Earlier this month, Mr. @John Burson sent me a private message and asked if I would be interested in reviewing his discrete op amps in my freshly build AD1862 DAC. Of course I would. He was kind enough to promptly send me two V6 Vivid op amps in exchange for photos and review. Thanks John!
Here’s the first part of my write-up.
My background
I have spent a number of years building up projects in hifi and guitar amplification contexts, and while I have a basic level of electronics literacy, I have not yet developed a circuit for myself and thus consider myself a diy builder, not designer. I am a simple hobbyist.
In general here are some of the biases I bring into a comparison (whether these biases are true are false is perhaps a different topic. I realize that they are my biases):
I am also running an NOS R2R DAC and a Pass F5 amplifier so clearly I am not immune to audiophile appeal and perhaps desirable distortion.
Op amps are interesting parts in the sense that they are monolithic devices which contain quite a bit of circuitry. Really you could run a simple and great sounding preamp with 1 dual op and a simple power supply. Do I believe that changing preamps would make my system sound different? Of course. So, it stands to reason that changing 2 op amp stages per channel would similarly make a recognizable difference.
Test
I am pulling in a good friend (Friend #1) who designs audio gear at a well-regarded major manufacturer to listen to op amps with me. I am also trying to convince my friend to run some basic tests on a pile of op amps to give us data regarding what we are hearing.
I will be testing ADA4627-1 (currently in my DAC) VS NE5534 VS Burson V6 Vivid (and possibly also possibly LM741 and TL071 if these two are stable in this circuit).
On an upcoming weekend, my audio design friend will swap op amps for me while I listen in a blind test with a second friend. Friend #2 has almost 0 exposure to audio comparisons but is a foody and generally detail oriented, so I am particularly interested in his impressions as he does not have a horse in the race.
The only op amp I will listen to before this blind test will be the ADA4627-1’s which are currently in my DAC.
Hypothesis
My hypothesis before blind testing:
I expect to be able to consistently identify op amps 1, 2, and 3. Going into the test, I do not expect to have a strong preference for any of the 3 op amps over any of the others, although I do expect to have favorable opinions of all. I expect that the Bursons will have the largest subjective soundstage because of their slightly increased harmonic distortion profile (discussed below). Finally, although the NE5534 is old and inexpensive, I expect that in a blind test it will be rated as a peer to the other two options.
Friend #1 (engineer) hypothesis:
“I think [all] sets of op amps will sound good as the circuit around them is good. That being said, discrete op amp will show more nonlinearities and harmonic distortion. Whether this is desirable or not is up to the ears of the listener”
And when questioned about harmonic distortion:
“I like it a lot in the recording process. But for playback, I think that being faster and more accurate to better hear what the engineers did is more important.”
Harmonic Distortion
User @Martigane tested Burson op amps in his Whammy in 2020 and provided some useful measurements. In particular, check out the harmonic distortion profile of the Burson V6 Vivid:
Here: https://martigane.blogspot.com/2020/02/objective-benchmark-part-1-op-amps.html
This looks similar to the distortions in many of Nelson Pass’s JFET circuits. I am assuming that most people on this forum are familiar with his amplifiers. In particular the SIT amplifiers and the H2 buffers feature 2nd and 3rd order harmonics as a major feature in the sound of the amplifier. I ran tube amplifiers in my home for a long time, including a JE Labs 2a3 stage, and felt that they often threw huge sound stages. Perhaps the Bursons will nudge my rig in that direction.
Notes
Here’s the first part of my write-up.
My background
I have spent a number of years building up projects in hifi and guitar amplification contexts, and while I have a basic level of electronics literacy, I have not yet developed a circuit for myself and thus consider myself a diy builder, not designer. I am a simple hobbyist.
In general here are some of the biases I bring into a comparison (whether these biases are true are false is perhaps a different topic. I realize that they are my biases):
- Circuit design matters more than the grade of individual parts
- Power supply matters more than the grade of individual parts
- People that spend lots of time swapping op amps, tubes, and capacitors should sometimes just build or buy a different piece of gear
- I generally do not believe in burn-in in active devices and am not interested in expensive cabling.
I am also running an NOS R2R DAC and a Pass F5 amplifier so clearly I am not immune to audiophile appeal and perhaps desirable distortion.
Op amps are interesting parts in the sense that they are monolithic devices which contain quite a bit of circuitry. Really you could run a simple and great sounding preamp with 1 dual op and a simple power supply. Do I believe that changing preamps would make my system sound different? Of course. So, it stands to reason that changing 2 op amp stages per channel would similarly make a recognizable difference.
Test
I am pulling in a good friend (Friend #1) who designs audio gear at a well-regarded major manufacturer to listen to op amps with me. I am also trying to convince my friend to run some basic tests on a pile of op amps to give us data regarding what we are hearing.
I will be testing ADA4627-1 (currently in my DAC) VS NE5534 VS Burson V6 Vivid (and possibly also possibly LM741 and TL071 if these two are stable in this circuit).
On an upcoming weekend, my audio design friend will swap op amps for me while I listen in a blind test with a second friend. Friend #2 has almost 0 exposure to audio comparisons but is a foody and generally detail oriented, so I am particularly interested in his impressions as he does not have a horse in the race.
The only op amp I will listen to before this blind test will be the ADA4627-1’s which are currently in my DAC.
Hypothesis
My hypothesis before blind testing:
I expect to be able to consistently identify op amps 1, 2, and 3. Going into the test, I do not expect to have a strong preference for any of the 3 op amps over any of the others, although I do expect to have favorable opinions of all. I expect that the Bursons will have the largest subjective soundstage because of their slightly increased harmonic distortion profile (discussed below). Finally, although the NE5534 is old and inexpensive, I expect that in a blind test it will be rated as a peer to the other two options.
Friend #1 (engineer) hypothesis:
“I think [all] sets of op amps will sound good as the circuit around them is good. That being said, discrete op amp will show more nonlinearities and harmonic distortion. Whether this is desirable or not is up to the ears of the listener”
And when questioned about harmonic distortion:
“I like it a lot in the recording process. But for playback, I think that being faster and more accurate to better hear what the engineers did is more important.”
Harmonic Distortion
User @Martigane tested Burson op amps in his Whammy in 2020 and provided some useful measurements. In particular, check out the harmonic distortion profile of the Burson V6 Vivid:
Here: https://martigane.blogspot.com/2020/02/objective-benchmark-part-1-op-amps.html
This looks similar to the distortions in many of Nelson Pass’s JFET circuits. I am assuming that most people on this forum are familiar with his amplifiers. In particular the SIT amplifiers and the H2 buffers feature 2nd and 3rd order harmonics as a major feature in the sound of the amplifier. I ran tube amplifiers in my home for a long time, including a JE Labs 2a3 stage, and felt that they often threw huge sound stages. Perhaps the Bursons will nudge my rig in that direction.
Notes
- @Mooly has a useful thread about making sure that op amps are stable https://www.diyaudio.com/community/...-checked-to-see-its-stable-havent-you.191389/
- @Vunce has used the V6 Vivids in the same Miro DAC that I built and had favorable things to say: https://www.diyaudio.com/community/...st-tht-i2s-input-nos-r-2r.354078/post-6283993 . If I remember correctly, Vunce has also toyed with building his own discrete op amps.
- In this thread, @EUVL describes his unique IV stage as “…something more similar to a discrete-but-simplified AD844…”
To me this sounds similar to the Burson concept.EUVL comments below that this is incorrect https://www.diyaudio.com/community/...urrent-mirror-iv-converter-a-la-ad844.360162/
Last edited:
Not wanting to upset things, but the Burson is quite expensive, ($85 each I think though could be wrong) the NE5534 is what many would call a "Jelly bean" op amp costing about $1, and the ADA4627 about $10 or so. Would it not be fair comparison to try one of the very best TI and possibly AD products? (They will always be much cheaper than $85) and are really great products. Otherwise what is the point of the exercise?
I am open to suggestions if you have a favorite model. Purpose of the comparison is specifically to compare 3 (or more) products across the price range. That is the explicit intent.
And the single V6 Vivids are sold $70 for a pair on the DIYaudio store. So $35/per
The best I have used are the BB OPA627 but these are old hat now (though still good I think) though as they are obsolete priced as Faberge Eggs! Maybe others could suggest a more up to date champion?
If I understand, the ADA4627-1 is comparable to an updated OPA627.
The stock at Mouser for OPA627 rapidly exceed the cost of the Bursons (over $30)
Here: https://www.mouser.com/c/semiconduc...9SWweNJnLc4NS2oipNY4MPumB1302tfoaAlsQEALw_wcB
The stock at Mouser for OPA627 rapidly exceed the cost of the Bursons (over $30)
Here: https://www.mouser.com/c/semiconduc...9SWweNJnLc4NS2oipNY4MPumB1302tfoaAlsQEALw_wcB
Yes you're right, as I said the 627 is so expensive I was thinking of de-soldering and selling mine! (joke) I'm sure others here have more up to date recommendations, maybe a look at op amp specific threads would help?
I think you'll find the datasheets of these three single-opamp chips, very intriguing
OPA1611
ADA4898-1
OPA1655
Also keep in mind that the NE5534 single-opamp chip is decompensated, and is only stable for gains of 3 or higher, unlike the NE5532 dual-opamp chip. The 5532 is unity gain stable.
OPA1611
ADA4898-1
OPA1655
Also keep in mind that the NE5534 single-opamp chip is decompensated, and is only stable for gains of 3 or higher, unlike the NE5532 dual-opamp chip. The 5532 is unity gain stable.
Important info as always thanks Mark
Question for you:
On some level swapping op amps seems to me comparable to the M2x board options. Does that seem similar to you too?
Question for you:
On some level swapping op amps seems to me comparable to the M2x board options. Does that seem similar to you too?
M2x daughter cards began life as easy-to-source alternatives that don't use obsolete & hard to find Toshiba JFETs. Also as ways to explore the sonic virtues or drawbacks of different circuit approaches: pure single ended ("Mountain View"), composite amplifier ("Norwood"), diamond buffer ("Austin"), transconductors ("IPS6"), bootstrapped power rails ("IPS7"), etc. I did intentionally make IPS7 very friendly to opamp swappers / experimenters: it even had two DIP sockets on the board, one for single opamps like the LT1122, and another for duals like the NE5532. Because some interesting opamps are ONLY available as duals, and other interesting opamps are only available as singles. I wanted people to have the ability to try all of them.
Checked the datasheet, you are correct - but it also says the 5534 can be unity gain stable with external compensation. If I understand correctly, this IV stage features a compensating capacitor (C31), so I think a 5534 should be fine
I am still trying to figure out if this sallen-key reconstruction filter features compensation or where I would add it. I am running config 1
No, the authors of the 5534 datasheet think "external compensation" is what they say it is, in their datasheet. They say "external compensation" is a capacitor connected between pin5 and pin8 of the NE5534. See second plot on page 6 and first schematic on page 7 of the Philips datasheet attached here. Philips designed the TDA1034 and let their American subsidiary Signetics sell it. Signetics renamed it the NE5534. Then TI copied the 5534 and sold their copy for considerably less. Which eventually helped convince Philips and Signetics to stop making it themselves.
@rayma @Mark Johnson Looks like I have some learning to do! Thanks for the corrections. Appreciate the engagement
@EUVL corrected my first post, pardon me!
@EUVL corrected my first post, pardon me!
Even in the "uncompensated" version, there's still a 12pF capacitor connected internally
between pins #5 and #8. See the schematic on page 9 of the NE5534 data sheet.
https://www.ti.com/lit/ds/slos070d/...86368&ref_url=https%3A%2F%2Fwww.google.com%2F
between pins #5 and #8. See the schematic on page 9 of the NE5534 data sheet.
https://www.ti.com/lit/ds/slos070d/...86368&ref_url=https%3A%2F%2Fwww.google.com%2F
22pf is referred to as the "Typical value". Is it safe to say that should be the value I use between pins 5-8? I am reading data sheets most days now but it's a steep learning curve so any input is appreciated
Config 1's resistor values are rather too high to suit an NE5534, they're better suited to a TL071. I suggest 2k4 for R1,2,5 & 6 and 330pF for C1-4. The lower R values reduce the noise.
The test will be worthless unless its double-blind in this scenario - so you either have to have the one conducting the tests in a separate room entirely incommunicado, or sand the markings off the opamps, have an entirely different friend remark them with random codes, and only reveal the coding after the test results are collected. Basically two people in the same room communicate expectation non-verbally by loads of non-verbal signals, which is why double-blind protocols had to be invented for getting meaningful results in clinical trials.
And of course you have to arrange that the circuit is equally compatible with each opamp (adequate decoupling, not overloading, sufficient compensation for stability...).
Well Mark... then it will be worthless 😛
I will be scoping the circuit with different op amps as well as measuring DC offset before we listen to them to check for stability / clipping / oscillation / etc etc as per the Mooly thread I linked above
I am not a senior DIY member, I am here to learn, listen, and have fun. This project is about me learning how to read spec sheets, learning how op amps behave, and having friends over for a good bowl of pasta (gnocchi preferably). I am introducing a level of blindness into the test for fun, but to your point even then anyone reading this thread should not take that as rock solid empirical collection. It is what it is - couple of folks getting together and tinkering
I will be scoping the circuit with different op amps as well as measuring DC offset before we listen to them to check for stability / clipping / oscillation / etc etc as per the Mooly thread I linked above
I am not a senior DIY member, I am here to learn, listen, and have fun. This project is about me learning how to read spec sheets, learning how op amps behave, and having friends over for a good bowl of pasta (gnocchi preferably). I am introducing a level of blindness into the test for fun, but to your point even then anyone reading this thread should not take that as rock solid empirical collection. It is what it is - couple of folks getting together and tinkering