With using smaller value, lower ESR caps, you will find a significant improvement. Initially, I had good results with improved bass and dynamic response using larger caps, but found the midrange to be grainy and strained. I switched from 1,500uF/25V Panasonic FM to 470uF/35V Panasonic FM and the results were very positive. Then I read wushuliu's post about using Panasonic hybrid Aluminum electrolytic-organic polymer caps. I read some Panasonic white papers and data sheets and decided to try the 330uF/25V Panasonic SEPF OSCON caps. In all three of my TPA31xx amps, switching to these OSCON caps has been very positive. The bass is better defined, the dynamic response is quick. Better yet, the midrange and highs are more fluid and relaxed. I mentioned before that using these low ESR (~ 14-16mohm) caps and X7R ceramic caps on the bootstrap positions per the Texas Instruments TPA data sheets (and Saturnus) are probably the best modifications you can make to these amps.
I recently switch to the OSCON caps for my new SMSL SA-36A Pro TPA3118D2 amp, and again, the amp sounds much better.
Last edited:
gain resistors for slave mode are different values, so placing the slave resistors befor the master amp will not have the result you expect, check values TI datasheet or in wiki here.
If you search this tread for "john8" you will find he discovered by using a frequency counter that there was a mistake on pcb and in datasheet, and how you could check your pcb for that mistake and how to solve this mistake. Has nothing to do with gain.
hybrfid aluminium electronic organic changed nothing to me....
I have to understand what is the source of this issue
If the hybrid 330uF sounds the same as 1500uF pan FC, then the problem should be easy to find
The gain is set by two resistors:
See table 1 in ti document.
MASTER / SLAVE GAIN R1 (to GND)(1) R2 (to GVDD)(1) INPUT IMPEDANCE
Master 20 dB 5.6 kΩ OPEN 60 kΩ
Master 26 dB 20 kΩ 100 kΩ 30 kΩ
Master 32 dB 39 kΩ 100 kΩ 15 kΩ
Master 36 dB 47 kΩ 75 kΩ 9 kΩ
The four recommended input lpf capacitors also change. See table 2.
GAIN INPUT IMPEDANCE INPUT CAPACITANCE HIGH-PASS FILTER
20 dB 60 kΩ 1.5 µF 1.8 Hz
26 dB 30 kΩ 3.3 µF 1.6 Hz
32 dB 15 kΩ 5.6 µF 2.3 Hz
36 dB 9 kΩ 10 µF 1.8 Hz
Use a multimeter to determine the correct components while the board is off.
(It might be easier to only change the 20 k resistor with a 39 k and see if it is loud enough?)
Last edited:
Well....
I found the issue....
I feel so stupid...
the quality of the record was bad. I had this bad feeling with a unmodified YJ board, but it was not so important.
Trying with good quality record was much much better.
Having a modded YJ board is not good in all situation....
Was this with the BG20 Visaton? Did you have a Zobel correction on the Visaton? Visaton has 10khz impedance ~20 ohm, 20khz impedance~30 ohm, filterfrequency impedance 50 ohm? Non of high frequency values even near 8 ohm the redboard was designed for, excessive peaking
I want to thank sharpi31 for the info posted on using isolation transformers instead of input caps. I posted the following on audiocircle but wanted to put it here too as I have not seen sharpi31 post on audiocircle:
I took the plunge and (following Sharpi31 on diyaudio) removed the input caps on my Audiobah board. I then connected (hardwired) the balanced inputs of the TPA3116 to the isolation transformer that is the output stage of my Buffalo II DAC. The DAC has firmware volume control, so as of now between the output pins of the ES9018 chip and the input pins of the TPA3116 there is only some wire and Sowter 3575 transformers wound with OCC copper (10K:10K isolating transformers).
It sounds very nice, considering that I am using a completely stock Audiobah board powered by a 12v adapter that came with a Yamaha keyboard well over a decade go. Bass is not as deep or powerful as I was getting with my TA2020 in a similar setup, but is wonderfully textured and the soundstage is huge. The sound overall is very warm, intimate and yet spacious - as you would expect. It shows enormous promise for when I sort the power supply and other details.
I measured DC offset at the speakers - getting 2.1mV and 0.2mV so that seems fine. No problems with background noise so long as the DAC is on. If I turn the DAC off but leave the amp on then there is a worrying hum coming through. It doesn't seem to cause a problem (so far!) but rattles my nerves a bit. It cannot be a ground loop because there is galvanic isolation, but I have had similar issues before using the isolating transformers in this way. Seems to come with the territory.
I took the plunge and (following Sharpi31 on diyaudio) removed the input caps on my Audiobah board. I then connected (hardwired) the balanced inputs of the TPA3116 to the isolation transformer that is the output stage of my Buffalo II DAC. The DAC has firmware volume control, so as of now between the output pins of the ES9018 chip and the input pins of the TPA3116 there is only some wire and Sowter 3575 transformers wound with OCC copper (10K:10K isolating transformers).
It sounds very nice, considering that I am using a completely stock Audiobah board powered by a 12v adapter that came with a Yamaha keyboard well over a decade go. Bass is not as deep or powerful as I was getting with my TA2020 in a similar setup, but is wonderfully textured and the soundstage is huge. The sound overall is very warm, intimate and yet spacious - as you would expect. It shows enormous promise for when I sort the power supply and other details.
I measured DC offset at the speakers - getting 2.1mV and 0.2mV so that seems fine. No problems with background noise so long as the DAC is on. If I turn the DAC off but leave the amp on then there is a worrying hum coming through. It doesn't seem to cause a problem (so far!) but rattles my nerves a bit. It cannot be a ground loop because there is galvanic isolation, but I have had similar issues before using the isolating transformers in this way. Seems to come with the territory.
Interesting mod!
Could you share more information?
Source for transformers, wiring, cost, pics, etc...?
Thanks!
Thanks Nick - There is not a lot to see as the Buffalo is enclosed in its own walnut box with wires coming out the back that I have soldered directly to the pads (nearest the chip) from which I removed C15, C16, C18 & C7. The instructions for this were given by Sharpi31:
Sowter 3575: 3575
It is awkward to get inside the Buffalo box at the moment for photos but the connections are straightforward - board to transformer to wire hanging out the back to Audiobah board as detailed above. The Buffalo has Paul Hynes regs, which really lift the sound quality.
irribeo - yes, by similar I meant with the BIASCAP mod applied, which allows removal of the input caps by connecting the input signal gnd to the BIASCAP pin on the TA2020 (if I remember correctly). Similar in that does not use input caps, but not the same. I only used one board for stereo though.
Sowter 3575: 3575
It is awkward to get inside the Buffalo box at the moment for photos but the connections are straightforward - board to transformer to wire hanging out the back to Audiobah board as detailed above. The Buffalo has Paul Hynes regs, which really lift the sound quality.
irribeo - yes, by similar I meant with the BIASCAP mod applied, which allows removal of the input caps by connecting the input signal gnd to the BIASCAP pin on the TA2020 (if I remember correctly). Similar in that does not use input caps, but not the same. I only used one board for stereo though.
Attachments
I should add that I am no electronics engineer so cannot guarantee that this is a safe setup. I have only applied mods that were suggested by others to my own goal of the purest signal path I can get. Based on what I have heard so far, I suspect that with a better power supply for the TPA3116 and possibly changing onboard inductors/electrolytics this solution may be a champ.
Thanks for sharing dboy. Eventually, I want to implement the full use of the differential (i.e., balanced) inputs of the TPA31xx amps to achieve the same goal as you. I looked into input balancing transformers from Lundahl and Jensen. I had contacted Kevin at K&K Audio, the U.S. reseller of Lundahl transformers. This is what he had to say:
That said, assuming that the DC voltage at each input terminal is identical for a channel (+ and -), then no input caps are required. I just don't know enough about the workings of the TPA31xx amps on whether the DC voltages at each terminal are indeed equal.
You mentioned that the bass is lacking, and I wonder if this might be the cause, or if it is simply an impedance matching/cut-off frequency issue.
That said, assuming that the DC voltage at each input terminal is identical for a channel (+ and -), then no input caps are required. I just don't know enough about the workings of the TPA31xx amps on whether the DC voltages at each terminal are indeed equal.
Hi Rhing - I would not say the bass is lacking, just slightly less of it than with the optimised setup I had with TA2020. This is not surprising given the less than ideal power supply on the other end of a long thin cable and only 2 x 470uF on the board itself. At this point I am only checking that the basic connection concept works - which it certainly seems to. I have thought about changing gain to improve impedance match with transformers, but I think one thing at a time.
I know a lot of DIY folks poo-poo the Bybee Music Rails, but as I wrote on audiocircle I had great joy using one to supply power to TA2020 amp. So I'm gonna try the same again with TPA3116 as soon as my delivery from Partsconnexion makes it over to me in the UK. Those Astron supplies don't seem to crop up on EBay in UK very much or I'd perhaps experiment with one of those. I think the key to good power for these amps is probably low output impedance of the supply, so suspect results will be similar. Interested in your thoughts though... or anyone else's.
I know a lot of DIY folks poo-poo the Bybee Music Rails, but as I wrote on audiocircle I had great joy using one to supply power to TA2020 amp. So I'm gonna try the same again with TPA3116 as soon as my delivery from Partsconnexion makes it over to me in the UK. Those Astron supplies don't seem to crop up on EBay in UK very much or I'd perhaps experiment with one of those. I think the key to good power for these amps is probably low output impedance of the supply, so suspect results will be similar. Interested in your thoughts though... or anyone else's.
Sounds quite nice though.
I received the blueboard 2.0 yesterday and it has the same distortion in the mid treble as the 2.0 redboard. Today I switched the bootstrap caps on a redboard with 250V X7R ceramics and treble(coarseness) distortion dissappeared
So I just wan't to say to all new on this thread these must be swapped, (even on the blueboard). It has been the best improvement so far
So I just wan't to say to all new on this thread these must be swapped, (even on the blueboard). It has been the best improvement so far
I received the blueboard 2.0 yesterday and it has the same distortion in the mid treble as the 2.0 redboard. Today I switched the bootstrap caps on a redboard with 250V X7R ceramics and treble(coarseness) distortion dissappeared
So I just wan't to say to all new on this thread these must be swapped, (even on the blueboard). It has been the best improvement so far
You mean the ceramic bootstraps used on YJ redboard distort in same way the TH filmcaps do on blue/black??? The ceramics are underspecced on redboard if 10times voltage rating is needed to avoid distortion ,films should not distort, but who knows what they are
Did you measure the removed ceramics btw? Just for capacitance?
I received the blueboard 2.0 yesterday and it has the same distortion in the mid treble as the 2.0 redboard. Today I switched the bootstrap caps on a redboard with 250V X7R ceramics and treble(coarseness) distortion dissappeared
So I just wan't to say to all new on this thread these must be swapped, (even on the blueboard). It has been the best improvement so far
Can you be more specific, what brand of caps; TDK, Vishay, or other? There are many X7R's.
Also, perhaps an image of your work, this one doesn't show how he soldered those on black/blue board, since TDK's are a bit different. Thanks