I am glad to hear there is good nonpolar!
I see some UUP at MOUSER. I assume you were speaking of SMD type?
Looking at the schematic with more understanding it looks pretty simple to bypass those ceramic caps simply by removing them and making your lead attachment to the correct side. All of the jumpers stay the same.
Easy enough to get those things off but I hope I do not have to put something back in. Certainly those little NICHICON cans would be much easier to replace than ceramics! If they are not too big.
I see some UUP at MOUSER. I assume you were speaking of SMD type?
Looking at the schematic with more understanding it looks pretty simple to bypass those ceramic caps simply by removing them and making your lead attachment to the correct side. All of the jumpers stay the same.
Easy enough to get those things off but I hope I do not have to put something back in. Certainly those little NICHICON cans would be much easier to replace than ceramics! If they are not too big.
If you can get a SMD radial that fits on to the 1206 pad of the x7r that's great, otherwise remove the x7r and short the pads. Some finagling of a non-SMD cap between your actual input and the TPA3255 will get you going. Take your pick. 🙂
I have the TPA3255EVM board and have been following this thread with interest. I’m pretty set on removing C17, C28, C55 & C63 and connecting two Lundahl LL1521 input transformers (like LL1540 but with a small step up). All opamps out the circuit, and no coupling caps. I did similar mods with a couple of TPA3116 boards and preferred the change. Hoping to find an hour or two over the next week or so to get it done 🙂
Yes, I want to do a similar thing, with my DAC's transformer output stage (Lundahl LL1684).
But now I'm concerned about the unused NE5532 op amps - apparently they may draw excessive current and/or generate noise -
http://www.ti.com/lit/an/sboa204/sboa204.pdf
https://www.maximintegrated.com/en/app-notes/index.mvp/id/1957
The termination methods recommended at those links might be awkward to implement on the EVM, so I wonder if the most straightforward solution might be to disable the 12V supply to the op amps - by removing inductor L6, or by scraping away a section of trace to the input of L6?
Can anyone comment please?
Just ground both inputs on them! If you use the inputs at the long header and the right jumpers, you bypass it all.
There is nothing wrong with bipolar electrolytics. It's pure nonsense. Wasn't BG NX HI-Q the pinnacle of electrolytic coupling cap according to the internet 10 years ago?
Sorry, should have mentioned that I have the 3250 EVM, not the 3255. The component numbering is mostly identical, but the 3250 EVM does not have the 3255's J28 header.
This is very interesting !
Right now I was also wondering if audio transformer is not the right solution after all for those like me having only single ended sources. But, is the same good sound quality achievable with a transformer then with a big quality film cap ? Having read about transformer distortion, but knowing nothing in fact, I am curious about some feedback. Going here straight to the chip pins with Jantzen and wima caps, the sound is so good and so nice lows going so deep and detailed. I don't want to lose this even if I recover again a full bridge amplification.
Last edited:
Very interesting indeed - and yes, on TPA3116 this was a big step forward - personally I did not hear any negative side effect of the transformer - even in comparison to already improved coupling caps. So please let us know how things develop.
Using the big header at the front of the board would make the connection easy if you were not worried about the cap that much of this discussion has been about. It is still there. No question the easiest way to bypass the op amps.
More I think about it the best way to get rid of the effects of the 5532s is to remove them.
It is not as if they require much power but it does seem ridiculous to leave unused components on the board.
Could there be some advantage of leaving the decoupling capacitance intact? Could this be of any benefit to the other devices using the 12 volts rail? As opposed to removing the choke.
I would think a high quality input transformer would have to sound better than a 5532. Transformers are just fine. Certainly the LUNDAHLs! I would tend to ask Dave Slagle to make me a pair if I was to do this. The LUNDAHLs have the advantage of shielding for those who do not want to deal with that.
More I think about it the best way to get rid of the effects of the 5532s is to remove them.
It is not as if they require much power but it does seem ridiculous to leave unused components on the board.
Could there be some advantage of leaving the decoupling capacitance intact? Could this be of any benefit to the other devices using the 12 volts rail? As opposed to removing the choke.
I would think a high quality input transformer would have to sound better than a 5532. Transformers are just fine. Certainly the LUNDAHLs! I would tend to ask Dave Slagle to make me a pair if I was to do this. The LUNDAHLs have the advantage of shielding for those who do not want to deal with that.
I am not sure if you are being satirical here.
In my limited experience there is no comparison to any bipolar cap against the Black Gate Ns. Not pretending I have tried many of them. Just the ones I have replaced. Mainly NICHICONs in my memory.
I am using the Ns as output couplers in they DSP/crossover that will precede this amplifier. The difference between the original, and I would think better than average (NICHICON) nonpolars installed by YAMAHA, was not subtle. I tend to think they were trying to make this unit good considering the use of PCM63s for the DACs. I see on their latest iteration of this device (mine are 20 years old) they use polarized electrolytics for the output couplers which surprises me. I have not heard the SP2060 but I have read good things about its sound quality despite those caps!
Maybe there is no difference between a WIMA MKP and more exotic capacitors either? I have used plenty of WIMAs and know they are good but for some applications there are much better choices. Or so I think ...
Rick, there's a huge advantage of reducing the chance of screwing up your board by leaving the 5532s in place. Frankly, grounding both inputs and then bypassing with the relevant jumper (may not work for the 3250, but I haven't studied that one) completely removes it from the discussion.
Similarly, I do want to keep this thread a little more focused on getting evm boards up and running/troubleshooting than hijacked by the vagaries of the mod/ cap rolling community. It'd really be best to separate the two so those that need help can find it rather than get mired in discussion of whether this or that cap sounds best at the anecdote level of rigor. In this sense I'm very much siding with Chris and pretty incredulous about the changes being made, and have been far more concerned with ensuring you and others don't burn up your boards and speakers than whether this or that mod helps.
Similarly, I do want to keep this thread a little more focused on getting evm boards up and running/troubleshooting than hijacked by the vagaries of the mod/ cap rolling community. It'd really be best to separate the two so those that need help can find it rather than get mired in discussion of whether this or that cap sounds best at the anecdote level of rigor. In this sense I'm very much siding with Chris and pretty incredulous about the changes being made, and have been far more concerned with ensuring you and others don't burn up your boards and speakers than whether this or that mod helps.
The guy who started did not make any mention of being dedicated to getting the module up and running.
But if this is now yours and it is going to be dedicated to answering the same questions over and over again, well, that's fine.
Take care,
But if this is now yours and it is going to be dedicated to answering the same questions over and over again, well, that's fine.
Take care,
Discussing "cap or transformer sound" isn't a board specific (TPA3255EVM) topic. You could discuss this with any amp-topic here in general and it would always be "off-topic".
If you can distinguish the difference between the caps you mentioned you should try changing the cap at the CStart pin (valid for all TPA32XX) as this affects THD performance in the lower frequency-range as well. (Because this pin is connected in some way to the internal compensation/loopback network. The CStart cap shifts one pole/zero within sup 2kHz range)
If you can distinguish the difference between the caps you mentioned you should try changing the cap at the CStart pin (valid for all TPA32XX) as this affects THD performance in the lower frequency-range as well. (Because this pin is connected in some way to the internal compensation/loopback network. The CStart cap shifts one pole/zero within sup 2kHz range)
I'm not wanting to discourage playing with the boards (and in the next couple months I hope to readily contribute in the form of a daughterboard for I/O), but answering the same question time and again is really what this thread is about, and let's let it do that job well.
How about a new thread that deals with modding this particular board? I would not be surprised if people were interested in this topic. With a dedicated thread such info might be easier to find.
I am looking for an easy solution to the pop-off sounds from my speakers when using the TPA3255EVM and Meanwell 350-48 combination. I came across this module on Audiophonics:
Stereo Protection & delay Module for Loudspeaker V1.1 - Audiophonics
Would this work?
Stereo Protection & delay Module for Loudspeaker V1.1 - Audiophonics
Would this work?
You'd need two mono boards because the amp is BTL.
You may manually mute the amp before switching off with a switch, will also do.
You may manually mute the amp before switching off with a switch, will also do.