Somehow my module has ak411 something which converts my toslink into I2S.
I was using the upsampler in right justified MSB first, did not work well haha, now back to normal, looks promising.
I was using the upsampler in right justified MSB first, did not work well haha, now back to normal, looks promising.
Looks good... different than the pcm-J... It's kinda ready for a listening, opa604 i/v , op1611 buffer filter!
As of today reference pcm58 dac: 97.7db 2nd THD , 0Vdb
at -60vdb : 2n -66db . 3d: -63 db
That is very very good... (the other channel can't get be adjusted as much, -103.4db, , and -51db)
at 20khz, there is a little bit of thd -85db at 28khz, , -88.3db 2nd...
Multiple signals dynamic range seems to diminish in the order of 70db, usually they are in the 90db to 100db
With all the mods I have made to the board and the best PCM I could find...
There is 90db range of dynamics even in the low registers... compared to 70 in the 'control'
the -3db point is ridiculous at ... 2.5hz something I am laughing. It is flat to 3 hertz... to 18450...
THD, at -0.69dbV , -110...(-107.1 other channel) ,
at -62.21dbV , -58.42nd -513rd (-55 -483rd other channel)
It certainly has less noise at full power, -110 and -107 is impressive, however I need to better understand the process of adjusting the mbs and 3 first bits, it is very challenging. Now I am going to listen!
at -60vdb : 2n -66db . 3d: -63 db
That is very very good... (the other channel can't get be adjusted as much, -103.4db, , and -51db)
at 20khz, there is a little bit of thd -85db at 28khz, , -88.3db 2nd...
Multiple signals dynamic range seems to diminish in the order of 70db, usually they are in the 90db to 100db
With all the mods I have made to the board and the best PCM I could find...
There is 90db range of dynamics even in the low registers... compared to 70 in the 'control'
the -3db point is ridiculous at ... 2.5hz something I am laughing. It is flat to 3 hertz... to 18450...
THD, at -0.69dbV , -110...(-107.1 other channel) ,
at -62.21dbV , -58.42nd -513rd (-55 -483rd other channel)
It certainly has less noise at full power, -110 and -107 is impressive, however I need to better understand the process of adjusting the mbs and 3 first bits, it is very challenging. Now I am going to listen!
There are a lot of SHHHH shhhhh szhdhfhhhhhhh in signers voices. maybe it is the filter? Maybe it is the 1611 opamp , maybe that tantalum resistor, maybe the regulator, the black gates, no idea 🙁
OS 4x sounds way better, but 8x kind of sounds more muffled with more bass, less soundstage depth. 4xOS is just perfect!
There is a huge difference with this modified version... It is outstanding... true HIFI. Thanks God for this!
opamp roll, ne5532>8597>132>1611 , the 1611 is very aggressive in high and lack definition in high frequency which is very annoying, in contrast NE5532 has more definition in high frequency and no aggressiveness, the 8597 is very musical but the NE5532 is more complete, the 132 is maybe some bridge between which some great resolution but no musicality beside a fake warmth which falls apart quickly.
I like the sound as is so op604 will stay and no roll for the i/v.
I have been listening to Ich hatte viel Bekümmernis, BWV 21, Berlin Classics, Hans-Joachim. especially no.6 which has everything to challenge a DAC.
There is a huge difference with this modified version... It is outstanding... true HIFI. Thanks God for this!
opamp roll, ne5532>8597>132>1611 , the 1611 is very aggressive in high and lack definition in high frequency which is very annoying, in contrast NE5532 has more definition in high frequency and no aggressiveness, the 8597 is very musical but the NE5532 is more complete, the 132 is maybe some bridge between which some great resolution but no musicality beside a fake warmth which falls apart quickly.
I like the sound as is so op604 will stay and no roll for the i/v.
I have been listening to Ich hatte viel Bekümmernis, BWV 21, Berlin Classics, Hans-Joachim. especially no.6 which has everything to challenge a DAC.
I'm not 100% following what modifications you have done, @gabdx , but what would bother me with using the OPA1611 to directly drive headphones is the huge output impedance variation over the audible frequency range:
This is directly from the OPA1611 datasheet. With a 300 Ohm impedance headphone, you are a factor two down in output voltage at 80 Hz (6 dB, factor 4 in power).
The spec for the NE5534 only says (on page 5 of the datasheet) '0.3 Ohm' at 10 kHz, without a graph, so on paper it should be a lot better for this purpose (directly driving a headphone).
This is directly from the OPA1611 datasheet. With a 300 Ohm impedance headphone, you are a factor two down in output voltage at 80 Hz (6 dB, factor 4 in power).
The spec for the NE5534 only says (on page 5 of the datasheet) '0.3 Ohm' at 10 kHz, without a graph, so on paper it should be a lot better for this purpose (directly driving a headphone).
Last edited:
I am running it as buffer, closed-loop , 101 db of feedback.
Z out is basically 0.0000000000000 lol , what set the output Z is only the resistor,
Many headphones amps have a 100, 120R output Z and drive 32 ohm cans.
I looked at many schematics and found I needed 15R.
Thanks! I hope I am right.
Z out is basically 0.0000000000000 lol , what set the output Z is only the resistor,
Many headphones amps have a 100, 120R output Z and drive 32 ohm cans.
I looked at many schematics and found I needed 15R.
Thanks! I hope I am right.
Many headphone amps are bad, in that they have too high an output impedance. So you gave up on the direct drive? As far as output power is concerned these opamps should be fine; I believe the HD6-something Sennheisers have sensitivity around 100 dB SPL per mW, so 0.15V output and 0.5mA should give you close to 80 dB SPL, plenty loud.
The only thing that will zap an opamp like that is either incorrect supply connection (reversed) or possibly fitting a dual for a single or vice versa. If anything like that happens then its possible for caps to get reverse biased or presented with to much voltage. Most opamps are short circuit proof (check the data sheet).
Yeah, but it makes no sense at all 🙂
I have the hd600 here, at 97db and they sound way louder than my new hd650, which are rated 103 db I think.
Maybe the drivers are stiff and will loosen up.
I gave up using a pcm58 board direct because the gain is too high, I keep using my tube headphone which is Direct after the AKM pins, balanced to single ended.
on My 600 I have the inner foam removed, my ear is directly with the driver, maybe this explains and the fact they are 15 years old.
I have the hd600 here, at 97db and they sound way louder than my new hd650, which are rated 103 db I think.
Maybe the drivers are stiff and will loosen up.
I gave up using a pcm58 board direct because the gain is too high, I keep using my tube headphone which is Direct after the AKM pins, balanced to single ended.
on My 600 I have the inner foam removed, my ear is directly with the driver, maybe this explains and the fact they are 15 years old.
opamp Cap 🙂 everything is fine, just a dump mistake! A fire-cracker.
Sorry, that was of course nonsense, comparing apples and oranges, or open loop vs 30 dB gain.
Checking with my scope now! Unfortunately channel 1 looks damaged on the scope from maybe amplifier tests, So I am stuck with one channel 🙁
Last edited:
I don't quite understand who to set up the MSB and 3 other bits of data with the trimpots.
2 can be done with THD tests , but 4 is too complicated..
I am not sure how to see the high frequency noise and if the filter works properly with the scope.
2 can be done with THD tests , but 4 is too complicated..
I am not sure how to see the high frequency noise and if the filter works properly with the scope.
The receiver is AK4118 , Filter SM5847,
There is increase digital noise in the audio band when I use anything but 44,48,88khz to feed the AK4118, 24 or 16 bits.
The dac creates square-waves if I feed it with more than 18bits from the filter.
There is increase digital noise in the audio band when I use anything but 44,48,88khz to feed the AK4118, 24 or 16 bits.
The dac creates square-waves if I feed it with more than 18bits from the filter.
I've no idea on that either tbh. Its not something I've ever encounterd.
I suspect you need something of far higher resolution than a scope to see a lot of minute differences.
Yep, with PCM-J which are quite rare, I had exceptional high -60db figures, and book figures for noise etc.
With this PCM-K version it is a different animal and I run it with 4 adjustments.
It looks genuinely selected for the low noise and high dynamic capabilities. It is leagues above the basic and J versions.
I readjusted as well as I could... it gives sub-level THD vs the plain P version...
-3db points with scope : 14-35479.
THD 990.5hz 0db and -60.
R: 2. -106.3, 3. -100.9, 4. -107.8, 5. -96.3, up to 10th harmonics.
R: 2. -60, 3. -44.2
L: 2. -106 no other harmonics.
L: 3. -40.4
The big difference is in square-waves etc. I can see the distance between peaks and dips in the lower registers with square waves, it has 20db more dynamic range than the two other chips.
I don't know if this is the chip or the board, I could swap one just to see but it is risky to damage the chip. I could insert a plain J and see if anything can be attributed to the 'modified' board.
With this PCM-K version it is a different animal and I run it with 4 adjustments.
It looks genuinely selected for the low noise and high dynamic capabilities. It is leagues above the basic and J versions.
I readjusted as well as I could... it gives sub-level THD vs the plain P version...
-3db points with scope : 14-35479.
THD 990.5hz 0db and -60.
R: 2. -106.3, 3. -100.9, 4. -107.8, 5. -96.3, up to 10th harmonics.
R: 2. -60, 3. -44.2
L: 2. -106 no other harmonics.
L: 3. -40.4
The big difference is in square-waves etc. I can see the distance between peaks and dips in the lower registers with square waves, it has 20db more dynamic range than the two other chips.
I don't know if this is the chip or the board, I could swap one just to see but it is risky to damage the chip. I could insert a plain J and see if anything can be attributed to the 'modified' board.
Two hypothesis:
the J is a genuine rare chip with better performances, however it did not performed as well as the modified DAC.
I can rule out the black gates, the special tantalum resistors , the type of IV (MAYBE), but I doubt, ill clear this one of the equation too.
The PCM K looks genuine but the K could have been faked, we will never know...
What is sure is that something is making that new board sound superb, and it could be the regulators. It has special dedicated external unit following the LM33x, and the -12V is also a very costly regulator.
The digital side has an extra custom made pre-regulator replacing the LM33x and I adjusted voltages higher.
Maybe it is only the Digital filter which is way superior to the 1704?
I bet it is either the filter of the power supply.
Lets listen with plain PCM-J and see if it sounds as good or get better specs even!
the J is a genuine rare chip with better performances, however it did not performed as well as the modified DAC.
I can rule out the black gates, the special tantalum resistors , the type of IV (MAYBE), but I doubt, ill clear this one of the equation too.
The PCM K looks genuine but the K could have been faked, we will never know...
What is sure is that something is making that new board sound superb, and it could be the regulators. It has special dedicated external unit following the LM33x, and the -12V is also a very costly regulator.
The digital side has an extra custom made pre-regulator replacing the LM33x and I adjusted voltages higher.
Maybe it is only the Digital filter which is way superior to the 1704?
I bet it is either the filter of the power supply.
Lets listen with plain PCM-J and see if it sounds as good or get better specs even!