Greetings Folks, a virgin posting. Could probably be equally applied to the PFM website)
Young nippers and using ELS63s (after very tip-uppable Lowther horns) have lead me to changing back to SS amps after many happy years of playing around with valve amps (usually with chokes and large caps hanging out the back in a way that improves sound and simultaneously attracts unwanted attention from the juveniles). Pre-valves, listening was based around a NAC62 and a Quad 606 driving whatever chimp coffins I had built at the time. Sadly a UB40 meant teh 606 went West several years ago, and the Nac62 was retained mothballed (minimal value for resale as had started skipping happily down McBride Rd). Of late, the 62 has had several updates applied to it (Tibbs Traco and internal LT1086 regs, OScons being the main things).
However, I think that HiFi DIY should be a secular area and we should stay away from any particular church (or is that "an ecumenical matter") in the interests in getting music into our living rooms with ultimate pragmatism.
To cut to the chase I've just recently constructed a Naim Clone with boards from ebay (+ a few component upgrades). It has been a bit of a struggle conceptually and I've began to realise some of the design problems as I've changed codes from a high impedance/low current valve mentality to low impedance/high current transistor circuits. I have been delighted with the results.
Being the saddo that I am, I have scrutinised the circuits of Naim, 😎Avondale, DiyHiFi (ebay = avondale with diff pcb layout), McBride 135 and some other (very new threads) with separated PSUs for the input and driver stages of Naimalike amplifier circuits.
The question that is now bugging me concerns the most expensive part of any of these designs - the power supply. Amplifier board apart there seems to be the following solutions:
1. Transformer(25-0-25)- fullwave rect - high C filter----V out 40V appprox
2. Transformer (25-0-25) - fullwave rect - CLCLC filter ----V out 40V appprox
3. Transformer- fullwave rect (40-0-40) - high C filter--- Vout 56V appprox--Reg circuit Vout (final)-40V approx
I have yet to see:
4. Transformer(40-0-40) - fullwave rect - CLCLC--- Vout 56V appprox--Reg circuit Vout (final)-40V approx.
Woukd there be any benefit in going all the way to 4 as with seperate transformers, capacitor banks etc, these become massively expensive payouts (espec as Im think Evox Rifa 47000/6800uF caps (63V or 100V dep on option)). Also andy other thoughts on the fina;l values of the =/1 rails?
Opinions and sharing of experience would be most warmly received.
Regards to all
Andrew
Young nippers and using ELS63s (after very tip-uppable Lowther horns) have lead me to changing back to SS amps after many happy years of playing around with valve amps (usually with chokes and large caps hanging out the back in a way that improves sound and simultaneously attracts unwanted attention from the juveniles). Pre-valves, listening was based around a NAC62 and a Quad 606 driving whatever chimp coffins I had built at the time. Sadly a UB40 meant teh 606 went West several years ago, and the Nac62 was retained mothballed (minimal value for resale as had started skipping happily down McBride Rd). Of late, the 62 has had several updates applied to it (Tibbs Traco and internal LT1086 regs, OScons being the main things).
However, I think that HiFi DIY should be a secular area and we should stay away from any particular church (or is that "an ecumenical matter") in the interests in getting music into our living rooms with ultimate pragmatism.
To cut to the chase I've just recently constructed a Naim Clone with boards from ebay (+ a few component upgrades). It has been a bit of a struggle conceptually and I've began to realise some of the design problems as I've changed codes from a high impedance/low current valve mentality to low impedance/high current transistor circuits. I have been delighted with the results.
Being the saddo that I am, I have scrutinised the circuits of Naim, 😎Avondale, DiyHiFi (ebay = avondale with diff pcb layout), McBride 135 and some other (very new threads) with separated PSUs for the input and driver stages of Naimalike amplifier circuits.
The question that is now bugging me concerns the most expensive part of any of these designs - the power supply. Amplifier board apart there seems to be the following solutions:
1. Transformer(25-0-25)- fullwave rect - high C filter----V out 40V appprox
2. Transformer (25-0-25) - fullwave rect - CLCLC filter ----V out 40V appprox
3. Transformer- fullwave rect (40-0-40) - high C filter--- Vout 56V appprox--Reg circuit Vout (final)-40V approx
I have yet to see:
4. Transformer(40-0-40) - fullwave rect - CLCLC--- Vout 56V appprox--Reg circuit Vout (final)-40V approx.
Woukd there be any benefit in going all the way to 4 as with seperate transformers, capacitor banks etc, these become massively expensive payouts (espec as Im think Evox Rifa 47000/6800uF caps (63V or 100V dep on option)). Also andy other thoughts on the fina;l values of the =/1 rails?
Opinions and sharing of experience would be most warmly received.
Regards to all
Andrew
Hi,
I am building a ebay clone too.
I have the same problem as you.
Like say, Nigelwright7557, the best solution is a transformer for each channel.
The drivers can be supply by unregulated supply (CLCLC filter) and the input stage regulate.
But with this solution is very expensif (tow tranformers, two rectifer filter).
=> see Teddy Reg for the regulation part :
Regulating NCC200 front end, and other tips 1
Regulating NCC200 front end, and other tips 85
I opte for a cheapest solutin how give equivalent result : one transformer, one simple filter and one or two regulated board (first one board, and if it's no good I build another board, it's very cheap).
My choose come form articles write during the 80's and 90's by a professionnel in a reputate magazine.
(Dominique Jacovopoulos, http://amplifredy.spaces.live.com). The site is in french and the articles too, but that is ones of the best articles I found about amplifer and power supply.)
You can see his regulated supplies here (for class AB amp)
http://www.diyaudio.com/forums/solid-state/164227-nap-140-clone-amp-kit-ebay-regulated-supply.html
The other source of inspiration come form the NAP 250 design. Nap 250 had a single big tranformer and two regulated board.
Modifying Naim Audio power amplifiers
http://www.avforums.com/forums/atta...413-naim-sound-vs-bryston-sound-r0011193e.jpg
The McBride design have a complex regulated board, with curent and voltage regulation.
The Jacovopoulos regulation is more simple with only voltage regulation.
I found 3 other articles about this type of regulation :
Capacitance Multiplier Power Supply Filter
The Class-A Amplifier Site - The Capacitance Multiplier
The Class-A Amplifier Site - A Simple Voltage Regulator
This regulation is build to class A, but it can be adapte to class AB (seen end of « Poject 15 » article).
The principal defaut of a regulated supply come from the speed response.
=> more simple is, more speed is.
Actually I design a regulated board (simple with only regulated voltage).
Moreover, which parts do you change in your amp ?
I am building a ebay clone too.
I have the same problem as you.
Like say, Nigelwright7557, the best solution is a transformer for each channel.
The drivers can be supply by unregulated supply (CLCLC filter) and the input stage regulate.
But with this solution is very expensif (tow tranformers, two rectifer filter).
=> see Teddy Reg for the regulation part :
Regulating NCC200 front end, and other tips 1
Regulating NCC200 front end, and other tips 85
I opte for a cheapest solutin how give equivalent result : one transformer, one simple filter and one or two regulated board (first one board, and if it's no good I build another board, it's very cheap).
My choose come form articles write during the 80's and 90's by a professionnel in a reputate magazine.
(Dominique Jacovopoulos, http://amplifredy.spaces.live.com). The site is in french and the articles too, but that is ones of the best articles I found about amplifer and power supply.)
You can see his regulated supplies here (for class AB amp)
http://www.diyaudio.com/forums/solid-state/164227-nap-140-clone-amp-kit-ebay-regulated-supply.html
The other source of inspiration come form the NAP 250 design. Nap 250 had a single big tranformer and two regulated board.
Modifying Naim Audio power amplifiers
http://www.avforums.com/forums/atta...413-naim-sound-vs-bryston-sound-r0011193e.jpg
The McBride design have a complex regulated board, with curent and voltage regulation.
The Jacovopoulos regulation is more simple with only voltage regulation.
I found 3 other articles about this type of regulation :
Capacitance Multiplier Power Supply Filter
The Class-A Amplifier Site - The Capacitance Multiplier
The Class-A Amplifier Site - A Simple Voltage Regulator
This regulation is build to class A, but it can be adapte to class AB (seen end of « Poject 15 » article).
The principal defaut of a regulated supply come from the speed response.
=> more simple is, more speed is.
Actually I design a regulated board (simple with only regulated voltage).
Moreover, which parts do you change in your amp ?
Another layer off the onion.
Greetings,
some good info to contemplate over breakfast. Thanks.
The clone I have made:
1. Single 300VA 25-0-25 transformer from RS. Output split and feeds two bridges each with 20mF per +/-rail and then straight into the amp boards. Caps are nothing special - Nover 50V "Audio" -As this was a first SS power build I thought I'd get my biggest mistakes out of the way before I started to pour disposable income into it.
2. Changes to parts (v.general - haven' kept a copy of the manifest) from ebay supplier:
(i) Output and driver transistors used where from Cricklewood.
(ii) All ceramic caps replaced by PS, polypropylene (a small cache from the days of active cross overs (Rod Elliot)), and silver mica.
(iii) Have yet to put some wet tantalum 47uF into the feedback cap position - one channel has a 50mV DC offset, the other 26mV. Should have payed some more attention to the hfe of the caps, aswell as "leakyness" of this cap. Will change the 1k feedback resistor to something fancy when I take it apart.
3. Casework made of Al stock (12mm sides), 5mm tops and bottom, with Cu front and backplates. Heatsink = Al channel, pasted and bolted onto the side panel.
Runs v.cool indeed.
4. Oversight 1. Should have split the baseplate into two parts so that I can access both sides of the PCB without worrying about the transformer.
Oversight 2. Not enough space to put in speaker protection modules - crikey, that would be tears before bedtime.
Next build (which is what this is all about) will be monoblocks (without oversight#1,2). So areas for consideration
Main PSU and caps with/without regulation
Front end - quite taken with teh Teddy Pardo monolith regulator idea.
Regarding PCB I'm quite happy to get my own together, using laserjet printing methods. Being a thieving magpie I dont have much anxiety about borrowing PDFs of standard layouts and stitching them together in powerpoint and overlaying big white shapes to block out the bits I dont want. As eacch amp will have its own bespoke box, I'd rather go down the motherboard route and avoid any multi-story lash ups which are apparent in my "Nac62"
Anon,
Andrew
Greetings,
some good info to contemplate over breakfast. Thanks.
The clone I have made:
1. Single 300VA 25-0-25 transformer from RS. Output split and feeds two bridges each with 20mF per +/-rail and then straight into the amp boards. Caps are nothing special - Nover 50V "Audio" -As this was a first SS power build I thought I'd get my biggest mistakes out of the way before I started to pour disposable income into it.
2. Changes to parts (v.general - haven' kept a copy of the manifest) from ebay supplier:
(i) Output and driver transistors used where from Cricklewood.
(ii) All ceramic caps replaced by PS, polypropylene (a small cache from the days of active cross overs (Rod Elliot)), and silver mica.
(iii) Have yet to put some wet tantalum 47uF into the feedback cap position - one channel has a 50mV DC offset, the other 26mV. Should have payed some more attention to the hfe of the caps, aswell as "leakyness" of this cap. Will change the 1k feedback resistor to something fancy when I take it apart.
3. Casework made of Al stock (12mm sides), 5mm tops and bottom, with Cu front and backplates. Heatsink = Al channel, pasted and bolted onto the side panel.
Runs v.cool indeed.
4. Oversight 1. Should have split the baseplate into two parts so that I can access both sides of the PCB without worrying about the transformer.
Oversight 2. Not enough space to put in speaker protection modules - crikey, that would be tears before bedtime.
Next build (which is what this is all about) will be monoblocks (without oversight#1,2). So areas for consideration
Main PSU and caps with/without regulation
Front end - quite taken with teh Teddy Pardo monolith regulator idea.
Regarding PCB I'm quite happy to get my own together, using laserjet printing methods. Being a thieving magpie I dont have much anxiety about borrowing PDFs of standard layouts and stitching them together in powerpoint and overlaying big white shapes to block out the bits I dont want. As eacch amp will have its own bespoke box, I'd rather go down the motherboard route and avoid any multi-story lash ups which are apparent in my "Nac62"
Anon,
Andrew
Hi,
Thanks for the informations
You can put a filter before the tranformer.
There is interesent filtre here (in french, but it's just for the shematic) :
Electronique - Théorie - Ferrite
Electronique - Realisations - Filtre secteur 001
Electronique - Realisations - Filtre secteur 002
Electronique - Realisations - Filtre secteur 003
(you can use google to translate Outils linguistiques)
and this article (in english) is very interesting :
Solid State Power Amplifier Supply
But if you put these kings of filters, be aware !! It's need speatial composants (for exemple, condo must be « X2 class »)
Thanks for the informations
You can put a filter before the tranformer.
There is interesent filtre here (in french, but it's just for the shematic) :
Electronique - Théorie - Ferrite
Electronique - Realisations - Filtre secteur 001
Electronique - Realisations - Filtre secteur 002
Electronique - Realisations - Filtre secteur 003
(you can use google to translate Outils linguistiques)
and this article (in english) is very interesting :
Solid State Power Amplifier Supply
But if you put these kings of filters, be aware !! It's need speatial composants (for exemple, condo must be « X2 class »)
Many thanks,
some seen, some new, some forgotten. The TNT link was a good refresher. Many things lost whilst in the land of valves. Some things always hold true and that is to stick the real money into the PSU and the rest can then work to the best of its limitations.
Have made 1 decision - pragmatism and a degree of fiscal prudence which relates to my 12.5% Yorkshire genetics.... following purchase of two 300VA dual 0-30V transformers (Farnell, going cheap) at the mo.
- scrapping concept of regulation for output transistors and will stick to twin bridge rectified CLCLC filtered cap stage. The caps are the part that will soak up the dosh - the German F&T caps ref in ref#5 TNT hyperlink do sound like specification porn and blow the socks off their plebian brethren as offered by some commercial tweak sites. The decision was based on complexity and cost:
i) I will rely on output devices (e.g Vellerman relay kits) to close down any gig current drains before the speakers vaporise)
ii) a regulator is not a trivial circuit and requires box space and heatsinking
iii) regulator circuits require higher voltage headroom ergo larger transformer voltages and by my own checking 100V rated power caps would be immediately mandatory for any degree of confidence
iv) Throwing some money at high ripple / low esr caps (and sensible film bypass caps) t 63V rating now appears to be rather prudent.
I now intend to take taps from the filtered HT and pass on to some regulation for the low power stages of the circuit. Time to start looking at the monolith regs as a cheap and cheerful route and then the vbe route. Complex decisions to be made with the proviso that it can be grafted into the basic circuit, etched and populated without a round trip to Cygnus X1 for parts.
And so to bed...
some seen, some new, some forgotten. The TNT link was a good refresher. Many things lost whilst in the land of valves. Some things always hold true and that is to stick the real money into the PSU and the rest can then work to the best of its limitations.
Have made 1 decision - pragmatism and a degree of fiscal prudence which relates to my 12.5% Yorkshire genetics.... following purchase of two 300VA dual 0-30V transformers (Farnell, going cheap) at the mo.
- scrapping concept of regulation for output transistors and will stick to twin bridge rectified CLCLC filtered cap stage. The caps are the part that will soak up the dosh - the German F&T caps ref in ref#5 TNT hyperlink do sound like specification porn and blow the socks off their plebian brethren as offered by some commercial tweak sites. The decision was based on complexity and cost:
i) I will rely on output devices (e.g Vellerman relay kits) to close down any gig current drains before the speakers vaporise)
ii) a regulator is not a trivial circuit and requires box space and heatsinking
iii) regulator circuits require higher voltage headroom ergo larger transformer voltages and by my own checking 100V rated power caps would be immediately mandatory for any degree of confidence
iv) Throwing some money at high ripple / low esr caps (and sensible film bypass caps) t 63V rating now appears to be rather prudent.
I now intend to take taps from the filtered HT and pass on to some regulation for the low power stages of the circuit. Time to start looking at the monolith regs as a cheap and cheerful route and then the vbe route. Complex decisions to be made with the proviso that it can be grafted into the basic circuit, etched and populated without a round trip to Cygnus X1 for parts.
And so to bed...
Hi,
About buying expensif transformer. You can use an old from an old amp.
I buy a JVC A-X2. (pictures here HT Forum - EQUIPAMENTOS VINTAGE)
It cost me 35 bollars (45 euros).
There is a 400 VA 33-0-33 transformer in it. (And I will use the box, the diode, the equalizer...)
Another think, you don't need a toroid transformer.
Look this amp, DENON PMA-2010AE :
New, it cost around 2000 dollars
DENON HI FI PMA2010AE DENON AMPLIFIER
DENON PMA2010AE Video Review
The heatsinking act like a shield and block the radations from the transformers.
This design is alos particulate because, the fields of the two transformers cancel each other.
So, the best way to prevent the perturbation from the transformer, is to make a shield around.
In old amp (like Marantz 1060), the transformer is put in a metal box (but i NOT recomand you that, because the actuals transformers are design to have air to cool down).
Note : toroid transformer has other avantage (like space reduce, less voltage augmentation with a empty load...). (But I not agree with space reduce, Ok, the volume is lower, but it take more surface and it's a circle)
i) output devices protection : I buy this one on eaby
http://shop2.hifidiy.net/index.php?main_page=product_info&cPath=8_11&products_id=34
(Not arrive, I still waiting for).
ii) regulator
« not a trivial circuit » => not agree, it's mush simple than a amp.
« requires box space » => not agree, it's depend the board design.
« requires heatsinking » => true, but, you can recycle old form PII or PII (it's free, you find them in the trash http://rdi.us/AK7-2B.jpg)
To avoid big heatsink, a solution is to double the ballast transistor (a transistor cost 2 dollars, so it's a cheap solution => +8 dollars for two regulated board => cheaper than big capacitor).
iii) Your 25-0-25 transformer is too short for a regulated board. The Nap 140 clone need beetwen 36 and 40 V supply.
So, 25*1,41 = 35v + the lost trouth the diode (around 1V) so 34 volt, and after you need to drop 5-8 Volts trought the regulated board. Arival = 28 Volts
By the way, the transformer are generaly 35-0-35 or 39-0-39 Volts for the NAP 140.
About buying expensif transformer. You can use an old from an old amp.
I buy a JVC A-X2. (pictures here HT Forum - EQUIPAMENTOS VINTAGE)
It cost me 35 bollars (45 euros).
There is a 400 VA 33-0-33 transformer in it. (And I will use the box, the diode, the equalizer...)
Another think, you don't need a toroid transformer.
Look this amp, DENON PMA-2010AE :
New, it cost around 2000 dollars
DENON HI FI PMA2010AE DENON AMPLIFIER
DENON PMA2010AE Video Review
The heatsinking act like a shield and block the radations from the transformers.
This design is alos particulate because, the fields of the two transformers cancel each other.
So, the best way to prevent the perturbation from the transformer, is to make a shield around.
In old amp (like Marantz 1060), the transformer is put in a metal box (but i NOT recomand you that, because the actuals transformers are design to have air to cool down).
Note : toroid transformer has other avantage (like space reduce, less voltage augmentation with a empty load...). (But I not agree with space reduce, Ok, the volume is lower, but it take more surface and it's a circle)
i) output devices protection : I buy this one on eaby
http://shop2.hifidiy.net/index.php?main_page=product_info&cPath=8_11&products_id=34
(Not arrive, I still waiting for).
ii) regulator
« not a trivial circuit » => not agree, it's mush simple than a amp.
« requires box space » => not agree, it's depend the board design.
« requires heatsinking » => true, but, you can recycle old form PII or PII (it's free, you find them in the trash http://rdi.us/AK7-2B.jpg)
To avoid big heatsink, a solution is to double the ballast transistor (a transistor cost 2 dollars, so it's a cheap solution => +8 dollars for two regulated board => cheaper than big capacitor).
iii) Your 25-0-25 transformer is too short for a regulated board. The Nap 140 clone need beetwen 36 and 40 V supply.
So, 25*1,41 = 35v + the lost trouth the diode (around 1V) so 34 volt, and after you need to drop 5-8 Volts trought the regulated board. Arival = 28 Volts
By the way, the transformer are generaly 35-0-35 or 39-0-39 Volts for the NAP 140.
- Status
- Not open for further replies.