NAD 319. Short circuit output transistor.

[Hunts smoothing bomb]

Hi all,

I have the titled amp and wondered if anyone could give some advice?

Please see attached document.

Many thanks

[PJL]

The service manual is on the internet FOC. Two of the NPN power transistors are monitored and switch off the PSU under fault conditions so you should not remove one of those. Otherwise I would have thought your suggestion is OK at least to prove the rest of the amp is working.

[Hunts smoothing bomb]

Thanks PJL, yes I have the service manual and will check on this.

Cheers

[Beobloke]

2SA1386s are £4.55 + VAT each from Farnell. Why not just stump up for one and replace the failed item? There's a good chance this is all that is wrong.

[Hunts smoothing bomb]

Well, this is exactly the info I was after, can I replace just one then?
I was worried that because they are all strapped in parallel that they need to be perfectly matched on gain otherwise one could end up sweating more than the others and going pop?


Many thanks
Lee

[PJL]

And I think you would be right in saying they need to be matched. Whats more they would also need to be matched with their NPN counterparts. It helps that there are small value resistors in each emitter providing a bit of -ve feedback but even so...

You could compare the voltage on the emitter of the new one with an old one under static and large signal conditions. A scope with the feature to invert and sum channels would be the way to go.

[Beobloke]

Originally Posted by Hunts smoothing bomb

Well, this is exactly the info I was after, can I replace just one then?
I was worried that because they are all strapped in parallel that they need to be perfectly matched on gain otherwise one could end up sweating more than the others and going pop?

Personally, I would indeed replace all four, but you've nothing to lose by changing the one that has gone and seeing what happens.

They won't all need to be perfectly matched by hand - the specification of the devices themselves will ensure this, as long as you buy genuine ones (which is why I'd go to Farnell). This is a mass market, decently powerful amplifier that is designed to be robust, not a fragile 'high end' audiophile nightmare in a gold-plated box!

[mhennessy]

It's a beast!

Audio power amplifiers are hard enough to fault-find at the best of times. With this one, you have a larger than average power supply, so you'll have to be extra careful if you wish to avoid fireworks.

The first alarm bell is the total lack of fuses on the secondary side, so the unit relies on the mains fuse blowing if there's a problem. This makes fault finding rather more tricky as there's no easy point to measure the current taken by each channel, unless the wiring between PSU and amp boards lends itself to that.

Personally, I don't think that matching the output transistors is an issue. The 0.22R emitter resistors are large enough to comfortably swamp the small variations in Vbe - and that's what matters, not Hfe. Selecting transistors on a production line would be suicide for a mass-market manufacturer like NAD - it simply wouldn't get past a design review.

Yes, it's fine to test with 1 missing. Frankly, it'll be fine to use it with 1 missing. That number of transistors is OTT for an amp of that power, and is part-marketing, part-insurance (given that they haven't implemented a sensible current limiting circuit, and they offer a bridged mode, where each half would "see" half of whatever load impedance is used).

The big question is the usual question: did the damage stop with the output transistor, or did we get the usual chain-reaction that happens in DC-coupled amplifiers?

With that in mind, I would not plug this thing into the mains. I would use a dual bench power supply to get just the power amp up and running first of all. Ensure it's a PSU with a current limit. And despite the +/-80V rails in the amp, it'll run quite happily at a lower power from +/-30V that a typical bench PSU supplies.

Good luck

[Craig Sawyers]

Since the output comprises complimentary emitter followers, the performance should be insensitive to device matching. Douglas Self's multi-transistor designs have no need of matching, and he has certainly done design consultancy for NAD - so this might be one of his designs.

But I completely agree - power up the channel with the dead transistor (almost certainly second breakdown) on a separate currently limited power supply and work upwards from +/-10V in modest increments. There can be collateral damage to drivers.

[mhennessy]

Minor correction - it appears that there are fuses in the rails. Great big 10A ones. That makes it easier to some extent. And I made a typo with the rails (60V, not 80V). Sorry!

I must admit that the front end is rather unlike Self's usual designs. He doesn't seem to be a fan of complimentary designs like that, and wouldn't like the fact the input transistors Q305 and Q307 aren't LTPs. Unlike his Cambridge Audio designs, which are instantly recognisable. Still, anything is possible

[mhennessy]

Oh dear! It looks like I was right about the HT fuses the first time. What's shown on the schematic as F301 and F303 is actually a pair of wire links in the photos. Oh well...

[Craig Sawyers]

Looking at the manual the 2SA1386 (and 2SC3519) are listed as O,P,Y, which is the HFE classification

O = 50-100
P = 70-140
Y = 90-180

In other words, it really does not matter what the HFE is.

[Hunts smoothing bomb]

This is really good stuff, thanks ever so much guys, you've all been a big help. Unfortunately I won't get a chance to work some more on the amp until Thusday night now but will concentrate on powering the defective channel from a handy 30-ve 0v 30+ve supply I have.

Cheers

[McMurdo]

as well as doing repairs for an Arcam agent, they are also NAD agents too.
Always worth checking the PSU electrolytics (all of them) on a NAD especially one of these. They dry up and lead to instability.

On the last example I did, every one of the smaller ones was open-circuit and the big 4 bus caps only read about 40% on the meter.

[dave cox]

Is that why they specified 'vented' caps on the main power-rail (10u/100V) ?

Such small capasitors, with so much energy available, can be really quite dangerous !

dc