300uF Rubycon ZLH caps. I've had noticeably worse results using one of the cheap 19V SMPS laptop chargers bought for the Amp Camp Amp (that amp can have a CRC supply so audible PSU noise is less of an issue, although better PSUs were an improvement there too).
After that I'd consider input coupling (single ended cap coupling noticeably inferior to a balanced input transformer), output filters (low ESL caps, coils that won't saturate and values optimised for the loudspeakers in use) and local decoupling (polymer/electrolytics on amp PCB plus low ESL bypasses as close to chip as possible).
This is my experience of course. The point I wanted to make was the order to address these issues (how each one is handled is a separate issue) - no point trying 10 different amp board electrolytic decoupling caps while you're using a £5 laptop charger as PSU.
Last edited: