There was just a small bit of irony cast into my phrase.
It is even more easier and cheaper to design mediocre amplifiers yielding in higher profit. You should leave the concept that people want to make good things and do right. If that was the case, there was no DIY either!
ok ! now that you provoke me
is this one so expensive to build ?
this should be the Aura va50 integrated amp ... sounds very good. For commercial reasons i guess they saved in the power supply (small transformer, small caps, ecc.) but the sound is excellent (opinion of a friend whose ears i trust completely).
The cost is not in the design ... is in the parts (transformers, caps, enclosures, volume control, even connectors ...). So i guess there is where savings can be done (i.e. parts)
In some case the design is even already available in the application notes for free
But the mastery is in the ability to separate the bad from the good
As i said at the beginning i thought that Rotel small amps were state of the art
... and now i know that they are flawed designs
Last edited:
Look at that HK Citation 12, it was ok in its day, but will post awful numbers by modern standards (Which may or may not have much to do with sound quality, but as a designer you have to meet marketings need for good looking numbers if you want to shift reasonable numbers of units)....
Non degenerated input pair with resistive load, so there is likely quite some distortion there, add a current source and current mirror to improve that and get (a lot) more gain.
Single transistor VAS, make that a darlington for more transconductance.
Feedback resistors are large, so a bit noisy, but I expect that is to try to match the (rather low) input impedance, still it likely dominates the noise performance.
Rather odd bias arrangement for the output stage, but whatever works I guess. I would expect a largeish cap across CE of Q8 would improve matters, but measurement might be indicated).
R19/C8 are IMHO the wrong side of the output inductor, which is odd.
150pF Dominant pole is MASSIVE by todays standards, cannot be arsed to figure the slew rate that implies, but it won't be pretty.
A more recent design would probably have gone for an EF output pair with NPN and PNP devices, but power PNP back then was problematic.
There is no output short circuit protection in that design, which I would have an issue with, but then I do pro audio where surviving a copper nail across the output is kind of basic.
Commercial design has way more constraints then are always obvious, everything from trying to re use a part that you buy in the tens of thousands for another product (Helps keep costs down) to marketing requirements to the cost of metal parts to import and export tariff codes (Sometimes you can tweak a product to fit a different code and save on tariffs into a specific market) to various safety and regulatory issues.
Non degenerated input pair with resistive load, so there is likely quite some distortion there, add a current source and current mirror to improve that and get (a lot) more gain.
Single transistor VAS, make that a darlington for more transconductance.
Feedback resistors are large, so a bit noisy, but I expect that is to try to match the (rather low) input impedance, still it likely dominates the noise performance.
Rather odd bias arrangement for the output stage, but whatever works I guess. I would expect a largeish cap across CE of Q8 would improve matters, but measurement might be indicated).
R19/C8 are IMHO the wrong side of the output inductor, which is odd.
150pF Dominant pole is MASSIVE by todays standards, cannot be arsed to figure the slew rate that implies, but it won't be pretty.
A more recent design would probably have gone for an EF output pair with NPN and PNP devices, but power PNP back then was problematic.
There is no output short circuit protection in that design, which I would have an issue with, but then I do pro audio where surviving a copper nail across the output is kind of basic.
Commercial design has way more constraints then are always obvious, everything from trying to re use a part that you buy in the tens of thousands for another product (Helps keep costs down) to marketing requirements to the cost of metal parts to import and export tariff codes (Sometimes you can tweak a product to fit a different code and save on tariffs into a specific market) to various safety and regulatory issues.
Ok, I take that back: there is a dc path from E1: through the volume pot.
And in this other picture the mosfets are placed as expected.
Still, as said, curious design.
But if your reliable friends ears tell you what amplifiers sounds great, why should you not accept that and enjoy your favourite music?
If Madame Butterfly (performed by Monserat Caballe) shreaks through a pocket radio in a slaughterhouse at top volume, I'm yelling in accord as loud as possible.
Even if I am completely mistaken by the true origin of the soundsource.
That will consolidate my dubious reputation even further!
And in this other picture the mosfets are placed as expected.
Still, as said, curious design.
But if your reliable friends ears tell you what amplifiers sounds great, why should you not accept that and enjoy your favourite music?
If Madame Butterfly (performed by Monserat Caballe) shreaks through a pocket radio in a slaughterhouse at top volume, I'm yelling in accord as loud as possible.
Even if I am completely mistaken by the true origin of the soundsource.
That will consolidate my dubious reputation even further!
Attachments
Hi thanks again. A design not impeccable that can trigger issues ...
anyway is that simple as possible that is stuck in my mind ... i do not like complexity in general. Live is already complex.
I believe in intelligent modding for instance. I read an article about a cd player i think from Denon The author a modder with good reputation was saying that the unit was overbuild, too many op-amps.
It achieved a better sound taking out parts !!!
unbelievable ... maybe they had stock of old opamps to finish ?Another explanation could be that showing a lot of stuff can provide a sense of consistency, the feeling that i am buying a lot of stuff for the money ?
To be honest i am looking for one also because in this forum i have found info that something can be done in order to increase current delivery, a weak point of the unit (i.e. a bigger mains transformer, bigger caps and different output mosfets)
i am not sure to understand this ... you mean that the performance is more important of the audio quality from the playback system ?
On one thing i am sure ... the better the playback system the bigger the pleasure, even with so so recordings.
Thanks for the very valuable advice. Again back to the starting point. I mentioned the Citation 12 because i read very positive opinions about its sound that i understand now do not are backed equally good instrumental evidences ?
It is a confusing situation ... we can agree i think of the lack of correlation between measurements and sound quality ... then why manufactures keep on showing very good measurements and magazines print lab reports ?
not all i have to say ... someone actually skips entirely the lab work
They base their judgements only on listening tests.
very very difficult but also very very interesting. One main issue that i have tested myself is stability. It is very important for a design.
Clearly when we talk of sound quality we enter the field of subjectivity.
What one likes can be unpleasant to another ...
Measurements are very objective ... but tell nothing about the sound
In conclusion stability first and second the sound
As i guess the Citation 12 must be quite stable (in an article Nelson Pass mentions its excellent reliability) and has a good sound it could be a very good choice for setting up a nice playback system
Last edited:
I would disagree that measurements and sound quality are unrelated...
I have never met someone who could identify sound quality issues in a properly done test that I could not then eventually measure (Sometimes it takes some work).
Now a lot of the datasheet stuff is of course not the numbers you want, because the datasheet stuff is the numbers that are easy to make look good, and are the things that are easy to reduce to a simple number...
THD @1kHz for example, measured (surprise, surprise) down where there is plenty of open loop gain so the feedback can be expected to cure most ills, show me the distortion Vs frequency and power 3D surface, not the data at a single point, it would be far more meaningful.
Damping factor? 1000 Vs 5000, I mean, who cares, really?
Nobody publishes dynamic test results, no two or three tone IMD plots, nothing about clipping recovery, no tone bursts, and these are where the interesting data lies.
Reactive loads, something else the datasheets don't talk about, here be bad sound if the VI limiter activates.
Hell, voltage and current headroom over the design power levels are actually more then slightly meaningful with real loads and real program content.
I have 20W amplifiers (Continuous) that can push 40A or 80V into a reactive load for 10s of ms at a time.
You could not market such a thing because the FTC rules say RMS 'POWER' is what you must specify and it costs real money to do it right, but this is DIY I can build whatever makes me happy.
Also, the performance is FAR more important then the technical quality, always has been, probably always will be. Consider say the Koln Concert, where they used the wrong Piano never mind the wrong mics, and it is still an awe inspiring performance.
I have never met someone who could identify sound quality issues in a properly done test that I could not then eventually measure (Sometimes it takes some work).
Now a lot of the datasheet stuff is of course not the numbers you want, because the datasheet stuff is the numbers that are easy to make look good, and are the things that are easy to reduce to a simple number...
THD @1kHz for example, measured (surprise, surprise) down where there is plenty of open loop gain so the feedback can be expected to cure most ills, show me the distortion Vs frequency and power 3D surface, not the data at a single point, it would be far more meaningful.
Damping factor? 1000 Vs 5000, I mean, who cares, really?
Nobody publishes dynamic test results, no two or three tone IMD plots, nothing about clipping recovery, no tone bursts, and these are where the interesting data lies.
Reactive loads, something else the datasheets don't talk about, here be bad sound if the VI limiter activates.
Hell, voltage and current headroom over the design power levels are actually more then slightly meaningful with real loads and real program content.
I have 20W amplifiers (Continuous) that can push 40A or 80V into a reactive load for 10s of ms at a time.
You could not market such a thing because the FTC rules say RMS 'POWER' is what you must specify and it costs real money to do it right, but this is DIY I can build whatever makes me happy.
Also, the performance is FAR more important then the technical quality, always has been, probably always will be. Consider say the Koln Concert, where they used the wrong Piano never mind the wrong mics, and it is still an awe inspiring performance.
Very interesting indeed. Thank you. The only reason that i can think of about why the magazines do not carry out the tests you mention is that many exotic products would come out quite badly.
And this would mean less advertisement i.e. less money for the magazine.
I have come to the conclusion that when DIY is done right can be very competitive with commercial products because even considering the scale just the cost of parts would make the commercial product very expensive.
If it is true that cost of parts amount for the 1/10th of the MSRP this should mean that the commercial equivalent of a diy unit should be lets say 5-6 times the diy unit cost.
I am quite convinced that very expensive commercial units can sound really good ... but the price to get them is also very high.
Last edited:
Rule of thumb for the pro market is that we expect a 70% gross margin.
1/10th MSRP (Nobody really pays list) seems about right, but don't forget that we pay Digikey/Mouser prices for stuff, if you can buy 10k units per year of say power transistors and place the contracts 6 months in advance then you find that Mouser and Digikey are working at probably a 25-50% gross margin, so your parts cost is a factor of at **least** two (and more often four) higher then the manufacturers who are doing bulk orders on long lead times.
This means that your actual parts cost is (even for the good stuff) only a factor of two or so less then the buy it off the shelf cost, and that assumes your first one works (Not a safe assumption).
1/10th MSRP (Nobody really pays list) seems about right, but don't forget that we pay Digikey/Mouser prices for stuff, if you can buy 10k units per year of say power transistors and place the contracts 6 months in advance then you find that Mouser and Digikey are working at probably a 25-50% gross margin, so your parts cost is a factor of at **least** two (and more often four) higher then the manufacturers who are doing bulk orders on long lead times.
This means that your actual parts cost is (even for the good stuff) only a factor of two or so less then the buy it off the shelf cost, and that assumes your first one works (Not a safe assumption).
Thanks for the info and sorry i have gone OT with this discussion on costs.
Fact is that i was hugely and stupidly underestimate the whole design process
And i was giving for granted the quality of some commercial units that have been sold for years
Surely it takes a very educated eye to spot flaws in a design
I surrender ... way beyond my understanding abilities
Thanks a lot again
Last edited:
ginetto61,
It took a very fine and nice discussion here just over 150 contributions and over 6000 views for an issue what is otherwise never addressed. Thank you for opening that up with a 'simple question'!
We all have our preferences and talents. They may or not coincide and sharing opinions and knowledge will bring everybody a step (or a leap) closer to what one aims to reach in life. There are other ways, and history has proven records of many such disastrous examples. Your talent and contributions has challanced a lot DIY-gurus and experts; the reached concensus, as far as I take it, has pacified the battefield of prejudices and biases (!).
This summer, I visited the ruins of Diana's temple at Nemi (just a bit south of Rome), a very remote and distant remnant of times when the order in the world was more grim and strict. I enjoyed it being there in our time, and not needing a golden bow for access (*). We have all gained a lot.
(*) Frazer, The Golden Bow.
It took a very fine and nice discussion here just over 150 contributions and over 6000 views for an issue what is otherwise never addressed. Thank you for opening that up with a 'simple question'!
We all have our preferences and talents. They may or not coincide and sharing opinions and knowledge will bring everybody a step (or a leap) closer to what one aims to reach in life. There are other ways, and history has proven records of many such disastrous examples. Your talent and contributions has challanced a lot DIY-gurus and experts; the reached concensus, as far as I take it, has pacified the battefield of prejudices and biases (!).
This summer, I visited the ruins of Diana's temple at Nemi (just a bit south of Rome), a very remote and distant remnant of times when the order in the world was more grim and strict. I enjoyed it being there in our time, and not needing a golden bow for access (*). We have all gained a lot.
(*) Frazer, The Golden Bow.
Hi ! thank you for your kind words
I have understood that i am prejudicial ... not a very intelligent way to approach problems. Mainly a result of ignorance i guess.
I was taken for granted things not true.
Thank you very much to you and all for the very valubale advice.
Kindest regards, gino
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.