the standalone resistor is the simpliest and cheapest method for I/V. But there are others parameter to consider hence the more complex circuits around you can see: traffos, transistor, oaps, filter, further buffer...
Note your dac has 1K resistor with a // as simple filter which seems to be a 1n5 MKP cap. Here is your simple I/V stage and the simplicity here gives an importance to the choice of the passive parts to tailor final sound signature but that's off topic here. But yes you have room for tweaking here.
If I were you I will try the recent chips of AK and ESS as advised already with reference of DACs advised in the post. Advantage of the old pcm dac chips is jitter has less importance than the speedy in frequency modern chips. And it's difficult to compare the both generation face to face because of oversampling and filter embeded in the todays dac chips. Your little TDA1387 is a NOS (non oversampling) one and it's endeed cheaper to make. Though some very good NOS dac with old pcm chips may cost expensive to target a good sound.
If I dare I would ask to Marcelivdg to design a TDA1541 but with a linear 21 bits instead the former 16 bits of this chip
... But due to such technologies it would cost too much to produce for the todays brands imo. It could be called TDA2141...
To make it short your I/V interogation is maybe not deserving all your consideration.
Note your dac has 1K resistor with a // as simple filter which seems to be a 1n5 MKP cap. Here is your simple I/V stage and the simplicity here gives an importance to the choice of the passive parts to tailor final sound signature but that's off topic here. But yes you have room for tweaking here.
If I were you I will try the recent chips of AK and ESS as advised already with reference of DACs advised in the post. Advantage of the old pcm dac chips is jitter has less importance than the speedy in frequency modern chips. And it's difficult to compare the both generation face to face because of oversampling and filter embeded in the todays dac chips. Your little TDA1387 is a NOS (non oversampling) one and it's endeed cheaper to make. Though some very good NOS dac with old pcm chips may cost expensive to target a good sound.
If I dare I would ask to Marcelivdg to design a TDA1541 but with a linear 21 bits instead the former 16 bits of this chip
... But due to such technologies it would cost too much to produce for the todays brands imo. It could be called TDA2141...To make it short your I/V interogation is maybe not deserving all your consideration.
Hi All
I said I will update, so here I am.
In this second round of DAC listening, I ordered just 2 DACs
(much less compared to the first round, which was 8 I think).
They are
- FX-Audio FX-01 DAC (PCM5102)
- SMSL M3 DAC (CS4398)
The FX-Audio FX-01 arrived 2 days ago.
I listened to it,
It's quite detailed actually,
but compared to my TDA1387, it's less good.
Why?
1) the FX-01 sounds less alive and less real than my current TDA1387
2) highs sound kind of weird on it
in general, the feel with the FX-01 is more synthetic,
where with the TDA1387 it feels so real and accurate, as if you have the musical instruments here.
Both were tested with the same speakers of course - Yamaha HS7 monitors which I was recommended about here (thank you all)
In short: no
Still waiting for the SMSL M3 which should arrive nowdays,
I will update after that too.
As I wrote in an earlier post in this thread, after seeing how good the TDA1387, I bought a second (spare) one,
should something ever happen to my first.
If after the SMSL the TDA1387 is still the winner, I will buy a third even,
and probably one of the more expensive models that Lee created, with more inputs and stuff.
Thank you again, to abraxalito, for a recommendation that has really helped me.
I said I will update, so here I am.
In this second round of DAC listening, I ordered just 2 DACs
(much less compared to the first round, which was 8 I think).
They are
- FX-Audio FX-01 DAC (PCM5102)
- SMSL M3 DAC (CS4398)
The FX-Audio FX-01 arrived 2 days ago.
I listened to it,
It's quite detailed actually,
but compared to my TDA1387, it's less good.
Why?
1) the FX-01 sounds less alive and less real than my current TDA1387
2) highs sound kind of weird on it
in general, the feel with the FX-01 is more synthetic,
where with the TDA1387 it feels so real and accurate, as if you have the musical instruments here.
Both were tested with the same speakers of course - Yamaha HS7 monitors which I was recommended about here
(thank you all)In short: no
Still waiting for the SMSL M3 which should arrive nowdays,
I will update after that too.
As I wrote in an earlier post in this thread, after seeing how good the TDA1387, I bought a second (spare) one,
should something ever happen to my first.
If after the SMSL the TDA1387 is still the winner, I will buy a third even,
and probably one of the more expensive models that Lee created, with more inputs and stuff.
Thank you again, to abraxalito, for a recommendation that has really helped me.
Last edited:
And now the second update:
The SMSL M3 arrived today!
It is CS4398 based, and the CS4398 is a Vout chip.
(The FX-Audio FX-01 is also based on a Vout chip - there it was PCM5102)
I am listening to it, and surprise
So far, it sounds better than my #1 DAC, which was the TDA1387 based made by Lee.
The sound is warm and deep,
it does not feel synthetic,
the Bass is also better (hence a very good implementation),
and I hear more details even.
Its sound really reminds a good record player I used to have during the 90s..
(and no, this DAC does not contain any tubes)
The reason I am really surprised is because it is a Vout DAC, Delta Sigma, Over Sampled - all the things that we should avoid when wanting a good sound.
BTW it is not playing at a higher Sample Rate/Bit Rate than the TDA1387.
I am still on 16bit 44.1KHz.
It costs 70$ (inc. shipping), and can be USB Powered, from the same cable that brings the data..
In addition, it has Optical and Coax inputs too (total of 3 inputs, in addition to the USB),
and when using the Optical/Coax, then there is a second USB connector, which is used only for power, for these last 2 inputs.
I am so glad I tried this DAC, because If I hadn't, I would have completely resisted any Vout/DeltaSigma DAC in the future.
What a surprise
I should also praise it for being compact (despite having 3 inputs), and amazing build quality.
This is my first SMSL product,
but after such a sound, there will definitely be more SMSL products that I buy in the near future.
The SMSL M3 arrived today!
It is CS4398 based, and the CS4398 is a Vout chip.
(The FX-Audio FX-01 is also based on a Vout chip - there it was PCM5102)
I am listening to it, and surprise
So far, it sounds better than my #1 DAC, which was the TDA1387 based made by Lee.
The sound is warm and deep,
it does not feel synthetic,
the Bass is also better (hence a very good implementation),
and I hear more details even.
Its sound really reminds a good record player I used to have during the 90s..
(and no, this DAC does not contain any tubes)
The reason I am really surprised is because it is a Vout DAC, Delta Sigma, Over Sampled - all the things that we should avoid when wanting a good sound.
BTW it is not playing at a higher Sample Rate/Bit Rate than the TDA1387.
I am still on 16bit 44.1KHz.
It costs 70$ (inc. shipping), and can be USB Powered, from the same cable that brings the data..
In addition, it has Optical and Coax inputs too (total of 3 inputs, in addition to the USB),
and when using the Optical/Coax, then there is a second USB connector, which is used only for power, for these last 2 inputs.
I am so glad I tried this DAC, because If I hadn't, I would have completely resisted any Vout/DeltaSigma DAC in the future.
What a surprise
I should also praise it for being compact (despite having 3 inputs), and amazing build quality.
This is my first SMSL product,
but after such a sound, there will definitely be more SMSL products that I buy in the near future.
Last edited:
And now the second update:
The SMSL M3 arrived today!
It is CS4398 based, and the CS4398 is a Vout chip.
(The FX-Audio FX-01 is also based on a Vout chip - there it was PCM5102)
I am listening to it, and surprise
So far, it sounds better than my #1 DAC, which was the TDA1387 based made by Lee.
The sound is warm and deep,
it does not feel synthetic,
the Bass is also better (hence a very good implementation),
and I hear more details even.
Its sound really reminds a good record player I used to have during the 90s..
(and no, this DAC does not contain any tubes)
The reason I am really surprised is because it is a Vout DAC, Delta Sigma, Over Sampled - all the things that we should avoid when wanting a good sound
BTW it is not playing at a higher Sample Rate/Bit Rate than the TDA1387.
I am still on 16bit 44.1KHz.
It costs 70$ (inc. shipping), and can be USB Powered, from the same cable that brings the data..
In addition, it has Optical and Coax inputs too (total of 3 inputs, in addition to the USB),
and when using the Optical/Coax, then there is a second USB connector, which is used only for power, for these last 2 inputs.
I am so glad I tried this DAC, because If I hadn't, I would have completely resisted any Vout/DeltaSigma DAC in the future.
What a surprise
I should also praise it for being compact (despite having 3 inputs), and amazing build quality.
This is my first SMSL product,
but after such a sound, there will definitely be more SMSL products that I buy in the near future.
The CS4398 is a pretty good converter. I am not surprised.
I didn't know that.The CS4398 is a pretty good converter. I am not surprised.
I knew that it's popular, but only when hearing it do I understand why..
BTW we should also give credit to SMSL for the implementation..
I found a picture of its internals,
here it is for anyone who is curious:
An externally hosted image should be here but it was not working when we last tested it.
The capacitors don't look anything special..
Yet the sound is the best one I had, from 10 different DACs.
Last edited:
The IC fab process for a fast sigma-delta chip is probably not great for analog amplifiers.
Current output is easier to make precise since electric charge is conserved, current switching can basically be made exact almost to the electron, whereas voltage is affected by stray resistance, capacitance, inductance and external magnetic fields. Or put another way electric energy can readily transform into other forms, and is less easy to account.
The first round was also fruitful.
The TDA1387 DAC is so good that after I heard it, I purchased another unit, to have as backup..
For anyone interested, it's this:
L1387DAC 4X Quad TDA1387 fever USB decoder beyond the TDA1543 | eBay
Amazing value you get, for 21$...
Of course
Looks like this is no longer on Ebay !....Or any other site !
Hi Alexandre
This is amazing, thank you for this link!
This explains why I am enjoying it so much..
It also shows that my taste is quite persistent,
since I really liked the TDA1387 by Philips,
and really like this CS4398, from the same creators, just under a different company..
Can anyone please explain what does the DEM part do?
Why not?
The link is still working..
It's also on eBay, with the same old link.
I highly recommend favoring eBay over AliExpress every time something is sold on both.
eBay (and Paypal) take care of customers when there's a problem.
On AliExpress, should you ever have a problem, and have to open a dispute,
you will see that AliExpress's system favors sellers over customers, even when the seller was mistaken.
I avoid AliExpress as much as possible, and will not buy there anything above a few $.
Earlier when I went to the link Alexandre gave, I was in a hurry, so only read the section regarding the CS DAC.
I now came back to it, and read the whole page.
I am curious what do they mean by the term "super DAC"?
Also, can you please tell what is a switched capacitor DAC circuit?
I now came back to it, and read the whole page.
I am curious what do they mean by the term "super DAC"?
Also, can you please tell what is a switched capacitor DAC circuit?
In general, dynamic element matching (DEM) reduces the effect of mismatch on DAC performance by switching between unit elements in such a way that the errors average out.
The oldest type of DEM that I know of, is the type used in the TDA1540 and TDA1541. Elements of current mirrors are periodically swapped to get precisely the correct average ratio and RC filters are used to get rid of the switching frequency. Hence the large number of external capacitors that these chips need.
Modern multibit sigma-delta DACs use a different type of DEM. For example, suppose you want to convert a five-bit noise-shaped signal to analogue form. You could do that with a binary to thermometer code converter and 31 unit DACs. When the code is 00000, you enable zero unit elements, when the code is 00001, you enable the first element, with code 00010 you enable the first and the second elements and so on, with code 11111, you enable all 31 of them.
So far so good, but the unit elements are never exactly equal. Therefore, the transfer from digital code to analogue output signal will never be a straight line. This causes distortion and in a noise-shaped DAC, it also converts out-of-band quantization noise into the audio band.
You could, however, decide to use different unit elements each time. For example, suppose you want to convert the code 00010. Instead of enabling the first and the second element, you can also enable the 12th and the 17th element, or the 5th and the 30th. When you randomly pick the desired number of elements out of the available elements, you on average get a straight line transfer from digital code to analogue output. This eliminates the distortion. Many smarter ways to do this have been devised, that not only eliminate the distortion but also much reduce the noise increase caused by mismatch between the unit elements.
To give you a single-bit example:
Suppose you have a small capacitor that you charge to some reference voltage in the first half of each clock cycle. In the second half you discharge it into the virtual ground input of a current-to-voltage converting amplifier. Depending on a data signal, you do or don't swap the connections of the capacitor during the discharge phase. You then have a basic single-bit switched-capacitor DAC.
Advantage compared to just switching on or off a current source with the data signal: much reduced jitter sensitivity, because the amount of charge coming out of the capacitor doesn't depend much on the exact length of the discharge phase.
Disadvantage: large discharge current peaks are hard to handle for the current-to-voltage converter. There are several tricks to solve this again.
You probably missed the previous page of this thread..
Please see this specific post: (#84)
Will Voltage-Out DACs Ever Be Good, Like Current-Out DACs?
I am on a ΔΣ DAC, and it is amazing,
it's even better (significantly) than the old one that I really liked.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.