Thank you.
I have had good results from changing input caps on my tube amp and 3 class D amps, so I am hoping I should hear some difference with this amp, too.
I just picked up some 330uF 25v Nippon Chemi-Con electrolytic (general purpose) to replace the Jamicons and some Kemet ceramics for the bootstraps. It was all the local electronics shop had in stock, but probably still far better than stock YJ components. Will see how it goes after some modding after work today. 🙂
Do 1 modification at a time and you might be able to tell if keramics are better then film for bootstrap🙂 I trust the people that said with Jamicons the 3116 didn't sound like a 3116 or 3110, not even close. On audiocircle 1 said they sucked energy out of music, they are black holes😀
A couple of pages ago someone said that the input caps should be a few millimeters away from the chip.. Has anyone tried to place the input caps closer to the chip on the YJ black board? What could cause ringing if the distance is too great? I checked TI's general suggestions and they also suggested that the input caps should be close to the chip to avoid distortion..
I don't have the same board as you but pretty much agree with your comments on soundstaging/imaging and lack of realism on HF on mine prior to adding the decoupling on pin17. Without this mod my board wouldn't be listenable longer term.
It will start to sound more like other 3116's or 3110 once Chemicons are in. After 30 minutes I guess you will know how board sounds then. That still will have flaws, bootstrap could be one, but through hole films or ceramics same distance from chip...blueboards I saw had Epcos there too, values were correct, inductance and esr will be same or maybe higher for ceramic?
you could ask TI what happens internally in this chip when avcc is lower then PVCC. they only answered somebody that asked if he could lower avcc that he could seperate avcc from pvcc, but voltage level should remain the same.
Putting in any series resistance between the two power pins will mean that the voltage level can never be truly identical. However TI do suggest 10ohms on another chip in the family so there is precedent. According to the DS, the AVCC goes to an internal voltage regulator - the existence of 3V on the input pins suggests this regulator is 6V. However since the minimum supply voltage is 4.5V, it could be the reg is set to 4V. Watching what happens to the DC voltage on the input pins as the supply is reduced to the minimum would confirm one or the other.
Isn't there already a 1nF, 100nF, and 220uF or bigger decoupling caps already on pins 17, 18, and 19? Are you saying disconnect 17 from 18 and 19 and give it it's own decoupling caps close to pin? I know in Ybdz wiener the SMT caps are as physically close as possible to the pin.
Yes I am saying that pin17 does need its own noise filter to protect it from the very noisy neighbouring pins. The PVCC pin is just to the output MOSFET in my understanding and has plenty of 400kHz switching noise on it. Why send all this hash into the analog rail? - voltage regs (speaking of the reg internal to the IC here) have very poor rejection at HF typically so the signal processing circuits get hit with a barrage of HF rubbish if there's no filter.
So do we need to break the trace connection between pin 17 and Pvcc pins? Then you are using another set of 1nF, 100nF, and 220uF cap just for Avcc?
Look at my earlier posts for the details, beginning here - http://www.diyaudio.com/forums/class-d/237086-tpa3116d2-amp-60.html#post3924399
No, I'd not recommend wasting time with the 1nF. That really can't contribute anything in such a low impedance circuit. Use a decent sized ceramic cap, at least 4u7 and add a parallel 'lytic to taste.
No, I'd not recommend wasting time with the 1nF. That really can't contribute anything in such a low impedance circuit. Use a decent sized ceramic cap, at least 4u7 and add a parallel 'lytic to taste.
Thanks Abraxalito for the post number regarding the Avcc decoupling mod. So much stuff happens on this thread I missed it. Will have to try this but I am not sure I have a problem with mine as is.
You are overgeneralizing. There have always been people who found their stock board lacking. That's why we are modding them. The component quality issue has come up quite a few times already. You simply cannot buy these boards stock and expect to be getting the best performance. Sometimes they sound okay, sometimes they don't. The YJ boards are for modding NOT for stock. You want stock go with hiampmini. There's a reason it costs more. With the cheaper there have been issues with poor quality PS caps, the inductors, wrong values for parts. On and on.
Do the mods that have been recommended. Buy the boards that have been recommended and then judge. Its that simple.
Last edited:
If you really want to get great performance then either be prepared to completely overhaul youtr YJ board or buy the hiampmini or DUGs bare pcb from the group buy.
Forget about the input caps and go straight for the PS caps and output filters in your case because it sounds like there is definitely an issue with your board/system matching. Also verify that your parts have correct values. The poor quality of the PS caps and subsequent sonic degradation has come up multiple times for cheap YJ boards and similar.
Last edited:
Well obviously I need to hear more about PIN17 decoupling! 🙂
But first an update; after work I:
- replaced the Jamicon 1000uF tanks with Nippon Chemi-Con AKG 330uF 25v (thinking of hanging a large cap off the P/S, but this amp is only fed >180Hz)
- replaced the big blue 1uF input caps with Russian Military 2.2uF 63v K73-16 PETP
- replaced the small blue "input"(?) caps with Roederstein 2.2uF 63v axial, non-polar "gold"
- replaced two of the blue bootstrap caps with .22uF Kemet monolithic ceramics
- replaced two of the blue bootstrap caps with a parallel configuration of 2x 0.1uF Kemet monolithics + 1x IC 0.01uF "green" ceramics (total ~.21uF, measured higher)
After the initial test quickly revealed a poor solder joint (to my credit) the subsequent listening has been very encouraging. 😀 😱
I was even inspired to play some vinyl - Astrud Gilberto, The Shadow of Your Smile - Verve - Mint copy
Last edited:
Obviously on chips like the tpa3110d2 TI recommends, even in datasheet, to have this added resistor/cap on AVCC. The difference is with tpa3116d2 it is not recommended by TI in datasheet and so somebody asked on their forum like I copy/pasted when you first posted about your recommendation. Even when asked the question on their online forum, they answer this chip needs same level voltage on AVCC and PVCC. So this is completely different for this 3116 chipfamily compared to chips like the tpa3110d2.
I think it was about decoupling caps. But everything on and off pcb can be an antenna for noise receiving and transmitting. The inputcaps might form a better antenna on YJblue then on other versions, I don't know, there are many loops and half loops on YJblue pcb that might act as antennas too, more then on other pcb's because the loops are larger. But I don't know if lowlevel loss is from these loops/antenna's.
I have the YJ Black/Blue board and I'm about to order parts for modifications.
I need to ask a few questions before ordering. Because shipping charge from Mouser is so high (20 EUR), I want to get it right at the first try.
For PS caps I will be using OSCON (667-25SEPF330M) as many of you have recomended.
For big input caps I will use 2,2 uF Jantzen caps that I already have. But which caps would be good replacement for those two little input caps?
Would some of these work well:
505-MKS42.2/50/5
505-MKS22.2/63/5
75-MKT1822522014
667-ECW-FD2W225J
For inductors I'm planing to use Bourns toroids, because of more easy installation compared to big coilcrafts. I have 8 Ohm speakers so, I should choose 22 uH?
542-2305-V-RC
652-2100HT-220-V-RC
The first one has bigger current rating, but will they fit on the amp board and do I have to change caps that are part of the LP Output Filter when changing coils from 10 uH to 22 uH? I have been reading what is said about the output filter in this thread, but I can't get my head around it.
For bootstraps I was going to use TDK X7R Ceramics (810-FK20X7R2E224K), but now some of you said film caps might be better.I found these:
667-ECW-FD2W224J4
or should I use some other film caps?
And finally a question about potentiometer. Panasonic EVJ-Y10F03A54 is said to be a good one, but Mouser doesn't have these. Is there a good alternative?
I know, that I had many questions, but I'm hoping to start modding soon and discovering TPA3116 chip's full potential.
PS. All part numbers are Mouser Part Numbers.
I need to ask a few questions before ordering. Because shipping charge from Mouser is so high (20 EUR), I want to get it right at the first try.
For PS caps I will be using OSCON (667-25SEPF330M) as many of you have recomended.
For big input caps I will use 2,2 uF Jantzen caps that I already have. But which caps would be good replacement for those two little input caps?
Would some of these work well:
505-MKS42.2/50/5
505-MKS22.2/63/5
75-MKT1822522014
667-ECW-FD2W225J
For inductors I'm planing to use Bourns toroids, because of more easy installation compared to big coilcrafts. I have 8 Ohm speakers so, I should choose 22 uH?
542-2305-V-RC
652-2100HT-220-V-RC
The first one has bigger current rating, but will they fit on the amp board and do I have to change caps that are part of the LP Output Filter when changing coils from 10 uH to 22 uH? I have been reading what is said about the output filter in this thread, but I can't get my head around it.
For bootstraps I was going to use TDK X7R Ceramics (810-FK20X7R2E224K), but now some of you said film caps might be better.I found these:
667-ECW-FD2W224J4
or should I use some other film caps?
And finally a question about potentiometer. Panasonic EVJ-Y10F03A54 is said to be a good one, but Mouser doesn't have these. Is there a good alternative?
I know, that I had many questions, but I'm hoping to start modding soon and discovering TPA3116 chip's full potential.
PS. All part numbers are Mouser Part Numbers.