NAP250 clone

I'm looking at building a NAP250 clone. Since this will mostly be a nostalgia project, I'm trying to be fairly faithful to the original.

However, I won't have the volumes Naim did (for parts matching), and some parts production has since been moved to China (so I'm probably looking at greater variability to start with). The knife-edge stability of the original therefore has me a bit worried.

Here's something pretty dang faithful, but with the standard Miller compensation replaced by transitional Miller compensation (which seems appropriate as it was first proposed by Baxandall, who also has his diode in the quasi-complementary circuit):

attachment.php


It serves its primary purpose of improving the stability without sacrificing the slew rate, although I couldn't get the stability up much without introducing a lot of distortion. So the phase margin is still, well, marginal. On the other hand, the slew rate has doubled, and the frequency response has improved considerably.

Any thoughts on this, including what it might do to the sound signature?

Thanks,
Jeff.
 

Attachments

  • nostalgiaNAP.jpg
    nostalgiaNAP.jpg
    116.6 KB · Views: 14,156
Member
Joined 2010
Paid Member
It's good to take an interest in ideas from Naim and other popular manufacturers of the seventies etc but I'm sure you're aware that it's not a straightforward, textbook design. Perhaps changing the compensation scheme this much would be too radical a change to what is already a heavily tweaked design - and also has a few weaknesses that are actually essential to its distinctive sound.

No doubt, a certain amount of instability arose from those modifications that might otherwise have been eliminated by "proper" design. You might also consider what TMC would do to Naim's Pace-Rhythm-Timing effect. That would certainly be interesting to investigate :)

Current versions such as the 250DR model, don't have quite the degree of sound character of the earlier models so maybe that's where you should look now for a reference point to Naim's sound quality.
 
Hmmm... I did a bit more digging and my "original" I had cobbled together from the McBride, HackerNAP and NC200 schematics appears to have been a bit off.

This one:

attachment.php


shows slightly lower values for the Miller integrator and the phase lead cap.

Plugged into SPICE it doesn't do all that bad:

Code:
                original     TMC version
Phase margin      24º           40º
Gain margin       11dB          13dB
Slew rate         9V/uS         16V/uS
THD @1watt        0.011%        0.015%

Only one number jumps out: how scared should I be of a 24º phase margin?

Thanks,
Jeff.
 

Attachments

  • NAP250 schematic.jpg
    NAP250 schematic.jpg
    92 KB · Views: 12,075
Someone on pinkfishmedia pointed out that it's going to take a lot of finesse in the board layout to keep a 5p capacitor at 5p. Definitely gives one pause for thought.

Here's a 3rd version with 39p in the TMC. Even better stability, but you can see that the THD suffers:

Code:
                original     TMC/3pF    TMC/39pF
Phase margin      24º         40º        50º
Gain margin       11dB        13dB       15dB
Slew rate         9V/uS       16V/uS     12V/uS
THD @1watt        0.011%      0.015%     0.028%
 
Member
Joined 2010
Paid Member
Don't be commited to reducing distortion with Naim designs. The main concern for users is the actual sound quality you hear. I realise you are trying to achieve some technical improvements through simulation but the significant levels of low order harmonics that all NAP models have to some degree is the desirable sound quality. Lose or mishandle that and it reverts to a mediocre sounding quasi-comp, probably as found in RCA's cookbook designs of the period.

Unless you are just interested in analysis, could I suggest you get yourself a few cheap but good enough 140/250 clone kits (some out there have omissions or may suit pinouts for Japanese transistors) and assemble one as original alongside a modified one and listen, comparing over a a week or so to lessen expectancy and other biases we can't avoid.

In direct comparisons, you can learn a few more things about audio and the importance of distortion to sound and spatial perception that you'll never find in simulation or texts of electronic theory. It isn't formal psycho-acoustics but it occupies many us working on perfecting our clone builds. Naim sound may have stood for what was possible and made audio headlines almost 50 years ago but it's still well worth pursuing, if just for the educational aspects of turning a handful of junk parts into a respectable amp with a very entertaining sound quality. Keep at it :)
 
  • Like
Reactions: 1 user
in fact Jeff, there has been a great job done here(on "nap140 clone thread") for years on this schematic.
I compare the NAP to jlh, we love it for what it is and every time we want to improve it, we lose our identity in favor of something different.
as the jlh is quite flexible in therm components, as the nap is very sensitive and even prisoner of some transitors.
there are only the output transistors that are "not too much"
critical.
Rensli has begun an interesting work on resistances and their quality, I will follow his story with interest.
 
@huggygood, indeed it was partly that thread that got me started on the "minimalist intervention" idea. (By minimalist I mean just enough to increase my odds of keeping the smoke in, rather than trying to "improve" the sound.)

Which also talks to @Ian's suggestion of building one of the clone kits: they're sort of in the opposite direction: most of them try to improve the sound (eg. LTP degeneration) while by and large not much improving the stability (aside from adding inductance to the output).

I'm probably answering my own question, though: just go build the original spec and see what happens. If nothing else, smoke can be a good learning tool. ;)
 
that's exactly what I meant.
already, building a clone is a priesthood in itself.
I bought my first clone kit about ten years ago, I assembled it as is, listened 1 day and then dismantled and stored.
then 4 or 5 years ago I came across this thread and it was a revelation for me.
I took months to search and bring together the right components for
put it back in its "state of origin" and the first listen, I fell in love, he had nothing to do with "the thing" that I had assembled before.
In fact at the time I had two naim to repair at the workshop, so I could make the direct comparison and I knew that I was in the right place.
Since then, I started looking for something to build a 90 or a 110 (mine being closer to a 110 than a 140)
I can only advise you to buy a kit 42.5 to go with it and make it undergo the same treatment as the nap.
for the price, it's the same as a zen garden and in addition it makes music :)
 
Member
Joined 2010
Paid Member
I looked over some current clones for a typical NAP build and the best for the difficult to find and expensive semis is still the GLZOZONE or ZEROZONE finished or kit sets for a NAP140 from several sellers (3 boards per set). There is still a possibility of fakes but I bought 3 pairs of finished and kit forms last year that all tested fine with my gear and I did curves with a DCA75 tester to compare the doubtful SS types. The lineup of transistors is spot-on: BC239C, MPSA06/56, ZTX653/753, ZTX384, MJE243/253, 2SC2922. This has a good chance of sounding great - My recent build which equates to probably NAP160, sounds as near as good as it gets for me. For the present, I choose it first every time. :)

You can buy the parts and PCBs individually but what a PITA trying to source that lot and the other components individually - also paying the sum of all individual freight costs. I should also mention that there are errors on the overlay for a 1.2k resistor and the trimpot in the bias chain. Be sure to check against the schematic as you'll get very little bias and a crossover crackle if it isn't fixed.

Hi Huggygood, did you locate a schematic for NAP 90? I know it is simpler than NAP models and is closer to a Nait design i.e. it uses something like BD743 - a T0220 power transistor and the drivers are also ZTX652/752. I'm just not certain of all details as I haven't seen the schematic other than a partial drawing posted originally at another forum. There is a reverse-engineered Nait2 schematic here but that's probably not quite the same.
 
Hi Huggygood, did you locate a schematic for NAP 90? I know it is simpler than NAP models and is closer to a Nait design i.e. it uses something like BD743 - a T0220 power transistor and the drivers are also ZTX652/752. I'm just not certain of all details as I haven't seen the schematic other than a partial drawing posted originally at another forum. There is a reverse-engineered Nait2 schematic here but that's probably not quite the same.
It's exactly that :)
i start from the nait 2 with to 220 power transistor and muscled power supply.
I always keep the idea that it remains a "clone like nap 90" but I will try to get closer to it.
at worst, I will find one to lend me one to raise
components and values, but I'm in no hurry, the path is as important as the destination
 
it's very pretty and very tidy, I'm impressed!
why the jlh?
because it's simple, funny, super nice to assemble
and above all, that he does not look like any other.
he is a faithful companion who never poses a problem,
and everything he does, he does it well.
and finally, you who likes things tidy is the perfect candidate for a very nice assembly p2p
 
I do enjoy p2p. Here's my Bottlehead S.E.X. about 3/4 done:

attachment.php



And finished (completely custom casework on this one):

attachment.php



But I haven't done a dead-bug yet, and the JLH would be a perfect candidate for that....

Cheers,
Jeff.
.
 

Attachments

  • nearing complete.jpg
    nearing complete.jpg
    263.7 KB · Views: 5,550
  • SEX front.jpg
    SEX front.jpg
    325.4 KB · Views: 5,497
Member
Joined 2010
Paid Member
I love how varied this hobby is. Doing my own PCB layout is my favourite part, and finding the parts ranks up there in the top 5 (particularly if they're really hard to find). :cheerful:
Actually, I share this sentiment of immersion in all areas of the hobby. Sourcing parts from remote locations is often prohibitive for me though, particularly when returning goods for credit is at the buyers cost. Do this a few times and you soon tire of spending much more on freight than goods value or ending up with almost nothing.

Not all parts suppliers are straight-up when their products are intended for different markets - maybe a market that is unaware of differences between OEM and generic semis. With on-line selling platforms, we can't often see what is offered and only guess the quality. So it's sad to find after a long wait, that your bargain low noise input transistors hiss like general purpose switch types even though the label looks original and says otherwise :(