no, but it's certainly one of my favorite.You meant that the Naim became your favorite amp after you installed the tantalum.
and yet amps, I really got a lot
Is this the kit this thread is all about:
Aiyima NAIM NAP140 AMP CLONE KIT 2SC2922 Amplifier board Kits For DIY 2 channels J163-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group
Aiyima NAIM NAP140 AMP CLONE KIT 2SC2922 Amplifier board Kits For DIY 2 channels J163-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group
This thread is years old and the kits have changed a lot over that time. If you inspect the components on this version, you can see that many are common Ebay type substitutes for cost or availability reasons. There are usually about 6 different types of NAP140 or similar model clone kits available but the schematic shown is for another model and some of the transistor substitutions degrade the sound quality to the point where the kit is a waste of time because the listening result is only fair and doesn't compare well with the original.
It's probably fine for adapting to similar quasi-complementary designs from diyAudio or something you designed yourself but not if you expected good Naim sound. Read more of the thread, where VAS transistor types and their effects are discussed. Check out Bigun's TGM10 thread for an easier search of the topic too.
It's probably fine for adapting to similar quasi-complementary designs from diyAudio or something you designed yourself but not if you expected good Naim sound. Read more of the thread, where VAS transistor types and their effects are discussed. Check out Bigun's TGM10 thread for an easier search of the topic too.
Yes. Suitable transistors are very very expensive and probably hard to find in some area. I don't know how much do the Zetex cost in the US. Here it is very expensive, more expensive than output transistor.
I think paralleling transistors is a nice trick. Another one I wanted to try was to lower power supply so transistors with low Vceo can be used.
My current stock is C1384/A684 (50V, 1A, 11pF, TO-59L).
TO-59 BC like BC556 is fine if only for the input stage where noise is more important. The ZTX is higher current than the BCs. Many of the ZTX are not complementary and not linear. My favorite high current TO-59 is 2N5401 and it's only 600mA. ZTX can go up to 5A.
The reason the particular ZTX653/753 types are used in the real thing, is largely because they have high capacitance (Cob). Normally for state-of-the-art designs, this is quite the wrong type of transistor to fit as the voltage amplifier but as said before, we aren't trying to clone a SOTA design with a kit like this one.
Because Cob is voltage dependent like a varicap, such a high capacitance with a signal applied, has the action of modulating the compensation cap, Cdom, to which it is in parallel. These transistors then, contribute significantly to the unique sound of Naim's NAP and Nait models.
I have seen plenty of kit versions fitted with 2SB647/D667C for the VAS and CS transistors TR4, TR6. These types are similar to BC639/640 in regard to capacitance (15pF) and also as 2SB649/D669 is to BD139/140.
In comparison to original types though, these ones just don't have enough Cob to raise Cdom to around 70pF and then modulate it enough to sound right. There are some other high capacitance types but they are hard to find or buy in through-hole form, unfortunately.
For the input stage, I think you need as high gain as possible, slightly mis-matched to give a low DC offset. BC239C is good, as is 2N5089 if selected highest gain examples are used. I doubt genuine BC239C will be available at low cost for much longer.
BC546B works ok, as will most small signal transistors but the noise level is a little higher and the gain is not as high as is desirable. I aim for hFE around 500 though this may prove less stable unless the associated parts combinations are as close to original as possible.
Because Cob is voltage dependent like a varicap, such a high capacitance with a signal applied, has the action of modulating the compensation cap, Cdom, to which it is in parallel. These transistors then, contribute significantly to the unique sound of Naim's NAP and Nait models.
I have seen plenty of kit versions fitted with 2SB647/D667C for the VAS and CS transistors TR4, TR6. These types are similar to BC639/640 in regard to capacitance (15pF) and also as 2SB649/D669 is to BD139/140.
In comparison to original types though, these ones just don't have enough Cob to raise Cdom to around 70pF and then modulate it enough to sound right. There are some other high capacitance types but they are hard to find or buy in through-hole form, unfortunately.
For the input stage, I think you need as high gain as possible, slightly mis-matched to give a low DC offset. BC239C is good, as is 2N5089 if selected highest gain examples are used. I doubt genuine BC239C will be available at low cost for much longer.
BC546B works ok, as will most small signal transistors but the noise level is a little higher and the gain is not as high as is desirable. I aim for hFE around 500 though this may prove less stable unless the associated parts combinations are as close to original as possible.
I think a max 30pF is not at the high side. There are plenty of choices if higher (or lower) Cob is wanted.
Normally, fast transistors will have low Cob. But 30pF@140MHz is not hard to find. Compare for example the ztx with one or two of my inventory:
ZTX753: 2A, 140M, 30pF, hfe=240
C4883: 2A, 120M, 30pF, hfe=240
A1930: 2A, 200M, 26pF, hfe=100-320
Harder is to find the proper complementary between NPN and PNP. But i speculate that the real reason people had to keep the same exact transistor is because people were not ready to change the compensation.
In passive crossover design, a fourth order slope can be achieved with many circuits. The capacitor at one end can be 1uf or 40uf, the sound will be similar as long as the total curve is similar. Same with compensation technique in feedback amplifier. Who cares with different miller value as long as the 'total compensation' is the same.
Last edited:
Member
Joined 2009
Paid Member
We certainly do care about the ratio of compensation capacitance which is variable as opposed to the amount of capacitance that is fixed. After all, it is the variable (Cob) capacitance that determines the degree of the sound effect. Are you aware of how varicap or varactor tuning in receivers works? As explained about the modulation effect induced by the varying audio signal, if you use a low capacitance VAS transistor pair, the effect disappears and the sound quality is not very good - I'd call it dull. Sure, it can be improved by increasing Cdom, the fixed amount of of capacitance, from 39pF to about 70 but that results in a standard quasi-complementary sound quality - so why bother?
I guess I could have written an even longer post covering every possible argument but it is hard to find transistor pairs with high Cob because they need to fit the clone kit PCB, since the kits are what this thread is about. No doubt we could always find some large TO220 or SMD parts that are intended for drivers or power switching applications and we could then make or modify our own PCB to suit them. The peak current rating of 2A though, is not essential. 0.6 - 1.0A max. Ic rating should be sufficient.
There are many more possibilities when we are not restricted to the existing Ebay kits or design. If you have ideas for semis that prove to sound more like the original Naim products, then I suggest you build and compare them with the real thing formally and post the comparison here or start another thread about it. Otherwise, discussing sound quality that we have never heard is rather a funny concept - virtual audio? That's something to discuss in the lounge forum, perhaps?
I guess I could have written an even longer post covering every possible argument but it is hard to find transistor pairs with high Cob because they need to fit the clone kit PCB, since the kits are what this thread is about. No doubt we could always find some large TO220 or SMD parts that are intended for drivers or power switching applications and we could then make or modify our own PCB to suit them. The peak current rating of 2A though, is not essential. 0.6 - 1.0A max. Ic rating should be sufficient.
There are many more possibilities when we are not restricted to the existing Ebay kits or design. If you have ideas for semis that prove to sound more like the original Naim products, then I suggest you build and compare them with the real thing formally and post the comparison here or start another thread about it. Otherwise, discussing sound quality that we have never heard is rather a funny concept - virtual audio? That's something to discuss in the lounge forum, perhaps?
Without prejudice to the correctness of the evaluations on the capacity of the transistors, however, I would also like to say to all those interested in this excellent and beautifully sounding circuit, not to be afraid of the task of making it and not to give excessive importance to the obsessive search for active components now obsolete if not unobtainable.
On the other hand, I am absolutely convinced that even the great Julian Vereker, if 40 years ago, had had at his disposal the current active components, with an unquestionable better linearity and a greater bandwidth, he would certainly not have refused to use it in his creations.
Over the years I have made four different types of NAP clones, with countless attempts to upgrade and refine and apart from the first realese (rigidly equal to the original and with the current protection net) especially the last type H-140 (BC550 for the input differential and the "current mirror", the VAS consisting of 2SA1145/2SC2705 and 2SC2240, MJE15030/15031 for the driver stage) and gave me a magnificent sound and absolutely convincing.
On the other hand, I am absolutely convinced that even the great Julian Vereker, if 40 years ago, had had at his disposal the current active components, with an unquestionable better linearity and a greater bandwidth, he would certainly not have refused to use it in his creations.
Over the years I have made four different types of NAP clones, with countless attempts to upgrade and refine and apart from the first realese (rigidly equal to the original and with the current protection net) especially the last type H-140 (BC550 for the input differential and the "current mirror", the VAS consisting of 2SA1145/2SC2705 and 2SC2240, MJE15030/15031 for the driver stage) and gave me a magnificent sound and absolutely convincing.
Last edited:
And the circumstance of the great sound I point out that at home I also own an original NAIM NAP 200, which compared with the last two versions of my clone, if you see it exactly on par in terms of seductive Timbre, but it remains a step behind in terms of "speed', sound-stage and driving skills, especially with my isodynamic magneto-planar MAGNEPAN.
However, it is very important and fundamental to foresee:
1) A generous and high-performance low impedance power supply stage that in my case includes a transformer of at least 500VA-30Vac, two or four robust double-diodes type Fast and 60,000 uF filtering,
2) The fundamental caution of mounting, in the input differential, the transistor with the highest hfe on the input side of the signal,
3) A careful selection of the electrolytic capacitors to be mounted on the PCBs. After various experiments and having used in the past also the excellent Philips/Vishay and ELNA, I strongly recommend the NICHICON Series Muse Gold UFG and UKZ. Leave everyone else behind.
4) The absolute originality of the TOSHIBA 2SC5200 power amps (there really is an exaggerated amount of Chinese fakes in circulation), including those that are sometimes sold with kits,
5) A scrupulous adjustment of the bias to 7.5/8.0 mV of the power amps and about 40 mA of Rest Current for each supply rail (positive and negative),
and your clone will not fail to offer you extraordinary satisfaction.
However, it is very important and fundamental to foresee:
1) A generous and high-performance low impedance power supply stage that in my case includes a transformer of at least 500VA-30Vac, two or four robust double-diodes type Fast and 60,000 uF filtering,
2) The fundamental caution of mounting, in the input differential, the transistor with the highest hfe on the input side of the signal,
3) A careful selection of the electrolytic capacitors to be mounted on the PCBs. After various experiments and having used in the past also the excellent Philips/Vishay and ELNA, I strongly recommend the NICHICON Series Muse Gold UFG and UKZ. Leave everyone else behind.
4) The absolute originality of the TOSHIBA 2SC5200 power amps (there really is an exaggerated amount of Chinese fakes in circulation), including those that are sometimes sold with kits,
5) A scrupulous adjustment of the bias to 7.5/8.0 mV of the power amps and about 40 mA of Rest Current for each supply rail (positive and negative),
and your clone will not fail to offer you extraordinary satisfaction.
Last edited:
In the photo my latest version of the NAIM H-140 clone compared to another excellent design which is the DartZeel NHB-108 clone, a rare and perhaps only circuit, apart from the Ayre electronics, with global Zero Feedback (discussed in another specific thread of the Forum) and which I have currently abandoned, moving to an improved release with a double pair of power amps and active/passive components selected.
Together they represent two extraordinary designs able to intimidate any amplifier even with great pretensions and certainly the best available in DIY kits.
Last edited:
I think you are referring to my statement about 'total compensation'.
I understand your view that internal capacitance is of lower quality than the external one (Miller). With 'total compensation' I didn't mean the sum of these two capacitance types. Rather, the total sum effect of all the compensation parts (resistance and capacitance) wherever you do it in the circuit.
Anyhow, Ian Finch has different view on the importance of high internal Cob.
We approach the topic differently but Bigun and I probably agree on the particular sound effect and what specific circuits and components make up the unique composite of Naim's sound quality. If the specific VAS design was not an essential part of this, why else would Naim retain just this part of the NAP circuit with its semis and components intact, yet change most other semis several times over about 45 years? I suggest Les Wolstenhome's NCC200 modules also retain that VAS section intact because he recognizes the merit too.
Different models may dictate scaling of some components to maintain performance with different supply voltages but otherwise, not much changes in the circuit over the range from NAP140 through to NAP250 and Nait models at least up to the 5i as well. The omission of the phase alignment networks in Nait models may be just a marketing decision but I don't hear much difference with or without them.
Member
Joined 2009
Paid Member
I haven't done many comparisons but I do like Nichicon gold capacitors and have used them extensively in my projects. In particular, the Nichicon MUSE bipolar cap of around 10uF makes, to my ears, the best input capacitor for an amplifier - to my ears better than any MKP. Presently, my NAP clone has a tantalum at the input but I may experiment...
Member
Joined 2009
Paid Member
Indeed, and I think quite a few people discovered early on (in my case my AKSA clone, TGM1) that the VAS compensation has quite a noticeable impact on the sound of an amplifier
I learned only after a kind poster demonstrated through simulations that these networks help with stability (large square waves in particular) but perhaps their need is marginal with most music and they can be eliminated when using appropriate output power devices, speaker cables etc.