Replacement transistors, NAIM NAP90

Member
Joined 2010
Paid Member
At typical listening levels, there may be little difference in "sound quality" between these parts and they may be safe even at full power. Any apparent difference will more likely have to do with with any "match" between the different architectures of the Quasi-complementary output stage. This does affect the distortion pattern but it's not always noticeable in Naim designs, probably due to the heavy tweaks used elsewhere in the circuit.
MJE 13009 series parts are 200V types but have almost the same DC SOA rating of ~3 amps at 30V. FJE 13009, though, seems to be Chinese origin, usually in a TO3P package and not as I first assumed, from Fairchild. I can't find much data on these and to be honest, I would not trust what seems to be a generic part there, possibly with several unknown sources, packages and specs.

Maybe other members have more information on these as TO220 parts.

I would seek out ST Micro parts. There should be several sources of genuine BD911 series parts in Europe, including Farnell, Conrad, TME, RS-online etc. BD743 was an early source of these, but essentially the same transistor. I don't see any advantage though in using original NOS parts. The first datasheet, looking at current BD743 products, shows minor differences to the original Texas part, so you have to be careful what you are comparing:
http://www.datasheetcatalog.org/datasheet/PowerInnovations/mXqqwtx.pdf
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00001277.pdf
bd743 transistor | Datasheet Archive
At less than $1 ea, it's hardly worth quibbling about using current standard types, as long as they do meet the ratings, with the necessary derating when used with the amplifier case as a heatsink, as do the NAP amplifiers. Buy several from the same pack, if you can, and try to measure with your meter for a best match of gain (Hfe) - not that it will improve performance but you may match the channels better and that will sound better. It could be worth experimenting but take care with damaging the PCB any more.
 
Naim Audio Nap 90 NA 005 (BD911 replacement

Hello...would I be right in thinking that all 4 output NA005 transistors are PNP as the BD911 is NPN....and I though stereo amps are usually npn/pnp pairs....
Also has anyone the circuit available...
My output transistors are dead short as heatsink compound melted away
can any one help..
 
Member
Joined 2010
Paid Member
Hi Mike, The output transistors in all the early NAP and Nait series amplifiers are both NPN type because throughout the NAP range, the traditional quasi-complementary design which dates back to about 1970, was simply a scaled up or down version and NPN power transistors were the much better and more reliable than PNP types.

Naim have never published their circuits for general distribution and their service agents have long-term confidentiality agreements on their IP, so you won't find any official schematic/circuit diagram or service manual for Naim products. The original NAP 250 power amplifier stage schematic, however, has escaped someone's files and along with some early preamp circuits, has been floating about everywhere on the net for many years. You also see it in modified form, as per Avondale Audio's clones. Just Google "Naim amplifier schematic" and there you should find both forms. NAP 90 and other smaller models though, have a number of different resistor values and power semis suiting lower voltages, output power and costs. Some semis may be just be older versions of common types but many will now be obsolete too.

I suggest you trace the design for yourself or check and correct the resistor values on a copy of the NAP250 schematic to get the correct component types used in your NAP 90. The 90 is also simplified a little around the VAS and VI limiter stages so take care with the actual resistors and circuit design there.
 
Last edited:
Member
Joined 2010
Paid Member
Re melted heatsink compound: With TO220 size mica insulators and plastic insulating bushes on the mounting bolts, the transistors should not be able to short to the heatsink, with or without the heat conductive white grease applied. If you used suitably thin silicone rubber pads instead of mica/grease, they should still not melt or short . Did you omit the bushes or was there another problem - like the wrong type (say computer CPU type) heatsink compound?
 
Member
Joined 2010
Paid Member
Yes, I've suggested that BD911 be used previously in this thread because the nearest commercial product to NA005 was suggested long ago by a Naim staffer as BD743 ( perhaps they are one and the same, maybe selected by gain in each pair? ) However, the type has been obsolete for many years now and seemed to be over-stressed in this model anyway. There are much better quality audio transistors now but there are a couple of problems with changing over, to say, a TO247 sized OnSemi MJW3281. First, though smaller than most audio power transistors, it probably still won't fit the available board space and pin spacings etc. Second is sound quality. The emphasis of Naim design and sound quality in the Julian Vereker period, was on "P.R.A.T" and the sound just isn't the same as todays very low distortion, "sanitized" sounding amps, as you may be aware. So then, I'd suggest you stick to the original concept and the similar high voltage BD series power transistors like the currently available BD911. At least give that a fair trial before tinkering with necessarily bigger semis and risking PCB damage by lifting copper pads and perhaps worse.

Where possible, avoid obsolete and NOS parts because even if they prove to be genuine, next time they likely won't be. Always look for current stock parts from authorized sellers as the first option, if you want to keep your amp. going.

Also regarding the schematic, it is basically the same circuit as the lower powered and simplified Nait but some resistor values and other components are somewhat different in order to work with the higher voltages necessary for a 30W amplifier rather than the 13W of a Nait2, for example. For interest, here's a interesting article on recapping the similar Nait 2 and the author has drawn up the schematic on LT SPICE, seemingly correct for that low power model. You might like to compare with your amp and other schematics referred to. Nobeone’s Guide to Servicing a Naim Nait 2 – hi-fi AF
 
Last edited:
Thankyou very much Ian....I will do that...I didn't know the tantalum caps and top hats can go leaky....but then when I do jukebox amps I only ever change the electrolytic and leave the tants......as they are not hifi....but I still check the biasing as Seeburg amps have problems with creating biasing. and blowing up as most of the amps I look at are over 50 years old....I should invest in a leakage tester as well...thankyou again
 
Member
Joined 2010
Paid Member
Re: the 8 pin chips...Yes, they are apparently reed relay banks and need a specified minimum DC voltage to operate each individual reed coil and hold reliable contact. However, I haven't seen inside an actual NAC92 but Naim did use a similar style 16 pin DIP chip lookalike even in the veteran NAC12 preamp. Inside the moulded package was a bank of miniature reed relay assemblies and these had the crazy-amusing flaw of "jingling" along with the the bass beat, like a poorly designed tube amp if you placed the preamp on say, a glass shelf or tabletop near the bass speakers, such that they were acoustically coupled. For all the great points, Naim engineering was often quite amateurish and there were lots of quirky, if endearing design features and flaws that were justified by loads of BS excuses and expensive extras as fixes for the flaws, before pro. engineers took over and changed the game in more recent times.

Reed relays would have been chosen to allow switching by remote logic yet avoid expensive, bulky signal relays and questionable solid state switches. You'd think it was an ideal arrangement but reeds are still not perfect and even though sealed and compact, the tiny contact surfaces can accumulate damage and behave erratically even at audio signal currents.

re: the start-up delay.
NAP90s that I have looked at didn't have any in-built auto delay. They were early models from the chrome bumper era and you were supposed to turn the volume down before switch-off and up after switch-on to avoid a thump or crack. I can ony guess that the delay is built into the signal source switch and the supply to the relay coils would be where the delay circuit was, likely timed by a small electrolytic cap and resistor there, in the supply to the relay coils.

If you are concerned about a lengthy 7 sec delay, so would I be and I'd re-cap the 17V supply completely, including any caps near the relays, just to be certain, since I don't know what each cap does there and how leakage affects operation. FWIW, I think 2-3 seconds would be plenty of delay. More than that and we get anxious, so I imagine its well worth fixing.
 
A 7 second delay does sound a little excessive - and the issue is that it might go from 7 to 70 to never turning on eventually. I design those small electrolytic timing caps OUT whenever possible. The rail caps might last 30 years, but the little ones never do. As long as it calls for something 100 uF or less you could use an X7R ceramic and never have to fool with recapping the protection circuit again.
 
Member
Joined 2010
Paid Member
That X7R cap is a good idea. I just checked the linked schematic for the similar Nait2 model and it seems that a 47uF 25V electrolytic identified as C75 does the delaying there by holding off the DC control signal voltage to the signal relay coil driver circuit. Assuming a similar cap is also in the Olive colour series NAP90, that would be the suspect and X7R type caps are cheap enough.
 
Last edited: