Please look at the attached file. This is the output of my DAC part of a recorder.
I have replaced all NJM4580 and NE5532 to LM4562, the result isn't what I like, too bright and on XLR output, I feel some distortion too.
I plan to:
1. Replace some or all to OPA2107,
2. Make some change to the circuit (maybe by pass the volume control CS3318, shot pin 22/24), and even by pass IC20 (which is ok to leave out RCA output and only use XLR)
I am no expert, and these are all SMDs, so soldering is a pain, I wish it could be at least good on paper and then I implement it.
If you great DIYs can give me some advise, that will be great! Thanks for your time.
I have replaced all NJM4580 and NE5532 to LM4562, the result isn't what I like, too bright and on XLR output, I feel some distortion too.
I plan to:
1. Replace some or all to OPA2107,
2. Make some change to the circuit (maybe by pass the volume control CS3318, shot pin 22/24), and even by pass IC20 (which is ok to leave out RCA output and only use XLR)
I am no expert, and these are all SMDs, so soldering is a pain, I wish it could be at least good on paper and then I implement it.
If you great DIYs can give me some advise, that will be great! Thanks for your time.
Attachments
If this were in my hands, and I was only concerned about balanced output with no need of the volume controller, I would bypass ic20 & ic26, and go balanced to balanced from ic21 to ic54. Delete ra19a/b/c/d, delete the join between ra-3d & pin3 of IC54 & pin24 of ic26, connect ic21 pin7 to ra3-d, connect ic21 pin1 to ic54 pin 3. I'd replace ic21 & ic54 with OPA2134ua(way better chip than 2107), with the critical addition of at least 220uf electrolytic in parallel with each c60, c65, c181 & c191, or else the chips will be starved for instant current & dynamics will suffer badly, as will overall sound. I would also delete C13 & C14, as they should not be needed, as no significant dc offset should be present at the chip output. Do all this, and it should be a terrific sounding output stage.
Thank you guys!
Too bad, before I read the reply, I have already installed OPA2107. The good news is it's so much much better than LM4562. I wonder why LM4562 doesn't work proper in this design, I am sure it's also a very good chip, it just doesn't work for me, it's very harsh. I will order some OPA2134ua to test, the price is right, I have no idea why OPA2107 cost that much than others.
So, if I still want to keep the RCA and headphone output, I can't delete ra19a/b/c/d, is it OK I just do the rest for XLR (delete the join between ra-3d & pin3 of IC54 & pin24 of ic26, connect ic21 pin7 to ra3-d, connect ic21 pin1 to ic54 pin 3.). One thing is that I think the DAC output is differential signal, so ic21 pin7 goes to ra3-d, but should ic21 pin1 goto ic54 pin2 instead of pin3? (negative instead of positive) so both get reversed output voltage, so the XLR is still differential, or it doesn't matter?
Too bad, before I read the reply, I have already installed OPA2107. The good news is it's so much much better than LM4562. I wonder why LM4562 doesn't work proper in this design, I am sure it's also a very good chip, it just doesn't work for me, it's very harsh. I will order some OPA2134ua to test, the price is right, I have no idea why OPA2107 cost that much than others.
So, if I still want to keep the RCA and headphone output, I can't delete ra19a/b/c/d, is it OK I just do the rest for XLR (delete the join between ra-3d & pin3 of IC54 & pin24 of ic26, connect ic21 pin7 to ra3-d, connect ic21 pin1 to ic54 pin 3.). One thing is that I think the DAC output is differential signal, so ic21 pin7 goes to ra3-d, but should ic21 pin1 goto ic54 pin2 instead of pin3? (negative instead of positive) so both get reversed output voltage, so the XLR is still differential, or it doesn't matter?
If you don't have ultrasonic oscillations (the caps on the power pins to ground should take care of that) and you cleaned the board with isopropyl alcohol, then... that's how is supposed to sound.
Probably your brain got used to a mellow sound
PS: OP2107 is so expensive because the input stage is Difet isolated (ultralow input bias) and laser trimmed (for perfect DC performance). You pay for something that you don't use in audio.
Probably your brain got used to a mellow sound
PS: OP2107 is so expensive because the input stage is Difet isolated (ultralow input bias) and laser trimmed (for perfect DC performance). You pay for something that you don't use in audio.
Last edited:
Yes, you can leave the phones/rca out feed lines connected via Ra19a/b/c/d & just do the rest of the xlr out bypass the same, although you are correct, the differential feed should go to the same phase inputs of ic54. My error. You would need to reconfigure ic54a the same as ic54b or vice versa. You know, I just noticed that the schematic shows x2 gain on ic54b and x1 unity gain on ic54a, which does not actually make sense. A proper balanced line driver should darn well have matched gain on both phases. So, you may want to adjust the feedback resistor value on one phase or the other.
I am not very surprised that the LM4562 sounded harsh with the existing 0.1uf caps on the supply pins, which are no doubt ceramic. Adding large value lytics in parallel with these would have made the 4562 sound a lot better, especially if you also replaced the ceramic caps with nice plastic films. The 2107's not doubt don't sound as good as they could if the caps have not been so altered. The 2107's are more expensive than the better sounding 2134 due to being made for instrumentation type work, where dc & gain precision are more important than sound.
I am not very surprised that the LM4562 sounded harsh with the existing 0.1uf caps on the supply pins, which are no doubt ceramic. Adding large value lytics in parallel with these would have made the 4562 sound a lot better, especially if you also replaced the ceramic caps with nice plastic films. The 2107's not doubt don't sound as good as they could if the caps have not been so altered. The 2107's are more expensive than the better sounding 2134 due to being made for instrumentation type work, where dc & gain precision are more important than sound.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.