Also note, that for TI and many other manufacturers, at one time it was in vogue to release the same IC in multiple temperature ranges -- 0->70C, -40->85C, -40->125C, etc., as different orderable items. Often the only thing in the specs that was different was the temperature. Maybe sometimes some DC specs were different (like DC offset or offset drift). Almost universally, the actual die in the package was the same. So, you'll see multiple versions with ancient op-amps like the TL071.
It took years, but someone finally figured that the cost of doing this (separate devices with just slightly differing manufacturing flows, separate inventory, etc.) for essentially the same product was expensive. Now, as John says, you'll see just one item specified over it's maximum recommended range, though you may see the same chip released separately automotive-qualified and commercial versions. The automotive likely has differences due to higher testing/qualification requirements, but ultimately still the same piece of silicon inside and either version will work the same for DIY audio.
It took years, but someone finally figured that the cost of doing this (separate devices with just slightly differing manufacturing flows, separate inventory, etc.) for essentially the same product was expensive. Now, as John says, you'll see just one item specified over it's maximum recommended range, though you may see the same chip released separately automotive-qualified and commercial versions. The automotive likely has differences due to higher testing/qualification requirements, but ultimately still the same piece of silicon inside and either version will work the same for DIY audio.
Couldn’t leave things as they were…
In particular there were these 2 excellent Op amps, having a lot of common but offering a slightly different music representation, perhaps the ADA4627-1 small additional airyness/lightness being a tad more seductive (while the OPA1656 groove can be addictive)… but which one is truer ?
I let the 2 op amps burn-in an additional 20h and decided to use my Ipod as source. Hardly HIFI, but the purpose was solely to be able to monitor things. The Ipod can just drive my Beyer to the right volume level, being the souce direct option. Then I can introduce the simple HP amp, adjust the volume on it, so any change to the music presentation comes from that test rig which is mainly a double op amp per channel.
After another 20h of burn-in sound is settled. TBH I can’t tell if it really improved after the initial 5h, as direct comparison is impossible, but my feeling is « negligeable if anything ». It used to burn-in initialy though, the first hour mainly and then slightly so.
This test revealed again how excellent both op amps are and that they have more similarities than differences. However, the ADA4627-1 sounded still a tad more open up, airy and with a higher tonality balance. But in comparison to the direct source it turned out it lightens and brightens things a tad. Ads sparkling, takes off weight. Sounds fun though, but less true to the source.
On the other hand, the OPA1656 turned out to be very very close to the source, neutral in fact. Its tonality and balance are true, details are what they are and its massive punch weight are the real ones. In fact it hasn’t really a dark balance : it has a neutral one and the ADA4627-1 chip is in fact a tad brighter, sparkles and shines a tad (and the ADA4625-1 I turned down becomes though quite bright in return).
This gives us options to adjust the tonal balance. If you want pure HIFI and an excellent component that is neutral and incredibly fun, then the OPA1656 is simply an amazing chip, probably state of the art. If you feel on the other hand your output stage needs a bit of extra sparkling, then the ADA4627-1 is a very valuable option… to help adjusting the tonal balance.
I am not sure one can do much better in terms of HIFI than the OPA1656, integrated chip or not (how the world has changed !) as long as it is correctly implemented and used, but let’s see if it reacts to Class A and gets even better, well bearing in mind that air for improvement is very very thin now, at least to my ears.
Claude
In particular there were these 2 excellent Op amps, having a lot of common but offering a slightly different music representation, perhaps the ADA4627-1 small additional airyness/lightness being a tad more seductive (while the OPA1656 groove can be addictive)… but which one is truer ?
I let the 2 op amps burn-in an additional 20h and decided to use my Ipod as source. Hardly HIFI, but the purpose was solely to be able to monitor things. The Ipod can just drive my Beyer to the right volume level, being the souce direct option. Then I can introduce the simple HP amp, adjust the volume on it, so any change to the music presentation comes from that test rig which is mainly a double op amp per channel.
After another 20h of burn-in sound is settled. TBH I can’t tell if it really improved after the initial 5h, as direct comparison is impossible, but my feeling is « negligeable if anything ». It used to burn-in initialy though, the first hour mainly and then slightly so.
This test revealed again how excellent both op amps are and that they have more similarities than differences. However, the ADA4627-1 sounded still a tad more open up, airy and with a higher tonality balance. But in comparison to the direct source it turned out it lightens and brightens things a tad. Ads sparkling, takes off weight. Sounds fun though, but less true to the source.
On the other hand, the OPA1656 turned out to be very very close to the source, neutral in fact. Its tonality and balance are true, details are what they are and its massive punch weight are the real ones. In fact it hasn’t really a dark balance : it has a neutral one and the ADA4627-1 chip is in fact a tad brighter, sparkles and shines a tad (and the ADA4625-1 I turned down becomes though quite bright in return).
This gives us options to adjust the tonal balance. If you want pure HIFI and an excellent component that is neutral and incredibly fun, then the OPA1656 is simply an amazing chip, probably state of the art. If you feel on the other hand your output stage needs a bit of extra sparkling, then the ADA4627-1 is a very valuable option… to help adjusting the tonal balance.
I am not sure one can do much better in terms of HIFI than the OPA1656, integrated chip or not (how the world has changed !) as long as it is correctly implemented and used, but let’s see if it reacts to Class A and gets even better, well bearing in mind that air for improvement is very very thin now, at least to my ears.
Claude
Big difference in price but not necessarily sound quality using your ears, as the test bed 
Factor in that the ADA4627-1 is a single & OPA1656 is a dual
Good info none the less, thanks, JohnC and the the TI/BB team should be very proud of their accomplishments.
1) ADA4627-1ARZ = $7.81 each
2) OPA1656IDR = $2.72 each
What gain are you running in your testing? Power Amp? Beyer = beyerdynamic, what model? headphones driven directly?
Read through the thread a bit, seems you have a few different test setups you are evaluating these devices in.
Factor in that the ADA4627-1 is a single & OPA1656 is a dual
Good info none the less, thanks, JohnC and the the TI/BB team should be very proud of their accomplishments.
1) ADA4627-1ARZ = $7.81 each
2) OPA1656IDR = $2.72 each
What gain are you running in your testing? Power Amp? Beyer = beyerdynamic, what model? headphones driven directly?
Read through the thread a bit, seems you have a few different test setups you are evaluating these devices in.
Hi
If my understanding of the schematic of the Classic 47 is right then gain is 3.1, so quite lowish. Well, just had a quick look at it, didn't play with that anyway, you to check if you want... and yes, driven directly in this 'test rig'.
This is all just for a convenient preselection (to avoid wrecking a delicate board), gain in MY application (eg my Chipo RIAA) will be the one I have chosen for the second stage, and that is at the moment x28 (29dB). No way higher, but I have the possibility to adjust downwards if needed (x15 or even x10). My highish gain didn't affect the noise and it could enable me to drive directly some power amps, in case of...
I understand that different gains (and set ups, not to mention I normaly always listen to loudspeakers) could lead to different findings, hence me just doing my preselection that way, but keeping all parts and trying first the 2 or 3 I retained to hear in what direction the wind blows and if general impressions are indeed confirmed... and what fine tuning might be required. But that's another story, for the other thread preferably, there are other parameters to consider for that RIAA amp
HP are Beyerdynamics DT990, the old and trustfull initial version from the late 80s, made for the pros (not the current DT990 PRO!) - the one with 600R impedance, 1/4" plug and a more linear frequency response than today. Perhaps not the most resolving nowadays but more than enough to assess these differences and still a nice and proven HP for that kind of stuff IMHO. Used it a lot a few decades ago when monitoring opera recordings, another story. Pity they tweaked it later for portable devices and a different market, reducing impedance to 250R and changing some quality bits on the rare 600Rversion, but well...
Sorry, OFF TOPIC!
Claude
If my understanding of the schematic of the Classic 47 is right then gain is 3.1, so quite lowish. Well, just had a quick look at it, didn't play with that anyway, you to check if you want... and yes, driven directly in this 'test rig'.
This is all just for a convenient preselection (to avoid wrecking a delicate board), gain in MY application (eg my Chipo RIAA) will be the one I have chosen for the second stage, and that is at the moment x28 (29dB). No way higher, but I have the possibility to adjust downwards if needed (x15 or even x10). My highish gain didn't affect the noise and it could enable me to drive directly some power amps, in case of...
I understand that different gains (and set ups, not to mention I normaly always listen to loudspeakers) could lead to different findings, hence me just doing my preselection that way, but keeping all parts and trying first the 2 or 3 I retained to hear in what direction the wind blows and if general impressions are indeed confirmed... and what fine tuning might be required. But that's another story, for the other thread preferably, there are other parameters to consider for that RIAA amp
HP are Beyerdynamics DT990, the old and trustfull initial version from the late 80s, made for the pros (not the current DT990 PRO!) - the one with 600R impedance, 1/4" plug and a more linear frequency response than today. Perhaps not the most resolving nowadays but more than enough to assess these differences and still a nice and proven HP for that kind of stuff IMHO. Used it a lot a few decades ago when monitoring opera recordings, another story. Pity they tweaked it later for portable devices and a different market, reducing impedance to 250R and changing some quality bits on the rare 600Rversion, but well...
Sorry, OFF TOPIC!
Claude
The OPA1656 is obviously a world-class performer in all aspects. The question is, due to its extremely fast nature (GBW=55 MHz), what's the best way to keep it from oscillating when used as a replacement IC in older designs? The usual 100nF cap across the power pins is a good start, but is that enough? Will a 100pF cap across the feedback resistor tame it sufficiently without sacrificing performance?
If the circuit itself is regular it's difficult to make it oscillate. However old designs seem to like using tricks that only works with certain opamps. In that sense, it may oscillate. The 1656 itself is not prone to oscillation.
OPA1656
I just chip rolled a Carver C-2 preamp with all OPA1656 (except the headphone amp which was OPA2156, because I wanted R-R input). Worked fantastically.
I added 0.1uF caps on the back side and had not stability problems. I have a 400MHz scope and saw nothing untoward. Used Browndog adaptors.
I just chip rolled a Carver C-2 preamp with all OPA1656 (except the headphone amp which was OPA2156, because I wanted R-R input). Worked fantastically.
I added 0.1uF caps on the back side and had not stability problems. I have a 400MHz scope and saw nothing untoward. Used Browndog adaptors.
ClaudeG:
I had a pair of the original DT-990, also.
They were indeed VERY nice AND very comfortable as well!
Thanks for reminder of good memories!
mlloyd1
I had a pair of the original DT-990, also.
They were indeed VERY nice AND very comfortable as well!
Thanks for reminder of good memories!
mlloyd1
Some progress on the Class A bias…
I did a very dirty build just to decide if it would be worth investigating this potential tweak further in depth or not, before going for the a proper set up, especialy as John said modern chips wouldn’t benefit from this tweak… And I have no reason to doubt him. But as time allows it…
This is a crude test rig in form of an additional socket/ adapter (I know it is not intended to be used like that !) containing 1 transistor and 1 resistor for the Class A biasing of one op amp (single or half a double, output stage). Any (double or 2 x singles) Op amp can be quickly inserted so that allows quick comparisons of the exact same device with and without Class A bias. It also enables biasing or not different op amps nearly on the fly. Downside is I end up with a pile of sockets, that one being sandwiched, but well, just to have an initial flavour of the sonic direction, quick and dirty…
I only tested the 3 contenders I liked best. Bias is 2.6mA: enough to stay in class A with my high impedance, high efficiency HP.
AD825
I am using it in Class A bias since 2001, that’s my long time favourite that way. I do know Class A really trenscends this part. However, finding out if the same magic would work is a good test to know if my additional Class A socket works and does more good than harm, as not a proper « directly sordered on a board » solution. And the magic worked exactly as expected. The sound began to flow, was airy, very much addictive with a lot of drive and a nice integration and consistant fusion of all registers inclusive treble. A very nice representation with a great soundstage, IMHO that Class A bias has here no downside. That sound is very likeable with nor real flaw or sharp corners. It improved a lot, but despite its Class A bias it doesn’t quite reach the level of a simple OPA1656 : it still lacks a bit of precision/transparency, low level noise, 3D soundstage in depth, and a non rolled off treble for that. In fact, it is not entirely neutral and reminds me of (good) old tube implementation – now that I played thanks to Papa with H2, I wouldn’t be surprised if AD825 has a lot of it and once biased in Class A it makes the most of it while diminishing its weakest points (due to its age).
I use AD825 in my 20y old DAC and it will stay there as it is still very good plus it rounds off the weaknesses of older DAC chips while flatering their sound. My guess is OPA1656 might not shine better (or be even too performing) in a weaker devices… while on the other side being the new king of the modern age as far as I am concerned.
ADA4627-1
It did not benefit at all from Class A biasing in my case. It lost completely all its magic and subtelity, sounding somewhat congested, muddy and flat / distorted. This one is IMHO best left alone.
OPA1656
I have no clear cut yet on this one and would tend to say I prefer so far this op amp not biased, so left alone. Note it is an excellent performer as it is, so no harm. IMHO the bias I used does quite affect the sonic signature of this op amp. On one hand, I did like the better bass, with more drive and impact while still managing to be less dry, but also the extra speed so nothing sounded washed off. In fact bass and speed across the register were what I liked. But on the other side I found some of the airyness and 3D was gone, sounds were less flowing and ending sharper, the entire representation had more density and was based on the bass, all was more to the ground, so some magic (subtlelity, agility, air) was lost and it is quite a different experience. I also found the extra speed and sparkle in the treble a tad more tiring long term. All these are nuances, not as big a change as for AD825, but still there. Hmm, I wonder if it asn’t all extra distorsions that I heard ? It also « seemed » to sound a tad louder…
So far OPA1656 benefited in some aeras of the bias, but IMHO in general it lost more than it gained. Now the big question is: did it lost due to the Class A bias… or due to me piling up sockets for this quick and dirty sound check ? Somewhere, I hesitate between leaving experiments as they are / using OPA1656 « stand alone »… and building a proper board this time without additional sockets, having the transistor and resistor directly soldered on the board (no extra adapter), to allow an apple to apple comparison ‘same board with and without Class A bias’. Just to see if what I missed wasn’t due in fact to the chip’s excellence… and possible sensitivity to non sense builds like my test rig (extra contact losses, extra stray capacitance…).
If you guys now better, just let me know, it could spare me some time although I will build another OPA1656 board anyway and adding a few parts directly on its top isn’t difficult… nor irreversible.
Claude
I did a very dirty build just to decide if it would be worth investigating this potential tweak further in depth or not, before going for the a proper set up, especialy as John said modern chips wouldn’t benefit from this tweak… And I have no reason to doubt him. But as time allows it…
This is a crude test rig in form of an additional socket/ adapter (I know it is not intended to be used like that !) containing 1 transistor and 1 resistor for the Class A biasing of one op amp (single or half a double, output stage). Any (double or 2 x singles) Op amp can be quickly inserted so that allows quick comparisons of the exact same device with and without Class A bias. It also enables biasing or not different op amps nearly on the fly. Downside is I end up with a pile of sockets, that one being sandwiched, but well, just to have an initial flavour of the sonic direction, quick and dirty…
I only tested the 3 contenders I liked best. Bias is 2.6mA: enough to stay in class A with my high impedance, high efficiency HP.
AD825
I am using it in Class A bias since 2001, that’s my long time favourite that way. I do know Class A really trenscends this part. However, finding out if the same magic would work is a good test to know if my additional Class A socket works and does more good than harm, as not a proper « directly sordered on a board » solution. And the magic worked exactly as expected. The sound began to flow, was airy, very much addictive with a lot of drive and a nice integration and consistant fusion of all registers inclusive treble. A very nice representation with a great soundstage, IMHO that Class A bias has here no downside. That sound is very likeable with nor real flaw or sharp corners. It improved a lot, but despite its Class A bias it doesn’t quite reach the level of a simple OPA1656 : it still lacks a bit of precision/transparency, low level noise, 3D soundstage in depth, and a non rolled off treble for that. In fact, it is not entirely neutral and reminds me of (good) old tube implementation – now that I played thanks to Papa with H2, I wouldn’t be surprised if AD825 has a lot of it and once biased in Class A it makes the most of it while diminishing its weakest points (due to its age).
I use AD825 in my 20y old DAC and it will stay there as it is still very good plus it rounds off the weaknesses of older DAC chips while flatering their sound. My guess is OPA1656 might not shine better (or be even too performing) in a weaker devices… while on the other side being the new king of the modern age as far as I am concerned.
ADA4627-1
It did not benefit at all from Class A biasing in my case. It lost completely all its magic and subtelity, sounding somewhat congested, muddy and flat / distorted. This one is IMHO best left alone.
OPA1656
I have no clear cut yet on this one and would tend to say I prefer so far this op amp not biased, so left alone. Note it is an excellent performer as it is, so no harm. IMHO the bias I used does quite affect the sonic signature of this op amp. On one hand, I did like the better bass, with more drive and impact while still managing to be less dry, but also the extra speed so nothing sounded washed off. In fact bass and speed across the register were what I liked. But on the other side I found some of the airyness and 3D was gone, sounds were less flowing and ending sharper, the entire representation had more density and was based on the bass, all was more to the ground, so some magic (subtlelity, agility, air) was lost and it is quite a different experience. I also found the extra speed and sparkle in the treble a tad more tiring long term. All these are nuances, not as big a change as for AD825, but still there. Hmm, I wonder if it asn’t all extra distorsions that I heard ? It also « seemed » to sound a tad louder…
So far OPA1656 benefited in some aeras of the bias, but IMHO in general it lost more than it gained. Now the big question is: did it lost due to the Class A bias… or due to me piling up sockets for this quick and dirty sound check ? Somewhere, I hesitate between leaving experiments as they are / using OPA1656 « stand alone »… and building a proper board this time without additional sockets, having the transistor and resistor directly soldered on the board (no extra adapter), to allow an apple to apple comparison ‘same board with and without Class A bias’. Just to see if what I missed wasn’t due in fact to the chip’s excellence… and possible sensitivity to non sense builds like my test rig (extra contact losses, extra stray capacitance…).
If you guys now better, just let me know, it could spare me some time although I will build another OPA1656 board anyway and adding a few parts directly on its top isn’t difficult… nor irreversible.
Claude
Attachments
Hi,
If you use a 2.6mA current source next to the output stage of the Op-Amps I’d say that you force them more into ClassB operation than ClassA.
The OPA1656 has a 3.9mA (typ) quiescent current per channel and per rail : by deducting the input/driver stages’ current consumption, the output stage Idle current will be somewhere around 2.5mA per rail. This leaves a maximum of 5mA peak available current for ClassA operation.
So I think that your 2.6mA current source will decrease the peak available current from the Idle for ClassA operation by more than 50% if you consider the OPA1656 instead of increasing it.
If I didn’t understand something, please explain.
Thanks.
If you use a 2.6mA current source next to the output stage of the Op-Amps I’d say that you force them more into ClassB operation than ClassA.
The OPA1656 has a 3.9mA (typ) quiescent current per channel and per rail : by deducting the input/driver stages’ current consumption, the output stage Idle current will be somewhere around 2.5mA per rail. This leaves a maximum of 5mA peak available current for ClassA operation.
So I think that your 2.6mA current source will decrease the peak available current from the Idle for ClassA operation by more than 50% if you consider the OPA1656 instead of increasing it.
If I didn’t understand something, please explain.
Thanks.
Hi,
Hmm... where to start...
You may want to read this, I guess it explains it pretty well:
Biasing Op-Amps into Class A
There are of course also other sources to explain the difference between Class A and Class B.
In short (and probably badly explained as my time and ability are limited), the idea is to make sure there is always a "current call" at the output stage eventhough music modulation would (in a ususal set up normaly) lead to 0V, hence never 0 crossing at the output stage. As for what the Op amp can deliver (its current delivery), that is far more, so we are far away from its max. / no saturation etc., so probably not the prob either
But at the end of the day, I don't know what is inside that chip (protections, limiters) and may force things, and don't know any other way to tweak it: so I just try it and if it sounds better, than fine... and if not, well... then I just tried
Thanks for replying and thinking about it!
Claude
Hmm... where to start...
You may want to read this, I guess it explains it pretty well:
Biasing Op-Amps into Class A
There are of course also other sources to explain the difference between Class A and Class B.
In short (and probably badly explained as my time and ability are limited), the idea is to make sure there is always a "current call" at the output stage eventhough music modulation would (in a ususal set up normaly) lead to 0V, hence never 0 crossing at the output stage. As for what the Op amp can deliver (its current delivery), that is far more, so we are far away from its max. / no saturation etc., so probably not the prob either
But at the end of the day, I don't know what is inside that chip (protections, limiters) and may force things, and don't know any other way to tweak it: so I just try it and if it sounds better, than fine... and if not, well... then I just tried
Thanks for replying and thinking about it!
Claude
Thank you to send this article which indeed seems interesting but if you read it carefully, the current source is inside the feedback loop with a buffer and not directly connected to the output...
Hi,
Hmm... where to start...
You may want to read this, I guess it explains it pretty well:
Biasing Op-Amps into Class A
There are of course also other sources to explain the difference between Class A and Class B.
In short (and probably badly explained as my time and ability are limited), the idea is to make sure there is always a "current call" at the output stage eventhough music modulation would (in a ususal set up normaly) lead to 0V, hence never 0 crossing at the output stage. As for what the Op amp can deliver (its current delivery), that is far more, so we are far away from its max. / no saturation etc., so probably not the prob either
But at the end of the day, I don't know what is inside that chip (protections, limiters) and may force things, and don't know any other way to tweak it: so I just try it and if it sounds better, than fine... and if not, well... then I just tried
Thanks for replying and thinking about it!
Claude
Yes but that's not really relevant for our purpose - they could have biased the output buffer if they wanted aswell.
It is just this article was handy (quick link for you instead of explaining it all) albeit indeed just an extension to comment an existing product (PIMETA). There are other articles on Class A bias, I believe the oldest I read was from the 70s...
Anyway, whatever we do and however it works (I guess I should switch off the engineer in me), all what matters is: does it sound better or not. In the past it did rather improve the op amps, maybe also because it helped their power supply rejection on the connected rail (look at other articles on the subject).
Anyone else trying this?
I may just spend my time doing the proper board and trying it out one last time instead of thinking and typing to try to fgure out, LOL
Claude
It is just this article was handy (quick link for you instead of explaining it all) albeit indeed just an extension to comment an existing product (PIMETA). There are other articles on Class A bias, I believe the oldest I read was from the 70s...
Anyway, whatever we do and however it works (I guess I should switch off the engineer in me), all what matters is: does it sound better or not. In the past it did rather improve the op amps, maybe also because it helped their power supply rejection on the connected rail (look at other articles on the subject).
Anyone else trying this?
I may just spend my time doing the proper board and trying it out one last time instead of thinking and typing to try to fgure out, LOL
Claude
Powerex, if the topic is of interest to you, you could read the Active Crossover book from Douglas Self, section "Reducing 5532 Distortion by Output Biasing", gives for sure more explanation. You can find the book pretty easily ;-).
Although other articles improved on this base with other trying (was quite popular 2 decades ago when I did it), it gives a pretty good starting point. Another bit is that usualy it is even better to do it from the rail that suffers the worst rejection, although probably not as bad if not despite what Self describes in his very specific application (5532 fan).
'Nough said, back to practice
Claude
Although other articles improved on this base with other trying (was quite popular 2 decades ago when I did it), it gives a pretty good starting point. Another bit is that usualy it is even better to do it from the rail that suffers the worst rejection, although probably not as bad if not despite what Self describes in his very specific application (5532 fan).
'Nough said, back to practice
Claude
Of course but it’s also interesting to know WHY your audio circuits explorations do sound good or not otherwise you’ll just have to note the results without a proper understanding. But this is your choice and it deserves respect...
Anyhow, my initial question remains the same : is it useful to add a current source to the output of an OpAmp if its output stage draws enough Idle current when driving high to medium load impedances which is generally the case in audio applications.
Perhaps that an experimented designer from TI could help ?
Anyhow, my initial question remains the same : is it useful to add a current source to the output of an OpAmp if its output stage draws enough Idle current when driving high to medium load impedances which is generally the case in audio applications.
Perhaps that an experimented designer from TI could help ?
I posted here only because OPA1656 shone, as a copy of what we did in the other thread. In fact these tests were just a first quick and dirty round with other op amps to help me selecting the ones I really wanted to test for my real life application, which is the second stage of the Chipo RIAA preamp and for which the dangerous "amp rolling" is very limited per construction (FET input, low DC offset, other parameters to consider given the filter and output stage etc.).
So for me indeed not worth investigating further here, this has never been intended to be a shoot out to rank op amps (wouldn't work as sensitive to application etc.), I just wanted to come back to John saying he did a vey good job and that this chip deserves indeed some consideration for future built. My day to day job takes me enough time improving technical processes, so back to hobby and enjoying music.
Regarding answers re Class A bias from qualified designers, I am surprised you missed it in this thread (you were an early poster with questions). Here we go, you will find one reply from the man himself, John, post N. 91. That post was the reason why I said we were not really expecting anything from this tweak on that chip. Having said that, I would welcome any feedback, as I woudl have expected 0 change and I have some change, hence me perhaps going identical boards next round, just to understand and for teh sake of the exercice. I posted because I though I might gain time, but by now I realize I should have spend my time trying instead of typing conjectures.
Anyway, I went for it because I couldn't be 100% sure if I had enough current draw for my application hence forcing a given known value myself, plus I don't have experimented if a Monticelli bias gives the same sonic results as my well-known transistor bias. TBH, the only Monticelli I know is an amazing painter and I felt the last sentence of John's post was opening a potential door to try (albeit not very encouraging), which I pushed LOL
But if someone has the time for the theory behind, more than welcome to read the WHY
Meanwhile enjoy music
Claude
So for me indeed not worth investigating further here, this has never been intended to be a shoot out to rank op amps (wouldn't work as sensitive to application etc.), I just wanted to come back to John saying he did a vey good job and that this chip deserves indeed some consideration for future built. My day to day job takes me enough time improving technical processes, so back to hobby and enjoying music.
Regarding answers re Class A bias from qualified designers, I am surprised you missed it in this thread (you were an early poster with questions). Here we go, you will find one reply from the man himself, John, post N. 91. That post was the reason why I said we were not really expecting anything from this tweak on that chip. Having said that, I would welcome any feedback, as I woudl have expected 0 change and I have some change, hence me perhaps going identical boards next round, just to understand and for teh sake of the exercice. I posted because I though I might gain time, but by now I realize I should have spend my time trying instead of typing conjectures.
Anyway, I went for it because I couldn't be 100% sure if I had enough current draw for my application hence forcing a given known value myself, plus I don't have experimented if a Monticelli bias gives the same sonic results as my well-known transistor bias. TBH, the only Monticelli I know is an amazing painter and I felt the last sentence of John's post was opening a potential door to try (albeit not very encouraging), which I pushed LOL
But if someone has the time for the theory behind, more than welcome to read the WHY
Meanwhile enjoy music
Claude
Last edited:
Yes and that’s your opinion and it deserves respect...
The question is :
Are recent Op-Amps such as the OPA1656, which is especially designed for audio, biased in ClassB ?! Or is the Idle current so low that they will leave ClassA (even at maximum Vout) when the load impedance is let’s say 10K at the lowest.
The question is :
Are recent Op-Amps such as the OPA1656, which is especially designed for audio, biased in ClassB ?! Or is the Idle current so low that they will leave ClassA (even at maximum Vout) when the load impedance is let’s say 10K at the lowest.
