I used to be an avid reader of Absolute Sound for many years, alas, sadly not the latest years. I remember a story about an ice cream vendor. Upon a question on why his products tasted so good, he merely answered: I always use good stuff.
Which begs my next question, why even try to clone a NAIM amp at all. Given that most of you already freely admit it`s not even good to start with.
I mean there must literally be a gazillion amps out there just begging to be built, and those don`t even require you to subscribe to the kind of religious mantra served by the same company's PR department.
I suppose we have all heard the stories about LP12 NOT sounding the best being the result of a incorrect set-up? Likewise NAIM, even though they had some nice ads. Remember the one with the amp halfway through the floor? "Sorry I didn`t catch your NAIM". Absolutely priceless.
Or maybe I`m just plain wrong??
Regards
R
Which begs my next question, why even try to clone a NAIM amp at all. Given that most of you already freely admit it`s not even good to start with.
I mean there must literally be a gazillion amps out there just begging to be built, and those don`t even require you to subscribe to the kind of religious mantra served by the same company's PR department.
I suppose we have all heard the stories about LP12 NOT sounding the best being the result of a incorrect set-up? Likewise NAIM, even though they had some nice ads. Remember the one with the amp halfway through the floor? "Sorry I didn`t catch your NAIM". Absolutely priceless.
Or maybe I`m just plain wrong??
Regards
R
I'll tell you mine
I have my own lineup with a motivated preference order
1) - clone Spectral or Norma, already cloned , years ago , in my parerer the best, or almost, but you can not do it because you can not find the right pieces, 2n 5566 first and without 5566 lose much , without forgetting at Norma the emitter follower after IN's cascode with dual fet channel P j 109 Toshiba
2) - the Dartzeel 108 that I have already cloned, but which can no longer be called Dartzeel. This scheme can be done quite easily because it is all, but it costs a lot in components and related selections that must be stringent (you discard a lot of stuff, too much)
3) - 140, o Albarry 408 e 1008 , simple, economical, takes up little space, the heat sinks are minimal. It 'a scheme of the pipe and not too well done (all those compensations confirm it), but that remains
PS
those with ideas are welcome
I have my own lineup with a motivated preference order
1) - clone Spectral or Norma, already cloned , years ago , in my parerer the best, or almost, but you can not do it because you can not find the right pieces, 2n 5566 first and without 5566 lose much , without forgetting at Norma the emitter follower after IN's cascode with dual fet channel P j 109 Toshiba
2) - the Dartzeel 108 that I have already cloned, but which can no longer be called Dartzeel. This scheme can be done quite easily because it is all, but it costs a lot in components and related selections that must be stringent (you discard a lot of stuff, too much)
3) - 140, o Albarry 408 e 1008 , simple, economical, takes up little space, the heat sinks are minimal. It 'a scheme of the pipe and not too well done (all those compensations confirm it), but that remains
PS
those with ideas are welcome
Last edited:
Back in the day I borrowed a lot of amps from my high-end dealer. Being American I wasn't particularly pre-disposed to Naim. I tried Linn, Naim, Perreaux, Hafler, Bryston, Conrad Johnson, McIntosh, etc. The Naim sounded the best. Sadly, it was out of my budget.
So this is sort of a trip down memory lane for me. My main system is a Pass design. One of the few advantages of living in Ireland -- Class A just means a slightly lower oil bill.
So this is sort of a trip down memory lane for me. My main system is a Pass design. One of the few advantages of living in Ireland -- Class A just means a slightly lower oil bill.
What was your simulation? Transistor, resistor values? Nominal current?
I've also heard the rumour that JV adjusted by listening. I don't believe it. That would have been quite a difficult exercise and not really his style. He was a measure everything sort of guy.
with respect, but Naim's antagonists are not great
listened to with my usual references the original 140 does not go down, has a not excellent medium and high range, limited dynamics and does not articulate on the medium-low
it can be improved enough
Last edited:
KSC1845s in the LTP, ZTX653/753 for VAS, 2N5551 for Vbe, MJE243/253 drivers.
LTSpice file attached.
Cheers,
Jeff.
Attachments
Any recomendations? I'm always open to suggestions - my speakers are 3 ohm 86dB. 50 watts is more than ample power.
Ditto.
Bang in some JFETs on the front, swap Q5 for a IRF9610, drop Q6 for a TL431 and add some nice n easy to drive FETs on the output... err, I think I've seen that somewhere before
Last edited:
Jeff,
I make the optimum 28 ohms for your circuit. I had to use a 2N5550 as my Spice doesn't have the 2N5551, but it should be the same.
Brian
Edit: Sorry, I was using 10mA current rather than your 8.5mA. Your optimum is 28 ohms.
I make the optimum 28 ohms for your circuit. I had to use a 2N5550 as my Spice doesn't have the 2N5551, but it should be the same.
Brian
Edit: Sorry, I was using 10mA current rather than your 8.5mA. Your optimum is 28 ohms.
Last edited:
Hi Brian,
Huh. So what am I doing wrong?
Ropt = (KT/qIe)(1 + R2/R1)
KT/q = 0.025V at room temp.
Spice shows me 7.64mA at the emitter of the 2N5551.
Ropt = (0.025/0.00764)(1 + 2000/987)
Ropt = (3.27)(3.03)
Ropt = 9.9
...Bang in some JFETs on the front, swap Q5 for a IRF9610, drop Q6 for a TL431 and add some nice n easy to drive FETs on the output... err, I think I've seen that somewhere before
The TL431s in the CCSes turned out to need stabilising which made them more trouble than they were worth. I've gone back to standard 2-BJT CCSes:
Cheers,
Jeff.
Attachments
Your only mistake is to use Hawksford's formula. It makes some simplifying assumptions and does not reflect real circuit behaviour.
He he... it was probably some of those simplifying assumptions that allowed me to understand it at all. 😉
Simplified formula were meant for those didn't have access to simulator. With simulator we do .step param 😀
Little Toy one lol...my Pyle rack I think sounds Lighttt...it takes BALLZ with the filtering with SS or it sounds Bubbly and BRIGHTTT...garbo capoo 6800uf's too...I'll drop in the 50v 10,000uf's soon here...that toroid is a little small for it's size but not too bad....should easyily handle two and a little better response for guitar.
I couldn't find the Norma Audio schematic. The zeners between output gates are strange. To bias the latfets two diodes are sufficient. The zener is usually used for gate (Vgs max) protection, but then the connection is wrong.
You don't have to use dual JFET for the input. Two single JFETs would be better than dual BJT (without modification to the schematic). Look up BF862 (its SMD). Also discontinued but I think can be found. BF862 is good when Vds is low.
I own the 2N5564 and many BF862 but my Goldmund clone uses different metal JFET that I don't remember the type. You must have heard the Goldmund, right? Follow the Goldmund Clone threads, you may be able to get the part you're looking for.
in hurry I may have drawn quickly and mistaken some connection
however, the standard Norma 8.7 scheme is this
It is still drawn quickly and some detail, marginal, is left hanging and circled circled.
Consider that there are 2 resistances to mass that, when you turn it on, remain connected to mass and then, after a few seconds, are disconnected
It had two levels of counter-reaction, first when power , second and after a few seconds to prevent it from oscillating
If somebody loves to know values, I can detect them
however, the standard Norma 8.7 scheme is this
It is still drawn quickly and some detail, marginal, is left hanging and circled circled.
Consider that there are 2 resistances to mass that, when you turn it on, remain connected to mass and then, after a few seconds, are disconnected
It had two levels of counter-reaction, first when power , second and after a few seconds to prevent it from oscillating
If somebody loves to know values, I can detect them
Attachments
Last edited:
seen the BF 862 SMD, but hold only 20 volts
then they must be absolutely coupled and I do not know how the SMD works:
are they made with minimum tolerances?
Goldmund has an electrolytic cap to signal mass in the feedback network and, that's not good.
But it is an interesting circuit, but always derived from the usual Hitachi application
2N 5566 , 5564 are very very good for input stage
Last edited:
Yep. I think this is a general problem with audio design. Most of us get taught a certain, generalized set of engineering rules and then apply these to audio and get frustrated. Others learn some superficial engineering, think they know it all and then deny audible reality in favour of these inadequate theories. The key is that hearing is believing; if your theory doesn't agree then your theory is not fit for purpose.
Next, you'll need to answer your own question about why the NAP seems to be using the Vbe multiplier in the wrong way. Not the Hawksford configuration.
- Status
- Not open for further replies.