Except for the opamp, does the 15.0.15 V line powers anything else?
Analogic +5v for AK4396
Analogic +3.3v for the CS8416
To improve the sound replace the LM1117-5v (Analogic +5v AK4396) by a SALAS SHUNT REG ........
Serge
You're welcome
1500pF? AK4393 or AK4396?
In this way you altered the LPF quite a bit, it's not a surprise that it sounds not so good...
I would search the raspiness in other places...
Exactly which parts are you using (capacitors and resistors)?
They can be problematic... do a favour to yourself and order from HiFiCollective some audiograde caps, shipping from UK is pretty low.
The + 15V is the source for DAC's analog 5V
I've done a Google Documents' folder with the last BOM, schematic, LPF schematic and signal and power paths and numbering.
Give it a look:
https://docs.google.com/open?id=0B-qKRBuaKtzYYWNkNzRmODktZTQyNS00YjA2LWIwNzItN2UzNzhjODI5MWRm
To improve the sound replace the LM1117-5v (Analogic +5v AK4396) by a SALAS SHUNT REG ........
Sure it will...
In the next weeks I'll post a new shunt PS schematic, simpler than Salas one, which gave me great results with the My_Ref.
Quote: Originally Posted by jimchan
The last change i have made was C40-43, 1500pF Wimas. Although this solved the raspiness at the treble, it also took away a lot of the realism and attach of the piano and other vocals.
1500pF? AK4393 or AK4396?
In this way you altered the LPF quite a bit, it's not a surprise that it sounds not so good... I would search the raspiness in other places... Exactly which parts are you using (capacitors and resistors)?
AK4396, 4562NA opamp, most other caps changed as advised, not touched C46,C47, not touched the resistors yet.
Changed C40-43 down to 100 pF now. Again nice treble, raspiness, but lose the bass.
I guess I need to use exactly 1000pF in this position, as you recommend (?) I have an MKP10 with this value I can try...(only have 1500pF FKP2 in stock ;-)
4396....4562NA....Silmic II decoupling....havn't touched any resistors yet.
The last change i have made was C40-43, 1500pF Wimas. Although this solved the raspiness at the treble, it also took away a lot of the realism and attach of the piano and other vocals.
1500pF? AK4393 or AK4396?
In this way you altered the LPF quite a bit, it's not a surprise that it sounds not so good... I would search the raspiness in other places... Exactly which parts are you using (capacitors and resistors)?
AK4396, 4562NA opamp, most other caps changed as advised, not touched C46,C47, not touched the resistors yet.
Changed C40-43 down to 100 pF now. Again nice treble, raspiness, but lose the bass.
I guess I need to use exactly 1000pF in this position, as you recommend (?) I have an MKP10 with this value I can try...(only have 1500pF FKP2 in stock ;-)
4396....4562NA....Silmic II decoupling....havn't touched any resistors yet.
Hey,
i've just taken a "couple" of macros of the board (warning, i've left them full-sized so you could zoom in and see a bit more, maybe):
dac-debug
Please don't laugh at my sometimes a little beginner~ish soldering. I've simply tried to stuff the most i could on the top of the board, maybe that was a mistake...
Anyways, i'd be really grateful if you could have a look at the board. Maybe i've done something very stupid wrong...
Also, i've noticed that receiving 96k as opposed to 44.1k audio seems to put a additional 4*C on the AK and ~7*C on the 7812.
@Fremen: I'm so grateful for all your comments and help, but still i somehow don't believe that my 1117-5v cap is the reason for this. I've put some el-cheapo cap in there which changed nothing. The el-cheapo *certainly* has enough ESR, if that should ever have been the problem. For my humble understand this rules the cap out, doesn't it? I'd rather suspect that maybe i've done something at the output of the AK wrong and it oscillates there (but nothing is audible...!) - also i've looked over the board *so* often trying to see whether everything was done correctly ;/.
Thanks again, all of you for your help. I'm really really counting on your further help with this! Thanks so much!
Regards
- Dario
i've just taken a "couple" of macros of the board (warning, i've left them full-sized so you could zoom in and see a bit more, maybe):
dac-debug
Please don't laugh at my sometimes a little beginner~ish soldering. I've simply tried to stuff the most i could on the top of the board, maybe that was a mistake...
Anyways, i'd be really grateful if you could have a look at the board. Maybe i've done something very stupid wrong...
Also, i've noticed that receiving 96k as opposed to 44.1k audio seems to put a additional 4*C on the AK and ~7*C on the 7812.
@Fremen: I'm so grateful for all your comments and help, but still i somehow don't believe that my 1117-5v cap is the reason for this. I've put some el-cheapo cap in there which changed nothing. The el-cheapo *certainly* has enough ESR, if that should ever have been the problem. For my humble understand this rules the cap out, doesn't it? I'd rather suspect that maybe i've done something at the output of the AK wrong and it oscillates there (but nothing is audible...!) - also i've looked over the board *so* often trying to see whether everything was done correctly ;/.
Thanks again, all of you for your help. I'm really really counting on your further help with this! Thanks so much!
Regards
- Dario
Last edited:
OK, two problems:
- You're still using the AK4393 LPF with the AK4396 so high frequencies are exacerbated, often resulting in harshness
- The LM4562 is overly precise and detailed on high freqs so every minimum problem with them is thrown in your face.
If, after that, highs are not as refined as you like you should also change the opamp.
OPA827 is as precise as LM4562 but it hardly get harsh.
Last edited:
Hi any veiws on this for an output stage please.
http://www.ebay.co.uk/itm/Dual-channels-HDAM-High-Definition-Analog-Module-Audio-Discrete-Opamp-Module-/260940042523?pt=LH_DefaultDomain_0&hash=item3cc13d111b#ht_700wt_1037
http://www.ebay.co.uk/itm/Dual-channels-HDAM-High-Definition-Analog-Module-Audio-Discrete-Opamp-Module-/260940042523?pt=LH_DefaultDomain_0&hash=item3cc13d111b#ht_700wt_1037
You need to clean off all that flux residue on your PCB. Use some isopropyl alcohol and a soft toothbrush to clean both sides of the PCB. Then you can inspect the PCB to check for dry joints.
Just from looking at the photos I am sure that you have a few dry joints there and some joints look as if the solder is missing.
Carefully re-solder all the suspect joints, especially around the regulators and PSU, but be careful not to leave the iron on the joint for more than a few seconds. The clean the PCB again.
I don't know whether a dry joint can lead to the heating you describe, but I have just had a single dry joint in the 'capacitance multiplier' PSU of my MOSFET amp which caused it to refuse to regulate.
Also, make sure there is no contact between any of the solder pads of the regulator pins. Clean the flux off and run the tip of a screwdriver between the pins to clear any possible short.
Have fun.
Harshness from a DAC can be caused by a poor source, but can also be worsened by a bad connection from RCA cable to the DAC board. The cable from the RCA to the board should be twisted evenly. Two strands of CAT5 cable neatly removed from the casing, and left twisted with their original twist, provide a reasonable connection from the RCA to the board. My friend has the AK4396 with the default (AK4393) output filter capacitors, and I don't notice any particular issue with hardness/harshness or sibilance. I suspect your source, or SPDIF termination to the board to be the issue.
You may wish to consider using a transformer to isolate the SPDIF.
You may wish to consider using a transformer to isolate the SPDIF.
Last edited:
Interesting
Did you take in account the peculiar LPF needed by AK4396 (fc=125kHz, Q=0.692)?
So, probably, FCs are not causing overheat but still, IMHO, is safer to remove them after LM1117s.
Hi Erin,
harshness could also be caused (exacerbated) by an unbalanced timbre and the AK4393 LPF, when used with AK4396, is tonally unbalanced.
Also the LM4562 often can get harsh.
Look at his first pic he is has an ak4393
One of the Caps de-coupling the Op-Amp is wrong polarity
Last edited:
My concern with it would be that there is no way to bias it to remove dc offset. It needs a trimmer pot-
Next to AR2s buffer with LL1674s trafos in 4 + 1 mode (which I'm in the process of getting PCBs made for) this is the best output I've seen yet.
I think you're talking about Nebuk pictures and yes one of the opamp caps is reversed.
But my post was for jimchan...
I think you're talking about Nebuk pictures and yes one of the opamp caps is reversed.
But my post was for jimchan...
My apologies, this thread is getting difficult to follow
