Did you make any modifications to the IV circuit at all?
I have replaced the classic OP by discrete OP .
Expensive but better for me .
Serge
I have replaced the classic OP by discrete OP .
Expensive but better for me .
Serge
U can check my photo. There is a lot of smd choke around the board labelled "L" next to it.
If u use passive transformer out, the board only needs 3.3V for input. I have tap into the with the cap after the 1117 regulator (basically I have removed those regulator and caps from the board).
For the oscillator, there is a smd choke before the 2 smd caps. Remove that to disconnect from the main 3.3v supply and tap in lifepo4 from there.
I don't use the optical input so I remove to smd choke to prevent noise from that part.
ES9038Q2M Board
Why did you remove the power supply channel 5V line that only supplies the display and commands?
I will try a combination of LME49720 and AD 797s for the op amps.
I might try the filtering suggested by ESS as the changes are easy to make as the IV circuit is nearly identical.
Exactly. We were even running tube heaters off batteries and it was the benchmark to compare non battery designs to. It was actually quite surprising how much difference the heater supplies made.
The only thing that can be tricky, especially with digital supplies is type of local bypassing. The battery itself being very low Z, but obviously you can't have it right at the DAC so there's a length of track (wire) and then local cap or series of parallel caps next to DAC which all interact with each other.
@Mikett, if considering LT3045 for AVCC, I would suggest using it as shown here: ES9038Q2M Board
Last edited by a moderator:
Why did you remove the power supply channel 5V line that only supplies the display and commands?
it's not really necessary ,I just want to have one transfo ......😀
The Op amps is a long story
Last edited:
I am using a current boosted LT3042 which will not overheat with high currents.
It also sounds better when used with a 3 term regulator preregulating the input. I will be trying some tantalums as the C set. So far I have determined that film caps sound the best.
I am still not convinced about the 3042/5 but I am asking myself this question. Why not boost the op amp AVCC current capability with a high speed buffer?...like the headphone amp Markw4 is using? The other question is why not high speed video buffers that can drive literally anything stably?
It also sounds better when used with a 3 term regulator preregulating the input. I will be trying some tantalums as the C set. So far I have determined that film caps sound the best.
I am still not convinced about the 3042/5 but I am asking myself this question. Why not boost the op amp AVCC current capability with a high speed buffer?...like the headphone amp Markw4 is using? The other question is why not high speed video buffers that can drive literally anything stably?
Hi checked also these, but finally the lp5907 seemed to me the best choice. Can be installed directly on the panel. Ultra low noise. High psrr even at 100khz. Really reasonable price. See my former posts with pictures.
Okay, for example the one you mentioned here:
LT3045 DC-DC 2-20V to 3V 3.3V 5V 9V 12V Linear Regulator Power Board Low Noise | eBay
Cset which usually is 4.7uf, should be changed to 22uf to reduce LF noise. The LT3045 data sheet says to use high quality low leakage ceramic caps for that because any leakage current can affect regulation. Also, they say Cset and Rset should be connected to ground at right the ground terminal of Cout. Also, maybe good to remember that high-k dielectrics like most ceramics except NPO or C0G, are piezoelectric and can pick up acoustic noise from vibrations, and thus are not recommended for sensitive circuits.
But, okay, let's say one can find some suitable caps. A couple of pics of the module are attached below. Cset is circled in red. Doesn't look like much room in there to put 5 times the amount of capacitance.
So, how are you guys doing that anyway?
Attachments
I am also wondering if I should disconnect the power line to optical receiver as I do not use it. Wonder if it would make sense to tie the according DAC input to ground then or if it should be kept floating? Usually free inputs are recommanded to tie to ground...also in this case?
Anybody looked at LT6655? http://www.analog.com/media/en/technical-documentation/data-sheets/6655ff.pdf
There are some circuits in the applications section to boost the output current.
Also, they can be used with LT3045 as a buffer to make an ultra low 1/f noise 500mA current regulator. Very strange people just ignore that.
No need to disconnect it. Won't help anything.
Also, if we are talking about the Chinese ES9038Q2M board, the optical TOSLINK receiver interfaces to one of the GPIO pins. Those particular DAC pins can be software configured as outputs, so not a good idea to short them to ground, say, if there were any thought of something like that. It shouldn't hurt anything to disconnect the TOSLINK receiver power input, or even to short that to ground. But again, little or virtually no effect on power consumption, so no benefit from doing it. Won't affect DAC noise.
Last edited:
Thank you Mark. I am so happy with the sound of my board now I really wonder if it can be improved further - in fakt I really can live with it now - and love it.
Next mod is the clock change and rectifier Diode change to Hexfreds of my analog PSU that supplies all opamps for AVCC and for output stage.
It would be easy to change the voltage from +-15V to +-16V as I remember Benchmark does - did you already try this and does it improve something?
Next mod is the clock change and rectifier Diode change to Hexfreds of my analog PSU that supplies all opamps for AVCC and for output stage.
It would be easy to change the voltage from +-15V to +-16V as I remember Benchmark does - did you already try this and does it improve something?
The problem is fabrication always has been for me.
Lots of ideas and tinkering but making the darn circuits. My level of tinkering kind of levels off at modifying to a small degree prefab stuff nowadays.
Last edited:
@freezebox, Benchmark runs opamps at +-18vdc. To do that they use LME49860 opamps. Those opamps use the same die as LME49720, but somehow they are rated for higher voltages. Maybe they select them for leakage currents for voltage breakdown or something, I don't know, and I don't know who knows.
Anyway, the LME49720 data sheet says +-17 volts is maximum, and never exceed +-18v or damage may occur. If you have adjustable regulators you can just turn up the +-15v to +-17v assuming your transformer puts out enough voltage for the regulators to work at that output level.
However, I don't think there is much to be gained by increasing the +-15v in most cases. Benchmark does it at least in part to be able to run at professional line levels such as may be needed in mastering rooms and recording studios.
Regarding whether or not your DAC can be improved any more, unless you have done all the recommended mods then I would say it can be improved. Maybe it could be improved even with all the mods done, as it still won't sound quite as good as, say, a Benchmark.
Anyway, the LME49720 data sheet says +-17 volts is maximum, and never exceed +-18v or damage may occur. If you have adjustable regulators you can just turn up the +-15v to +-17v assuming your transformer puts out enough voltage for the regulators to work at that output level.
However, I don't think there is much to be gained by increasing the +-15v in most cases. Benchmark does it at least in part to be able to run at professional line levels such as may be needed in mastering rooms and recording studios.
Regarding whether or not your DAC can be improved any more, unless you have done all the recommended mods then I would say it can be improved. Maybe it could be improved even with all the mods done, as it still won't sound quite as good as, say, a Benchmark.
Last edited:
"AVCC is a purely analog supply. Low impedance is actually far more critical than absolute (low) noise. It is not feeding a clock or something where we have jitter effects to worry about."
The best choice is to provide a microprocessor off the DAC board, because there is a high noise travelling across the I2C lines. Providing its own power source yields for much lower noise, no doubt about, but for further improvement, we may isolate the I2C lines with an I2C isolator as well.
We want low noise and low Z. If considering an LT3045, better to have it with low Z and low noise because you will hear the noise as noise in the analog output, and the noise will be low frequency rumbling. Using low 1/f reference for the LT3045 should give much less audible noise, and output Z is still a function of LT3045 loop gain.
To be clear, I am not suggesting using any type of LT3045 based AVCC supply would be a good idea without testing and comparing with something like the ESS opamp AVCC circuit.
However, if the DAC is an ES9038PRO, an opamp is probably not going to be able to supply sufficient current to be able to use for comparison purposes. In that case maybe LifePo4 batteries with the shortest possible wires and whatever caps seem to work with it might be a good thing to compare active supplies with. Don't know, but would be inclined to try it.