About design philosophy _ are minimalists right or not ?

[gino1961]

It can be as easy as duplicating the chip manufacturer's circuit suggestions from datasheets such as this one for Texas Instrument's DRV603.

hi thank you very much for this very helpful link I think i will try this for sure
Considering the small power consumption and the low voltage a power supply based on batteries could be a very interesting option
I am using small portable headphone amps powered by a battery to watch the TV and the absence of noise impresses me a lot
Unfortunately the crcuits are so small that is difficult to work on them I was thinking to cannibalize their guts put them in a box and make a line preamp with a better pot and panel connectors
The chip you mention has excellent performance indeed
I will try it soon

 

[gino1961]

I like simple.

hi congratulations for your spectacular system this is the approach i love at least for speakers They are really really nice indeed
Have you ever tried a solid state amplification with it ? for sure these speakers must be exceptionally revealing of any issue with the amplification
i would try a simple but good sounding class A integrated maybe ? a one box solution
I love tubes for mids and highs
For the bass Hz i prefer the damping provided by a solid state amp

 

[gino1961]

I owned one of the Zen power amps. https://www.passdiy.com/project/amplifiers/the-zen-amplifier

Quite the simple circuit for the output stage.
View attachment 364733

hi thanks a lot for the kind suggestion Did you enjoy its sound ? still using these days ? was the heat so high ?
i have seen the old Pass Aleph 1 and i am surprised to see no ventilation holes in the box I am sure that with some big holes in the bottom and on top the temperature could rise less Maybe also adding a low noise fan above the output stages
i am very curious about the advice of persons who have owned very good sounding units and what they ended with
I am always trying to get profit from other people experience and knowledge

 

[gino1961]

Me too in 1997. I tried a similar approach with bipolar transistors and found out fast that to get sufficient gain for negative feedback you have to add current sources, for better performance get a differential input stage, then improve it with Wilson current mirrors, and still the DC performance was meagre and very temperature dependent.

Hi thanks a lot I am attaching the circuit i simulated It is just a gain block The recommended Vsupply is +60VDC but also with +48VDC the distortion spectrum of a 1kHz sine was very satisfying with a 2nd order harmonic at -90dB
I used other bjts because i did not find the models of the originals But the option of different bjts pairs is endless
The original preamp is the Radford ZD22 The line stage is made complex for the presence of a tone control stage between an input amp and an output amp that share the same topology This preamp was used as a reference by an Italian audio magazine in the late '70s
if i can get sufficient performance from a 2 bjts circuit i would not be motivated to try a 20 bjts design for sure
Regarding current sources i understand it is possible to get very low distortion also without them I read a lab report confirming this
Again the problem will be to check if the instruments will confirm the good results obtained with the simulation

Looking at the last circuit I'd tried I realized that it makes much more sense to use an opamp which contains all this and more in a thermocoupled package with tightly matched components, and all this for almost no money. I just kept a power Mosfet for sufficient output current.

Yes i will try also the opamp i have been suggested I like the fact that it works also with low voltage and low consumption
I could use batteries and therefore avoid the power supply design and build

 

[RayDunzl]

A complex discrete (for audio) + digital control (for remote, volume, switching, display) preamplifier from 20 years ago.

All those little black things are transistors.

Single ended and balanced in and out with voltage or current drive outouts with two zones for simultaneous separate (or same) input and output.

I wouldn't bother but I have one so might as well show.

 

[gino1961]

A complex discrete (for audio) + digital control (for remote, volume, switching, display) preamplifier from 20 years ago.

All those little black things are transistors.

Single ended and balanced in and out with voltage or current drive outouts with two zones for simultaneous separate (or same) input and output.

View attachment 364860

I wouldn't bother but I have one so might as well show.

Hi wow this is electronic art at its best
Just to be clear I admire deeply this type of units that could end the game
To be perfectly honest I like a little more the two boxes solution with the Power supply separated from the amplification circuits
But this is a complete control unit Very nice Gorgeous

 

[Blumlein 88]

hi thanks a lot for the kind suggestion Did you enjoy its sound ? still using these days ? was the heat so high ?
i have seen the old Pass Aleph 1 and i am surprised to see no ventilation holes in the box I am sure that with some big holes in the bottom and on top the temperature could rise less Maybe also adding a low noise fan above the output stages
i am very curious about the advice of persons who have owned very good sounding units and what they ended with
I am always trying to get profit from other people experience and knowledge

What I had was the commercial version, the Aleph 0. It was one big cubic heat sink. It got very, very hot. 30 wpc @ 8 ohms, single ended MOSFET. It sounded good on my Quad ESL63 speakers. It was okay, but underpowered on a couple different Magnepans. It seemed to fall apart on any ported speaker. Lent it to friends. It was terrible on Thiels, Mirage M3si, even fell apart on some little Radio Shack Linaeum speakers. On ported speakers it wasn't just power, the Aleph sounded bad at any volume. After satisfying my curiosity I sold it to someone who had LS3/5a speakers. Near that time I had a 35 watt triode amp. It was much more capable on most speakers if limited in power itself.

Pass Laboratories Aleph 0 monoblock power amplifier

The Aleph Null, or 0, represents Nelson Pass's maiden product under the Pass Laboratories banner. When he left Threshold several years ago, Pass had the luxury of starting over with a clean slate, and decided immediately that he wanted to design a single-ended MOSFET amp. The result is aptly...

www.stereophile.com

If you read this you'll note the reviewer used it with ESLs, Maggies and his own design of a sealed box speaker. Did he try it with ported units and find the same problem? Don't know.

 

[gino1961]

Hi thank you for the very interesting reply
I like a lot this kind of aesthetic Very like industrial but attractive design Even if the best way to listen music is in complete darkness
I understand that today extremely high power MOSFET are available Like 300w MOSFET
Maybe a different choice of output device could improve performance ?
Or maybe using more output devices in parallel
I have seen this solution adoted in some high end tube preamps from Conrad & Johnson
It's all beyond my knowledge by the way
I would have chosen the right speakers after the amp
Others prefer to choose speakers first and the an amp that can make them sing

 

[rank]

A straight wire with gain.

 

[roog]

Douglas Self’s book ‘Small Signal Audio Design’ gives a good explanation as to why the simplest solution is not best.

 

[Barrelhouse Solly]

My design experience is mostly with programming. There the simplest design that meets all of the requirements is the best. Some times a complicated design is necessary. Some people seem to be predisposed toward complexity because it seems clever to them. A general principle is that the fewer moving parts, the fewer points of failure. The thing to watch in any design is that the components and interfaces between components need to be as robust as possible. A single point of failure that's fragile is no good.

 

[gino1961]

A straight wire with gain.

hi yes that should be the goal The problem is that no wire has gain So the comparison is impossible
Personally i would look at lab test like Mr Amir is normally doing
The APx555 is a phenomenal tool to check how much an active device adds to a signal

 

[gino1961]

Douglas Self’s book ‘Small Signal Audio Design’ gives a good explanation as to why the simplest solution is not best.

Hi thank you very much for the kind and valuable advice
Usually minimalist circuits have worse measurable performance compares to more complex design
Once agreed on the validity of lab testing subjective opinions lose significance

 

[gino1961]

My design experience is mostly with programming. There the simplest design that meets all of the requirements is the best. Some times a complicated design is necessary. Some people seem to be predisposed toward complexity because it seems clever to them. A general principle is that the fewer moving parts, the fewer points of failure. The thing to watch in any design is that the components and interfaces between components need to be as robust as possible. A single point of failure that's fragile is no good.

hi thank you very much for your precious advice But lets take opamps They are single chips but extremely complex internally
Some can have tenth of transistors The very best opamps provide measurable performance almost impossible to achieve with discrete designs
in terms of THD and noise suppression for instance

 

[fatoldgit]

My design experience is mostly with programming. There the simplest design that meets all of the requirements is the best. Some times a complicated design is necessary. Some people seem to be predisposed toward complexity because it seems clever to them. A general principle is that the fewer moving parts, the fewer points of failure. The thing to watch in any design is that the components and interfaces between components need to be as robust as possible. A single point of failure that's fragile is no good.

The other point about software design/implementation is its mission critical'ness***

You can have a simple design but if its mission critical then you can add 50% or more code on top that is "simply" there to ensure errors are caught, self corrected/retried if possible and if not, that the process is shutdown in an elegant way such that it can be restarted from its failure point with zero loss of integrity.

It's this lack of attention/adversion to even the most basic level of security/error/process checking that is at the root cause of most large scale hacks these days.

As an old grey beard I wont start a "kids today" diatribe but to simply note that back in my day, before world wide networks become ubiquitous, commercial software needed to be as correct as possible from day one because the ability to quickly/easily send a patch out to 100's/1000's/10,000's customers didnt exist which I think today (given this ablity now exists) leads to a lack of focus on some of the error checking/security basics safe in the knowledge that fixes can be deployed very easily.

And of course modern development methods such as Agile dont help... time to market is all important so get something out the door and worry about blowback later.


Peter

*** noting mission critical is relative to the application... a banking transaction back end is just as critical to the business as a car control bus is to the driver as is car dealership software to the profit of 15,000 car dealers(see CDK Global hack)

 

[Waxx]

I had some discussion about this with an software architect i know (who works in banking software for years), and he said;

"it need to be as simple as possible to work within the desired parameters, which sometimes require very complex structures. But every line of code must have a purpose and be efficient in getting there, or it needs to be removed or replaced"

And i think he is right, and it should be a general principle for all engineering. Keep it simple if you can, but never compromise on quality or miss the required functionality on all levels. And sometimes that requires a very complex device.

 

[rank]

Hi thank you very much for the kind and valuable advice
Usually minimalist circuits have worse measurable performance compares to more complex design
Once agreed on the validity of lab testing subjective opinions lose significance

I admit that I haven't read the 700+ pages of Self's book. I understand that he uses 37 5532s per channel (unsoldered for ease of maintenance) and that he's a big fan of negative feedback, doesn't care much about damping factor... OK.
You know, each designer goes in his own path (Jean Hiraga, Charles Henri Delaleu, Nelson Pass or Akito Kaneda, to name just a few). At our level, what role will this play?
I no longer build electronics (in the past I have used two module-based amps, still working perfectly).
But I think now it's of no interest because I only use solid state.
And you can find them cheaply at huge discounts or second-hand.

It's different for "tubeys" amateurs though, primarily to reduce the cost and use wired cabling like Jadis gear -but 15 times cheaper-.

The problem is that no wire has gain So the comparison is impossible
Personally i would look at lab test like Mr Amir is normally doing
The APx555 is a phenomenal tool to check how much an active device adds to a signal

I should have added "all things being equal". You didn't understand, no matter.

 

[gino1961]

I admit that I haven't read the 700+ pages of Self's book. I understand that he uses 37 5532s per channel (unsoldered for ease of maintenance) and that he's a big fan of negative feedback, doesn't care much about damping factor... OK.

hi and thanks again for the very valuable advice I understand that there is a point after which any further improvements will not provide and audible difference I did not understand completely which is this point
But for instance i would like to start my own measurements of commercial preamps and some kits that i am going to put together
I am looking for a signal generator and i realize that cheap ones have much high THD of the units that i should measure
A good entry point for very low THD would be the QA403 audio analyzer but it costs about 800 euro
I am not a pro so that price is a little too much for me I am sure that is a very high quality device

You know, each designer goes in his own path (Jean Hiraga, Charles Henri Delaleu, Nelson Pass or Akito Kaneda, to name just a few). At our level, what role will this play?

yes i understand that They have followers like it were a religion some times The main discussion is about the importance of measurements
Measurements are very objective while listening impressions are not That is what generates discussions and debates

I no longer build electronics (in the past I have used two module-based amps, still working perfectly).
But I think now it's of no interest because I only use solid state.

And you can find them cheaply at huge discounts or second-hand.
It's different for "tubeys" amateurs though, primarily to reduce the cost and use wired cabling like Jadis gear -but 15 times cheaper-.
I should have added "all things being equal". You didn't understand, no matter.

me too I could only experiment a tube in the VAS and stop I have a friend who tried a tube power amp with issues that i would like to avoid
Solid sate when done well could be perfectly adequate And long lasting I have an old Sansui with 50 years still working
Maybe i should recap it to be safer

 

[Barrelhouse Solly]

I had some discussion about this with an software architect i know (who works in banking software for years), and he said;

"it need to be as simple as possible to work within the desired parameters, which sometimes require very complex structures. But every line of code must have a purpose and be efficient in getting there, or it needs to be removed or replaced"

And i think he is right, and it should be a general principle for all engineering. Keep it simple if you can, but never compromise on quality or miss the required functionality on all levels. And sometimes that requires a very complex device.

The required level of complexity is directly related to the complexity of the problem. The thing to avoid is complexity for complexity's sake.

 

[fatoldgit]

I had some discussion about this with an software architect i know (who works in banking software for years), and he said;

"it need to be as simple as possible to work within the desired parameters, which sometimes require very complex structures. But every line of code must have a purpose and be efficient in getting there, or it needs to be removed or replaced"

And i think he is right, and it should be a general principle for all engineering. Keep it simple if you can, but never compromise on quality or miss the required functionality on all levels. And sometimes that requires a very complex device.

The "issue" (as opposed to the old days) is hardware today is not constrained in the way it was (memory, cpu, bus speeds etc) so programmers today generally dont worry about code efficiency (except in specialist low level software such as database engines etc).

Thus the art of writing "lean and mean" code has been lost to a high degree cause the hardware constraints that forced this on us back in the day no longer exists.

Plus on top of that, lots of software is abstracted up into tool chains/runtime engines that are many levels removed from say a 'C' program that compiles directly into an executable.

This is not to say that say using a GUI engine is bad... if you want a modern interface its needed but each layer of abstraction, obviously, leads to inefficiencies.

I am very glad I could have a good career leveraging mainly 'C/C++' skills..I look at programming job requirements today and you need a huge list of product skills/knowledge to even remotely qualify.

Peter "Get off my Lawn" Grey Beard.