bits are sent in the form of square waves. And square waves, just like any other wave, are never perfect. Converters (some more than other) are sensible to the quality of what it is sent. They interpret them a certain way with all the tools we have given them. For example, take a look at the first page where you can see the analog square wave of the SEN,CEN. That is a nice attempt at a perfect square wave but it isnt.
i also doubt that your computer never misinterpreted a bit but it is possible
i also doubt that your computer never misinterpreted a bit but it is possible
Last edited:
I am sorry to say that I do not agree with the above arguments.
Firstly an IV converter is a 100% analogue circuit. It receives current signals (from either a MC phono pick-up, or the analogue current output of a DAC). Thus the power supply for the circuit is not too different from those feeding any other analogue circuits.
Secondly, those who read the article will know that the bias circuit, driven by the floating supply, runs in a completely different current loop than the signal current loop. Since current can only flow in a loop, the two are independent of each other. There are no electronic components or circuits that I know of, however imperfect, that do not obey Kirchoff's Law. And Kirchoff's law makes sure that the bias current does not affect the signal current loop, and vice versa.
Thirdly, the square wave as shown in the first page of this thread is feed by a functions generator and has nothing to do with digital signal. The output square wave is "not perfect" because of parasitic capacitances in the JFETs. There are ways and means to reduce the effects of these capacitances to get a more "perfect" square wave, and we do know how. But there is little point in doing so, because at most DAC outputs, some form of LP or reconstruction filter is usually implemented. That is also why there are provisions for Civ on the PCB, to filter off some HF content of the current signal from the DAC.
My 2 cents,
Patrick
.
Firstly an IV converter is a 100% analogue circuit. It receives current signals (from either a MC phono pick-up, or the analogue current output of a DAC). Thus the power supply for the circuit is not too different from those feeding any other analogue circuits.
Secondly, those who read the article will know that the bias circuit, driven by the floating supply, runs in a completely different current loop than the signal current loop. Since current can only flow in a loop, the two are independent of each other. There are no electronic components or circuits that I know of, however imperfect, that do not obey Kirchoff's Law. And Kirchoff's law makes sure that the bias current does not affect the signal current loop, and vice versa.
Thirdly, the square wave as shown in the first page of this thread is feed by a functions generator and has nothing to do with digital signal. The output square wave is "not perfect" because of parasitic capacitances in the JFETs. There are ways and means to reduce the effects of these capacitances to get a more "perfect" square wave, and we do know how. But there is little point in doing so, because at most DAC outputs, some form of LP or reconstruction filter is usually implemented. That is also why there are provisions for Civ on the PCB, to filter off some HF content of the current signal from the DAC.
My 2 cents,
Patrick
.
Last edited:
maybe i am m wrong, but what i wanted to say is that that not every power supply is suitable for good signal (digital) integrity. that a digital signal is not only one and zeros, it is an non-perfect square wave that can be misinterpreted, depending on the dac and related components. At least that how i understand it.
by the way your square waves looks darn good if you ask me
looking foward to build it.
by the way your square waves looks darn good if you ask me
looking foward to build it.
yes, the posting on the previous page is one of many other naive statements written just like it but people who honestly think they know what they are talking about (this is said not in a derogatory way), the whole bits is bits thing has been run over and over and over, in this case the poster even seems to be unaware of the fact that the dac from the point it becomes analogue, to the very point we are dealing with it here, is about as vulnerable low level analogue current conveyed signal as is possible.
Last edited:
jitter, until till you have heard a manster clock powered by low noise power supply vs. smps you wouldn't understand, I didn't either.
Last edited:
I think the original messge got lost. I said that in Digital circuits, I did not beleive batteries would be The Great Improovement and hence my PC running for 2 years analogy. This comment was in response to
'I work with robots and vfds in my real job and you wouldn't believe all the issues from poor quality mains.'
As I would be suprised a standard (well designed) power supply would have trouble with digital circuits.
'I work with robots and vfds in my real job and you wouldn't believe all the issues from poor quality mains.'
As I would be suprised a standard (well designed) power supply would have trouble with digital circuits.
In most "standard" digital circuits such as a computer, the regularity of the timing of the signal (jitter) is not critical for the computation to function.
In digital audio, the digital circuit still functions with jitter, but the results in the audio domain suffers.
If you wish to discuss this further, I think there are many threads in this forum which is a more suitable platform than this particular thread.
Best Regards,
Patrick
In digital audio, the digital circuit still functions with jitter, but the results in the audio domain suffers.
If you wish to discuss this further, I think there are many threads in this forum which is a more suitable platform than this particular thread.
Best Regards,
Patrick
true but well designed power supplies are a rarity , applies to both sections of the dac, remember the "digial section is still analog signals (no infinite bw in the real world.)
Sorry to get off topic, but running the Cen on batteries only makes sense to lead one use for them the whole dac (of corse with proper post regulation.) I'll shut up.
I've been using the cen for about 2 weeks now, and just about on to my third set of batteties. Its running very well in my system. It added more transparency and detail, which was just what I was after.
With my current buffer, im now using the +5V supply from the dac instead of the 9V battery. It seems more stable but its still drifting a little bit, within +-25mV as per TDA1541 specs. So maybe not worth worrying about. The only issue I don't like is the small amount of dc going in to my f5, does anybody think that would cause any problems?
Ryan
With my current buffer, im now using the +5V supply from the dac instead of the 9V battery. It seems more stable but its still drifting a little bit, within +-25mV as per TDA1541 specs. So maybe not worth worrying about. The only issue I don't like is the small amount of dc going in to my f5, does anybody think that would cause any problems?
Ryan
> It added more transparency and detail, which was just what I was after.
Compared to what ?
> The only issue I don't like is the small amount of dc going in to my f5
A DC offset of about +/-20mV at the F5 input is IMHO no issue for the speakers.
But if you are still worried, you can put a servo at one of the following locations :
1) the biasing current source for the TDA1541
2) the lower current source of a JFET follower (or DC-B1, if you prefer) after Riv
Of course you can always use a decoupling cap at the F5 input.
But that would not be the solution I want to use for myself.
I would try to stabilise the biasing current source for the TDA1541 to start with, which is the source of your offset problem.
But pleased to hear you are enjoying the results of this very simple but elegant circuit.
Patrick
Compared to what ?
> The only issue I don't like is the small amount of dc going in to my f5
A DC offset of about +/-20mV at the F5 input is IMHO no issue for the speakers.
But if you are still worried, you can put a servo at one of the following locations :
1) the biasing current source for the TDA1541
2) the lower current source of a JFET follower (or DC-B1, if you prefer) after Riv
Of course you can always use a decoupling cap at the F5 input.
But that would not be the solution I want to use for myself.
I would try to stabilise the biasing current source for the TDA1541 to start with, which is the source of your offset problem.
But pleased to hear you are enjoying the results of this very simple but elegant circuit.
Patrick
I have not tried this myself, but I think it is probably a good solution for a stable CCS :
http://datasheets.maxim-ic.com/en/ds/MAX6126.pdf
See figure 2 on P.15.
In your case, use MAX6126 2.048V, Grade A, and say MPSA18 or BC550 as driver.
Of course you also need a low tempco resistor of say 0.25W nominal.
I would first use a trimmer to find the right value for your DAC.
Then replace it with (multiple if necessary) fixed resistors (wired in parallel) to get the exact value as the trimmer to be replaced.
Patrick
.
http://datasheets.maxim-ic.com/en/ds/MAX6126.pdf
See figure 2 on P.15.
In your case, use MAX6126 2.048V, Grade A, and say MPSA18 or BC550 as driver.
Of course you also need a low tempco resistor of say 0.25W nominal.
I would first use a trimmer to find the right value for your DAC.
Then replace it with (multiple if necessary) fixed resistors (wired in parallel) to get the exact value as the trimmer to be replaced.
Patrick
.
Last edited:
Thanks for the suggestion for an alternative current source.
>Compared to what ?
Better compared to my SRPP valve circuit. Mind you it had no current buffer so the voltage on the DAC was way over spec. I'm done with tubes for a while i think, not to say there are no good tube circuit designs.
This weekend i'm going to try swapping the decoupling caps in the Cen circuit with some better quality film caps. I've already had much success with replacing the IV resistor to a homemade Honeycomb mobius loop resistor using CuNi44 resistance wire, which i belive is John Browns design, from his thread described here about half way down the page:
http://www.diyaudio.com/forums/digital-line-level/79452-building-ultimate-nos-dac-using-tda1541a-357.html
>Compared to what ?
Better compared to my SRPP valve circuit. Mind you it had no current buffer so the voltage on the DAC was way over spec. I'm done with tubes for a while i think, not to say there are no good tube circuit designs.
This weekend i'm going to try swapping the decoupling caps in the Cen circuit with some better quality film caps. I've already had much success with replacing the IV resistor to a homemade Honeycomb mobius loop resistor using CuNi44 resistance wire, which i belive is John Browns design, from his thread described here about half way down the page:
http://www.diyaudio.com/forums/digital-line-level/79452-building-ultimate-nos-dac-using-tda1541a-357.html
> a homemade Honeycomb mobius loop resistor
The ideal came from a German DIY tube audio site, if I am not wrong.
http://www.jogis-roehrenbude.de/Spulenwickeln.htm
Patrick
The ideal came from a German DIY tube audio site, if I am not wrong.
http://www.jogis-roehrenbude.de/Spulenwickeln.htm
Patrick
Thanks for the link.
One other thing I'd like to point out about the sound quality I'm experiencing is; warmup time is extremely critical, it can make or break your listening experience.
It seems as though the more one upgrades there system the more it matters. Another thing it is bringing out is weaknesses in other parts of my system, I'm noticing subtle frequency response peaks in my speakers for example.
One other thing I'd like to point out about the sound quality I'm experiencing is; warmup time is extremely critical, it can make or break your listening experience.
It seems as though the more one upgrades there system the more it matters. Another thing it is bringing out is weaknesses in other parts of my system, I'm noticing subtle frequency response peaks in my speakers for example.
IMHO your warm-up issue is more related to the current bias of the TDA1541 DAC, rather than the IV circuit itself.
When using DACs like PCM1704 and AD1865, where such bias is taken care of and stabilised internally inside the IC, we have experienced no such warm up issues. The IV circuits play to their full glory from T=0s, and we have no drift or output DC issues at all.
This is why I suggested you to try out the MAX6126A based current source. The MAX6126A is well compensated for thermal drift. You only need to make sure that you do not spoil that stability by using a drifting resistor. If cost is no issue, I suggest you use the likes of Vishay S102 or Caddock TF020. But at least Vishay Dale CMF60-T2 (50ppm/°C).
(I am assuming of course that the TDA1541 itself is thermal-drift stable.)
Patrick
.
When using DACs like PCM1704 and AD1865, where such bias is taken care of and stabilised internally inside the IC, we have experienced no such warm up issues. The IV circuits play to their full glory from T=0s, and we have no drift or output DC issues at all.
This is why I suggested you to try out the MAX6126A based current source. The MAX6126A is well compensated for thermal drift. You only need to make sure that you do not spoil that stability by using a drifting resistor. If cost is no issue, I suggest you use the likes of Vishay S102 or Caddock TF020. But at least Vishay Dale CMF60-T2 (50ppm/°C).
(I am assuming of course that the TDA1541 itself is thermal-drift stable.)
Patrick
.
Last edited:
>IMHO your warm-up issue is more related to the current bias of the TDA1541 DAC, rather than the IV circuit itself.
Yes, but i was referring to my system as a whole. The Cen circuit exposes the warmup time in my system more so than ever. Sorry, i should be more descriptive.
But yes, i will look into the device that you recommend, thanks.
Yes, but i was referring to my system as a whole. The Cen circuit exposes the warmup time in my system more so than ever. Sorry, i should be more descriptive.
But yes, i will look into the device that you recommend, thanks.