Sound Quality Vs. Measurements

Status
Not open for further replies.
diyAudio Member RIP
Joined 2005
This wiki is o.k., although I dissent that Nelson really invented the concept, as such feedback goes way back (a tube amp with feedback to the input device cathode for example) --- and I believe that Comlinear lost some battles after that patent issued.

Current-feedback operational amplifier - Wikipedia, the free encyclopedia

Good for high slew rate, usually not so good for d.c. properties and distortion. Would represent a huge sea change in the context of the Hafler at this point.

Does anyone know of a high-end power amp that uses this approach? I vaguely recall some such but forget who it was, maybe dating back to the '80's when fast slewing was all the rage. On a bog-standard level the NAD3020 power amp section uses current feedback, and an ancient IC from STMicro, the TDA2007A, tried to have a current-feedback front end and additional voltage gain following, and was the devil to stabilize.
 
That's what I'd do. Sharing of resources somehow always brings its own baggage, better not play with it, especially since the added complication and cost is truly negligible.

Simple enough. I had it drawn that way originally, but something keeps bugging me. There was an explanation somewhere why one actually did want to share, but I can't find it anywhere. Whatever it was, it is probably in the weeds anyway.
 
Not really topology, but doing the transient response, it is clear I need to up the feedback coupling cap to 220u, replace most of the resistors with 1%, and add diodes with larger caps on the rails for the VAS and IPS, with a rail to rail bleeder. Electro on the input will be replaced by a large poly. Obvious stuff.

If one built the amp as a current source rather than a voltage source, a-la Quad, than would one not want current feedback? I thought that argument kind of went flat not because it was not a good idea, but because most speakers are designed assuming a voltage source.
 
Simple enough. I had it drawn that way originally, but something keeps bugging me. There was an explanation somewhere why one actually did want to share, but I can't find it anywhere. Whatever it was, it is probably in the weeds anyway.

Frankly, for the life of me, I can't think of any more but two reasons:

1. Less components, less real estate on the PC borad required, and

2. Less components, smaller price for materials and workmanship.

We often laugh at the price savings, and point out that at the volumes factories do their buying, a resistor, zener diode and capacitor probably don't cost them more $0.05 to buy - if that.

What we forget is that this saving is multiplied by the number of units produced. Say 50,000 - that works out at (50,000 * 0.05) $ 2,500.

Now, what do you think, off hand, what would a company rather do, lay down the money for extra parts which no-one can even see, or "invest" that money in advertising?

We here are in many ways the antipods of the industry. Just look at Thorsten and myself, here we are, using as many transistors as we feel like, in pursuit of what we hope will turn out be some cools sounds, but we are being damn wasteful from an industrial point of view. We do it because we are working for ourselves here, giving no thought to series production, and having no bean counter peering over our back.

In the actual industry, such things are reserved for only the most exalted of brands, whose price stickers NEVER ever have less than four figures and more often than not, five figures.

So, as I understand it, you are doing this strictly for yourself - indulge yourself then, if you have it, flaunt it, do it RIGHT or don't bother.
 
Not really topology, but doing the transient response, it is clear I need to up the feedback coupling cap to 220u, replace most of the resistors with 1%, and add diodes with larger caps on the rails for the VAS and IPS, with a rail to rail bleeder. Electro on the input will be replaced by a large poly. Obvious stuff.

If one built the amp as a current source rather than a voltage source, a-la Quad, than would one not want current feedback? I thought that argument kind of went flat not because it was not a good idea, but because most speakers are designed assuming a voltage source.

As I see it, current feedback, like any approach, has its ups and downs.

Up is the fact that it is more or less independent of voltage swing, works equally well with 1V and 20V output. Thus, your full bandwidth is preserved at all times.

Down is the fact that it is not easy to temper its DC aspects, needs more care and possibly a few more parts.

Voltage feedback is about the same, but in opposineg direction. Relatively easy to adjust a circuit's DC parameters, but dependant on voltage swing.

Don't worry about it. If you do, you will soon be asking yourself what the hell are you doing, modifying an amp well into radical waters, it would have been easier to build another one from scratch.
 
diyAudio Member RIP
Joined 2005
If one built the amp as a current source rather than a voltage source, a-la Quad, than would one not want current feedback? I thought that argument kind of went flat not because it was not a good idea, but because most speakers are designed assuming a voltage source.

Be careful not to confuse "current feedback" with "current source" or "current dumping" a la Quad.
 
Frankly, for the life of me, I can't think of any more but two reasons: ......
So, as I understand it, you are doing this strictly for yourself - indulge yourself then, if you have it, flaunt it, do it RIGHT or don't bother.

Having been through the Troy Institute for Maneuverability, and my first career as a manufacturing quality engineer, I don't forget the fractions of a cent. ( One could have a field day on cost of manufacturing of any amp I have seen. Little screws, bolts and nuts, too many assemblies) For the hobby side, I just think if it will be hard to bugger the board. There was some subtle technical reason. I will post it if I ever find it. Could have been somewhere in 20 years of AA.

I have another philosophy that has served me well. Fix the first 95% of you biggest problem and move on. Eventually, that last 5% will be your biggest problem. If it needs fixing, do so then. Why leave a problem? Well, if one gets carried away with tiny details, they may never get to the next really big problem. I see it all the time.
 
This wiki is o.k., although I dissent that Nelson really invented the concept, as such feedback goes way back (a tube amp with feedback to the input device cathode for example) --- and I believe that Comlinear lost some battles after that patent issued.

Current-feedback operational amplifier - Wikipedia, the free encyclopedia

Good for high slew rate, usually not so good for d.c. properties and distortion. Would represent a huge sea change in the context of the Hafler at this point.

Does anyone know of a high-end power amp that uses this approach? I vaguely recall some such but forget who it was, maybe dating back to the '80's when fast slewing was all the rage. On a bog-standard level the NAD3020 power amp section uses current feedback, and an ancient IC from STMicro, the TDA2007A, tried to have a current-feedback front end and additional voltage gain following, and was the devil to stabilize.


Accuphase did ......!
 
This wiki is o.k., although I dissent that Nelson really invented the concept, as such feedback goes way back (a tube amp with feedback to the input device cathode for example) --- and I believe that Comlinear lost some battles after that patent issued.

You are right, but at the time those pointing this out were considered envious neighbobs. I can't recall any IC company paying royalties on these patents.
 
You are right, but at the time those pointing this out were considered envious neighbobs. I can't recall any IC company paying royalties on these patents.

Perhaps it's more about the ego than the money.

I have always suspected Nelson Pass of a whopping ego. I can't legally prove it, but I resent his ideas peddled to the DIY community that one MOSFET can do it all, yet he builds his own amps with a multitude of output devices. His Threshold models used to have 30+ output transfitors.
 
diyAudio Member RIP
Joined 2005
Perhaps it's more about the ego than the money.

I have always suspected Nelson Pass of a whopping ego. I can't legally prove it, but I resent his ideas peddled to the DIY community that one MOSFET can do it all, yet he builds his own amps with a multitude of output devices. His Threshold models used to have 30+ output transfitors.

Different Nelson --- last name not first. See the wiki.
 
Perhaps it's more about the ego than the money.

I have always suspected Nelson Pass of a whopping ego. I can't legally prove it, but I resent his ideas peddled to the DIY community that one MOSFET can do it all, yet he builds his own amps with a multitude of output devices. His Threshold models used to have 30+ output transfitors.

Yes it's Carl Nelson of Comlinear, no need to vent at NP it will make you no friends here.
 
Perhaps it's more about the ego than the money.

I have always suspected Nelson Pass of a whopping ego. I can't legally prove it, but I resent his ideas peddled to the DIY community that one MOSFET can do it all, yet he builds his own amps with a multitude of output devices. His Threshold models used to have 30+ output transfitors.

Way off base on this one DVV... :rolleyes:
 
There is some irony here, as I put in my order to Digi-Key I was thinking of all the help I have received in an effort to "POOGE" an old Hafler. A process coined by, Mr. Pass.

His one transistor wonders are for those that need a 6W amp for their Louthers. He serves a market. If you need 100W, I am sure he can accommodate that after you get the mains to your house beefed up. If you read his papers, he is big on not making distortion in the first place, so many outputs are typical of his work. Even the original 60W Adcom 353 had triples. Actually, I am really tempted to snag a PA-5 or Forte off the Web.
 
His one transistor wonders are for those that need a 6W amp for their Louthers. He serves a market.

I personally was an active supporter of one transistor concept, and struggled for it at russian forums. Zen9 and FirstWatt F3 are not bad at all. But this concept has revealed its inherent limits to me, after several years of efforts. Not because of low-power. Even with best Louthers, this concept does not allow to approach top leage sound, in resolution, microdetails and bass articulation. I could understand, why some designers could wonder about amount of supporters and real merits of the approach. But this does not mean any claim to Nelson, I am thankful to him for his role in DIY community. Everybody should go his own route.
 
I personally was an active supporter of one transistor concept, and struggled for it at russian forums. Zen9 and FirstWatt F3 are not bad at all. But this concept has revealed its inherent limits to me, after several years of efforts. Not because of low-power. Even with best Louthers, this concept does not allow to approach top leage sound, in resolution, microdetails and bass articulation. I could understand, why some designers could wonder about amount of supporters and real merits of the approach. But this does not mean any claim to Nelson, I am thankful to him for his role in DIY community. Everybody should go his own route.

Vladimir, far from me that I shoud pronounce anyone to the Devil, and Nelson Pass has been very kind to me personally on two or three occasions I contacted him. In terms of personal honour, I owe him.

But I do NOT believe the "single transistor" approach can provide truly satisfactory and consistent results. Locally, people went literally mad over Nelson's concept, so I have had many opportunities to hear his ideas at work. At best, it was all right, but one thing remained constant - his designs are rather sensitive to their work loads.

A more difficult speaker will deliver VERY different results from a tame one. As a firm believer that the amp should always be a constant voltage device, to me, this is completely unacceptable.

But if their owners are happy, who am I to complain?

It's just not my cup of coffee, that's all.
 
Status
Not open for further replies.