Hello Nigel,
Reading you is always a pleasure, and I'm sure nobody would object if you indulge more in British hifi history
I found that matching the drivers in a CFP is very critical, how about the Naim's quasi?
Hi Telstar ,
The whole thing about all conventional output stages is they seem less trouble than people would expect . This is because it is a current amplifier with a voltage gain of between 78% MOSFET and 97% CfbP . The 94% of Darlington's is especailly fortunate I have found for the more old fashioned designs ( Velleman sell one ) . The Naim is one side CfbP so it sort of answers the question . In recent times I have become very brave and wire my outputs away from the bias stage on long wires . No real problems . This means I don't have to ruin PCB's when trying new things . When building a CfbP I solder the transistors together so as to be one device . Why we can not buy a super transistor like this I do not know ? The stability can be enhanced by soldering a small capacitor between base and collector of the input device . This will mimic the slightly slower transistors of old , it will have superb linearity . Some people are noticing medium speed transistors sound best . It is probable that they have a better balance of gain , stability and bandwidth ?
About hearing . 200 Hz to 10 000 Hz is not bad ( a good TV ) . Some research shows 2 MHz has some relevance to our hearing ( waveform shape ) . I supplied some KEF T27 tweeters to Oxford University guaranteed to give 40kHz . The guys said we humans seem to hear mostly as a digital system , perhaps we detect digital distortion easily because of this ( his words ) . The ear also is a crude analogue device of about 30% THD . On top of that a pulse seems to come back to the ear from the brain which is very fast ( phase lock loop was suggested ) . One other conjecture . The ear produces harmonics very like SE valve amps ( the internal resonance of the chamber ) . Thus as long as it is below 1% THD the ear says that is zero distortion . If the pattern changes slightly we say it is coloured . Thus 1 % or below with linear harmonics zero distortion will be heard . The Naim distortion is almost that and seems contrived ( 1 K + 22K LTP ) . The Naim has slight more 1/3 rd harmonic which gives a dry/ fast sound , my suspicion is not a choice . The Quad 303 seems almost perfect . The 303 is a Quasi design also .
The whole thing about all conventional output stages is they seem less trouble than people would expect . This is because it is a current amplifier with a voltage gain of between 78% MOSFET and 97% CfbP . The 94% of Darlington's is especailly fortunate I have found for the more old fashioned designs ( Velleman sell one ) . The Naim is one side CfbP so it sort of answers the question . In recent times I have become very brave and wire my outputs away from the bias stage on long wires . No real problems . This means I don't have to ruin PCB's when trying new things . When building a CfbP I solder the transistors together so as to be one device . Why we can not buy a super transistor like this I do not know ? The stability can be enhanced by soldering a small capacitor between base and collector of the input device . This will mimic the slightly slower transistors of old , it will have superb linearity . Some people are noticing medium speed transistors sound best . It is probable that they have a better balance of gain , stability and bandwidth ?
About hearing . 200 Hz to 10 000 Hz is not bad ( a good TV ) . Some research shows 2 MHz has some relevance to our hearing ( waveform shape ) . I supplied some KEF T27 tweeters to Oxford University guaranteed to give 40kHz . The guys said we humans seem to hear mostly as a digital system , perhaps we detect digital distortion easily because of this ( his words ) . The ear also is a crude analogue device of about 30% THD . On top of that a pulse seems to come back to the ear from the brain which is very fast ( phase lock loop was suggested ) . One other conjecture . The ear produces harmonics very like SE valve amps ( the internal resonance of the chamber ) . Thus as long as it is below 1% THD the ear says that is zero distortion . If the pattern changes slightly we say it is coloured . Thus 1 % or below with linear harmonics zero distortion will be heard . The Naim distortion is almost that and seems contrived ( 1 K + 22K LTP ) . The Naim has slight more 1/3 rd harmonic which gives a dry/ fast sound , my suspicion is not a choice . The Quad 303 seems almost perfect . The 303 is a Quasi design also .
I think they dont make CFP dual tranasistors because there isn't enough demand? People wants super high gain so they make darlingtons. Idk really, but sounds like a good idea for diy. The discrete semiconductor industry is so f-cked up today...
idk... I like good tube amps, but mostly push-pull of triodes with serious iron -the stuff i cannot afford nor dare to build myself. I had a few tech chats with friends and the conclusion seemed to be that vacuum tubes are better because of physics, because the electrons travels much faster in there.
I like fast amps, but I hope for true speed not the fake PRAT induced by 3rd/5th harmonic excess (I'm sure you understand what i mean, NVA for instance).
I remember somebody did some tests and THD of about 0,1% can be heard, less than that guess not, unless is alot of high-order garbage (i believe in the weighted harmonics order).
I do not have much experience in naim sound, besides their cd-players. Have heard only recent top of the line amps and they do sound good, but nothing to yell miracle. Probably more neutral and less fussy with cables than the old ones? Lots of PRAT and slightly dark yet timbrically correct sound is how i describe Naim house sound. I think JV had good ears.
I want to build this "naim NAP clone" because it's a simple circuit, and allows me to try quasi sound. I'd rather use a reputable circuit than some random one found on the net. I'm still young and I'm not sure where my preference will land in the end, but I like to find an explanation of what I hear to the circuitry and the measurements.
I've read all this thread, and it's obvious that somebody liking is not somebody else's (i.e. the inductor at the output -which is something i must try with and without).
It's so true the thing that you said about using cables to connect the output transistors!
I have a mockup switching circuit on breadboard to test 4 different output transistors (yes, CFP, but the drivers stay same, as I dont want to get crazy) in my headamp. Zero noise pickup or oscillations. Among other things, I hear a different armonic spectrum with different pairs.
uh sorry for long post, this happens when i have lunchbreak early and still some free time
idk... I like good tube amps, but mostly push-pull of triodes with serious iron -the stuff i cannot afford nor dare to build myself. I had a few tech chats with friends and the conclusion seemed to be that vacuum tubes are better because of physics, because the electrons travels much faster in there.
I like fast amps, but I hope for true speed not the fake PRAT induced by 3rd/5th harmonic excess (I'm sure you understand what i mean, NVA for instance).
I remember somebody did some tests and THD of about 0,1% can be heard, less than that guess not, unless is alot of high-order garbage (i believe in the weighted harmonics order).
I do not have much experience in naim sound, besides their cd-players. Have heard only recent top of the line amps and they do sound good, but nothing to yell miracle. Probably more neutral and less fussy with cables than the old ones? Lots of PRAT and slightly dark yet timbrically correct sound is how i describe Naim house sound. I think JV had good ears.
I want to build this "naim NAP clone" because it's a simple circuit, and allows me to try quasi sound. I'd rather use a reputable circuit than some random one found on the net. I'm still young and I'm not sure where my preference will land in the end, but I like to find an explanation of what I hear to the circuitry and the measurements.
I've read all this thread, and it's obvious that somebody liking is not somebody else's (i.e. the inductor at the output -which is something i must try with and without).
It's so true the thing that you said about using cables to connect the output transistors!
I have a mockup switching circuit on breadboard to test 4 different output transistors (yes, CFP, but the drivers stay same, as I dont want to get crazy
) in my headamp. Zero noise pickup or oscillations. Among other things, I hear a different armonic spectrum with different pairs.uh sorry for long post, this happens when i have lunchbreak early and still some free time
Last edited:
What I like about this tread is people seem to have enough knowledge . They ask themselves what sound they like . Some seem to know too much and are listening without the need to listen when reading other threads in other places .
I went in to our local Maplin store and met a very nice chap called John . He said they have a customer called Mr Know It All . From John's description he seems to know all the cliches . I said I am Mr Know Nothing Much because I just about know enough to get along .
With valve amps the transformers are a problem .
I went in to our local Maplin store and met a very nice chap called John . He said they have a customer called Mr Know It All . From John's description he seems to know all the cliches . I said I am Mr Know Nothing Much because I just about know enough to get along .
With valve amps the transformers are a problem .
Yes please, I know one guy that can etch a good pcb (and with silkscreen too that would help me not making mistakes
I'm not in a hurry to build it, though, and it will be interesting to hear your opinion on Greg's mods.
Hello
I do not have silkscreen since all my PC boards are hand drew-it.
I can give you the size of the PC board, the color codes, I can make a picture from my stuffed board etc.
That is the best I can do. Also you can ask me in case something not clear to you.
Again I do not have silkscreen , I'm sorry.
Please let me know if that good enough to you.
Thank you
Greetings Gabor
My almost completed amp.
An externally hosted image should be here but it was not working when we last tested it.
Nice, short wiring there. Have you powered it up yet and checked for hum, noise?
It is never simple to do your own arrangement of boards and power supply so it would
be interesting and helpful to others to hear how successful you were mounting the
amplifiers on the transformers and also squeezing the valve preamp. section supply in.
Good work and BTW, what TO3 transistors are you using for power output?
It is never simple to do your own arrangement of boards and power supply so it would
be interesting and helpful to others to hear how successful you were mounting the
amplifiers on the transformers and also squeezing the valve preamp. section supply in.
Good work and BTW, what TO3 transistors are you using for power output?
Powered up and the only noise it had was a some nasty valve hiss. That's all gone since I added the volume control.
The previous way I had it wired, I could hear my hear beat coming through each time I brought my hands near the signal wires.
Overall it's now VERY quiet, but will still be adding some additional shielding for the transformers and PSU's.
I'm using the BUV20's for the output transistors.
Thank you.
I am currently busy with 2 of these, hopefully they're both going to be done by the end of January.
Hi Nigel 
I was looking again at Semelab's range of power transistors and note that BDY56 & 58 are still in production. The problem is, does anyone distribute them in small quantities?
Magnatech certainly distribute a great range of Semelab's high power audio transistors for pro. audio in all popular case types (The MG633X/MG941X types even give Sankens some stiff competition for gain linearity and Ft). Profusion also sell a few of these types.
Yes, I have built a clone NAP140 using some very fast LAPT Sankens and drivers to suit but I can't say it sounded as good as I hoped. It ended up with MJ21193/4 outputs driven by the original MJE243/53 which I thought, resulted in better sound quality. I guess that's a personal view but I agree with your earlier statement that high Ft doesn't necessarily suit this design.
I was looking again at Semelab's range of power transistors and note that BDY56 & 58 are still in production. The problem is, does anyone distribute them in small quantities?
Magnatech certainly distribute a great range of Semelab's high power audio transistors for pro. audio in all popular case types (The MG633X/MG941X types even give Sankens some stiff competition for gain linearity and Ft). Profusion also sell a few of these types.
Yes, I have built a clone NAP140 using some very fast LAPT Sankens and drivers to suit but I can't say it sounded as good as I hoped. It ended up with MJ21193/4 outputs driven by the original MJE243/53 which I thought, resulted in better sound quality. I guess that's a personal view but I agree with your earlier statement that high Ft doesn't necessarily suit this design.
Hi Ian , This is the best I found .
http://www.cricklewoodelectronics.com/Cricklewood/search.php?mode=search&page=1
I wonder if a small capacitor between base and collector on fast devices might help ? It might only be 22pF . Perhaps win / win ? Resemble the old device and might have slightly better measurements . Douglas Self says external capacitance is better than internal when talking about the voltage amplifier stage of a typical amp ( VAS ) . The VAS cap is the one that keep the amplifier stable . Could be the older transistors had slightly larger Cob ?
http://www.cricklewoodelectronics.com/Cricklewood/search.php?mode=search&page=1
I wonder if a small capacitor between base and collector on fast devices might help ? It might only be 22pF . Perhaps win / win ? Resemble the old device and might have slightly better measurements . Douglas Self says external capacitance is better than internal when talking about the voltage amplifier stage of a typical amp ( VAS ) . The VAS cap is the one that keep the amplifier stable . Could be the older transistors had slightly larger Cob ?
I haven't seriously messed with the VAS in NAP designs. The Ferranti/Zetex parts are such a fixture there that it seems pointless when their Cob is 30pF and way out line with current design trends. If you saw what NAP80 used for VAS transistors, you would surely blanch.
I have substituted BC639/40 and KSC3503/A1381 for a bit of fun but realised that the simple swaps then needed tidying up with more compensation. This led to a different "sound" which wasn't the object, unfortunately. Maybe it's worth a closer look now, in respect of your comments.
Thanks for the link. Cricklewood don't seem to offer quantity pricing but at least there is a credible source for BDY56/58, even if they don't disclose it. Strange, not to say anything about their products apart from a few words of very basic description.
Regarding the output stage; even on the CRO, in comparing 2 differently fitted output stages, you could see a different switching behaviour. I guess this characterised the crossover harmonics and hence the sound quality with some programme - I noticed it pretty easily on small classical ensemble recordings so given my mature earsight, it would be apparent to others too.
I have substituted BC639/40 and KSC3503/A1381 for a bit of fun but realised that the simple swaps then needed tidying up with more compensation. This led to a different "sound" which wasn't the object, unfortunately. Maybe it's worth a closer look now, in respect of your comments.
Thanks for the link. Cricklewood don't seem to offer quantity pricing but at least there is a credible source for BDY56/58, even if they don't disclose it. Strange, not to say anything about their products apart from a few words of very basic description.
Regarding the output stage; even on the CRO, in comparing 2 differently fitted output stages, you could see a different switching behaviour. I guess this characterised the crossover harmonics and hence the sound quality with some programme - I noticed it pretty easily on small classical ensemble recordings so given my mature earsight, it would be apparent to others too.
Hi all,
I started following this thread some time ago and I even bought the kit (from hifidiy.net). I now have some spare time and enthusiasm to start building it. Resistors have been soldered in and I would like to ask a question about capacitors. I want to use polystyrenes instead of the ones supplied but I don't have the exact values on hand.
Is it critical if I use slightly higher values than specified: e.g.
47pf instead of 39pf;
270pf instead of 220pf
560pf instead of 470pf.
Will such changes upset the networks they're used in or in any way affect stability, bias, sound, etc?
I ask because I'm not technical.
Thanks and happy Christmas holidays.
I started following this thread some time ago and I even bought the kit (from hifidiy.net). I now have some spare time and enthusiasm to start building it. Resistors have been soldered in and I would like to ask a question about capacitors. I want to use polystyrenes instead of the ones supplied but I don't have the exact values on hand.
Is it critical if I use slightly higher values than specified: e.g.
47pf instead of 39pf;
270pf instead of 220pf
560pf instead of 470pf.
Will such changes upset the networks they're used in or in any way affect stability, bias, sound, etc?
I ask because I'm not technical.
Thanks and happy Christmas holidays.
Bare boards:
After spraying:
An externally hosted image should be here but it was not working when we last tested it.
After spraying:
An externally hosted image should be here but it was not working when we last tested it.
All good choices because they are cautious ones by going up a notch . Polystyrene fails when old due to foil to wire corrosion . Philips used lead foils to overcome this , now banned . COG ceramic are very good ( look in specs ) . Stability should be enhanced by what you have done . Some people go the other way , that is risky . Don't use non GOG( NPO ) ceramics . Mylar are OK .
Last edited:
Usually that is correct . In the old days components were to +/- 10 % so that almost says if going 10 % up it is the same but the cautious choice . - 10 % probably is OK if carefully checked with an oscilloscope . Problem with the clone kits is they sometimes are very far removed from the original design . The cautious approach seems best . Remember exact values can be made by parallel capacitors . I have a design that uses 6n8 , I find I have to use 4n7 + 2n2 as 6n8 in the quality I want is rare .
If high grade transistors are used go up rather than down . High grades often have very little Cob or intrinsic capacitance . For example a typical small transistor 6 pF , high grade 1.2 pF . Some respected types were more like 50 pF . These amps often have no apparent compensation . If going to a high grade replacement smoke from tweeters might result . A 47 pF added where none was before would almost certainly work . Upgrading amps is a bit of a minefield .