Repairing a HK1400 Amplifier

[Old Sparky]

I'm in the process of repairing a Harman-Kardon 1400 in which the upper (NPN)o/p transistor in one channel has failed s/c base to collector, and also reads about 350 ohms each way, base to emitter on my DMM.  Having sourced a pair of the correct o/p transistors and fitted them, I now feel that some or all of the transistors upstream of the o/p pair should be replaced even though they check out good in 'cold' tests.  THis is a bit of a daunting task, since not only is this amplifier not the easiest thing to work on, but there are over a dozen transistors in the circuit of various types, and whilst I must have several hundred transistors of different types in my spares boxes, I have none of any of the types it uses (2SD888A, 2SD686,  2SA970B, 2SB648A, 2SC2240, & 2sc1775, amongst others, so it looks like I'll have to google data sheets on all of them to check if I have any equivalents, which, in itself, will be time consuming.  Much as I dislike abandoning things, I'm beginning to wonder if repairing this amp. is worthwhile.  What do other members think, and fdoes anyone know if (a) the transistors I mention are available, or (b) know of any equivalents.

[AlanBeckett]

I don't know this Amp but bitter experience says that if it's DC Coupled you're just as likely to write them off again if any of the others are shaky.

Alan

[Old Sparky]

My feelings exactly, Alan. I also have in the 'to do' list, a NAD 3020 or similar with the same problem. I put that on one side quite a while ago, then foolishly agreed to see if I could fix the HK1400 - sucker for punishment, me!

[Murphyv310]

Hi.
I agree with Alan. The other thing is in DC coupled circuits with three legged fuses is the possibility of out of value resistors and leaky caps upsetting the DC conditions.
You might be lucky with equivalents but one mistake with different pin outs and its bang.
A better bet might be to put a £1 on the bookies for a white Christmas.
Have fun.

[Mark Hennessy]

When asking for help, it's surely good etiquette to provide (links to) service data where available?

http://www.hifiengine.com/manual_library...1400.shtml

(you have to register, but it's free and well worth the effort).

This amp is conventional in design, runs from modest supply rails, and lends itself to logical fault finding. Is it worth repairing? Only the OP can answer this. How much is the customer willing to pay, how long will it take, how long is any guarantee, will you cover consequential damages (as there appears to be no DC offset protection to save the loudspeakers), etc, etc... We can't tell you how to run your business.



Forgive me while I climb on my soapbox

I've never understood why people get so nervous around audio amplifiers.

Sure, if you just plug it in and hope for the best, you're likely to waste transistors. But why would anyone do that? It just baffles me...

Quite simply, a pair of bench PSUs (or a dual model) with variable current limiting is all you need. Most people already have these, but if not, they are cheap and plentiful second hand. Something like a Farnell L series, or the classic Thurlby-Thandar PL320 or similar.

If your bench power supplies aren't current limited, or - inexplicably - you don't have bench power supplies, then build a couple of 50mA current sources. Just 2 transistors and 2 resistors per rail, plus a scrap heat sink. Or, a pair of 24V light bulbs. Even a pair of 100 ohm resistors! Most output transistors are rated to at least 10 amps - usually more.

At the very least, a lamp limiter on the mains would save the new output devices. Many folk have no qualms using these on old radios, so why wouldn't you use them on an amplifier?

Once you can power the thing up without fireworks, you're then able to apply logic. The circuits are very straightforward to understand, so after suitable measurements and a bit of thought, you've found the root cause - usually in less time than it takes to research and source and fit substitute transistors. And by avoiding the "blind" blanket-change approach, you've avoided the risks of introducing new faults along the way.

Beyond a quick check for shorted transistors, I hardly ever bother with "cold" checks. Get the thing safely powered and follow your nose. Simple

[ppppenguin]

Like Mark, I'm a great beleiver in current limited supplies. Whether that's lamp limiters, bench PSUs or just good old resistors. Plenty of kit will run after a fashion on very limited power while you find out what's gone wrong.

What would really scare me is some power electronics where it can't run without some very hairy voltages and currents. The guys who maintain electronic controls for locomotives, power inverters and such like must have their methods of avoiding death and destruction.

[Murphyv310]

Hi.
Yes I am with Mark & Jeffrey. Never skimp on test equipment in a workshop, stabbing in the dark is not the logical way to go around things.
I gave up full time repairs 10 years ago but I still do amps as I enjoy it. I never sold on or threw out any of my hard earned test gear as it's paid for itself many many times.
Repairing DC coupled amps is nigh on impossible without a current limited supply or something improvised.
I just take it for granted these pieces of equipment will be used by any engineer doing repairs.

[Old Sparky]

I'm a member of HiFi Engine and have already downloaded the service manual, studying the circuit. AS for current limited PSUs, I agree that they are very useful, but I only have two, one goes up to +22v @ 2A max, the other up to ca. +18v @ 10A Max. The latter is an ex car radio demo stand PSU, the former I built many years ago to a design published in PW. Since I mainly work on portables & car radios, these are adequate for those jobs, though I have succesfully repaired quite a few HiFi amps. over the last 10 years or so without the problems I'm having with the HK1400. As to what is termed 'blanket replacement', I'm sure I read somewhere, maybe on this forum or on UKVRRR that, because, as I well know, an o/p stage failure can damage transistors upstream of the output, so replacement of all Transistors in the defective channel is the way to go. I'm sure someone said that, in his days as a service engineer, this was less time consuming than any other way to proceed.

[Mark Hennessy]

I'm a member of HiFi Engine and have already downloaded the service manual, studying the circuit.

Perhaps my hint was too subtle

I wasn't proving the link for your benefit; rather, for anyone else who fancied getting involved.

So, speaking very broadly, I was trying to suggest that it would be nice if the OP could provide the link (or attachment) in these sorts of threads. I'm finding that this is a recurring theme across forums in general. In short, when asking for help, providing easy access to the relevant data - along with anything else that might tempt someone to get involved - will maximise the chance of receiving help. Yes, we can all use Google, but why not help people to help you?

AS for current limited PSUs, I agree that they are very useful, but I only have two, one goes up to +22v @ 2A max, the other up to ca. +18v @ 10A Max. The latter is an ex car radio demo stand PSU, the former I built many years ago to a design published in PW. Since I mainly work on portables & car radios, these are adequate for those jobs, though I have succesfully repaired quite a few HiFi amps. over the last 10 years or so without the problems I'm having with the HK1400.

I would not describe them as "useful". They are essential. Absolutely essential.

And I should clarify what I mean by "current limited". Any half-decent power supply will have current limiting, but that value of current is fixed - perhaps with "foldback" - and is there to protect the PSU from abuse like a short-circuit across the output.

But that's not what we need here. Rather, we need power supplies that have an adjustable current limit, right down to milliamps. A current limit that you have complete control of from the front panel. To protect the device under test.

These are the norm in electronics bench work. There are exceptions (e.g., the Farnell E series), but the vast majority should be properly adjustable, like the examples I gave earlier.

No idea if your 2 supplies fit that category, but if they don't, current sources or light bulbs can easily be added.

As for the voltage, audio amplifiers built around the basic "Lin" topology are pretty tolerant of varying supply voltages, and will happily work down at a much lower level than usual. I usually start at around +/-15V (current limit set to 50 or 100mA).

As to what is termed 'blanket replacement', I'm sure I read somewhere, maybe on this forum or on UKVRRR that, because, as I well know, an o/p stage failure can damage transistors upstream of the output, so replacement of all Transistors in the defective channel is the way to go.  I'm sure someone said that, in his days as a service engineer, this was less time consuming than any other way to proceed.

It's a commercial decision. I've done it in the past when devices were readily available and time was short. But here, you have to factor in the time it'll take you to identify and source the equivalents...

With current-limiting provided (in whatever form you can - even via a lamp limiter), you are able to analyse the problem and fix the thing that started it off, along with all the casualties taken along the way, and perhaps a few other parts that you fear might have suffered during the incident. That's the intellectually satisfying approach, and quite possibly the most time effective approach in this situation. Or perhaps not? Only you can decide, as I said before. I'm not against "blanket replacements" as such, but it carries the risk of creating more problems and is not intellectually satisfying. But each to their own - every situation is different.

Whatever route you take, good luck and keep us updated

[Murphyv310]

Blanket replacements were done when a common fault was detected and the only answer was say from a kit of parts.
In this case the amp has failed and there is no manufacturers back up so the process of elimination and test is required. Firing a load of transistors will never do and you'll never know what the actual fault was.