Not really sure what you mean. Most DC voltages in amps are fixed by the topology:
LTP - Base is near 0V (affected by base bias currents), emitters are Vbe away from base voltage. Collector(s) are usually one Vbe from the rail.
VAS - Emitter near/at rail. Base is one Vbe from the rail. Collector close to 0V (approx 1/2 the o/p stage bias voltage away).
Driver and output devices - collectors at rail. Base volages set by Vbe's compared to the output voltage
So how can these be changed, for better performance, without changing the topology?
I said, we have to measure our parts, if we want good sounding amplifier. A model only represent one 'typical' value. For example, open the datasheet for 2SK135, check the Vgs cutoff, it says minimum 0.15v and maximum 1.45v (no typical value). I have never found one above 0.6 but imagine if it is 1.45v, we're in trouble (but of course it will sound good enough for everyone
)AFAIK, Nait XS still use the JV base series compensation, and the LTP is more to JV's style too. But the base series compensation is not necessary with modern topology used by Nait. I was surprised they used it but I thought it was for psychology and marketing purposes. Removing it entirely is the proper decision as it does nothing but present stability issue with capacitive load.
The CFP topology does not sound conventional. It is very low distortion and the bass control is unique due to local feedback in the CFP.
You cannot assume that all transistors have 0.6 Vbe. You can match your LTP transistors' Vbe but once you put it in a circuit, their Vbe will depend on the circuit.
You can adjust current to put transistor A in its sweetspot, but by doing that you may put transistor B out of its sweetspot.
Diodes? LEDs? Theory say voltage drop is around 1.8v for green led. What if it is not? It may not work as intended. In a current source people used to use diode or transistor internal diodes to drop voltage around Vbe but if the drop voltage is insufficient then there won't be Vbe drop (but of course the amp will still work).
I used to own a Nait 2 (1988) and it didn't have base networks.
How much difference do they make to the output impedance?
Here's a review of the Nait 2: Stereophile
The hyperbole used in this review is interesting in that the Nait 2 didn't sound anything like as good as they suggest. Pleasant to listen to but poor detail, fuzzy imaging, weird bass, not musically convincing. Which is why I continued to design my own.
This quote from JV (or YV in this article) about not being able to make a more powerful Nait that sounded good is surprising to me. Either this is not true or Naim did not know why their low power Nait sounded good.
The HiFi marketing world is full of crap, in general.
The hyperbole used in this review is interesting in that the Nait 2 didn't sound anything like as good as they suggest. Pleasant to listen to but poor detail, fuzzy imaging, weird bass, not musically convincing. Which is why I continued to design my own.
This quote from JV (or YV in this article) about not being able to make a more powerful Nait that sounded good is surprising to me. Either this is not true or Naim did not know why their low power Nait sounded good.
The HiFi marketing world is full of crap, in general.
Nait 2 was produced before 2000 and Nait XS was in 2008. So using base compensation in modern CFP may not be JV's idea (see: Naim XS2 power section schematics). If Nait 2 is a quasi, the absence of the base compensation is a good thing if it is possible. It is probably easier if the voltage is low so that there are more options with transistor choice. I wont be surprised if the vas and/or driver is using small signal transistors?
Never checked about that. But stability does not have to relate with that.
Pleasant to listen to but poor detail, fuzzy imaging, weird bass, not musically convincing. Which is why I continued to design my own.
Is that your description of the Nait 2? That sounds a lot like CFP amp
The Nait 2 is quasi and I think used E-lines for the VAS and drivers.
One thing to bear in mind is that Naim has to make sure it's Naits don't sound as good as their flagship amps, like the 250, 300, 500. So they will be using inferior circuits or components in the Naits, in addition to smaller psus.
It is easier to make a low power circuit sound good than to make a higher power one sound as good. Naim does not want its Naits sounding better than its 250s.
AFAIK(???) the NAP250DR and above still use quasi outputs and base networks, aside from The Statement which has gone completely rogue.
One thing to bear in mind is that Naim has to make sure it's Naits don't sound as good as their flagship amps, like the 250, 300, 500. So they will be using inferior circuits or components in the Naits, in addition to smaller psus.
It is easier to make a low power circuit sound good than to make a higher power one sound as good. Naim does not want its Naits sounding better than its 250s.
AFAIK(???) the NAP250DR and above still use quasi outputs and base networks, aside from The Statement which has gone completely rogue.
Member
Joined 2009
Paid Member
Well, even the lower NAIT's have competitors to face so I would be surprised if they had any intent on dumbing down their sound quality - isn't it more likely they had to design to a budget and then proceeded to make the best sounding amp that was reliable for that price ? The higher models having a more generous budget would provide the designers for additional scope for better sound but with the constraint that they must deliver more power.
I do not agree about Nait 2 and Nait 3.
I did not have the 2 but I had the 3 for 3 years and for me, it remains one of the best integrated that I had.
On the other hand, it does not support multichannel speakers and complex charges, it requires speakers two voices with simple filtering and there, it gives all its potential.
That's what drove me to mount my clone in BD911 / BD912 with a good power supply
I did not have the 2 but I had the 3 for 3 years and for me, it remains one of the best integrated that I had.
On the other hand, it does not support multichannel speakers and complex charges, it requires speakers two voices with simple filtering and there, it gives all its potential.
That's what drove me to mount my clone in BD911 / BD912 with a good power supply
No. The material cost is not well correlated to the sound quality aside perhaps from the power supply. The smaller psu and case saved the most cost. If you look at the component count in the power amp section of a Nait 2 and a NAP110 there isn't that much in it.
What they did was to extend their range to a cheaper entry level to attract younger buyers. The NAD3020 (20W/ch) was introduced in 1978 and became a best seller. Creek 4040 (30W/ch) was around by 1982; very cheap, very musical. The NAIT1 (13W/ch) was introduced in 1983. The Mission Cyrus 2 (50W/ch) in 1984. This was a boom time for low cost, integrated amplifiers. These models very popular among college students.
It is an anomaly that the NAIT1 and the 1988 NAIT2 were such low power (13W/ch) and more expensive than all the others. This tends to support the story that Naim was unable to increase the power without unduly compromising the SQ although it could have been a deliberate decision to avoid canabilizing their other products. The NAITs must have sold on the back of Naim's audiophile reputation. The NAIT1 didn't even have a pre-amp output and so had no upgrade path. By 1993, the NAIT3 (image below) finally achieved 30W/ch and was based on the NAP90 according to Wikipedia, but the circuit topology is pretty much the same from eye-balling it (quasi without base networks).
Attachments
Last edited:
Things started to change more with the NAIT5 (30W/ch). The NAIT4 was skipped? I don't know how much JV had to do with the 5 as he checked out in 2000, the same year the 5 was launched. Largely a Roy George project, I suspect.
The 5 has the same quasi topology but has dropped the tantalum caps in favour of electrolytics and has encased the LTP in what look like RS232 plug shields with plastic caps! (surely not). There seem to be more diodes and resistors in there too and a couple of unpleasant disk ceramics (hopefully in the protection circuit). I've read that the 5 starts to use anti-vibration mounts for the PCBs which sounds like a George thing.
The 5 has the same quasi topology but has dropped the tantalum caps in favour of electrolytics and has encased the LTP in what look like RS232 plug shields with plastic caps! (surely not). There seem to be more diodes and resistors in there too and a couple of unpleasant disk ceramics (hopefully in the protection circuit). I've read that the 5 starts to use anti-vibration mounts for the PCBs which sounds like a George thing.
Attachments
Steve Sells joins Naim in Nov 2001. The NAIT5i (50W/ch) launches in 2003 and is a radical topology change (3779). CFP, lots of surface mount, VAS change, cut-out channels around the drivers (finned) and Vbe multiplier to confine heat, what appear to be electrolytics in the NFB paths. Seems that Sells redesigned it and changed the JV topology that had existed since the 1970s.
NAIT5i rev 3 image (I think):
NAIT5i rev 3 image (I think):
Attachments
Member
Joined 2009
Paid Member
Oddly, the 5i has a single tantalum that seems to be the Vbe multiplier bypass. Not the place I would have expected the only tantalum.
Speaking of unlucky. When I was in Japan my kind colleagues advised me never to present a Japanese date with 4 white chrysanthemums. All are symbols related to death!
Speaking of unlucky. When I was in Japan my kind colleagues advised me never to present a Japanese date with 4 white chrysanthemums. All are symbols related to death!
Last edited:
With all the scrutiny, Mr Sells might be feeling a little uncomfortable by now 
Perhaps I should mention that he did include an essential flaw from JV's adaptations, i.e. the high capacitance VAS transistors, though they are cascoded here in a somewhat fancy arrangement which offers some protection, as they were vulnerable to output shorts in earlier NAP90/Nait models at least.
It's amusing that he also performs the JV trick of using the same VAS transistor types for driver transistors too. I wonder if it is just coincidental or a gratuitous gesture?
,
Perhaps I should mention that he did include an essential flaw from JV's adaptations, i.e. the high capacitance VAS transistors, though they are cascoded here in a somewhat fancy arrangement which offers some protection, as they were vulnerable to output shorts in earlier NAP90/Nait models at least.It's amusing that he also performs the JV trick of using the same VAS transistor types for driver transistors too. I wonder if it is just coincidental or a gratuitous gesture?
,