This thread started as an OT from the TPA3116 where jean-paul had said:
I certainly accept that not all TPA3116 amps are identical, layout and power supply details are a major part of how the chip will sound. However I'd also venture that there's a characteristic of that chip too - its relatively poor (noisy) performance in the bass, for which reason I passed over it and have settled on TDA8932 for now as offering the bass quality I'm after. I'm still hopeful though that some of the newer TI chips will turn out not to have this issue and when I can get hold of them (thinking here of TPA5766/8) I'm enthusiastic to have a listen.
(The part where you wrote 'no-one can say all TPA3116 amps sound the same' is a straw man as I've not claimed that they do. I've never heard anyone else say it either.)
As it goes with amp chips, so it is with DACs - whilst implementation details matter for sure, S-D type DACs suffer from intrinsic noise modulation. If anyone's in any doubt at all about this fact, go look at the ES9018 thread where measurements are showing this - here: http://www.diyaudio.com/forums/digi...e-reference-dac-8-channel-47.html#post4001439
This is a detail that cannot be overcome by fastidious attention to power supplies since its an aspect of the hardware architecture or perhaps digital processing. The noisiness of S-D DACs manifests in various ways, generally in the reduction of dynamics. DACs with opamps as their output stage (such as ES9023) suffer more than those without (ES9018), because their on-chip opamps are susceptible to the HF noise being fed in from the DAC part, but also they're power supply sensitive too.
So in conclusion, yep I am comparing how DACs sound in real life where details matter. Some details of particular chips cannot be undone by implementations, no matter how ingenious. That's not to say other aspects cannot be optimized into the stratosphere.
I certainly accept that not all TPA3116 amps are identical, layout and power supply details are a major part of how the chip will sound. However I'd also venture that there's a characteristic of that chip too - its relatively poor (noisy) performance in the bass, for which reason I passed over it and have settled on TDA8932 for now as offering the bass quality I'm after. I'm still hopeful though that some of the newer TI chips will turn out not to have this issue and when I can get hold of them (thinking here of TPA5766/8) I'm enthusiastic to have a listen.
(The part where you wrote 'no-one can say all TPA3116 amps sound the same' is a straw man as I've not claimed that they do. I've never heard anyone else say it either.)
As it goes with amp chips, so it is with DACs - whilst implementation details matter for sure, S-D type DACs suffer from intrinsic noise modulation. If anyone's in any doubt at all about this fact, go look at the ES9018 thread where measurements are showing this - here: http://www.diyaudio.com/forums/digi...e-reference-dac-8-channel-47.html#post4001439
This is a detail that cannot be overcome by fastidious attention to power supplies since its an aspect of the hardware architecture or perhaps digital processing. The noisiness of S-D DACs manifests in various ways, generally in the reduction of dynamics. DACs with opamps as their output stage (such as ES9023) suffer more than those without (ES9018), because their on-chip opamps are susceptible to the HF noise being fed in from the DAC part, but also they're power supply sensitive too.
So in conclusion, yep I am comparing how DACs sound in real life where details matter. Some details of particular chips cannot be undone by implementations, no matter how ingenious. That's not to say other aspects cannot be optimized into the stratosphere.
You bring up a very interesting point - are measurements or sound the more important aspect? S-D DACs are designed to give good results on the standard measurements (THD with FFT). So for example they trade off distortion for noise, by design so the FFT looks 'clean'. If you build a DAC with TDA1387 you'll find it measures a whole heap better than TDA1543. One of my early prototypes did quite a bit better than I was lead to expect from the DS. I think the improvements in measurement stem from the 'continuous calibration' innovation that constantly adjusts the bit weightings in the DAC - the TDA1543 has linearity problems arising from bit weight errors as its using fixed resistors.
Last edited:
matt_garman wrote:
I'll put forward my own Ozone DAC as fitting the bill for this - certainly one can be assembled for considerably less than $100 in parts cost. Its using a TDA1387 (or stack of them) in NOS followed by an LC passive filter for image rejection, ending up in an AD815 (or discrete) balanced buffer which then feeds a hand-wound ferrite cored transformer.
I don't have PCBs for these elements yet as I've built all my prototypes on proto board. But if there's sufficient demand then I'll get to work on the boards....
I'll put forward my own Ozone DAC as fitting the bill for this - certainly one can be assembled for considerably less than $100 in parts cost. Its using a TDA1387 (or stack of them) in NOS followed by an LC passive filter for image rejection, ending up in an AD815 (or discrete) balanced buffer which then feeds a hand-wound ferrite cored transformer.
I don't have PCBs for these elements yet as I've built all my prototypes on proto board. But if there's sufficient demand then I'll get to work on the boards....
Last edited:
Hi
I am a great fan of TDA 1543 and 1545 and matched to tpa 3116 with markaudio 12p fullranger, the sound is fantastic. The icing on the cake is that it is so cheap !!
i have the TDA 1541 chip waiting to be build but the simplicity of the single power supply in TDA 1543 and resulting sound is hard to resist.
Yes i am interested to build TDA1387 dac. i have AD 815 also.
can you post a no frills schematic for me to follow ? A pcb would be a bonus.
I am not aiming for the ultimate but a modest system to play with .
thanks
kp93300
ChrisMmm wrote:
So have you found that your Subbu beats out the TDA1541A-based DAC? Does the Pedja I/V stage have any anti-imaging filter? If not it mightn't be a great match for the TPA3116 - I'm not sure how that chip will handle the considerable ultrasonic content emanating from an unfiltered NOS DAC.
Given one is a CD player and one a USB DAC the comparison would not just be the DAC obviously. However, as I have never done a side by side compare of the 2, I have just done such a comparison. Again, I don't see my ears as being fine tuned instruments at all but my preference, slight tho it was, is for the 9023. Just a little more detailed and airy - but I would be happy with both. I rarely play CDs these days but I would never part with the CD880.
And yes, I did modify the player to be NOS, can't say about the I/V stage, been a while since I saw the circuit but I doubt it has such a filter.
Ozone single TDA1387 with Fastron LC filter
I've been working on the filter design, to use standard Mouser (Fastron) inductors. The wound components are by far the most expensive aspect of this design - but they do come down nicely (to around $0.60 each) in reasonable volumes.
The caps here are C0G/NP0 types for lowest loss.
The power supply is intended to be fed to a hexacap which sits under the DAC PCB - you can decide how big to make that, the higher the capacitance the better the LF ambience retrieval.
I'll work on the AD815 back-end and post it up later. Since its NOS, the output buffer provides the droop correction.
I've been working on the filter design, to use standard Mouser (Fastron) inductors. The wound components are by far the most expensive aspect of this design - but they do come down nicely (to around $0.60 each) in reasonable volumes.
The caps here are C0G/NP0 types for lowest loss.
The power supply is intended to be fed to a hexacap which sits under the DAC PCB - you can decide how big to make that, the higher the capacitance the better the LF ambience retrieval.
I'll work on the AD815 back-end and post it up later. Since its NOS, the output buffer provides the droop correction.
Do we compare DAC or DAC chip ?
How to know about two dac chips if the the input & output stage are different for the two DACs ?
It could be nice to compare with ears two dac chips using two core boards with same layout for input stages, same local decoupling caps, same PS design (e.g. Audial use in the AYA2 shunts because the noise floor of the tDA1541 is high enough to live with it; Subbu DAC use a very low noise PS, ferrite beads, inductors... Audial didn't in its old design.
But the outputstages is always different ! So what you compare ? OK let compare the subjective sound of both, maybe one will win or maybe on several system and with the culture of each listeners some will find qualities and defaults to be different for each DAC !
Or one could winn by a large margin ? I surmise many can already answer to this !
One thing more and I know many like JP will not completly agree (but a little as like P. Rogic) parts are importants : change some passive parts and you can tailor each DAC with a total different sound ! I maid it with the Subbu : you can swap for a evil dancer to a metronom with both qualities & defaults with two set up...
Well Advantage to the TDA1541 here : 30 years of setup and tweak. Some men like Pedja, Thorsten, Jocko have a long experience. JP is not a bird of the year (and my understanding he played a little with the TDA1541 but I'm sure the ES9023 didn't say all he knows.... (here JP will say the caps just change the sounds and tells off me but just some played with the supplys and the passive parts and maid this chip with the same layout -ES9023 SUBBU boards- very different soundings. I repeat it : sound can be calm to demoniac just playing with the passive parts (and given different subjectibve qualities-
So fellows, with what do you listen your DACS ? Just with scopes, but if one has a pleasant distorsion and winn the match !!!!!!
Just two cents as I'm just a listener and not a designer. But at the end I doon't know how to compare a dac chip like the ES9023 and a TDA 1541 which need also a I/V conversion ? Do we have some knowledge with the embeded "opa" of the ES9023.
Could we not compare the ES9018 in I/Vmode with the TDA 1541 ? The ES chip will have certainly more transparency, for the rest it could be also about the tastes of the listeners ?.... But the large marginn winner effect ?!
sorry for two cents... I even read all the ES9023 DACs sound the same... so...
Hoax to do that ? Compare apple with plane ?
How to know about two dac chips if the the input & output stage are different for the two DACs ?
It could be nice to compare with ears two dac chips using two core boards with same layout for input stages, same local decoupling caps, same PS design (e.g. Audial use in the AYA2 shunts because the noise floor of the tDA1541 is high enough to live with it; Subbu DAC use a very low noise PS, ferrite beads, inductors... Audial didn't in its old design.
But the outputstages is always different ! So what you compare ? OK let compare the subjective sound of both, maybe one will win or maybe on several system and with the culture of each listeners some will find qualities and defaults to be different for each DAC !
Or one could winn by a large margin ? I surmise many can already answer to this !
One thing more and I know many like JP will not completly agree (but a little as like P. Rogic) parts are importants : change some passive parts and you can tailor each DAC with a total different sound ! I maid it with the Subbu : you can swap for a evil dancer to a metronom with both qualities & defaults with two set up...
Well Advantage to the TDA1541 here : 30 years of setup and tweak. Some men like Pedja, Thorsten, Jocko have a long experience. JP is not a bird of the year (and my understanding he played a little with the TDA1541 but I'm sure the ES9023 didn't say all he knows.... (here JP will say the caps just change the sounds and tells off me but just some played with the supplys and the passive parts and maid this chip with the same layout -ES9023 SUBBU boards- very different soundings. I repeat it : sound can be calm to demoniac just playing with the passive parts (and given different subjectibve qualities-
So fellows, with what do you listen your DACS ? Just with scopes, but if one has a pleasant distorsion and winn the match !!!!!!
Just two cents as I'm just a listener and not a designer. But at the end I doon't know how to compare a dac chip like the ES9023 and a TDA 1541 which need also a I/V conversion ? Do we have some knowledge with the embeded "opa" of the ES9023.
Could we not compare the ES9018 in I/Vmode with the TDA 1541 ? The ES chip will have certainly more transparency, for the rest it could be also about the tastes of the listeners ?.... But the large marginn winner effect ?!
sorry for two cents... I even read all the ES9023 DACs sound the same... so...
Hoax to do that ? Compare apple with plane ?
Last edited:
You are right of course. So, if you want to compare DACS, it is so important to state the environment in which you compare. This, as you say, includes personal taste and any bias you have to a particular chip
But lets not allow that to stop us comparing and stating preferences.
You are right of course. So, if you want to compare DACS, it is so important to state the environment in which you compare. This, as you say, includes personal taste and any bias you have to a particular chip
But lets not allow that to stop us comparing and stating preferences.
I'm a little worried you could think such a thing. If you would know me better I'am certainly not the one who want to stop anybody intteluctually honest and goodwilled. To be shorter where do you you see above I want people to stop ? I just say again to those two giants of DIYAUDIO to compare apples to apples or you are gliding...knowing myself they know it far better than me as i'm not a technician in anyway.
Abrax and JP know me a little and know that I often listen to their advises. But sometimes I suffered myself from bad judgement on a thread about tweaking the Subbu, especially when I wrote just few caps could transform totaly the sound charachter of this dac. Ask to JP
! So here i have to read the opposite with less concistency : one dac could not change with some caps and the other lived 30 years of tweaks where each layout, resistor, cap were discussed? And I can not say it ! Are you kidding ? And I add they never listen the dac of the other... so are we as usual listen a dac with eyes on a scope 
whichis just 10% of what could be "seen" ! As I respect a lot JP and Abraxalito who share a lot, I just repeat the rule as I saw a little of bad faith above, or biased is maybe a better word.
I have hear a lot of TDA1545/43/41, tweaked some cd player with its. Have share my modest and long tweaking cap experience with the Subbu with often a bad return as some people prefer always listen the prophets whic can be wrong sometimes and also because they are lazy... I have two Subbu also. No ES9018, some say to me they are close (as transparent and clear).
And I have not a clue about what is the best dac chip to rule them all as people say where you live, so I wanted just to repeat: let compare apple to apple, not about the dac chips... which I'am the first to find it inteeresting to compare, but about the way to do it. You must do everything to try to hear the dac chip and not the layout around and the better way is to try to share same typology of PS and Layout when it's possible. Some say they will just compare same scope goals, but I believe they are wrong, their eyes and technical mind are not their ears.... the first can not replace the last totaly even if you are Einstein ! With a same scope measurement, just change a cap and a Dac will sound different and the most bad faithed will say to you : you are biased because it's about taste ! LOL !
This is not because someone disagree he want to stop the others...
Last edited:
I never said you want to stop people comparing. I am agreeing with you that its is difficult to compare DAC chip to DAC chip and get a genuine apples to apples compare. So we need to be careful and state the environments when we make comparisons.
To be honest I have difficulties with such comparisons in that unless there is a clear difference that I can hear in the first few minutes I will no longer trust my ears/brain. I find it easier to determine over long term listening when a change is made in my system - as in a certain detail is heard that I wasn't aware of before, improved dynamics etc..
As to my brief compare of the CD880/TDA1541 and Subbu - yes I was somewhat surprised and a little embarrassed I could not easily determine any obvious differences but I tell it like I hear it. I am sure there are differences but as I said I just don't use the CD player any more so no point pursuing.
To be honest I have difficulties with such comparisons in that unless there is a clear difference that I can hear in the first few minutes I will no longer trust my ears/brain. I find it easier to determine over long term listening when a change is made in my system - as in a certain detail is heard that I wasn't aware of before, improved dynamics etc..
As to my brief compare of the CD880/TDA1541 and Subbu - yes I was somewhat surprised and a little embarrassed I could not easily determine any obvious differences but I tell it like I hear it. I am sure there are differences but as I said I just don't use the CD player any more so no point pursuing.
Re-reading my post I have to apologise.
Where I said "any bias you have to a particular chip" I wan't referring to you specifically, I have posted like this before and upset somebody so apologies. By "you" in that post I actually mean anybody, maybe I should say "one" or word it differently.
Once again, I do agree with you!
Where I said "any bias you have to a particular chip" I wan't referring to you specifically, I have posted like this before and upset somebody so apologies. By "you" in that post I actually mean anybody, maybe I should say "one" or word it differently.
Once again, I do agree with you!
I can't find much if anything to disagree with you Eldam Comparing DAC chips is context dependent for sure, even more so than say comparing opamps. Then we have to factor in biasses as some of us prefer particular 'sounds' (like 'detail').
Eventually there will be some standardized measurements which correlate well with listening - the best guess at what those will be for now is 'noise loading'. Thorsten has talked about that, its also the method that Belcher used decades ago to get correlation with perceived SQ of amps.
Comparing DAC chips is context dependent for sure, even more so than say comparing opamps. Then we have to factor in biasses as some of us prefer particular 'sounds' (like 'detail').Eventually there will be some standardized measurements which correlate well with listening - the best guess at what those will be for now is 'noise loading'. Thorsten has talked about that, its also the method that Belcher used decades ago to get correlation with perceived SQ of amps.
Ozone DAC output buffer
Here's how the AD815 buffer looks. Not all component values have been designed in, some - particularly around the AD815 feedback network - are placeholders until the preceding LC filter has been chosen.
The power supply is typically 5-6V (so I can use the smallest physical size caps in the hexacap passive shunt which this sits atop of).
Here's how the AD815 buffer looks. Not all component values have been designed in, some - particularly around the AD815 feedback network - are placeholders until the preceding LC filter has been chosen.
The power supply is typically 5-6V (so I can use the smallest physical size caps in the hexacap passive shunt which this sits atop of).
So now you are opening a new thread on my remarks. How nice
The parallel between both DAC chips is of course that they're both among the cheapest the manufacturer produces/produced while the performance does not disappoint (at least soundwise).The message is still that you can't compare a product you haven't heard to one you have heard even if they have the same DAC chip. Many words don't cover up that fact.
The sentence 'no-one can say all TPA3116 amps sound the same' is not a a straw man. I think it is called a "statement" in english. Or an opinion, if you like.
Last edited:
Must be a language difference. A cultural difference as well as IMO this all distracts from the issue itself. Your remark is in the same ball park. You compared a certain (no brand or type was mentioned) ES9023 device with a TDA1543 device and told a member that he would likely like TDA1543 designs better with some suggested changes in power supply and filtering. This without knowing the somewhat better designs with the ES9023. No blah blah can camouflage that.This thread is going nowhere.
Ontwijkend taalgebruik.
Ontwijkend taalgebruik.
Last edited:
Now you're resorting to misrepresentation of what I said. To wit
....told a member that he would likely like TDA1543 designs better.
whereas I recall I made a very specific suggestion which involved taking the power supply from a Subbu DAC and applying that to a TDA1543 with a passive filter applied to the output. So nothing at all about 'TDA1543 designs' as you've erroneously claimed, rather a very particular implementation. As far as I'm aware no such implementation yet exists in an available design.
And of course I'm willing to be wrong and learn if he didn't in fact like my suggestion better.
The thread is 'going nowhere' simply because your arguments aren't going anywhere perchance?
....told a member that he would likely like TDA1543 designs better.
whereas I recall I made a very specific suggestion which involved taking the power supply from a Subbu DAC and applying that to a TDA1543 with a passive filter applied to the output. So nothing at all about 'TDA1543 designs' as you've erroneously claimed, rather a very particular implementation. As far as I'm aware no such implementation yet exists in an available design.
And of course I'm willing to be wrong and learn if he didn't in fact like my suggestion better.
The thread is 'going nowhere' simply because your arguments aren't going anywhere perchance?
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.