that would be why I disagreed and explained in the first post (above^^) and then again in the last one. yes I get uptight about your assertions and tutorials on how not to do things, its not helpful, but they do wonders for your mod guru status, which makes the rules of the natural world not apply to you.
when you post so full of expert advise and then do the above, sorry I cant help a little haha, your questions and photo illustrated further that you didnt really know what the HF bypasses might be needed for in the first place .. large factor; they are all connected to the same ground plane (as they should be). I said nothing about arbitrary 100nf… I wonder why using larger bulk capacitance and no small decoupling is less a bandaid? a correctly calculated and with efforts to improve the power supply, just the small cap should be needed, but it should be closer.
I have no need to hold court here as you seem to, but I was saddened to see another set of instructions and foresaw it turning into yet another one of these threads. i'll leave you to it
no, they arent; certainly not how youve applied them (like a dipole aerial) and no, its not. again, inductance and physical size is the enemy of HF decoupling. if talking about HF decoupling, your sockets will render anything you put there ineffective.
so good luck guys, looking forward to another BOM that costs potentially many times the price of the board.
I guess we have a different definition of respect ...
Last edited:
one last thing, the random value caps with random value caps comment wasnt actually meant to describe you, just describing what is most likely to cause resonance, rather than your sweeping statement of all bypass caps 'sucking the life out of the sound' (a technical term). it seems you thought it applied to you.
Guru status?
I'm not a guru and I don't need or want to.
What I do is simply sharing my findings and trying to help others with my (limited) experience.
Maybe, if you don't consider kind and necessary to establish a communication link based on mutual recognition...
Describing me? Don't worry, I know you were talking about the method...
BTW, never said that bypasses 'suck the life out of the sound', this is an example of distortion...
The datasheet recommends both.
When you see that all bypasses have the same standardized value (100nF) those caps are band aid, IMHO, certainly they were not calculated for in circuit use...
I suppose that if you have small enough PS output impendance and short enough traces you can avoid decoupling at all.
But those choice have to be made on design phase, here we're talking about a ready made design and if we can improve/tune it with minor mods and/or different parts.
No, no, no... You love that word, isn't it?
BTW your approach seem to me a bit 'digital', or black or white if you want.
Reading your comments seems you consider correct only one decoupling approach and all the rest is ineffective/wrong.
You already stated that you don't like this design since you don't like several aspects of it, fine, we got it.
Me too I would have made several things differently but the design of an entire DAC is over my knowledge.
So socketed ICs can't be decoupled then?
Last edited:
Clave, interesting that you pick and choose what you can use from the datasheet only when it backs up your argument. there has been much talk over the years about the arbitrary recommendation for placement of 100uf and 100nf caps, its not a great idea either, usually the application notes will give a bit more information, but you still have to use your brain.
there have been articles at linear tech and Ti to cover this in detail and regulator application notes from linear and Ti are pretty decent too
far from digital approach, I look at the board and the components not for the simple sticker value as you seem to, which seems a bit digital to me. each component and its placement on the board is a lumped element, containing resistance, capacitance and inductance, whether it is a resistor, a capacitor, or an inductor. these factors are impacted again by board layout, ground and power planes, trace thickness, substrate thickness and frequency. impedance is a complex term, that is nothing like DC resistance.
they interact with other parts on the board and both common mode and differential mode impedance. the source and load impedance of individual parts like dacs, can usually only be found experimentally unless you are buying a heap from them and you can perhaps find out, but they impact on the effectiveness of the filters.
I recommend you read some Henry Ott and Howard Johnson, you may find it enlightening
not effectively ...
luckily most opamps have pretty good PSRR, so you wont notice your error so much. dac analogue supply pins on the other hand, often have effectively zero psrr
there have been articles at linear tech and Ti to cover this in detail and regulator application notes from linear and Ti are pretty decent too
far from digital approach, I look at the board and the components not for the simple sticker value as you seem to, which seems a bit digital to me. each component and its placement on the board is a lumped element, containing resistance, capacitance and inductance, whether it is a resistor, a capacitor, or an inductor. these factors are impacted again by board layout, ground and power planes, trace thickness, substrate thickness and frequency. impedance is a complex term, that is nothing like DC resistance.
they interact with other parts on the board and both common mode and differential mode impedance. the source and load impedance of individual parts like dacs, can usually only be found experimentally unless you are buying a heap from them and you can perhaps find out, but they impact on the effectiveness of the filters.
I recommend you read some Henry Ott and Howard Johnson, you may find it enlightening
not effectively ...
luckily most opamps have pretty good PSRR, so you wont notice your error so much. dac analogue supply pins on the other hand, often have effectively zero psrr
Last edited:
when asked questions that can be answered with a yes, or no, I use yes, or no when appropriate.
at the risk of repeating myself AGAIN, the main enemy of decoupling caps is inductance. of the lumped elements it is inductance that dominates at HF and lowers the effectiveness (raises the impedance) of caps of all types. adapter pins add inductance, PCB traces add inductance, vias add inductance, capacitor leads and capacitor construction add inductance.
at the risk of repeating myself AGAIN, the main enemy of decoupling caps is inductance. of the lumped elements it is inductance that dominates at HF and lowers the effectiveness (raises the impedance) of caps of all types. adapter pins add inductance, PCB traces add inductance, vias add inductance, capacitor leads and capacitor construction add inductance.
Just a quick note of encouragement to Mr. Inserra. I built the stock Mini DAC and upgraded to the published better BOM. It worked and I liked what I heard. I also worked along side with the "trial and try another" process used for tuning the Fremen Edition power amp. At some points I had 14+ sockets on each PCB. Their usefulness was in the ability to hear differences in parts selection and combinations. There was never an assumption that those sockets were a "best practice" approach - rather a tool of convenience that reduced soldering/desoldering. When the sockets were removed, there was a general audible improvement that reflects the negatives of their use as mentioned earlier. Their inclusion and expected negative influences was openly and honestly discussed over several months and on more than one thread. The process concluded with "One Sweet Sounding Power Amp" - sockets and all.
Many of us wish we had the knowledge, equipment and background to be able predict and guarantee the performance of items included in a circuit. Lacking those, what can be heard is what we must rely on. It's worked out quite well so far, so I'm comfortable adopting Dario's (and other's) "suggestions" as a path to finding what sounds best in my system and to my preferences. The addition of as much practical technical information as possible can only enhance the process.
I sincerely hope this thread can turn the corner away from personalized comments and zero in on those useful mods and additions that we can try for ourselves, in our own settings, in the search for "AUDIO NIRVANA"
Many of us wish we had the knowledge, equipment and background to be able predict and guarantee the performance of items included in a circuit. Lacking those, what can be heard is what we must rely on. It's worked out quite well so far, so I'm comfortable adopting Dario's (and other's) "suggestions" as a path to finding what sounds best in my system and to my preferences. The addition of as much practical technical information as possible can only enhance the process.
I sincerely hope this thread can turn the corner away from personalized comments and zero in on those useful mods and additions that we can try for ourselves, in our own settings, in the search for "AUDIO NIRVANA"
Last edited:
Just a quick note of encouragement to Mr. Inserra. I built the stock Mini DAC and upgraded to the published better BOM. It worked and I liked what I heard. I also worked along side with the "trial and try another" process used for tuning the Fremen Edition power amp. At some points I had 14+ sockets on each PCB. Their usefulness was in the ability to hear differences in parts selection and combinations. There was never an assumption that those sockets were a "best practice" approach - rather a tool of convenience that reduced soldering/desoldering. When the sockets were removed, there was a general audible improvement that reflects the negatives of their use as mentioned earlier. Their inclusion and expected negative influences was openly and honestly discussed over several months and on more than one thread. The process concluded with "One Sweet Sounding Power Amp" - sockets and all.
Many of us wish we had the knowledge, equipment and background to be able predict and guarantee the performance of items included in a circuit. Lacking those, what can be heard is what we must rely on. It's worked out quite well so far, so I'm comfortable adopting Dario's (and other's) "suggestions" as a path to finding what sounds best in my system and to my preferences. The addition of as much practical technical information as possible can only enhance the process.
I sincerely hope this thread can turn the corner away from personalized comments and zero in on those useful mods and additions that we can try for ourselves, in our own settings, in the search for "AUDIO NIRVANA"
At last - some positive input!
I'm interested in using the AK4399, but not with op-amp output. I'd be looking to use either a transformer output or run the output directly into the grids of a tube stage. Can anyone help with examples of this, and a wiring diagram?
Here is one example of transformer output - just says "Lundahl amorphous transformer", no mention of which.
APL HiFi – DAC-S | DACinfo
http://www.aplhifi.com/index.php?op=DAC_S
Here is one example of transformer output - just says "Lundahl amorphous transformer", no mention of which.
APL HiFi – DAC-S | DACinfo
http://www.aplhifi.com/index.php?op=DAC_S
Last edited:
The DAC9 (as it arrives from vendor) sound pretty similar to the modded MINI2496, with OPA827 it sound a bit better.
Futher mods enlarge the difference but we're not talking of night and day ones.
eBay seller minishow0328 has a Asynchronous XMOS USB module compatible for AK4399 / DAC7 / ES9018, http://http://www.ebay.com/itm/Asynchronous-XMOS-USB-module-compatible-for-AK4399-DAC7-ES9018-by-Weiliang-/130956281202?pt=US_Amplifier_Parts_Components&hash=item1e7d9a4572 it looks like it only works with the AK4399 DAC that has 'software control' Assemble AK4399 AK4113 DAC Decoder 192K Software Control by Weiliang | eBay
I'm interested, but wouldnt mind some feedback before jumping in, anybody tried this one?
I'm interested, but wouldnt mind some feedback before jumping in, anybody tried this one?
This thread has gone a bit quiet? I'm still interested in purchasing a AK4399 DAC and have seen a new one featuring two AK4399 chips
http://http://www.85china.com/product/19427381978/Dual%20AK4399%20AK4113%20HIFI%20decoder%20soft%20control%20XMOS%20USB%20DAC%20DAC
Being a newbie, I'm not 100% sure why two chips would be used? I would assume one for each stereo channel, but what are the benefits of this? Stereo signal separation maybe?
http://http://www.85china.com/product/19427381978/Dual%20AK4399%20AK4113%20HIFI%20decoder%20soft%20control%20XMOS%20USB%20DAC%20DAC
Being a newbie, I'm not 100% sure why two chips would be used? I would assume one for each stereo channel, but what are the benefits of this? Stereo signal separation maybe?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.