Technics SA-5760 - one channel repair

[ViperZ]

I have no way to match small signal transistors unfortunately. I agree, it's a good possibility that transistors from the same batch might be matched close enough. I just want to make sure first that the amp works as it is before starting any upgrades/updates.

A quick update.
* Turned on the power supply, voltages for amp were spot on (+/- 73VDC); amp not installed; protection relay clicked.
* Reinstalled amp board. Output boards not installed yet (and small board that sits on the heasinks). Turned on the receiver. Protection relay clicked. All four driver transistors got very hot after 1 minute. Turned it off. (other parts were cool to the touch on the board - electrolytics and large ceramic resistors).

Trying to figure why drivers would get hot. Is that because the output stage is not installed? But at the same time, why would they fully open with no outputs connected?

Should I continue and install the output stages, or start troubleshooting at this stage?

 

[ViperZ]

Installed both output sections today back in the amp. Driver transistors are cold now, which is a good sign. Still have a problem in the amp section.

Relay clicks on, then a few seconds later clicks off. DC offset is ok on both channels (20mV on good channel, 50mV on rebuilt channel - so easily adjustable if needed).

Now onto Icq. Bias current. It's a few mV at a measurement point, which is good, but not good on a rebuilt channel.

Looks like this problem was there to begin with - VR604 was dialed down to a minimum. VR604 adjusts Icq on a new (right) channel. I originally set it in the middle, and got 400mV at test points.

I have about 130mV when VR604 is dialed all the way down (and protection relay stays engaged). The expected value with VR604 in the middle is 7mV.

Driver transistors are cool to the touch, output transistors get a bit warm.

 

[ViperZ]

So with my limited knowledge... VR604 regulates current between two pre-driver transistors TR606 and TR608. But there is also TR604, which regulates how much TR608 opens, and it is listed as pre-drive current protection / control transistor in the parts list.

I do have all replacement parts, but I don't want to blanket replace all three transistors without figuring out, which one of them is causing the problem.

I guess, I'm pulling the amp board out again. Hopefully, I won't need to pull out output stages again.



Just to document further:

* TR604 - 2SC1509 NPN - replaceable KSC2383 - ECB to ECB
* TR606 - 2SB536 PNP - replaceable with KSA940 - BCE to BCE
* TR608 - 2SC1628 NPN - replaceable with KSC3503, except original is EBC, new one is ECB.

Since I'll be there, I might as well replace D602 and D604 - MA150 - to be replaced with 1N4148.


Variable resistor information, in case someone needs to replace those:
* VR601/602 - DC offset - 100 KOhm
* VR603/604 - Icq bias current - 1 KOhm
* VR701 - V adjustment on power supply - 2 KOhm

All five of mine seem to work smoothly after some deoxit treatment, but I do have replacements from Bourne just in case.

 

[ViperZ]

I think I'm at a point where I need more qualified help. I checked TR620 connections to the amp through wires just to make sure - all solid. Then I removed the board out of the chassis.

Every semiconductor (transistors and diodes) in that section of the amp looks identical when checking junctions compared to a good channel. I checked TR602, 604, 606, 608, 610, 612 and 614. No shorted capacitors on the board. I checked the resistors that are shown as 'critical' (shaded ones) - the values are spot-on.

I do not have a transistor tester to start removing and checking the parts out of circuit... Anything else I can do short of doing a shotgun approach?

 

[Oldsansui441]

Hi, if it were me l would be removing all transistors on that board and performing a 6 way diode test on each, esp the pre drivers. There is no accurate way to test them "in circuit". Same with the diodes, sometimes they cannot be checked properly in circuit either depending where they are. Do you have a diode function on your multimeter?
Those drivers shouldn't have been hot with the outputs disconnected, l am betting the other channel drivers were cool with the outputs removed?

Check ALL resistors on the board also, you may have to lift one leg on many to check accurately.

Don't power up again until you sort it out as the 130mV bias will likely destroy your new outputs and drivers if it hasn't already.

Unfortunately there is no easy way to diagnose problem without going through and manually checking all individual parts.

 

[ViperZ]

I pulled TR602, 604, 606, 608 out - did 6-way diode/junction test - all good, put them back. Removed D602, 604, 606, 608, 610 - all good - replaced them with 1N4148s (D612 already replaced since it was blown).

Drivers and outputs are ok.

(Drivers initially were hot because I didn't have outputs installed; they were cold when I installed outputs. Also, outputs barely got warm with this nasty bias current, but I only turned it on a couple of times for 10-20 seconds to check for DC offset and Icq on each channel. I guess, I got lucky there).

 

[Oldsansui441]

Ok, if all of the transistors, diodes and resistors check out the next thing to do would be to inspect all tracks and solder joints just to be sure there are no small breaks or solder bridges etc. Also confirm power supply values and earthing/grounding to that board. Might be a good idea to remove the bias trim pot and confirm resistance values on this also. I had one similar a while back that would adjust perfectly except for down the lower end, it was clean and unbroken but just wouldn't go down to lower values.

Are you 100% certain all transistors are installed with correct pinout? Guess you have confirmed this though when you took the voltage readings at the transistor legs.

 

[ViperZ]

Yeah, it doesn't help that output section is on separate boards, as well as that little board with a pair of transistors on the heatsink. So every wire needs to be checked. I still feel the problem is somewhere on this amp board.

I only changed 7 transistors (4 outputs, 2 drivers and one current limiter). I'll double check the datasheet on current limiter, but I am very sure all new transistors had identical pinouts to the originals.

I only checked Icq and DC offset on each channel. I didn't check voltages on individual transistors - too risky with this high Icq.

I did sweep the Icq trip pot (with power off, and it sweeps uniformly from one end to another. I had a similar problem where an open in Icq pot in another receiver took out pretty much every transistor in that channel. I can adjust the current with that pot quite uniformly (i.e. it was 400mV reading in the middle, 130mV reading at minimum), but of course it doesn't help if I'm looking for 0-7mV reading.

I will go over all resistors and solder joints this week. Back to work tomorrow...

 

[Oldsansui441]

Let us know how you go, must be something small NQR, just takes time to find sometimes .

 

[ViperZ]

Swapped out the protection relay. Interesting observation - little metal shield around the relay also holds two super-strong magnets (one on each side of the contacts). I wouldn't think that modern relays require those, but still put them back there when I installed new LY2-0-DC24 relay. Will measure the coil voltage next time I'll power up the receiver, since old coil voltage is 680 Ohms, and new is 600 Ohms. 1.2K 2W resistor in series with relay coil. So this resistor will be getting 1.846 Watt, assuming 71VDC across the whole circuit.

 

[quaddriver]

<insert eating popcorn emojii here>

interested 'cuz I had a customer drop 2 5770's off (which are slightly larger, consume more power and take many of the same subs)

A few notes on what I read:

if you read a resistor in circuit and it reads HIGHER than it should (like way higher, outside its tolerance) then it is bad. since you are measuring it, it cannot be in series, and any two resistors in parallel will always read lower than by themselves. however, you noted you should have 2.7K but got 3.9K. Ok, that sounds bad, but are you sure it should be 2.7K? 3.9K is suspiciously close to a valid, common value and I see a bunch of 3.9Ks in the 5770. If the body is visible and it is red purple red, then yep, should be 2.7K, otherwise can those 3.9K resistors be thought to be orange white red?

also, the MA150 is a strangely named 1N914. (its 35v) If you have 1N914 ok to use,( I got a 1000 of them for a buck....) otherwise 1n4148 is a good sub.

on the amp, replace as is, but if you are in the fm section, often around the ratio detector and AGC's those 'pairs' should be matched up to dead on balls accurate (just make sure they aint Ge, the Vf is about half...)

Lastly, Id need to check the SCM for topology, or else bypass the relay, but you generally can run the headphones on the drivers, which helps you proof out your inputs, mirrors, vas, drivers etc.. clean and verify the DC balance and bias pots or just flat out replace. too much bias at power on is a fast way to see what color smoke your new subs make (hence the typical DBT recommendation)

since this is a meaty receiver, I would agree that you need to pay attention to the power supply. If any part was bad, assume ALL bad and whack all the 'voltage output' units as they call em and change the zeners and any critical resistors and do a voltage test FIRST. A machine like this, will weld stuff when it goes bad and it would suck if the stuff it was welding was your living room.

the thread has been quiet for a month and a half so Im assuming you are yanking hair out by the handful at this time. dont fret, smart cookies here and every piece of silicon can be subbed around.

 

[ViperZ]

Thank you for the feedback. I ran out of ideas and pulling out my hair, so put it back together for now.
I checked all resistors, swapped all diodes, etc etc in that rebuilt channel. It even turns on the protection relay, but eventually bias current in that channel climbs to an unacceptable level.

This is basically where I got stuck:
I can adjust the current with that pot uniformly (i.e. it is 400mV reading in the middle, 130mV reading at minimum), but of course it doesn't help if I'm looking for 0-7mV reading.

The bias current readout is 130-400mV whereas it is supposed to be 7mV...

I did replace that one transistor - TR614 - 2SA777 with KSA1013. It is a current protection transistor, but I'm sure the replacement should be a proper one...

 

[quaddriver]

ok been peeking at the SCM from hifi-e and the 5760 is scarily close to the 5770. first off lets ignore the fact that to make the power, the machines used 2 sets of outputs in parallel. which means if one fails, its like a cascade effect and I bet an entertaining light show.

the 614 you yanked 2SA777 is subbed close by the KA1013. in early datasheets of the A777 it had Pc at 750mw and the A1013 is 900mw, but later A777 datasheets have it at 1w. this might be a problem. other uses of the A777 and C1509 pairs in the amp are the protection relay driver board and the +/-40v power supply. There,, there is enough current limiting devices around of such much lower wattage I aint scared (in fact, I just updated the 5770 +/-40v supply with the A1013/C2383 subs)

in the amp driver section, these appear to be such that if turned on by runaway bias, hence output, hence more power across the 390 ohm resistors I could see them going thermonuclear within the limits of the 330 ohm resistors and the 900mw vs 1w could be a deal breaker...in that case, A1220/C2690 subs are meatier.

anyways, you have 1 good channel and 1 bad right? the good channel sets good and stays good?

pull the bias resistors (4.7ohm) for all the outputs and all the emitter resistors (.33ohm). this isolates the outputs and using the good side, verify the bad side piece by piece, and if need be, inject a 1khz 150mv signal on either side and trace its behaviour.

Im of the mindset that if ONE of the limiters was bad, the other is too, 614's compliment is 610 no? start there.

 

[tarior]

A number of people have had trouble with excessive bias current when using MJE15030/31 for drivers..
Try 2SC4883/A1859 (Digi-Key) for the drivers. They have considerably better gain than MJE15030/31, and I think, better Vbe characteristics.

 

[ViperZ]

I really appreciate both of your inputs. I will reopen this receiver next week again and will try to follow your steps and suggestions. Will keep everyone updated. I really want to fix and keep this receiver for myself.

EDIT: so I am ordering a couple of pairs of 2SC4883 and 2SA1859 (not A versions) to replace MJE15030/MJE15031, which I used in place of original blown 2SB537 and 2SD382. The electrical properties of 2883/1859 are definitely much closer to 537/382 than 15030/15031.

The only other transistor substitution that I have done was 2SA777 replaced with KSA1013. Would it still be ok to keep ksa1013, or is there a better replacement?

Thanks again!

 

[ViperZ]

I didn't forget about this receiver! I received 2SC4883 and 2SA1859 transistors. Hopefully will find time between work and family in the next couple of weeks to take it apart, pull that board out, replace transistors, etc. I also bought a NOS 2SA777 transistor on ebay from a local Canadian seller, so if needed, I can pull out new KSA1013 and put this 2SA777 in the circuit.

 

[ViperZ]

FFS I feel really dumb... Pulled out 15030/15031, and then opened the bag with new 2sc4883/2sa1859... Back to square one. 4883 and 1859 are plastic package transistor (not metal back). SA-5760 uses METAL BACKS OF TRANSISTORS as collector connections (middle leg is cut off and there is no hole in the board for it).

 

[tarior]

You'll just have to make a connection. Get creative. There aren't a lot of connected tab transistors that are also good drivers these days.

 

[ViperZ]

There is no connection on the top side of the board - the heatsink is aluminum - can't solder to it. This would look really ugly if I ran the wire around the whole board to the back side...

 

[tarior]

Can you drill a 1/16" hole through the board?