All the smokescreens, mysteries and "NDAs" cannot hide the fact that these suggested "improvements" have been verified by subjective sighted evaluation only. Unless controlled listening tests or measurements are performed the results will be as useless as in the multitude of similar minded threads here. I prefer my own subjective view and have never found any claimed improvements or suggestions based on audiophile dogmas worthwhile.
Some things are harder than others to measure. Some things have not been studied in terms of human perception. For example, there is no threshold of audibility of sound stage. There is no standard measurement for it either. Thus you can have a well equipped lab and still have to do some listening tests. Whether or not you decide publish, and or whether or not skeptics decide they believe you, there isn't much else you can at some point. Since this is for human entertainment, ultimately the customers will decide if you chose well or not. The 'customers' in this diy forum are those who spend they money and time building dacs and or other audio gear. I simply want to know if there is any perceptual consensus among those who try the mods I have suggested so far. I will assume everyone is going to have their own opinion and not parrot whatever the first guy says.
The above having been said, it does matter somewhat the order in which mods are performed. Isolation of power supplies and of USB is the first thing to do. So is bypassing the last stage of the filter board. At that point further mods should be less masked by junk.
The above having been said, it does matter somewhat the order in which mods are performed. Isolation of power supplies and of USB is the first thing to do. So is bypassing the last stage of the filter board. At that point further mods should be less masked by junk.
Last edited:
E.g. making modifications without ensuring that performance is not impacted. Ears are not analyzers.
dog·ma
/ˈdôɡmə/
noun
That word describes you more than it describes me. You are the believer that the FFT analyzer tells everything worth knowing, which you call "performance"
/ˈdôɡmə/
noun
- a principle or set of principles laid down by an authority as incontrovertibly true.
"the rejection of political dogma"
That word describes you more than it describes me. You are the believer that the FFT analyzer tells everything worth knowing, which you call "performance"
No. But I'm a firm believer that your unsubstantiated subjective claims are not worth knowing.
In the process… TBA
Some context: ~DAC designers here have moved away from Shunts to series types. e.g. TPA Tridents shunts to Trident SR (using ADM7151). Now DDDAC (Doede) new version eschewing the original Tentlabs shunts for something similar, some custom series types- again indicating better performance.
We will see how the Nazar type turns out…
@MarcelvdG, thanks a lot for the concept circuit. I will get back to you later if I need any help
Shunt or series regulators, either way they have to be tuned/matched to the load. They can be designed well or poorly, given the load. For example, for the new Andrea Mori Sonic Empire DSD dac prototype they tested in the range of 10 - 20 different shunt designs in an effort to find the optimum one for their needs. Soon, I hope to have one here to see for myself. Maybe by the end of the week, but then I have to set it up. And this is the same type of thing I hear from other professional audio designers. Shunts are not a given for good sound. They have to be designed to work optimally with the load.
Last edited:
Couldn’t agree more, interesting to see progress here!
Btw, also those who have gone for batteries or these so called UltraCaps are beginning to have doubts. By the time it reaches the the target over long wires and connectors, the intrinsic benefits possibly all negated
Btw, also those who have gone for batteries or these so called UltraCaps are beginning to have doubts. By the time it reaches the the target over long wires and connectors, the intrinsic benefits possibly all negated
Please keep us updated Acko
I’ve tried shunt and series regs for various dacs including twisted Pear buffalo’s, so far I’ve always preferred the shunts . I did not like TPA Tridents so use alternative ones .
I’m currently using shunts with very high bandwidth error amps for the +5v , +/-15v FIRDAC board and +/-15v for the active filter board .
I’ve tried shunt and series regs for various dacs including twisted Pear buffalo’s, so far I’ve always preferred the shunts . I did not like TPA Tridents so use alternative ones .
I’m currently using shunts with very high bandwidth error amps for the +5v , +/-15v FIRDAC board and +/-15v for the active filter board .
Looks like you have already winnowed regs for the best possible outcome but out of curiosity have you compared the shunts with typical series regs especially for the Vref side of this DAC? My understanding is that Marcel’s on-board reg performance dominate and anything before is like a pre-reg only?
Attachments
Hi Acko , I’ve only used the shunts before the onboard vref regs for marcels dac .It would be interesting to try them direct to the vref if possible. With TPA Buffalo I tried the series and shunts to feed the AVCC pins direct , both types of regs sounded great but gave a little different result . I found the bass from shunts was more solid , series was slightly warmer . I tried various other regs too but not the LT3045 yet , those are supposed to be very good . Probably be cheaper and easier than the shunts to try
Last edited:
A lot of people work on "fashion", meaning something new is needed.
One big advantage of a shunt regulator is that current loops are clearly defined, at least for anything downstream from the shunt. If the shunt is then made fast to match up with a clean handover to a capacitor which handles the RF loop we pretty much know exactly where current will flow and in what frequency range.
Additionally, the net current draw is constant.
The downside is a lot of waste heat, no ready off the shelf chips (other than 431) for this job and of course, it's no longer fashionable.
We have seen "Jung Super Regulators" be fashionable, with extreme objective specification. Then we got shunts, then we got push-pull regs that never really got fashionable.
In reality we need to consider what regulator principle delivers the correct results to match the requirements of the circuit supplied. That of course means we must understand how the circuit actually behaves and what matters to that circuit.
There is no "universally good" option.
Some circuits for example are most sensitive to supply impedance, others to noise, still others need a regulator that can not only source current but sink it too. Some circuits tend to be sensitive to dynamic rail impedance and so on.
If we want to make a "almost ideal" regulator, it needs minimal noise, very low output impedance until handover to output capacitors, it needs to sink and source significant current as well as having very good load-step behaviour. That is quite an engineering challenge. And worse, most circuits will likely only need one or two of this grab bag list of features.
Thor
I'm aware that measurements are now out of fashion in this thread but I've noticed that "stock" filter has somewhat elevated distortion at 10kHz.
This is stock filter (without 3rd stage).
With a simple modification I could improve this by almost 20dB.
Since I'm under NDA I may be able to share what I find with Marcel, but probably can't share details publicly. However anybody with a pair of ears should easily find out the solution just by listening.
This is stock filter (without 3rd stage).
With a simple modification I could improve this by almost 20dB.
Since I'm under NDA I may be able to share what I find with Marcel, but probably can't share details publicly. However anybody with a pair of ears should easily find out the solution just by listening.
Last edited: