From my experience one should never by-pass coupling caps with smaller ones as this leads to two different time constants in the audio band and where they overlap the sound becomes smeared.
One high quality polyprop is the best if a cap must be used large enough to get down to 1hz. But the best cap is no cap if you can control the dc offset.
Cheers George
WRT time constants, I don't think this is the case George.
I think it is more that they can form a HF resonant circuit that has a peak
somewhere up there in HF land.
Regardless, agreed usually not a good idea.
Maybe someone should try series resistive dampers and see if that helps.
Yes... I can hear the echo.... the only good cap is no caaaaaap
L (and some R I suppose) filter...
http://www.diyaudio.com/forums/group-buys/216669-joachim-gerhard-filter-buffer-es9022.html
http://www.diyaudio.com/forums/group-buys/216669-joachim-gerhard-filter-buffer-es9022.html
Oppo will come out soon with an HA-1 Headphone Amplifier based on asynchronous USB with the same ES9018.
This device it will be interesting to mod it! I think more interesting and quite simple to deal with, than the player...
Waiting for the details and pictures first... and then open a thread for it...
BTW, Happy Christmas!
This device it will be interesting to mod it! I think more interesting and quite simple to deal with, than the player...
Waiting for the details and pictures first... and then open a thread for it...
BTW, Happy Christmas!
Last edited:
EMI absorbing material
Does anyone have experience using the 3M EMI absorbing material such as EMI Absorber AB5030S? This material will work down to 50 MHz. A slightly less expensive version which works down to 200 MHz is the 5000 series. Apparently you can place it over noisy chips such as DAC's, or on a flat part of the chassis. Apparently these materials can be used in cell phones for example to isolate the antenna from noisy chips.
From the datasheet it appears to be much more effective in transmission than in reflection.
There is a datasheet here For the 5000 series:
http://multimedia.3m.com/mws/mediawebserver?66666UuZjcFSLXTtmX&tnX&aEVuQEcuZgVs6EVs6E666666--
Does anyone have experience using the 3M EMI absorbing material such as EMI Absorber AB5030S? This material will work down to 50 MHz. A slightly less expensive version which works down to 200 MHz is the 5000 series. Apparently you can place it over noisy chips such as DAC's, or on a flat part of the chassis. Apparently these materials can be used in cell phones for example to isolate the antenna from noisy chips.
From the datasheet it appears to be much more effective in transmission than in reflection.
There is a datasheet here For the 5000 series:
http://multimedia.3m.com/mws/mediawebserver?66666UuZjcFSLXTtmX&tnX&aEVuQEcuZgVs6EVs6E666666--
These look very appealing:
EMI Absorber
3M
AB5000 series EMI absorbers consists of flexible soft metal flakes filler in polymer resin with acrylic pressure sensitive adhesive.3M EMI absorbers are near field absorbers that are specifically designed to eliminate noise from a few hundred megahertz to multiple gigahertz frequencies. They work by absorbing the EMI field generated by today's high speed electronics, eliminating emissions, crosstalk, and oscillations. They are flexible polymer sheets that are loaded with specific magnetic flake particles and are nonconductive, so they can be directly applied to circuitry and noisy traces without fear of shorting out traces and IC pins.
EMI Absorber
3M
AB5000 series EMI absorbers consists of flexible soft metal flakes filler in polymer resin with acrylic pressure sensitive adhesive.3M EMI absorbers are near field absorbers that are specifically designed to eliminate noise from a few hundred megahertz to multiple gigahertz frequencies. They work by absorbing the EMI field generated by today's high speed electronics, eliminating emissions, crosstalk, and oscillations. They are flexible polymer sheets that are loaded with specific magnetic flake particles and are nonconductive, so they can be directly applied to circuitry and noisy traces without fear of shorting out traces and IC pins.
Yes, have done it already for a while ago... To shield the DAC chip... using a different hi-tech material. Not very sure yet about improvements, to be honest. But anyway it may not have negative impact using such. It may be beneficial in conjunction with some another improvements in existing circuits...
Shielding the DAC chip is in my opinion, something very reasonable one may want to do anyway.
Doing some reasonable small improvements in a system/device, or take some precautions, even one may not observe at once a dramatic result in functioning of that device, the amount of all these it will lead to a general improvement, which it may be observed as dramatic in the end. This is my principal way of thinking...
Shielding the DAC chip is in my opinion, something very reasonable one may want to do anyway.
Doing some reasonable small improvements in a system/device, or take some precautions, even one may not observe at once a dramatic result in functioning of that device, the amount of all these it will lead to a general improvement, which it may be observed as dramatic in the end. This is my principal way of thinking...
Last edited:
Yes... I can hear the echo.... the only good cap is no caaaaaap
Caaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaap!
The never-ending-echo.
But some caps do sound nice... oh oh, they might think I am a heretic.
Cheers, Joe
.
JOE, none of them sound better than no coupling cap. if your circuit is thought out well enough to eliminate it.
Last edited:
Right in this big "dilemma" (caps - no caps), I think to give you the link for the new modding object/device, Oppo`s Headphone Amplifier (USB DAC):
http://www.diyaudio.com/forums/digi...r-discussions-upgrading-mods.html#post3752456
http://www.diyaudio.com/forums/digi...r-discussions-upgrading-mods.html#post3752456
Well, I have tried the trick with the cap between output phases of a DAC.
I have chosen almost by chance a 100pF film cap. I was too lazy to measure something. Just solder the caps and start the playback. And, voila! There is improvement.
Sorry for out of thread, but I did the experiment on a PCM1792 DAC system (which I`m working on right now...). BTW, I never head before a such high quality sound from a ES9018 system, as I could get out of this old now 1792 chip. But I will implement quite soon this my post DAC processing circuit on ES9018... It should be anyway even better...
I am not "a cap in the signal path" supporter, but in this case the cap it works in a very positive way. Thanks Joe for sharing the trick.
If I should use 100pF to parallel the I/V resistor, then it should be in my opinion a catastrophe (perceptual) muffed audio signal, for not talking more about the sound stage quality... But this 100pF cap used in between phases it bring a real improvement. How it may be the results using another cap value, I can not say for the moment, but I think to try it... I may come back with details when I will bother to do some measurements...
In my perception, this trick improve quite substantial the sound stage depth (3D to say so...), and it bring it all the stage a step more in front of the speakers, in the room. The speakers become even more transparent. I can clearly hear two instruments or vocal on the same axe, as one is more back than the other. There is obviously that the sound stage is improved. Also the low end spectre is improved too, as Joe has observed.
This it works!
Else, I highly recommend using of the TPS7A49 regulators with film cap on NR pin, for DAC chip power rails (and yes, I still use large decoupling capacities, and no any other filtering in signal path...). I could observe (in my setup) a quite dramatic improvement for the DAC system outputted audio signal, when I got down from a 12µV noise regulators (to power the DAC chip) to this 4µV noise TPS7A49.
I have chosen almost by chance a 100pF film cap. I was too lazy to measure something. Just solder the caps and start the playback. And, voila! There is improvement.
Sorry for out of thread, but I did the experiment on a PCM1792 DAC system (which I`m working on right now...). BTW, I never head before a such high quality sound from a ES9018 system, as I could get out of this old now 1792 chip. But I will implement quite soon this my post DAC processing circuit on ES9018... It should be anyway even better...
I am not "a cap in the signal path" supporter, but in this case the cap it works in a very positive way. Thanks Joe for sharing the trick.
If I should use 100pF to parallel the I/V resistor, then it should be in my opinion a catastrophe (perceptual) muffed audio signal, for not talking more about the sound stage quality... But this 100pF cap used in between phases it bring a real improvement. How it may be the results using another cap value, I can not say for the moment, but I think to try it... I may come back with details when I will bother to do some measurements...
In my perception, this trick improve quite substantial the sound stage depth (3D to say so...), and it bring it all the stage a step more in front of the speakers, in the room. The speakers become even more transparent. I can clearly hear two instruments or vocal on the same axe, as one is more back than the other. There is obviously that the sound stage is improved. Also the low end spectre is improved too, as Joe has observed.
This it works!
Else, I highly recommend using of the TPS7A49 regulators with film cap on NR pin, for DAC chip power rails (and yes, I still use large decoupling capacities, and no any other filtering in signal path...). I could observe (in my setup) a quite dramatic improvement for the DAC system outputted audio signal, when I got down from a 12µV noise regulators (to power the DAC chip) to this 4µV noise TPS7A49.
Last edited:
George, I just KNEW that would flush you out.
No cap in my feather or no feather in my cap?
Cheers, Joe
I think we may clarify a little bit...
A coupling cap is not one may really want between stages, when is about higher/highest signal quality transfer. A coupling cap is a compromise, specially when a separation AC/DC is necessary. I fully agree that none sound better than no coupling cap.
This is about a coupling cap.
When about this trick to use a cap between the channel`s phases of a DAC there is not about coupling something. Is about filtering... Or it is to "coupling" in between the two phases
At least, what for using a coupling cap in this circuit (when use of a such filtering cap), if the output offset is very low, and no separation AC/DC is necessary?
So, I think these two things are different as definition and area of use.
A coupling cap is not one may really want between stages, when is about higher/highest signal quality transfer. A coupling cap is a compromise, specially when a separation AC/DC is necessary. I fully agree that none sound better than no coupling cap.
This is about a coupling cap.
When about this trick to use a cap between the channel`s phases of a DAC there is not about coupling something. Is about filtering... Or it is to "coupling" in between the two phases
At least, what for using a coupling cap in this circuit (when use of a such filtering cap), if the output offset is very low, and no separation AC/DC is necessary?
So, I think these two things are different as definition and area of use.
Last edited:
Filtering by way of cancellation.
But you won't get the full benefit until you check what is does @ 20KHz relative to 1KHz and aim to be down 0.85x or -1.3dB to -1.5dB. Or tune by ear.
I just received this after I did to a '105 that I upgraded a few months ago and a few days ago did above.
"The Oppo is amazingly wonderful I did not think it could get much better than it was..., you would be crazy not to do this upgrade if you have [one of your] Level 3 players - it’s a no brainier!"
Cheers, Joe
Last edited:
OK, I took a closer look at this filtering way. As I used for my experiences another type DAC than the subject of this thread is, I think to not extend too much the details.
But I have observed some strange things. First the HF outputted noises (for the 1792 chip these are quite high and well defined) are undisturbed by this filter. I have increased the (film) cap value from 100pF to 500pF, and then to 1nF. I have used both 1Khz and 20Khz wav files for testing. I couldn`t affect the 20Khz signal to get it lower (-1,3dB) even with a 1nF cap placed between phases. It may be the value of my cap yet too low? It could be nice if Joe can tell us something about a value range for the caps to be used in this trick.
I could see that the taken FFT get worse as the cap value increase. More harmonics are to be seen...
But on the top of all these "beads", the sound get better and better as the cap value increase (in my experiments up to 1nF)... I do not understand. Or is me who do something not right in all this...
It could be both nice and usefully if many other members here will take a closer look on this, will do some experiments and measurements, and then discuss this quite unusual way of filtering, which actually improve the sound out of a DAC.
But I have observed some strange things. First the HF outputted noises (for the 1792 chip these are quite high and well defined) are undisturbed by this filter. I have increased the (film) cap value from 100pF to 500pF, and then to 1nF. I have used both 1Khz and 20Khz wav files for testing. I couldn`t affect the 20Khz signal to get it lower (-1,3dB) even with a 1nF cap placed between phases. It may be the value of my cap yet too low? It could be nice if Joe can tell us something about a value range for the caps to be used in this trick.
I could see that the taken FFT get worse as the cap value increase. More harmonics are to be seen...
But on the top of all these "beads", the sound get better and better as the cap value increase (in my experiments up to 1nF)... I do not understand. Or is me who do something not right in all this...
It could be both nice and usefully if many other members here will take a closer look on this, will do some experiments and measurements, and then discuss this quite unusual way of filtering, which actually improve the sound out of a DAC.
Last edited:
The value will depend on a number of things and to get around -1.3dB or slightly more, the cap value actually becomes quite critical. In the case of my Level 3, using 0.33 uF or 0.39uF, the response relative to 1KHz @ 20KHz is less than half a dB down. Double that, and 2 x 0.39uF = 0.78uF makes a big difference and the response @ 20KHz drops dramatically.
Sonically, going from 0.39uF to 0.78uF makes a very obvious difference. So you get to a certain point where improvements are incremental and then suddenly the value increase makes a big difference, a kind of 'knee' is reached and you have to adjust that 'knee' more critically - and when you know you have found the 'knee' then by all means tune it by ear, but you will be more than 1dB down from the experiences so far. That is where it really makes a difference.
BTW, this seems to work with all kinds of delta-sigma DACs, even whether they are "voltage" of "current" types. But with "voltage" types I recommend a small value series resistors on each phase, say 15R to 47R, and then a cap across after those resistors. The reason is that "voltage" DAC can have quite low output Z and the series Rs makes the cap value lower and easier to deal with.
In my case, with Sabre DAC, the termination impedance is 3.3R per phase, and the series R is already 390R per phase (two paralled phases) so a large value cap is required, mainly because of the very low termination Z (the 390R and 3.3R are effectively in parallel and totally dominated by 3.3R)
If no termination impedance is seen (as when Sabre DAC is used in "voltage" mode and it then has a large source Z, 780R per phase, 390R for two phases and 195R for four phases), then the cap value can be a lot smaller. But 1nF is still far near the required value to get that negative response down -1.3dB at 20KHz.
Everybody who has heard it done, the ones that I have done here and one more done by David in KL for a friend, have been totally convinced about the result.
So I am rather keen that you do it as well.
I you are using the Sabre DAC into a Virtual Earth, then I have yet had to deal with that. This means that the feedback/opamp is holding it to a theoretical zero Ohm and hence the cap value of infinite size would make no difference, in theory. Then I would recommend a pair of 3.3R in series with the output, but this time, unlike the "voltage" DAC above, put the cap before the two series 3.3R resistors. Then the predicted value would be around 0.68uF to 0.82uF.
Hope that helps.
Cheers, Joe
.