Sony HP-511 - Blowing Fuses

[GP49000]

I'm afraid that since the rectifier diodes and smoothing capacitors are OK, one of the power amplifier output circuits has shorted transistors and probably the direct-coupled driver transistors, too (if even ONE of the outputs has shorted, the other one and both drivers really should be replaced, because even if they haven't failed, they've taken a big jolt).

This is a quasi-complimentary output circuit whose "centerpoint" operates at 1/2 the B+ voltage; so there is a big, 2200UF/35VDC isolating capacitor between the output of the amplifier and the speakers to block that DC voltage. If that capacitor shorts. the output stage blows instantly and catastrophically, and often so does the loudspeaker. More often it goes open, resulting in no signal.

[gazrawly]

Hi GP49000

Thanks for helping..

This sounds a little terminal but hopefully something that can be saved? I can see two 2200uf 35v capacitors (C658 and C558). I guess I need to remove the board and remove them both from the board to test them? Would it be possible to point me in the direction of the direct-coupled driver transistors? I am guessing the other transistor would be part of the D291 collection?

Do you think that the output speaker glass fuses would have saved the drivers? I have just checked these and they have both blown also. I guess this would tally up with your post

Thanks
Gareth

[GP49000]

Factory-installed driver transistors are specified as 2SC1124 and 2SA706; output transistors are 2SD291. The transistors' markings will assume the "2S."

If unit has been serviced, other transistors may have been substituted. It's not beyond repair but the fix could get complicated. I've fixed a lot of this type amplifier; it was a very commonly-used circuit in the late 1960s/early 1970s but could be troublesome to fix because of how many associated parts could fail if something let go. Sometimes already-weakened parts would fail on power-up, or I'd miss one that had failed. Monitoring current draw and the centerline voltage (which should be approximately half the B+ power supply DC voltage) are essential when trying to fix these.

Unfortunately the fuses probably took longer to blow than the transistors.

You can check the output coupling capacitors the same way you checked the power supply reservoir capacitors after removing them from the circuit.

But you never know how something will turn out. I worked on a Dynaco Stereo 120, same basic circuit topology but much higher power (60 watts/channel) that had one output transistor shorted and nothing else...and it worked when that one transistor was replaced and I powered it up...still, I replaced all the Class B transistors (drivers and outputs) anyway since I didn't want it coming back for rework.

[GeoffK]

Originally Posted by gazrawly

This sounds a little terminal but hopefully something that can be saved? I can see two 2200uf 35v capacitors (C658 and C558).

One of those 2200uF 35V capacitors looks swollen with a mark on top, so faulty, both would need replacing. The driver transistors are probably the ones with the heatsink tag in front of the capacitors.

[julie_m]

If you were lucky, the output transistors, when they failed, would have held the supply voltage to low for anything else to be damaged by the bad DC conditions until the supply fuse blew. This tends to happen about four times out of every five. (And that fifth time can be a nightmare, so let's hope we don't have to go there.)

A faulty speaker coupling capacitor could certainly blow the output transistors in that channel. And its neighbour is the same age, so it would be prudent to replace both. The other channel may well be OK for now, but a new capacitor is still cheaper than two new output transistors and a new capacitor .....

You can do a crude test on the output transistors without unsoldering anything; just probe collector to emitter, both ways around, on a low resistance range. If they read dead short both ways, they have definitely had it. If you get high resistances (the same or different in each polarity) then you will have to unsolder the transistor to investigate further. If one of the pair is short C-E but the other tests normally, replace both anyway.

The odds are in your favour that the two output transistors on one channel have failed short-circuit (possibly due to a dead speaker coupling capacitor, which also should be replaced), and everything else except the fuse has survived.

Just another thought: That capacitor failing could conceivably have taken out the loudspeaker with it. Even if it still shows continuity, a speaker can be damaged in ways that will only show up when tested with music. Can you test your speakers on another, known good amplifier?

[Boater Sam]

There is another single 1.6A fuse the other side of the main set of fuses, is that blown? What is it connected to, sorry, no diagram?

[gazrawly]

Originally Posted by Boater Sam

There is another single 1.6A fuse the other side of the main set of fuses, is that blown? What is it connected to, sorry, no diagram?

I've just had a look inside and this fuse goes to the main power switch on the front of the unit and to the 110v/240v switch on the back of the case. This fuse is not blown

[gazrawly]

Originally Posted by Hetrodyne

One of those 2200uF 35V capacitors looks swollen with a mark on top, so faulty, both would need replacing. The driver transistors are probably the ones with the heatsink tag in front of the capacitors.

Closer inspections, it looks as if both capacitors are slightly swollen (watching the top and running a light around the capacitors, the reflection is slightly distorted from the middle...

Going to remove the amp board now.

Thanks everyone

[Techman]

Bad luck that it's not a simple diode rectifier fault after all - dirt, and pictures can be confusing sometimes.

I would definitely check the speaker connections for shorting wires, particularly if they're those 'din' plugs, as shorts in those are the fastest way to blow an output stage.

[Diabolical Artificer]

In short what you need to do is check the 2SD291 OP (output) transistors with your DMM diode checker by touching the red lead to pin 1 (see datasheet attached)then touch the case and pin 2 in turn. If they're ok you should get 0.6v ish, if they're dead you'll get the beep of death.

Same with the driver transistor's. the datasheet can be found by googling. Other search engine's are available.

Replace the two 2200u caps with something decent, IE a capacitor made by Rubycon, Nichicon, Nippon Chemicon etc. Panasonic do some nice gold coloured jobbie's which have good specs.Make sure you put them in the right way, IE negative to negative etc.

DO NOT TURN IT ON Build a lamp limiter or solder two 470 ohm resistor's in place of the fuses that blew. Before that set your DMM to resistance and check the resistance from ground to the positive of one 2200u cap, then the negative of the other. You should get a resistance over 1k or 1000 ohm's or thereabouts. If you get a low reading say 2 ohm's ish, you have a SC (short circuit) and will need to find it's cause.

Once you think there are no short's, turn on either using the LL or R's replacing the fuses. Keep you your finger on the switch in case of smoke.

As other's have said above better than I can, there is a lot that can go wrong and these arn't the easiest of amp's to fix, but you may be lucky. Nothing ventured and all that.

Andy.

[gazrawly]

I had some troubles removing one of the amp board screws but finally managed to remove it.

I've measured the OP transistors and as you all expected, I have two dead transistors (possibly three)..The readings (measuring on the lid of the transistor to Pin 1 and then Pin 2);

1. 0.582v / 0.503v
2. 0.575v / 0.500v
3. 0v / 0v
4. 0v / 0.374v
5. 0.387v / 0.550v

I coudn't seem to find an exact datasheet for the driver transistor's. So I followed an online guide using 200k Ohms reading on the DMM as a crude measurement. 0507 2SA706 and 0508 2SA706 read 0.00. I checked this against another transistor on the circuit board that appeared to be identical (0608 2SA706) and this measured an increase of resistance in one direction (0.20+) and a decline in resistance in the other direction (measuring both outside pins of the transistor). The middle pin to outside pins (swapping polarity) delivered the same results on the driver transistor's (no reading)... I measured these whilst still soldered to the board.

I am going to remove the 2200u capacitors and grab the parts to make an LL this weekend hopefully. Fingers crossed.

[Diabolical Artificer]

Oh dear, as suspected two or three dead transistor's. Just had a quick look on ebay, Littediode has them at
£22!!? Hells teeth, that's expensive. Someone here may be able to recommend a cheaper alternative, (maybe 2N3584?) but I'm afraid TO-66 trannie's aint cheap.

That said your description of testing confused me, if you put the red lead on the case (collector) you won't get the correct reading. A transistor is effectively two diodes in one package, the base - pin one, is the common anode for both.

The 2SA706 is a PNP trannie, so you use your black lead on the base, so without a pinout, IE knowing which pin is which you use your diode checker as before, muck about till you get a reading of 0.5v ish, leave the black where it is, then touch the other pin's.

Transistor pin out's follow no known logic AFAIK, there is this here which may help - http://www.zen22142.zen.co.uk/ronj/cpt.html and of coarse there's a sticky thread up top about how to test transistor's.

In light of your finding's you might want to try and find some transistor's before spending any more time on this amp, unless your rich that is.

Andy.

[Boater Sam]

There are some from Greece cheaper. There is probably a lower cost equivalent, Sony bits tend to be a bit OTT price wise

[julie_m]

There's almost certain to be a lower cost equivalent (same polarity NPN or PNP, same or higher voltage and current, gain within 3 times either way, fT not much higher or else it may break into oscillation) in a TO-220 package, and there ought to be room to mount it on the heatsink.

At least if it doesn't blow up with inexpensive transistors, then you could feel more confident replacing them with the exact part later (or alternatively, upgrading the other channel with new transistors, if the one you just fixed sounds better). And if it does blow up (if I was a gamblin' gal, I'd reckon the odds as being about 4:1 against that), you haven't wasted a fortune on rare transistors to destroy .....

One other thing ..... Amplifiers with two transistors of the same kind sometimes have a preset potentiometer to adjust the standing current. Has yours? If so, and barring knowing the correct value, set this to minimum current, then adjust upwards for no audible distortion on quiet passages (if you have a signal generator and oscilloscope, feed in a sine wave and watch the zero crossings.)

[gazrawly]

Thanks again for all your help...

I've just had a chance to take another look/read into the problem - I did spot a transistor called the TO66, aesthetically this transistor looked identical. I will try and figure out if it is a suitable replacement. As you said julie, I don't want to blow £50+ worth of transistors into orbit if I can help it! ... The TO-220 transistors seem to be completely different with 3 pin layout. Please excuse my novice question, but how would I mount this to the board given that the current D291's have 2 pin's and are mounted with a unique looking heatsink?

Thanks Wobble, I will look into retesting the transistors again... I thought £20 per D291 was expensive too, crazy eh! But this record player is sentimental so regardless of cost (to a degree) I would love to get it working again and have it running in the living room.

[TonyDuell]

All (bipolar) transistors have 3 connections called emitter, base, collector. In the TO66 package (TO66 is the name of the package/housing the transistor is contained in, and has nothing to do with the electrical properties), the collector is connected to the metal case, the base and emitter are the 2 pins.

In the TO220 package, there are 3 pins. In most cases the collector is the middle one, and it's connected to the metal tab with the screw hole at the other end of the package. The base and emitter are the outside 2 pins of the TO220 package.

So you can, in most cases, fit the outside 2 pins of the TO220 package into the PCB holes for the pins of the TO66 package. And screw the TO220 package down with one of the screws that held the TO66 package in place, this will make the collector connection.

There are, of course, some 'gotchas'. The first is that there is likely to be an insulator, a thin pice of mica, under the TO66 package to prevent the connector being connected to the metal heatsink/chassis. You need to do something similar with the TO220 package, the 'mounting kit' is a standard part and includes that mica insulator. You should put a smear of heat transfer compound on each side of the mica. Make sure the collector is connected where it should be and nowhere else.

The other main 'gotcha' is much rarer. The TO66 package is metal and has 2 screws. Therefore these screws are electrically connected together when the transistor is fitted. A very few units use that fact to complete part of the circuit. There will be a PCB track to each screw and said tracks are not connected other than via the transistor casing. If you have something like that you need to add a bit of wire to complete the circuit.

[gazrawly]

I think I understand this Wobble (please shoot me if I don't!!). I have attached a picture of the probes on the OP transistor's - I am hoping this is the correct way of testing them? If so, the readings are pretty much identical to the readings from the casing lid;
1. 0.583v / 0.503v
2. 0.572v / 0.504v
3. 0.000v / 0.003v
4. 0.000v / 0.375v
5. 0.402v / 0.547v
(using the diode setting on the DMM). Transistor's 1 & 2 are close to the 0.5v figure, 3 & 4 seem to short? Transistor 5 is suspect with a 0.4v reading on pin 1?


Tony, is there any downside to using the TO220 transistor over the genuine D291? I am picturing the TO220 transistor's middle pin being bent at a right angle (to some degree) in order to screw it onto one of the D291 mounting screws. I found a 'fitting kit' for the TO220 transistor http://www.hobbytronics.co.uk/to220-mounting-kit does this (mica?) lay between the fitted TO220 and the PCB after being cut to size and the middle pin leg is then wrapped around the D291 screw thread and then screwed down in place?

[Diabolical Artificer]

Yep, same as before Q (short for transistor) 1 & 2 look ok 3 & 4 are dead, deceased, ex - transistor's. 5 might work, will have to look at the circuit.

You could try bunging in a pair of 2N3055's on one channel if there's room, they should work after a fashion. I would put your good pair of 291's in a very safe place and try and get one channel working with whatever replacement Q's. You have a steep learning curve ahead of you.

The TO220 packaged transistor is completely different to the TO66, for a minute there I had to check Julie and Tony weren't mistaken. I would have thought something in a TO3 package would fit better but they would be a lot deare, if as I said there's room. To get a TO220 Q to fit, say a BD139 you would have to bend the leads ( the first two) and snip off the third. The third leg is the collector which is also connected to the body. You will have to make sure the legs that go through the hole don't touch the heatsink. I would have thought a TO220 package won't sit well on the heatsink though, it'll be a bodge but may work to get the amp up and running.

The hard bit is getting the rest of the circuit working. If there are no more damaged component's you'll be ok, if there are it ain't going to be easy for a novice.

Andy.

[julie_m]

TO-66 and TO-220 are transistor package styles, not actual transistors. TO-66 is like a baby version of the famous TO-3 power transistor package.

Although they look very different at first glance, you will find that if you snip the collector lead, the base and emitter leads can be bent to drop straight into the original holes (use sleeving where they pass through the metal heatsink) with the tab lying flat over one of the mounting holes. It could almost have been designed that way. And since your amplifier makes the collector connections by means of a metal ring with tapped "ears" connected to the PCB traces, you don't even need to do anything creative there; just make sure the mounting bolt is in good contact with the metal tab, and apply a tiny smear of Thermpath to each side of the mica insulator.

BD139 is actually in a TO-126 package, which is different again but could still be persuaded to fit (she says, wielding a pair of pliers in a slightly worrying way). It has a lower maximum working current than 2SD291 (1.5A as opposed to 3A), so might reduce the available power output. BD439 is also in a TO-126, but rated 4A.

[Diabolical Artificer]

What would be the TO220 jelly bean equivalent of the BD139 Julie? Can't think off the top of my head.
A.