Hello all,
I need some help deciding what to do with the output stage of my pre-amp. Seems their are a lot of people in this forum who know about properly utilizing opamps, and that's really what this is about.
Below is the schematic, with some notes. This is a pre-built circuit that was supposed to have dual AD797 opamps, but the manufacturer pulled a fast one and installed dual L49710 opamps instead. Furthermore, caps C402 and C405 were not WIMA, but some generic junk. I have replaced those with WIMA FKP3 and the sound quality improved dramatically. However, I suspect that this circuit is still not what it could be.
Can anyone offer any guidance on a couple of points? I have spent some time with the datasheets of the various parts, but cannot wrap my brain around it all.
1) Since the schematic specifies AD797, is it safe to assume the circuit is designed well to work with that opamp? I do know that the AD797 requires special care to implement (esp. power supply bipassing) correctly.
2) Does a simple replacement of L49710 even make sense here? Sure, it is a good opamp, but different opamps require different circuits, no? (perhaps leads to the next question)
3) Why have they omitted C404 and C408? Could it be that they are not required, due to the different opamp being installed?
4) Even if you can't or don't want to answer the above questions, do you have any advice on what you would do to improve the circuit?
The simplest thing for me to do would be to replace the L49710 opamps with AD797 opamps and install 22pF caps at C404 and C408 (as per schematic), but I will only do this if I can have someone confirm that it is a good circuit.
Thanks in advance for any advice and opinions offered.
I need some help deciding what to do with the output stage of my pre-amp. Seems their are a lot of people in this forum who know about properly utilizing opamps, and that's really what this is about.
Below is the schematic, with some notes. This is a pre-built circuit that was supposed to have dual AD797 opamps, but the manufacturer pulled a fast one and installed dual L49710 opamps instead. Furthermore, caps C402 and C405 were not WIMA, but some generic junk. I have replaced those with WIMA FKP3 and the sound quality improved dramatically. However, I suspect that this circuit is still not what it could be.
Can anyone offer any guidance on a couple of points? I have spent some time with the datasheets of the various parts, but cannot wrap my brain around it all.
1) Since the schematic specifies AD797, is it safe to assume the circuit is designed well to work with that opamp? I do know that the AD797 requires special care to implement (esp. power supply bipassing) correctly.
2) Does a simple replacement of L49710 even make sense here? Sure, it is a good opamp, but different opamps require different circuits, no? (perhaps leads to the next question)
3) Why have they omitted C404 and C408? Could it be that they are not required, due to the different opamp being installed?
4) Even if you can't or don't want to answer the above questions, do you have any advice on what you would do to improve the circuit?
The simplest thing for me to do would be to replace the L49710 opamps with AD797 opamps and install 22pF caps at C404 and C408 (as per schematic), but I will only do this if I can have someone confirm that it is a good circuit.
Thanks in advance for any advice and opinions offered.
An externally hosted image should be here but it was not working when we last tested it.
You probably dont want to downgrade from a real diamond to a sapphire?
If you want to upgrade to something, consider lme49990.
C404/408 are to prevent hf oscillations. AD797 might need that...???
IMHO...
If you want to upgrade to something, consider lme49990.
C404/408 are to prevent hf oscillations. AD797 might need that...???
IMHO...
Isn't that the single version of the LM4562 ? Its probably as good as anything in my experience. As the stage is 100% DC coupled, have you checked for DC offset at the output. The 49710 has low input bias currents though, and so even though they may be unmatched (can't say for sure without knowing the circuitry to the left) it shouldn't be an issue. Just measure it for confirmation... and enjoy.
Thank you for replying, palstanturhin. Given that the only Finnish person I know is a genius, and also given that Linus Torvalds created Linux, I must take your word for it. 😀
Perhaps I am being silly, but certainly you know more about this than I do. 🙁
Any other opinions? ...
Perhaps I am being silly, but certainly you know more about this than I do. 🙁
Any other opinions? ...
Thanks. I will check DC offset.
The circuitry to the left is direct output from CS3310, pins 11 and 14, respectively.
Have not had a chance to check DC offset yet, but I did spend some time listening more last night while doing other tasks. I have to say that I think the new caps (@ C402 and C405) have settled in nicely now and it really does sound quite good. Still not as good as direct source to amp, but certainly very good.
After some reading, I have ordered a PGA2311, which I will use to replace the CS3310. Perhaps that will give me the final increase in quality that I am looking for.
I will also try rolling various opamps in my DAC again, after the PGA2311 goes in the pre. I suspect I will find synergy there somehow.
After some reading, I have ordered a PGA2311, which I will use to replace the CS3310. Perhaps that will give me the final increase in quality that I am looking for.
I will also try rolling various opamps in my DAC again, after the PGA2311 goes in the pre. I suspect I will find synergy there somehow.
The LM4562 is pretty good... in fact better than that, its excellent. If you are looking for something genuinely "different" then try an OPA2604.
I've got a lot of options for the DAC. So far I have tried:
- OPA2132
- OPA2107
- OPA275 😀
- LM6172
- AD8599 (SOIC on DIP8 adapter)
- AD8066 (SOIC on DIP8 adapter)
My 2 favourites from that list are LM6172 and AD8599.
I also have the following in various states of shipping/assembly:
- AD8620 (SOIC on DIP8 adapter)
- AD823 (SOIC on DIP8 adapter)
- dual OPA627 (in dual DIP8 to single DIP8 adapter)
- dual LME49710HA (cans, in dual-can to DIP8 adapter) (~ LM4562)
- LME49720HA (can, in can to DIP8 adapter) (~ LM4562)
Last edited:
You'll just have to try them and see what you think yourself. Some pretty serious listening tests on my part have confirmed the LM4562 really is very transparent. The OPA2604 remains a favourite of mine and one that really does have a slightly different sound to it... some love it and others don't.
The circuit looks OK
I would look at the power supply, regulation, and decoupling circuits for ways to make improvements to the performance of the preamp.
I would look at the power supply, regulation, and decoupling circuits for ways to make improvements to the performance of the preamp.
Yes, it does seem that the 2604 is like Marmite. Based on what I have read, I have avoided it. I really should just get one and hear for myself. I do like Marmite, after all.
Thanks for your response. I am not educated in electronics enough to look at a schematic and make a lot of sense of it, particularly in the areas you mentioned. I could, however, at least check the installed components and make sure they match the specs on the schematic and replace anything that is not to spec (preferably with higher quality components). After all, I did already find some things that were not according to the schematic, or of low/questionable quality.
Unrelated update: My dual TO99 to DIP8 adapter arrived today. I soldered in two LME49710, then soldered the 8 pins in place. Checked all the solder joints. Installed it in my DAC.
Holy crap! Incredible! Best opamp(s) I have heard, by far.
I suspect once I get the PGA2311 into the pre, I'll be satisfied for quite some time.
Holy crap! Incredible! Best opamp(s) I have heard, by far.
I suspect once I get the PGA2311 into the pre, I'll be satisfied for quite some time.
Last edited:
Keep in mind that the PGA2311 is only good for a +/- 5V analog swing.
If you need more than that then consider a PGA2310 (PDIP) or a PGA2320 (SOIC) That handles +/- 15V analog swing for stereo.
FWIW
jer 🙂
If you need more than that then consider a PGA2310 (PDIP) or a PGA2320 (SOIC) That handles +/- 15V analog swing for stereo.
FWIW
jer 🙂
I actually don't know what analog swing is. What I did notice is that the PGA2311 supply voltage is +/-5v, which is the same for CS3310, whereas the PGA2310 supply voltage is +\-15v. I chose the 2311 because it seems like it is a drop-in replacement. Please correct me if I am wrong. Thank you.
It means that the signals voltage peak to peak level cannot exceed the chips supply voltage.
Else, you will either clip the signal or maybe risk damaging the chip.
jer 🙂
Else, you will either clip the signal or maybe risk damaging the chip.
jer 🙂
I think +\-5v is enough headroom, as my DAC is pumping no more than 2v as far as I know... Also the CS3310 has the same limitation and is working fine.
Thanks for the explanation.
Okay, It will be fine then. 😉
I was just something that I want to point out as many are not aware of this, as most opamp circuits are operated at higher voltages.
Cheers!!!
jer 🙂
I was just something that I want to point out as many are not aware of this, as most opamp circuits are operated at higher voltages.
Cheers!!!
jer 🙂
Last edited:
I think +\-5v is enough headroom, as my DAC is pumping no more than 2v as far as I know... Also the CS3310 has the same limitation and is working fine.
Thanks for the explanation.
2 Vrms actually means +/- 2.83 V peak, but still, +/- 5 V ought to do. The CS3310 datasheet says, "Signals approaching the analog supply voltages may be applied to the AIN pins if the internal attenuator limits the output signal to within 1.25 volts of the analog supply rails."
Are the opamps supplied with the same or do they get higher voltages? Mooly recently bumped into input common-mode range limitations when trying different opamps in a CDP-790, though since the circuit here has appreciable gain, only the limitations on the output side are of interest (along with common-mode distortion performance). Expect up to about +/- 4 V peak with the LME49710.
Since the CS3310 is not a low-noise king to begin with and has plenty of gain if needed, I would have liked to see lower gain and matched impedances at the opamp inputs as well. Try swapping R408 for a 3k3; the old 1k part can then replace R403. (A 250 ohm impedance imbalance is plenty good enough, and noise is not an issue.) Same goes for the other channel.
The stock preamp would typically output about 18 µV of noise, which with a typical power amp gain of 26 dB becomes a rather audible 360 µV. If 6 dB less (after the mod as suggested) still is too much for your tastes, remove R408 and replace the feedback resistor R406 by 100 ohms, this time leaving the 100 ohm R403 in place. 90 µV of resulting output noise (assuming the power amp is not limiting) would be very good integrated amp level and ought to be inaudible with normal speakers.
As far as the OPA(2)604 is concerned, I think Samuel Groner's measurements have shown pretty clearly that its area of expertise is higher than usual supply voltages (up to +/-24 V). It's pointless at +/-15 V, and pretty much useless at +/- 4.5 V (managing little more than 4.5 Vpp output even at no load).
Are the opamps supplied with the same or do they get higher voltages? Mooly recently bumped into input common-mode range limitations when trying different opamps in a CDP-790, though since the circuit here has appreciable gain, only the limitations on the output side are of interest (along with common-mode distortion performance). Expect up to about +/- 4 V peak with the LME49710.
Since the CS3310 is not a low-noise king to begin with and has plenty of gain if needed, I would have liked to see lower gain and matched impedances at the opamp inputs as well. Try swapping R408 for a 3k3; the old 1k part can then replace R403. (A 250 ohm impedance imbalance is plenty good enough, and noise is not an issue.) Same goes for the other channel.
The stock preamp would typically output about 18 µV of noise, which with a typical power amp gain of 26 dB becomes a rather audible 360 µV. If 6 dB less (after the mod as suggested) still is too much for your tastes, remove R408 and replace the feedback resistor R406 by 100 ohms, this time leaving the 100 ohm R403 in place. 90 µV of resulting output noise (assuming the power amp is not limiting) would be very good integrated amp level and ought to be inaudible with normal speakers.
As far as the OPA(2)604 is concerned, I think Samuel Groner's measurements have shown pretty clearly that its area of expertise is higher than usual supply voltages (up to +/-24 V). It's pointless at +/-15 V, and pretty much useless at +/- 4.5 V (managing little more than 4.5 Vpp output even at no load).
Last edited: