img 6840

Musical Fidelity NuVista M3 Amplifier Repair

I recently repaired and serviced this monster Musical Fidelity NuVista M3 integrated amplifier. Come and take a look inside.

The Musical Fidelity NuVista M3 Integrated Amplifier is a true two-box monster amplifier. The NuVista M3 was one of Musical Fidelity’s statement products back in 2001. Back then it cost $4500 USD and that was a heck of a lot of money! If you’d like more of a review article, try this, from Stereophile.


Unlike conventional designs, the Musical Fidelity NuVista M3 consists of two large and heavy boxes. The smaller box is the power supply. It actually houses transformers and that’s it, so it isn’t really a power supply, more a transformer cabinet. Three heavy-duty cables link these transformers – left power, right power and preamplifier – to the main amplifier chassis.

The second, larger cabinet holds the rectifiers and filter capacitors, preamplifier circuit and power amplifiers. The circuitry is straightforward and the quality of the parts is quite basic, in typical Musical Fidelity style. Board quality is good and the modular design is again typical of MF. There is nothing unusual here, except maybe the Nuvistor tube buffer implementation, which is cool.

Musical Fidelity NuVista M3

In terms of other design elements, the Musical Fidelity NuVista M3 uses Sanken SAP15 Darlington output devices. Each unit consists of two bipolar transistors, a diode and an emitter resistor in a high-gain configuration. This reduces the number of components needed in the drive circuitry but creates its own issues.

These devices are no longer available though and hard to match due to their inconsistently high gain. If this thing blows up, it could be a pain to fix now, especially if you can’t get access to the large number of devices you’ll need to select an adequately matched set.

Musical Fidelity NuVista M3 Specifications

Power Output: 250 watts per channel into 8 Ohms
THD: (Typically) < 0.003% at 1kHz (‘A’ weighted)
Frequency response: (Typically) 20Hz – 60kHz +0.25 / -3dB
Inputs: 1 phono 5 line level
Outputs: Speakers, Tape out, Preamp out.
Input sensitivity: (line) 300mV,  (MM) 3.5mV
Input impedance: (line) 100KOhm, (MM) 47KOhm
S/N Ratio: (line) (Typically) 102dB ‘A’ weighted
S/N Ratio: (MM) (Typically) 83dB ‘A’ weighted
Power consumption: 1200W max
Power requirements: 100/115/230V AC 50/60Hz
Dimensions: Amplifier 482 x 145 x 470mm (W x H x D)
Dimensions: PSU 365 x 145 x 250mm (W x H x D)
Weight: Amplifier – 23kg (un-boxed); PSU – 16kg (un-boxed)

Musical Fidelity NuVista M3
Looking inside, we see densely-packed circuitry and a less than ideal heat-sinking arrangement for the output devices. Mating them to a small aluminium block, mounted to a larger heat-sink through a tiny footprint creates a thermal management problem.
img 6811
Norte the modular layout, typical of Musical Fidelity. Of significance is the tiny footprint of the output device mounting blocks, screwed to the main heatsinks. This is not a sensible thermal interface and I would improve this with lapping and thermal paste during overhaul.
img 6814
Protection relay and internal-RCA cable detail

Previous Work…

This late serial numbered unit had visited a technician prior to its arrival here. This technician had replaced parts as part of an expensive repair job. In this case though and much to my relief, things were, for the most part, neat and tidy.

The issue, and I’m not sure if this had occurred to the owner, is that this work didn’t resolve the problem with the amplifier and was therefore completely unnecessary and expensive, with no positive result. The original channel drop-out problem with this NuVista M3 returned shortly after the work done a few years ago. My customer lived with this for some time before bringing the unit to me. Can you pick the new parts in the image below…?

img 6826 1
One non-original group of SAP15N Darlington transistor packages. All the other output devices are also new. The clue lies in the date codes, plus also some witness marks on the fasteners and excess flux. These amplifiers were manufactured in 2001. They can’t contain devices from 2009 unless those devices are non-original.

So, I found a set of 16 SAP15 Darlington devices from 2009 in this amplifier from 2001. As I said, the work is neat so I have no problem with it from that perspective. The problem is that it didn’t need new output devices. This work cost the owner well north of $1000 AUD.

The original fault returned shortly after the output devices were replaced. This tells us that they were absolutely OK. The amp still ran at this point and these things usually explode when they fail, leaving signs that are visible even after replacement. These devices are used across a range of MF amplifiers and they generally fail catastrophically.


One of the first things I do with any piece of equipment exhibiting a recurring fault is to carefully examine every component possible. I’m looking for clues (thanks, Robert Palmer), signs of damage or overheating, poor soldering or anything that might exhibit intermittent poor contact, like a switch or relay.

I found three critical and two potential issues in this NuVista M3:

  • A damaged TO-126 device (inflicted during previous repair)
  • Poor soldering on a diode bridge
  • Excessive idle power consumption and asymmetrical bias
  • A remote-controlled mechanical input selector – potentially flaky
  • A full complement of crappy original capacitors
img 6817
This immediately caught my eye, but obviously, no one else’s before now.
img 6818
The leg of this diode bridge might technically be connected to the board, but I’m not happy with the soldering. This has to be resolved.
img 6815
This driver device also caught my eye. When I found it, the transistor lay almost at right angles to this correct alignment. The strain on the device when pushed (previous repairer) from above has bent it, causing the legs to almost fracture. This can be repaired in place. Removing this board is painful and I wanted to get the thing running reliably and then speak with my customer about further work.
img 6821
The NuVista M3 contains all its original Jamicon capacitors. These are not great or really appropriate for an amplifier like this. You can see from the next few shots that the filter capacitors have come away from the boards slightly. They don’t appear to be leaking but they are free to move and may have suffered heat stress. I told my customer that it is best to replace them.

img 6823

img 6825
Lack of bench space and my messy main repair area makes working on the NuVista M3 challenging…!
59584156842 600aeafb 3640 4e83 aaed 09702d90ce7d
Idle power consumption is high. The real problem though is that one channel is dissipating most of this power, indicating an asymmetrical quiescent current condition. This may lead to one channel overheating, possibly causing a problem and definitely harming the sound.

Repair & Service

First, I repaired the poorly soldered diode bridge and the TO-126 transistor with the fractured legs.

img 6819
I hit a couple of the legs on this bridge with fresh solder and a nice, hot iron, correcting the marginal soldering. There can be no thermal issues with the joints of this bridge now.
img 6816
Likewise this driver transistor. Either of these issues could have lead to reliability problems. The device itself tested well and proper replacement is only possible by removing the entire board, hence this strengthening exercise. I’ll replace this device as part of a major overhaul to replace other worn parts.
img 6811
I cleaned and re-made all internal RCA cable connections and serviced the input selector switch. These can be problematic and it’s very likely related to the problems my customer experienced.
img 6836
I then ran the Musical Fidelity NuVista M3 through an extended test cycle and very carefully adjusted the quiescent current to reduce power consumption to around 95W at idle, and completely symmetrical left and right channel current draw.

So what were the main issues here? When the output devices were replaced in the previous repair attempt, the technician failed to correctly re-bias them. He either misadjusted the quiescent current or never looked at it at all. This caused a thermal imbalance and poor sound and this, combined with a flakey input selector switch were the likely culprits. The damage he caused to the transistor was incidental but not a good thing.

Results and Suggestions

After addressing the problems with this Musical Fidelity NuVista M3, including the poor soldering and fractured transistor legs, extensive bench testing and adjustment, I gave her a clean bill of health. Listening to the M3, she sounded lovely, particularly BIG, powerful and is clearly capable of driving some big speakers in larger rooms.

The only other thing I recommended to her owner is an overhaul, using much better parts. This will transform the performance of this unit in sonic terms. To anyone owning a Musical Fidelity NuVista M3 integrated amplifier and wanting the best performance and reliability, a full overhaul including new filter capacitors, improved film capacitors and critical bypass and decoupling is the prudent way to proceed.

Feedback from her owner is very positive:

Dear Mike

I thought I would let you know that my amplifier has never missed a beat since getting it home, thanks so much I really appreciate it.


My customer is bringing her back next year for this work and if you or anyone you know has one of these or any of the other big Musical Fidelity amps, I am very happy to offer service, repair and overhaul for these units.

img 6839
The Musical Fidelity NuVista M3 is a seriously big unit
img 6840
Looking lovely after a thorough detail and now working properly, to boot!

5 thoughts on “Musical Fidelity NuVista M3 Amplifier Repair”

  1. Hi Mike, interesting article, I own a Nu Vista M3 as reviewed.
    great sounding amp. I love it !
    Where do you service these amps ?
    I’m in South Australia & would like to up-spec my unit,
    it works beautifully at present. its just had the original selector switch refurbished. Couldn’t find a replacement selector switch. would do if I could ?
    Anyway its rare & good gear that I would like to preserve & may improve.
    Even If I could buy just a replacement selector switch ?

    Cheers Drew email:

    1. Hi Drew, thanks for your comment! All customer equipment is serviced by me, here in the Liquid Audio workshop. This includes gear from WA and around Australia, plus a tiny bit from overseas. There are many things that can be done to improve these amps but good to hear yours is running well. I’m helping a technician in Canada with one as we speak.

  2. I am familiar with this model. It is yet another Musical Fidelity abortion. Here is a further list of design flaws:

    * The use of NLA Nuvistor valves. Dumb.

    * The method of attaching the NLA output devices relies on an EDGE mounted slab of aluminium plate. Thermal transfer is particularly poor. I measured an almost 20 degree C difference between the temperatures on the output devices, relative to the main heat sinks. Not unusual for MF. They have a poor reputation for getting thermal issues right.

    * The use of Jamicon main filter capacitors. These are cheapest, poorest quality caps commonly available.

    * Agreed with all the other findings. The M3 is a terribly designed product.

    * A personal issue, to be sure, but the M3 is seriously garish and horrible to look at.

    The reason I happened on your page is because I am working on an ancient MF A-200DM. Looking at the A-200DM, it is clear that MF have learned very little from their earlier mistakes and cheap-arsed approach to design and manufacture.

    1. Hi Trevor and thank you for your assessment of this classic MF amplifier. There are some typical MF design issues here and many owners of these amplifiers have been hurt by reliability issues. That being said, many other amplifiers are worse than this and these at least sound nice, when they are working. My comments on your assessment:

      * MF liked using weird NLA parts. It’s not the smartest design decision but they are at least reliable in this role. The output devices are a bigger problem.

      * I addressed the thermal and heatsinking issues and yes, the heatsink arrangement is poor. I disagree with your assessment of other MF amplifiers, thermally speaking.

      * I disagree in terms of caps, Jamicons are FAR from the worst out there. Suntan, Samxon, Lelon and many others are way worse and all too common, unfortunately.

      * I also dislike the gold trim and garish styling.

      1. Agreed with the comment about Jamicon caps. They are certainly horrible and there are worse caps. Consumers should expect more when paying big Bucks for such a product. In fact, in your area, where mains Voltages can be (ahem) patchy, there is one model MF amp I worked on a few years back (sorry, I don’t recall the model number), where 8 or 12 Jamicon main filter caps were used. The caps were rated at 63 Volts. The rail Volts (at a regulated 230VAC) was 60 Volts! Almost all the main filter caps were dead. Terrible design. I seem to recall the heat sink fins were horizontal too. Made it easier to lift. Shame about the cooling.

        Also agreed with the comment about output devices. I like Sanken semiconductors. They’re pretty good, BUT, Sanken have a most unpleasant habit of deleting products from their range and frequently no substitute is available.

        BTW: I gave up trying to re-engineer the A-200DM. I just fixed the amp and issued the client with a warning. When clipped briefly, the amp exhibits severe crossover distortion for several second and the level must be reduced to around 10% of maximum power for bias stability to return. Weird AND horrible.

        FWIW: Back when I was with Marantz, we used to send out the most spare parts (given the number of products sold) to our Perth agent. When I questioned him, he explained just how horrible the mains supply could get back then.

Thanks for visiting, feel free to leave a comment. Advice is accessible via our contact page.

%d bloggers like this: