Yes I take a measure with one probe between the diodes DW1/2 on either
pin 4 (-) or pin 7 (+). The trimmers controls either pin4 V- or pin7 V+.
Set them from +-12volt to +-15volt. Most Op Amps can handle 15 volt, but
please check it before you do the adjusting.
I have a advice here from a guy in Denmark. If you can, try to replace caps
C222 , from 2.2nf to 470pf , It gave me a huge improvement in the treble.
Highs is much cleaner now.
It's something with the filtering to do, and It's way to conservative the stock 2.2 nF.
/Michael.
Hi Michael,
I am curious how you wire without opamp. Just take out the opamp? What happen than to the differential output of the DAC?
I am collecting the materials to build a discrete output stage like the one of Rays (Ray's Audio Page)
see picture. Probably I need to increase the Fc to 70-100k.
What do you think?
Good advise to decrease the C222. I will try in the next days.
During my first checks, I putted a 1kHz (32 bit 192 k generator) on the dac. See attached the output. I was puzzled how to get this one better. Probably, reducing the filter will help. I will check in the next days.
Regards,
Olaf
I am curious how you wire without opamp. Just take out the opamp? What happen than to the differential output of the DAC?
I am collecting the materials to build a discrete output stage like the one of Rays (Ray's Audio Page)
see picture. Probably I need to increase the Fc to 70-100k.
What do you think?
Good advise to decrease the C222. I will try in the next days.
During my first checks, I putted a 1kHz (32 bit 192 k generator) on the dac. See attached the output. I was puzzled how to get this one better. Probably, reducing the filter will help. I will check in the next days.
Regards,
Olaf
Attachments
Another question,
My Dac 9, V12 (assembled by supplier). has different resistors. RR1 = 1k, RR2 = 3k. In all the ref schematics these are equal, being 1k5. See also pictures.
I wonder now, is this on purpose, or would it be an improvement to change.
What are your experiences?
Thanks,
Olaf
My Dac 9, V12 (assembled by supplier). has different resistors. RR1 = 1k, RR2 = 3k. In all the ref schematics these are equal, being 1k5. See also pictures.
I wonder now, is this on purpose, or would it be an improvement to change.
What are your experiences?
Thanks,
Olaf
Attachments
Hi Olaf..
Just take out the op Amps.. There could be some problems with Impedance
load from your preamp. It is not a problem in my Setup, with my 4P1L preamp.
The advice of altering things in the DAC , is not my expertice at all. I got the tips lowering the C222, from a Danish designer of discrete op amps.
/Michael
Hi Olaf
Give it a try. Haven't got the slightest Clue.
I have been a little around changing op Amps. There is quite some difference here.
Now even if it's WRONG , knows it can play without any op Amps. Maybe someone which the right knowledge, would like to try it with transformer output only.
CFP discrete output stage
Hello Olaf
I'm using Rays output stage with my AK4399 DAC of Weiliang, dual version. It's working great. I adjusted the resistors/caps of the passive filter to fc of 125kHz and with Bessel characteristic (for constant group delay). The switched capacitor filtering of the analog outputs of the AK4399 allows for a mild filtering, so I wouldn't concentrate too much on fc, just keep transfer function as flat as possible in audible range.
The class A stage of the DOS is very good, but it also needs a good power supply with good PSRR to perform at its best, so don't overlook that (as I did at first).
Regards,
Nino
Hello Olaf
I'm using Rays output stage with my AK4399 DAC of Weiliang, dual version. It's working great. I adjusted the resistors/caps of the passive filter to fc of 125kHz and with Bessel characteristic (for constant group delay). The switched capacitor filtering of the analog outputs of the AK4399 allows for a mild filtering, so I wouldn't concentrate too much on fc, just keep transfer function as flat as possible in audible range.
The class A stage of the DOS is very good, but it also needs a good power supply with good PSRR to perform at its best, so don't overlook that (as I did at first).
Regards,
Nino
Last edited:
Hi Olaf, most of that ringing is caused by the digital filter. Unless there's a slow roll-off function in the DAC, there's nothing you can do about it.
There does appear to be a little overshoot caused by the analog filter - it's probably a Butterworth function. Changing cap values so that it's closer to a Bessel function will help here, but you need a simulator or proper filter design software to calculate them.
Best decouplig AK4399
summing up around the chip for the coming members:
1. DVDD § BlackGate STD or Niobium AVX NOJ or Sanyo OSCON
21. VCMR § BlackGate STD or Niobium AVX NOJ or Sanyo OSCON
25. VDDR § Elna Silmics II Black/Gold
26. VREFHR § Elna Silmics II Black/Gold
30. VREFHL § Elna Silmics II Black/Gold
31. VDDL § Elna Silmics II Black/Gold
35. VCML § BlackGate STD or Niobium AVX NOJ or Sanyo OSCON
41. AVDD § Elna Silmics II Black/Gold
all 10uf as the AKM recomedation but without bypass.
thanx to ClaveFremen and others in the thread.
those niobium caps may be found here.
will appreciate any recommendation to complete the list for the future.
summing up around the chip for the coming members:
1. DVDD § BlackGate STD or Niobium AVX NOJ or Sanyo OSCON
21. VCMR § BlackGate STD or Niobium AVX NOJ or Sanyo OSCON
25. VDDR § Elna Silmics II Black/Gold
26. VREFHR § Elna Silmics II Black/Gold
30. VREFHL § Elna Silmics II Black/Gold
31. VDDL § Elna Silmics II Black/Gold
35. VCML § BlackGate STD or Niobium AVX NOJ or Sanyo OSCON
41. AVDD § Elna Silmics II Black/Gold
all 10uf as the AKM recomedation but without bypass.
thanx to ClaveFremen and others in the thread.
those niobium caps may be found here.
will appreciate any recommendation to complete the list for the future.
Last edited:
DOS-modified for AK 4399
Here my approach to the output stage of the Dac. The circuit is to a large extent a copy of Rays DOS circuit. The modification I did is based upon the approach: to first subtract the two differential signals, than filter.
To handle the +/- 2.5 V DAC output, I changed a bit the values of the differential part.
Also I modified the Bessel filter, Fc 100 kHz and placed this after the subtraction.
From square wave measurements, I can not to see a difference with the original/on board circuit. For me a bit frustrating, that I do not any measurement prove of the performance.
(Anyone has some suggestions on measurement techniques?)
However, during listening testing, it is very clear, that this DOS -substract-filter-approach is much more crispy-clear. Also the imaging is better.
By listening testing, I tested also a filter-than subtraction approach (original DOS). This resulted in less crispy-clear and not so good imaging. The mechanism behind this, I do not know.
For me (and the family) this is the output circuit I will keep.
Regards,
Olaf
Here my approach to the output stage of the Dac. The circuit is to a large extent a copy of Rays DOS circuit. The modification I did is based upon the approach: to first subtract the two differential signals, than filter.
To handle the +/- 2.5 V DAC output, I changed a bit the values of the differential part.
Also I modified the Bessel filter, Fc 100 kHz and placed this after the subtraction.
From square wave measurements, I can not to see a difference with the original/on board circuit. For me a bit frustrating, that I do not any measurement prove of the performance.
(Anyone has some suggestions on measurement techniques?)
However, during listening testing, it is very clear, that this DOS -substract-filter-approach is much more crispy-clear. Also the imaging is better.
By listening testing, I tested also a filter-than subtraction approach (original DOS). This resulted in less crispy-clear and not so good imaging. The mechanism behind this, I do not know.
For me (and the family) this is the output circuit I will keep.
Regards,
Olaf
Hi Olaf
Nice work! I like your approach, I once thought the same, so I'm glad you did this. I was afraid the circuit would become unstable by unfiltered input, but according to your experience it works very well with the AK4399. Nice! You reduced passive filter by 2, which is good in terms of component count also
One question: As this is an output stage, isn't the passive filter circuit's transmission being influenced by the input impedance of the following amplifier this way?
Regards
Nino
Nice work! I like your approach, I once thought the same, so I'm glad you did this. I was afraid the circuit would become unstable by unfiltered input, but according to your experience it works very well with the AK4399. Nice! You reduced passive filter by 2, which is good in terms of component count also
One question: As this is an output stage, isn't the passive filter circuit's transmission being influenced by the input impedance of the following amplifier this way?
Regards
Nino
Hi Nino,
To your question: Indeed, I assumed next stage impedance above 10k, than the effect is minimal. (I tested with step response on filter, gave a nice "Bessel" response)
In the meantime, I start testing with Fc up to 160 kHz, this gives a bit better result. Probably, the optimal is pending the other parts of the set-up. (as the next parts will filter also these frequencies). I will stick to the 160 kHz for the moment.
Regards,
Olaf
To your question: Indeed, I assumed next stage impedance above 10k, than the effect is minimal. (I tested with step response on filter, gave a nice "Bessel" response)
In the meantime, I start testing with Fc up to 160 kHz, this gives a bit better result. Probably, the optimal is pending the other parts of the set-up. (as the next parts will filter also these frequencies). I will stick to the 160 kHz for the moment.
Regards,
Olaf
To all of you guys that have tried different (single) opamps on the Weiliang DAC9, here is an eyeopener for you. If you want firm bass and neutral sound, the way to go is Linear Technology´s LT1122. Yes, it´s expensive, but here it´s a keeper. Beats anything else I´ve tried. You really should try this one out 
I find the LME49710 is great in the midrange and shy on the very top and bottom. If you want more bandwidth the LME49990 is much better. Many headphone fanatics use it and so do I. Its a great Op-amp for I/V. If you want something that might be better you can look at muses O2 (stupid $$$$) or building a NTD1. Im using active crossovers with a total of 10- LME499990's and 4- opa2134's so an NTD1 for I/V would be overkill. It's also fun to try different Op-amps but gets depressing after you've Implemented the LME49990 as the reference.
I'd like to hear what others think of the LME49990 in comparison.
I'd like to hear what others think of the LME49990 in comparison.
Yesterday, I have tested in the same time
OPA627BP (original from 2007 not from ebay)
LT1122
LME49710
THS4061
NE5534 (from Kit)
OPA827
My list is
1. OPA627BP - very good sound. Not enough air a little bit. Old reliable horse.
2. THS4061 - comfortable for listening. Symphonic orchestra is Ok for me. Very speed sound and the best bass and backvocal in Octopus Garden (The Beatles). Some empty distance on border midrange and trebles
2. LME49710 - a lot of details. Good bass, good midrange. Soundstage is ok but space is a little bit empty (typically for all LME - IMHO).
3. LT1122 - really I don`t like sound of LT. A lot of details but everything sounds by own way. No music just good sounds.
4. OPA827 - bass is on the second place after THS, midrange is Ok but without air.
Not bad against price.
4. NE5534 (stock) - really not bad chip. Good for listening.
OPA627BP (original from 2007 not from ebay)
LT1122
LME49710
THS4061
NE5534 (from Kit)
OPA827
My list is
1. OPA627BP - very good sound. Not enough air a little bit. Old reliable horse.
2. THS4061 - comfortable for listening. Symphonic orchestra is Ok for me. Very speed sound and the best bass and backvocal in Octopus Garden (The Beatles). Some empty distance on border midrange and trebles
2. LME49710 - a lot of details. Good bass, good midrange. Soundstage is ok but space is a little bit empty (typically for all LME - IMHO).
3. LT1122 - really I don`t like sound of LT. A lot of details but everything sounds by own way. No music just good sounds.
4. OPA827 - bass is on the second place after THS, midrange is Ok but without air.
Not bad against price.
4. NE5534 (stock) - really not bad chip. Good for listening.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.