I am happy you can report this. The OPA1612 is very transparent and also measures greatly - except for phase rotations. In such a simple buffer (which is a single ended implementation with a trivial feedback loop) it is the best choice. If the circuitry gets a bit more complex to push noise, distortion, stability, and slew rates to a higher level, you are probably best served by the LME49720 (as in the Neurochrome buffers), with has less phase rotation kinks.
Some people swear by discrete opamps because they "must" be better than ICs - not to speak that they can be made to work in pure class A!!! The only ones I had that I could consider good sounding are the NewClassD Ultimate Edition opamps (absolutely transparent, brutally in-the-face in their way to display stuff), the Sonic Imagery ones (a tad warmer, but still absolutely transparent). Sparkos and Burson I consider really really bad. I haven't tried the Weiss OP2, which even measure better than the best integrated ones (which is an amazing feat, as ICs can be made to a much higher precision than the little laser cut components on discrete designs) and is apparently totally transparent. But the Weiss OP2 also costs an arm and a leg.
WTF? They added a huge amount of iron to make it so expensive?
I wonder how it compares to my little baby Purifi-Neurochrome Build — ImgBB — Page 1
RE discrete...
I suspect that what many may be missing here is that the IC opamps used here are being pushed very close to their limits for both current capability and voltage swing. while the ICs theoretically, from a technical perspective, have "enough" of each to do the job, they are operating with very little headroom in terms of current and voltage swing (if playing at high levels). The discrete options, on the other hand, often have >5x the current capability of the ICs. In my experience, a stouter input stage, which can operate both fully in class A and with a lot of current headroom, can sound better. The best commercial class D amplifier I have ever heard is the Mola Mola, and it uses a discrete input stage...
At some point I want to try a S.I. or Weiss based discrete input stage with the Purifi modules for these reasons.
I suspect that what many may be missing here is that the IC opamps used here are being pushed very close to their limits for both current capability and voltage swing. while the ICs theoretically, from a technical perspective, have "enough" of each to do the job, they are operating with very little headroom in terms of current and voltage swing (if playing at high levels). The discrete options, on the other hand, often have >5x the current capability of the ICs. In my experience, a stouter input stage, which can operate both fully in class A and with a lot of current headroom, can sound better. The best commercial class D amplifier I have ever heard is the Mola Mola, and it uses a discrete input stage...
At some point I want to try a S.I. or Weiss based discrete input stage with the Purifi modules for these reasons.
I agree with you. I was suffering since I got these VTV amps with Sparkos OPA not able to get decent sound from my speakers. In fact they were no where near to my Red wine Audio Singature 70.2 or the Quick silver KT120 tube based monoblocks. I was getting frustrated as I was reading lots of good thing about Purifi based amps. Could it be Input buffer quality also.
Anyway I am finding 1612 much better .
In my system with SMPS1200, Purifi module, VTV input buffer and OPA1612 I am finally getting a decent sound. Sparkos was total disappointment.
1612 sounding very clean and transparent with warmth and huge stage. Lows and highs are good. May need few hours for mids to open up.
I am sure Purify must have tried various options before putting 1612 in their input buffer. To me they look like made for each other.
Thanks for all the information.
Thanks barrows for recommending OPA1612. They are making the sound that I was looking for. Sparkos were real disappointment.
As I am a non technical person, I am learning in the process. An interesting hobby to follow and trying various options.
I would also like to try Weiss at some stage and these are rather expensive. Do you think VTV input buffer will be ok to trying Weiss.
Thanks again. I will inform VTV also about 1612 .
It depends on the opamp. The OPA1612 may be strained close to its limit, but the LME479xx family exists even in a variant that can output 100mA. But in general once you give these opamps +-15V then they can output about +-14.5V continuously, and since +-10V is sufficient to drive the Purifi modules at their max output (which they cannot do continuously, mind you), I would not be worried. But you are right, precision op amps can get up to VCC-VEE=50V and more, whereas this is usually tops at 30 and 40 for the smaller integrated ones. And current car go up to 250mA instead of around 25mA – which is less worrisome because they are usually used to drive in voltage not in current.
There may be a more serious matter however: Both Bruno Putzeys and NwAvGuy have complained that the pinout of the IC opamps is bad if you use an opamp in high gain mode, as the V- is very close to in_A+ and there is this inductive coupling between these two parallel and close pins. Putzeys clearly said that until there isn't an alternative standard he will continue to design discrete opamps (OTOH NwAvGuy did not seem to care much). This may be one of the reasons the "pro" opamps sound better, as the distances between the pins are larger and the input is on the opposite side from VCC and VEE.
My takeaway is: if you want to go for discrete opamps, go for the large, discrete precision op amps with their proper sockets.
yes...
"My takeaway is: if you want to go for discrete opamps, go for the large, discrete precision op amps with their proper sockets."
Agreed with the above for sure, as the need to fit them in a small footprint is a constraint on the design of the smaller discretes.
I am not familiar with the variants on LM 4562/4792 which have 100 mA current capability?
I find the OPA 1632 is a good option when looking for more current drive. It might be interesting to try the alpha 24 (from AMB) to drive the Purifi modules, these use an instrumentation amplifier approach with a OPA 1612 for the input and all of the gain, and then a OPA 1632 to drive the output. As i recall the OPA 1632 is pretty stout, with 150 mA of current capability if I remember correctly.
"My takeaway is: if you want to go for discrete opamps, go for the large, discrete precision op amps with their proper sockets."
Agreed with the above for sure, as the need to fit them in a small footprint is a constraint on the design of the smaller discretes.
I am not familiar with the variants on LM 4562/4792 which have 100 mA current capability?
I find the OPA 1632 is a good option when looking for more current drive. It might be interesting to try the alpha 24 (from AMB) to drive the Purifi modules, these use an instrumentation amplifier approach with a OPA 1612 for the input and all of the gain, and then a OPA 1632 to drive the output. As i recall the OPA 1632 is pretty stout, with 150 mA of current capability if I remember correctly.
options...
The SMPS 1200 offers unregulated ~ -/+ 18 VDC rails (or regulated at -/+ 12 VDC). Most of the optional input buffer board designs feature onboard regulators, meant to be used with the unregulated supply from the SMPS 1200.
In my Purifi build, I use a separate, linear, regulated, power supply at -/+ 18 VDC to feed the input board(s).
There are a lot of options and different ways to power the input board(s) if one is crafty about it.
The SMPS 1200 offers unregulated ~ -/+ 18 VDC rails (or regulated at -/+ 12 VDC). Most of the optional input buffer board designs feature onboard regulators, meant to be used with the unregulated supply from the SMPS 1200.
In my Purifi build, I use a separate, linear, regulated, power supply at -/+ 18 VDC to feed the input board(s).
There are a lot of options and different ways to power the input board(s) if one is crafty about it.
Yes, if you are doing your own design, that is definitely an option.
As far as I know, all the commercially available input boards use the same on-board regulator (or regulated input) to send to the Purifi module which requires +/-12V.
If you know of any commercially available input boards that are more flexible than this, can you post links. I'm assuming I have to make my own boards if I want to use different supplies for the buffer circuitry and the Purifi op amp supply.
As far as I know, all the commercially available input boards use the same on-board regulator (or regulated input) to send to the Purifi module which requires +/-12V.
If you know of any commercially available input boards that are more flexible than this, can you post links. I'm assuming I have to make my own boards if I want to use different supplies for the buffer circuitry and the Purifi op amp supply.
...
The Purifi modules have a max voltage for the supply rails you are referring to at +/- 20 VDC (I just double checked this on the Purifi amp data sheet). Therefore one can raise the voltage higher for the OPA supply voltages if there is an advantage to doing so. For example, Nord uses +/- VDC regulators on their input boards for their Purifi module based amplifiers. Higher voltage supplies to the input stage opamps can be an advantage for some parts.
The Purifi modules have a max voltage for the supply rails you are referring to at +/- 20 VDC (I just double checked this on the Purifi amp data sheet). Therefore one can raise the voltage higher for the OPA supply voltages if there is an advantage to doing so. For example, Nord uses +/- VDC regulators on their input boards for their Purifi module based amplifiers. Higher voltage supplies to the input stage opamps can be an advantage for some parts.
Do you think changing to Neurochrome input buffer is a better option in place of VTV buffer to go with my SMPS1200 and Purify module.. Just a thought.
I do not know the VTV buffer, but I expect it to be like all other ones around, only less optimised than others (for instance, I know that Apollon went to great lengths to optimise the values of the components and the routing to reduce noise and distortion). They are all derived from the standard buffer described in the hypex documentation, only with a socket in place of the original op amp. The circuit has then been optimised by (forgot his name) the designer of Sonic Imagery.
The Neurochrome products are in general excellent examples of engineering. I have both the input buffers (I helped Tom Christiansen with the testing) and the Universal Buffer and I can attest that they are splendid boards. I prefer them over the aforementioned buffer.
I think Neurochrome input buffer is using OPAmp 4562 if I read correctly on the image on their web site.
VTV has asked me to try 4652 also and see how it compares with 1612.
I am seeing most of the Purifi amplifier manufacturers are offering their Amps with options of Sonic Imagery and Sparkos. My experience about Sparkos has been very disappointing and there is no comparison with 1612 which is much cheaper and much better.
I am not sure this is specific to my system or amplifier, but I am wondering have these manufacturers really tested all their offerings and how they perform sonically. No body is offering 1612 except probably Appolon.
Because in my case there is not a subtle but a huge improvement using 1612 over Sparkos which are much more expensive also.
VTV has asked me to try 4652 also and see how it compares with 1612.
I am seeing most of the Purifi amplifier manufacturers are offering their Amps with options of Sonic Imagery and Sparkos. My experience about Sparkos has been very disappointing and there is no comparison with 1612 which is much cheaper and much better.
I am not sure this is specific to my system or amplifier, but I am wondering have these manufacturers really tested all their offerings and how they perform sonically. No body is offering 1612 except probably Appolon.
Because in my case there is not a subtle but a huge improvement using 1612 over Sparkos which are much more expensive also.