I do find it rather disappointing that a thread such as this has managed to descend in parts to a point that is rather distasteful.
It is quite clear what the intention of the original post is, to show an example of a well built amplifier in an era where profit margins and compact form factors seem to dictate more about how an amplifier is designed.
"How to make a proper amplifier" isn't a thread title that's supposed to be taken too seriously, it was written half in jest, it isn't supposed to be definitive and is most certainly a large generalisation. The point of the thread was simply to show off some good construction and build techniques inside a commercial product as a means of being instructive, yes, but mainly just for fun. It isn't to be taken too seriously.
The one thing that puzzles me however is why some of you seem to think that reliability comes at the expense of sound quality.
It is quite clear what the intention of the original post is, to show an example of a well built amplifier in an era where profit margins and compact form factors seem to dictate more about how an amplifier is designed.
"How to make a proper amplifier" isn't a thread title that's supposed to be taken too seriously, it was written half in jest, it isn't supposed to be definitive and is most certainly a large generalisation. The point of the thread was simply to show off some good construction and build techniques inside a commercial product as a means of being instructive, yes, but mainly just for fun. It isn't to be taken too seriously.
The one thing that puzzles me however is why some of you seem to think that reliability comes at the expense of sound quality.
ok then ...
In the age of 16 i constructed an amplifier made from smart kit ( since you are Greek i presume that you know what i am talking about ) though the same amplifier is also sold in UK from quasar electronics . here is a link
http://www.quasarelectronics.com/kit-files/smart-kit/1033.pdf
The amplifier was constructed for a costumer and i was privileged enough to have a teacher at 16 to explain to me what to truss from the kit and what to keep
--replace the resistors with precision ones
--truss the capacitors and place better ones
--form a bit better input filter
-- truss the ceramic miller caps and use styroflex
--form a better vbe multiplier
--add a zobel
--use a beefy power supply + heatsinks
--construct everything clean and in order
--scope the blody thing after finished the construction
The amplifier found its place in a seasonal shop located in Athens Filolaou 124 Pagrati area, called "Market " phone 210-7016499 and the duty was to play background music loaded with 12 small 6'' 2way speakers in combination of series and parallel and working hours was often something like 7-11
In the first years of operation they used tuners or tapes for a source right after used CD players and now days they use computer .In the first years the one that shut down the system shuts also down the amplifier ....
Now days the blonde that is in the office when shut down the PC thinks that the all system is also shutting down ( hell yeah there is no music any more so it is shut down enough for the blonde )
I can assure you that the amplifier is still working ...how well i dont know but still working ( you may as well check for your shelf )
Now days i am 47 and still learning about amps ...make the math ...the amp operates for more than 30 years and also has almost 10 years to shut down completely...
To my understanding the specific amplifier already covered 100 years of domestic use and hell yeah i am proud of it !!!!!
In the age of 16 i constructed an amplifier made from smart kit ( since you are Greek i presume that you know what i am talking about ) though the same amplifier is also sold in UK from quasar electronics . here is a link
http://www.quasarelectronics.com/kit-files/smart-kit/1033.pdf
The amplifier was constructed for a costumer and i was privileged enough to have a teacher at 16 to explain to me what to truss from the kit and what to keep
--replace the resistors with precision ones
--truss the capacitors and place better ones
--form a bit better input filter
-- truss the ceramic miller caps and use styroflex
--form a better vbe multiplier
--add a zobel
--use a beefy power supply + heatsinks
--construct everything clean and in order
--scope the blody thing after finished the construction
The amplifier found its place in a seasonal shop located in Athens Filolaou 124 Pagrati area, called "Market " phone 210-7016499 and the duty was to play background music loaded with 12 small 6'' 2way speakers in combination of series and parallel and working hours was often something like 7-11
In the first years of operation they used tuners or tapes for a source right after used CD players and now days they use computer .In the first years the one that shut down the system shuts also down the amplifier ....
Now days the blonde that is in the office when shut down the PC thinks that the all system is also shutting down ( hell yeah there is no music any more so it is shut down enough for the blonde )
I can assure you that the amplifier is still working ...how well i dont know but still working ( you may as well check for your shelf )
Now days i am 47 and still learning about amps ...make the math ...the amp operates for more than 30 years and also has almost 10 years to shut down completely...
To my understanding the specific amplifier already covered 100 years of domestic use and hell yeah i am proud of it !!!!!
Last edited:
Did you have to mount all the electrolytics vertically to get such a long life?
Attachments
Ahahahhaahaa!!!! 200 meters from my place
I'll go there to check the condition of the unit! (Screwdriver in hand)
You want me to slap the blonde too?
Sakis, pro audio amplifiers do not always sacrifice audio quality for a surplus of power or an increase in durability.
Please see the schematic and service manual for my Com-tech under the "Com-tech series" on this web page: Discontinued Amplifier Products
See CT400 for the schematic and CT400B for the service manual.
According to the following website, anything above 16,000uf in filter capacitance for a 250 watt amplifier into 4 ohms is really unnecessary. Elliott Sound Products - Linear Power Supply Design While this is the ratio that the author recommends, a 4,700uf capacitor filtering 5 amps of current is still perfectly sufficient (on his 100 watt amplifier)! Please see the tests that he has conducted to reach these conclusions.
Why on earth would Perreaux put 30,000 (or 60,000?) microfarads on each channel of a 60 + 60 watt amplifier? Wouldn't it have been more cost-effective to invest that money in other areas of the amplifier when the extra filter capacitance isn't doing anything? I would rather a company invest this wasted money in more features or higher quality components elsewhere.
Please see the schematic and service manual for my Com-tech under the "Com-tech series" on this web page: Discontinued Amplifier Products
See CT400 for the schematic and CT400B for the service manual.
According to the following website, anything above 16,000uf in filter capacitance for a 250 watt amplifier into 4 ohms is really unnecessary. Elliott Sound Products - Linear Power Supply Design While this is the ratio that the author recommends, a 4,700uf capacitor filtering 5 amps of current is still perfectly sufficient (on his 100 watt amplifier)! Please see the tests that he has conducted to reach these conclusions.
Why on earth would Perreaux put 30,000 (or 60,000?) microfarads on each channel of a 60 + 60 watt amplifier? Wouldn't it have been more cost-effective to invest that money in other areas of the amplifier when the extra filter capacitance isn't doing anything? I would rather a company invest this wasted money in more features or higher quality components elsewhere.
Because it's a 350 watt per channel amplifier? You seem to have misread an earlier post where a very different amplifier was incidentally mentioned. The values are not too far from Rod's recommendations.
Hope that explanation can help the thread avoid emotive distractions because I valued the observations of practical experience, especially the extended experience that is hard to replace.
Best wishes
David
Last edited:
Assuming that capacitor is cheap, I prefer to put 1000x Black Gate 1000uF in my 200W class-B amp. But capacitors aren't cheap. And small "good" capacitance is better than big "bad" capacitance soundwise.
Big capacitance means better ripple regulation and lower bass extension. Ripple regulation is not a big issue as long as we can judge by ears whether we can accept the sound of limited capacitance. Bass extension is another issue. This depends on the rest of the system. PA amps for example, especially for the bass will require more capacitance than a home amplifier system driving a pair of LS3/5A.
But too much, tho not too much theoretically, may give hard sounding bass, and the compromise of good lows is usually bad highs (so bass guy versus tube guy here). This is I think not capacitance fault but only side effects. I think mostly due to high ESR of cheap caps. But 10x4700uF=47000uF is I believe better than 1x62000uF (same capacitor brand).
David, thanks for pointing that out. I must have misread one of his posts. I'm sorry if I unfairly criticized the Perreaux on that topic. I do not believe however that pro audio generally or systematically compromises sound quality for power ratings. Sakis, is there anything you've read that indicates this? I also don't believe that my amplifier produces "dirty power". It's unwise to criticize the final sound produced by an amplifier if you haven't heard or tested it before.
It will never puzzled you if you try to be an open mind person, trying to understand what is in everybody's minds. Many debates are actually not necessary, I mean there are often situations where nobody is wrong, it is just that everyone want to bring across what they have in their mind to be understood or heard.
A complex approach: "I disagree that reliability comes at the expense of sound quality"
A simple approach: "There are many ways unknown to most people where reliability of an amp can be achieved without sacrificing sound quality. For example: a)... b)... z)..."
We can still turn it into something decent now that some people have got their grievances off their chests.
So here is what I think you need in order to make a 'proper' high end amplifer.
A decent power supply. I am with JC on this one. That means the transformer is adequately rated. I take the total output power into 8 Ohms (both channels) and multiply by 3 to get the transformer rating - minimum. Some designers go to 4x.
Filter capacitors. A lot of opinions on this one. 20 000uF per 100W per channel on a really high end amp is about right, up to a max of 50000uF
Going from 8 to 4 Ohms, you should see a nearly doubling in power. My Ovation 250 amp does 280 into 8 Ohms, and 480 into 4 Ohms. Some amps show exaclty double power into half load - the designers cheat by measuring at low load resistances and then extrapolating backwards to show exactly half power at double load resistance. Even in amps that use regulated power supplies, exacly double is highly unlikely.
Construction. Should be solid - no flimsy 1mm panels anywhere. I am not fixed on all aluminium construction, but if this is a design choice, make sure the panels are solid enough. That generally means nothing will be less than 3mm for aluminium, and 2mm for steel. A 10mm front plate also looks decent and sends the right kind of message to everyone: I mean business.
Heatsinking. Too many amps skimp here. After 1 Hour at 1/3 rated power (this is the worst case disspation condition for a class AB), the heatsinks need to be no hotter than 60C. If they are, you are skimping. Try to avoid the use of a fan - but in some commercial applications, it may be unavoidable.
Output Devices. One pair of 20A bipolar devices for every 50W. So, a 80W amp will have 2 devices, a 120W 3 devices. For mosfets, 1 pair of 20A devices for every 60W.
Device Vce rating should be at least 25% ABOVE the expected absolute maximum supply rail Voltage.
Load Drive capability: 60 degress into 3 Ohms for short periods (thermal limitation primarily)
Peak output current - lot of opinions here. Speaker loads can and do dip to very low values - maybe 2 Ohms or mless at specific frequencies - the amp should be able to handle these types of events without any serious current limiting issues (might mean delay in current limit or SOA protection, or oversized output stage).
PCB's. FR4 at minimum. No SBRB type PCB's.
No components run at over 65C on the board
On a big amp, decent overall protection - thermal, soft start, current limiting, DC offset etc.
Component selection - designers usually settle on a bunch on components that they are familiar with (same case BTW in all electronics designs). But generally, good quality branded metal film resistors; electrolytics - same story. No ceramic capacitors on any signal carrying parts of the circuit.
A decent 200W amp will weigh about 30Kgs. This tells you the transformer, power supply and heatsinking are all pretty large and adequatley sized (provided of course the manufacturer hasn't built the thing in a cake tin and bolted a brick to the bottom of course ;-)'. I've seen class AB amps that weigh 10 or 12Kg's and they are spec'd at 200W RMS per channel. That kind of design has far too many compromses to be taken in seriously in my view.
So here is what I think you need in order to make a 'proper' high end amplifer.
A decent power supply. I am with JC on this one. That means the transformer is adequately rated. I take the total output power into 8 Ohms (both channels) and multiply by 3 to get the transformer rating - minimum. Some designers go to 4x.
Filter capacitors. A lot of opinions on this one. 20 000uF per 100W per channel on a really high end amp is about right, up to a max of 50000uF
Going from 8 to 4 Ohms, you should see a nearly doubling in power. My Ovation 250 amp does 280 into 8 Ohms, and 480 into 4 Ohms. Some amps show exaclty double power into half load - the designers cheat by measuring at low load resistances and then extrapolating backwards to show exactly half power at double load resistance. Even in amps that use regulated power supplies, exacly double is highly unlikely.
Construction. Should be solid - no flimsy 1mm panels anywhere. I am not fixed on all aluminium construction, but if this is a design choice, make sure the panels are solid enough. That generally means nothing will be less than 3mm for aluminium, and 2mm for steel. A 10mm front plate also looks decent and sends the right kind of message to everyone: I mean business.
Heatsinking. Too many amps skimp here. After 1 Hour at 1/3 rated power (this is the worst case disspation condition for a class AB), the heatsinks need to be no hotter than 60C. If they are, you are skimping. Try to avoid the use of a fan - but in some commercial applications, it may be unavoidable.
Output Devices. One pair of 20A bipolar devices for every 50W. So, a 80W amp will have 2 devices, a 120W 3 devices. For mosfets, 1 pair of 20A devices for every 60W.
Device Vce rating should be at least 25% ABOVE the expected absolute maximum supply rail Voltage.
Load Drive capability: 60 degress into 3 Ohms for short periods (thermal limitation primarily)
Peak output current - lot of opinions here. Speaker loads can and do dip to very low values - maybe 2 Ohms or mless at specific frequencies - the amp should be able to handle these types of events without any serious current limiting issues (might mean delay in current limit or SOA protection, or oversized output stage).
PCB's. FR4 at minimum. No SBRB type PCB's.
No components run at over 65C on the board
On a big amp, decent overall protection - thermal, soft start, current limiting, DC offset etc.
Component selection - designers usually settle on a bunch on components that they are familiar with (same case BTW in all electronics designs). But generally, good quality branded metal film resistors; electrolytics - same story. No ceramic capacitors on any signal carrying parts of the circuit.
A decent 200W amp will weigh about 30Kgs. This tells you the transformer, power supply and heatsinking are all pretty large and adequatley sized (provided of course the manufacturer hasn't built the thing in a cake tin and bolted a brick to the bottom of course ;-)'. I've seen class AB amps that weigh 10 or 12Kg's and they are spec'd at 200W RMS per channel. That kind of design has far too many compromses to be taken in seriously in my view.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.