Harman Kardon HK775 repair help needed

[mbot]

hi, i've been working on fixing this HK 775 for a while now. i got this unit from someone who had tried to fix it himself but aside from screwing up a few traces on the boards hadn't really dome much but replace a few resistors only to watch them burn up again.

I began by replacing the burnt out resistor (R481) and fixing the traces on the board. i then checked around and replaced a few suspicious transistors (Q409-412). after checking what seemed like every component i powered it up using that lightbulb current limiter trick and the 100W bulb stayed bright so i shut it off and decided to check that the power supply was working right. disconnecting the power section from the rest of the circuit i tested the voltages and it all semmed right.

I then assumed the problem was with the output transistors (Q427-434) so i disconnected them and powered it up and the bulb dimmed after a brief glow. measuring the transistors it seemed most were open (no resistance E-C) so i ordered a new set and waited like 2 months for them and popped them in.

because i felt i had pretty thoroughly checked around for other faults i decided to turn it on and see what i got. i thought that the lightbulb trick would be sufficient to save the transistors should there be something else wrong. when i turned it on the bulb glowed for a second and then dimmed (hooray!). so i checked for dc on the speaker terminals and saw 55V which was descenging and went down to about 35V in about 10 seconds. shortly after that the dc voltage dropped to 0V and the bulb went bright. checking the transistors now in circuit shows they are all shorted E-C.

so, now i'm just wasting money trying to fix this thing but i am determined as i really enjoy a problem and i have to learn somehow. i guess my questions are:

1. any idea whats going on here? why would the voltage on the speaker terminals suddenly drop like that? i had thought that if the output transistors were shorted that i'd see the rail voltage on the speaker.

2. what could i have missed? what should i be looking for now? nothing seems to be heating up, and i've checked all resistors and caps. i haven't checked every transistor out of circuit because i'd rather not just start pulling out components randomly. where should i be looking?

3. I thought wrongly that the lightbulb limiter would save the output transistors but they appear to be dead again. is it a matter of using a lightbulb that allowed too much current? or is there something else i'm missing about how this is supposed to work?

Thanks in advance for your help.

mbot.

 

[(*steve*)]

Any DC voltage on the output is a bad thing and in retrospect a good sign to remove the power before something else happens.

It is possible that some part of the circuit driving the outputs is damaged and turning one output transistor hard on. The amplifier tries to get the output back to 0v by switching on the other output transistor. Eventually both die.

I would consider a lower wattage light bulb and/or adding some resistors in series with the collectors of the output transistors (say 10 to 100 ohms). If there is a short, the bulb will limit the current over time, and the resistor will both allow you to measure it and it will protect the output transistor. The resistor may smoke -- but it's cheaper than the transistors!

Check the transistors Q424 and Q425 and make sure one of them isn't shorted.

You should be able to get away with replacing a single pair of output transistors (with additional collector resistors) for testing. Only replace the rest when the amplifier is working correctly (i.e. it doesn't blow the output transistors *and* the quiescent output voltage is 0V, and you don't need resistors in the collectors).

The circuit diagram gives you voltages all over the place. Once you have the amplifier stable enough to test (perhaps with 100R resistors -- maybe even higher if these keep smoking -- up to 10k even!) measure the voltages. Look for asymmetry or wildly wrong voltages (note that the supply rails may be significantly lower due to the bulb you're using (15W?). Find where the deviation starts and that may be approximately where the fault lies.

The hardest part will be to differentiate between an error caused by a fault, and an error cause by the feedback loop trying to pull the output back to 0V. You could disconnect the feedback loop (lift one end of R423), but then you'd be straying into a different territory...

If you can get a slightly higher resolution scan of that circuit diagram it would make it much easier for us to read it too

 

[mbot]

thanks for the advice, i'll have a look in the the next day or so. meanwhile, here is a better version of the schematic.

thanks.

 

[mbot]

hello again, so here's where i've gotten to. i went through the driver and predriver stages checking each transistor with a dmm and then my 'new' ancient transistor tester i got working recently and i found about 5 transistors testing badly. so i replaced those and put in one set of the driver transistors with 100ohm resistors on the collectors and fired it up with a 15 watt bulb in place of the 100w one i used earlier. my main concern was to see really quickly if there was any dc on the speaker terminals and shut it off immediately so i might not toast the drivers again. it powered up, the bulb never went dim though, and there was only 5 volts across the speaker terminals instead of the previous 50v.

so, progress i guess? what i'm wondering is if the 5v might just be there because i need to do the dc offset and idle current adjustments in the service manual or if that is too large a voltage for that to be the case. also, if it isn't simple that it need some adjustment now then would that 5v be dangerous for the driver transistors? is this stable enough that i can leave it on and go poking around? i'd really like to not have to buy anymore of these transistors...

thanks again!

 

[(*steve*)]

5V is quite a lot.

Check the supply voltage and the voltage across those 100 ohm resistors.

The first will give an indication of how far the supply rails have sagged, the latter will indicate how much current is passing through the output transistors.

oh, and did your wire up all 8 output transistors, or are you running with just a pair at the moment? (I think I can read your comments as you only having one pair in now and them with the 100R collector resistors).

 

[mbot]

ya, i only have 2 of the output transistors in at the moment, Q427 and Q431.

so, i measured the voltages. the supply rails are at +7 and -7.4. the voltages across the 100 ohm resistors are 1.5v (Q427) and 0V (Q431). i guess that means there is 15mA going through the transistor. the Q427 side is where the majority if the dead transistors were in my last check through.

that amount of current seems well within the maximum ratings on the datasheet so does that means it would be safe to leave this on for a while to poke around? i've been taking measurements fairly quickly to try to avoid burning out the transistors again.

thanks.

 

[(*steve*)]

OK, what you have is the transistors on one side are being quite strongly turned on. to have the output at 5V when the supply is 7V and 1.5V is dropped across a resistor should tell you something...

I don't have a lot of time at the moment. I'll get back to you later.

The next step is to determine WHY the output transistors on one side are turned on. (Their failure is a symptom rather than the cause of the fault)

 

[(*steve*)]

There are voltage measurements on the circuit diagram. Starting with those on the output transistors and moving to the left, how do they compare to what is printed (they will likely be different due to the supply rail differences).

If you can give me the values on the circuit diagram and those that you read, I'll see if I can spot anything obvious.

 

[mbot]

Its been quite a long time...

Hello again, this is really old now. I moved a while back and have spent a year or so renovating and rebuilding have recently got a workbench back in my life again. so i pulled out some old projects that i had stored away and am back on this one now. The good news is i`ve got the amp stable and `working` (main problem seemed to be that D403 failed open). The bad news is it sounds like crap.

Every voltage on the schematic is fairly colse to the printed value except the +/- 0.61V around Q413. The best they get with the adjustment there is +0.35 -0.85. According to the manual this is to set the idle current by adjusting the voltage on the output resistors to 36mV, they are stuck at 0-0.8mv.

Q411 and Q412 get a bit warm while its on but i've read that this amp runs kinda hot anyway.

Putting a sine wave signal in shows that the positive side is squished on the output, not cutoff flat but a short fat curve where it should be taller. I assume this is because the +0.61V is not there on Q413.

Does anyone have an idea of what i should be looking for now? I realize this repair has ended up costing more in time and parts than its worth but its kinda my white whale right now....

Thanks in advance,

 

[(*steve*)]

Yeah, something is going on with the biasing of the amplifier.

Maybe someone else can take a look and see if they can spot what it might be.

 

[mbot]

Updated Schematic

Hi Again, I just went through and marked all the actual voltages on the schematic. They are in red near the normal ones. Hopefully this can help someone out there to point me in the right direction (or, i suppose any direction)

 

[mbot]

better quality

hmmm, this will be better quality....

http://i46.tinypic.com/rba2dv.jpg