Pioneer SX-580 Receiver Repair

[Smoky]

I pulled that transistor (Q502) and it does test bad on the workbench. When I touched it with a wooden stick while it was in circuit, I could see the front panel lights go dim and flicker.

*7 April 2020

The new transistor finally arrived! It looks different than what was posted on Ebay (no bent leads) but it fit perfectly.

Now after I powered the receiver back up I noticed that the "tape monitor" and "source" LED's would not light up. I started checking voltages and found no power at the base and the emitter of Q401 (2SD525). Q401 tested good out of circuit so I pulled Zener diode (D408) for an out of circuit test too. It is marked as an MZ-140 (WM-140) and it turns out to be a 14v Zener. It tested having continuity both ways using my DMM set for a diode test.

I then gave it a test using my power supply set to 18VDC along with a 1k resistor. As you can see, the DMM measures around 115mV across it. I just read that a Zener diode will read a lower voltage if it is shorted:

After the new Zener diodes were installed, the LED's came back to life on the front panel and the voltages within the receiver went back to normal!

Next was to connect the new Power Pack modules. Guess what?

I was able to dial in 14 mV DC of Bias on both modules as required. When I put together the RCS16 Circuit, I distanced the board from the transistors, doing this gives great access to the emitter resistors from above with grabber hooks when making the Bias adjustment.

Here's the surprise of the night though, I turned the receiver up-side down to measure the "DC Offset" from the soldered-side of the PC board and here's what I got (also at the speaker terminals):

Left channel:

Pin 3, 231mV
Pin 8, 234mV

Right channel:

Pin 3, 44mV
Pin 8, 44mV

Quite a disparity between the two channels. Funny thing is, the service manual doesn't explain how to make adjustments between the channels. I come to find out that the lower-end Pioneer receivers have no adjustment for "DC Offset."

But wait...

[Smoky]

It was hinted to me that the problem with the "DC Offset" is due to two of these 2SA798's. They are two transistors in one and they leak as they get older.

I was told to buy a handful of KSA992FBU's and solder two together by joining the emitters. Will this work?

So I made a little test pad since I don't own a breadboard yet and soldered it into the PC board holes of the removed 2SA798's. Both emitters of each pair of new resistors are tied to one green wire. I plugged in the first four KSA992's:

Here is the result from the Right Channel (55 mV) and the Left Channel (~38 mV).

Notice how the Left channel came down considerably from around ~230 mV!

For the next hour or so, I'm going to play "musical chairs" with 25 new transistors to see how low, and even, I can get the "DC Offset" 

[andy3055]

Those "dual" transistors will have the same variations of gain etc. due to temperature variations as they are in one enclosure so to speak.  This configuration is called a "long tailed pair." If you are using 2 individual transistors in each case, you may have to keep them close together by some means. I think you could use a piece of heat shrink so that the casings are in contact and any temperature variation is equal in them. If one of them is warmer than the other, they will heat and compensate each other. Not sure if what I say is making sense!

Make sure to tie the emitters together before the 13k resistor. Below is a nice write up on the "Long tailed pair."

https://www.electronics-notes.com/articles/analogue_circuits/transistor/long-tailed-pair-circuit.php

[Smoky]

I see exactly what you mean Andy.

Here is a closely matching pair. Not the lowest mV but even.

Most of them seem to linger around 50-ish mV until I decided to touch one of them with my finger. Immediately, the mV's dropped like a rock!

I then pulled out two of these Radio Shack clip-on heat sinks for one transistor while gripping the other with my finger. This gave me the best overall readings.

Here's the weird part, if I grabbed two transistors of the same pair with my fingers, the mV's would climb.

One other thing, while on the Variac at 117 VAC, I could get a pair down to 5.8 mV. But when I plug the receiver straight into the wall outlet in my house (~122 VAC) the same set of transistors is at 28-32 mV.

I will try a variety of heat sink combinations. Btw, inside of a Sencore oscilloscope I can see how they bonded two transistors with epoxy onto a tiny bar of aluminum.

Good stuff!

[andy3055]

When you touch them with your fingers you will be introducing noise because this is a high gain stage. Best is to use a piece of Aluminum or heat shrink and hold them together. Keep the leads as short as possible to minimize pick up of stray signals. I would not use any kind of glue but you could use some heat sink compound between them so that any imperfections get filled up with it that will make better heat transfer and keep them at the same temperature.

[Smoky]

And yes, the emitters were tied together under the sockets. The green wires tie the emitters. The black wires are the bases, and the white and violet colors are the collectors. The heat shrink identifies the pairs.

I will be able to put every transistor into its corresponding hole on the PC board once the pairs are selected.

I bet a lot of the noise is from the 6" length of wire I used too. I still see variations between transistors though.

[andy3055]

Transistors will have differences, no doubt. If you have a transistor checker that shows you hfe etc. you will be able to narrow down some more.

[aqibi2000]

Transistors will have differences, no doubt. If you have a transistor checker that shows you hfe etc. you will be able to narrow down some more.

This is where the curve tracer mentioned earlier would be handy

[Smoky]

I tell you, behind the scenes, I've been trying to find a curve tracer. Someone locally has a Tektronix 575 but he stopped communicating with me by e-mail. I'll get one soon enough though.

Anyway, I'm going against what's recommended and went ahead and bonded two transistors to a small bar of anodized aluminum. I used Super Glue Gel and it grabbed like crazy. I still have some of the black epoxy from a previous project and I may dip these assemblies in it

Btw, I was thinking, would anyone know if there is a newer SOIC package that is similar to a 2SA798 with matched gain built into it? Why I ask is, it wouldn't be too hard to produce a small board that pins could be added to and inserted  into the main Pioneer PC board.

I'm just a weekend-warrior when it comes to electronics so I'm just throwing stuff at the wall to see if it sticks!

[Smoky]

The reason I off-set the transistors on the bar is because it allows the legs to align to the board with just slight bending. I clipped one emitter leg and swung it over to the other and it will be silver-soldered at that joint.

Does it matter which side of the board the transistors are mounted on?

Here is a transistor being mocked-up.

*Update*

I installed the new transistors and plugged the receiver into my home's wall outlet (~122.5 VAC). Here are the "DC Offset" results:

64.9 mV at the Left Channel speaker terminals
62.3 mV at the Right Channel speaker terminals

Voltages at the new transistors:

Right Channel:

B  +0.2 VDC
C  -29.4
E  +0.7
C  -28.1
B  +0.2

Left Channel:


B  +0.2 VDC
C  -29.4
E  +0.7
C  -28.1
B  +0.2

They are identical. But, the collectors should measure a little lower in voltage per the schematcs (-26.5 VDC and -25.3 VDC).

So I put the receiver on the Variac and lowered the AC voltage until the voltages at the transistors aligned to the schematics. When the voltages at the transistors aligned to the schematics, the AC voltage at the Variac was at 111.8 VAC output.

But this is the most important part:

With the voltages correct at the transistor's collectors (-26.5 VDC and -25.3 VDC), and the Variac at 111.8 VAC, the DC Offset dropped to 33.4 mV for the Right Channel and 35.8 mV on the Left Channel!

Guys, we need to find a way to lower the voltage to the transistors while the Pioneer receiver is plugged into a 122.5 VAC outlet.

...and I make many edits to my posts, so please check them regularly

[andy3055]

If the space is there, mounting under the board is no harm as long as you don't mix up the legs looking at it from the wrong side!

[feedback.loop]

I tell you, behind the scenes, I've been trying to find a curve tracer. Someone locally has a Tektronix 575 but he stopped communicating with me by e-mail. I'll get one soon enough though.

Do you really need a curve? Perhaps some particular parameter should be matched.
How about a combination of a programmable power supply to sweep through some range you want, a meter with a computer interface too, and a little program, say, in Python?

[Smoky]

Thank you Feedback.Loop!

I'm not too familiar with the proper testing of transistors, so for the moment, I used the Pioneer receiver to determine a matched pair, at least in the "DC Offset" category.

I purchased 25 of these transistors for $5 from Digi-Key. The four I ultimately installed, tested the best that way.

The DC Offset voltages remain stable so I imagine that's a good thing?

Anyhoo, I hope to own one of those nice oscilloscopes like you have that does Frequency Response too 

[Smoky]

Guys, as you can tell, this receiver was a total basket-case, and it's not even close to being over!

There is another problem that surfaced when the new Power Pack modules were installed. There was little volume coming out of the Bose speakers that were attached at the time.

The good thing was, I was able to hear FM Stereo. Music was coming out of the Left and Right channels but at about .01 Watt per channel.

I was also hinted that there is an issue somewhere in the muting circuit.

I'm going to start digging in to that area to find the problem. I see two 2SK117's, and plenty of other things to check and to take voltage measurements at.

Let's keep it going

[Smoky]

You're not going to believe this!

If you go back to reply #10, remember resistor R410 trapped under the glue while being cooked by that 150 Ohm 5 Watt cement resistor? Well, that R410 is the 100 Ohm resistor that drops the voltage to those dual-transistors!!! When the R410's value went haywire due to the heat, the voltage rose on that rail and did the damage to those dual-transistors!

I just measured the voltage at the new 100 Ohm R410 that I installed during the re-capping process. It measures -29.6 VDC. I want to bring it down to -26.6 VDC as noted on the schematics.

I measure -31.1 VDC feeding the other end of R410 (-28v Rail, higher due to the voltage at my home's outlets ~122.5 VAC).

So we need to go from -31.1 VDC to -26.6 VDC.

If we can adjust this voltage, it will correct the voltage going to four very important transistors in this receiver, and in turn, the DC Offset should drop to the 30 mV range.

This is good stuff 

I'm guessing the current is around 15 mA.

So would the resistor be around 300 Ohms?

I just found a 1/2 Watt 330 Ohm resistor in my stash that measures 319 Ohms. I'm calculating that it should bring the voltage to -26.3 VDC. Should I try it?

[andy3055]

At this moment are you connecting the speakers? Providing there is no damaging DC on the speaker terminals, I would put it on the Aux input and feed a 150 mV signal and see if there is sufficient volume. If there isn't, then you still have a problem in the power amps. If there is good volume, then you can look towards the Preamp area.

If you don't have an audio generator, you can use this site: http://onlinetonegenerator.com/ You can take the output from your sound card and hook up to a pair of RCA jacks.

[Smoky]

Thanks for hanging in there Andy!

After the new dual-transistors were installed, there is still ~60 mV DC of noise at the speaker terminals.

I do get FM Stereo from both channels but at a very low volume (~.01 Watts).

I'm told there may be a problem in the "muting" section.

I first wanted to address the higher than normal voltages at the dual-transistor's collectors. I want to take this step by step.

[Smoky]

You Guys are not going to believe this!

I changed out the 1/2 Watt 100 Ohm R410 resistor for a 330 Ohm type (319 Ohms actual).

Can you see what I see?

It dropped the voltage rail to absolute schematic specification!!! -26.6 VDC

I also mounted it on the foil side of the board to be away from the heat of the big 5 Watt cement resistor it was next to.

[andy3055]

Great job! The numbers might change when you find the issue prior to the power amp section.

[Smoky]

Andy, would you mind taking a look at these IC's? They are supposedly matched double PNP transistors in a 6-TSOP package. I'm not sure the gain is adequate though compared to the 2SA798 dual-transistors. Anyway, I see little PC boards are available for them that can be adapted to the Pioneer receiver by using jumpers:

https://www.digikey.com/product-detail/en/nexperia-usa-inc/BCM856DS,115/1727-5391-1-ND/2676874


I'm not too sure these paired-up KSA992FBU's are the best option with 60 mV of noise at the speaker terminals.

You know me by now, I'm going to chip away at this until I get silence

[andy3055]

The specs are quite close but check this out! You may be in luck here: https://www.ebay.com/itm/2SA798-Original-New-Mitsubishi-Transistor-A798/121668651619

The seller is Dalbani corp. a well known supplier and they claim these are original Mitsubishi products which is great. Years back I have bought parts from them and they are reliable as far as I know. Here is the we site: http://www.ebaystores.com/dalbanicorporation/sch/_i.html?_nkw=2SA798&

You might as well get the originals and avoid the glue/thermal binding process. Also they will be matched. At $8.95 each and free shipping, you can't beat that. They dare not claim original if they are not as eBay will be behind them for sure.

Good luck.

Edit: They even have the STK 0029 for $15 and free shipping!

http://www.ebaystores.com/dalbanicorporation/sch/_i.html?_nkw=STK+0029&

[Martin Hodge]

I made a little smt adapter board to replace the 2SA798's with 2SA1873's in my SX-750. They have worked great ever since! And I have some extras...

[Smoky]

...and after two long years of electronics, I finally put together my first 8 Ohm dummy load!

Tomorrow I'm going fasten Hard Maple cleats to the 3-1/2 pounds of aluminum to suspend it off of the table

[Rollin Hand]

I need 4 pc boards ship to 33183.Can you tell me costs?Thanks

[Smoky]

Rollin Hand, I changed my mind and decided to buy two 2SA798's so I didn't buy Mr. Hodge's boards.

I didn't want to add any more "custom" parts at the moment.

If the new 2SA798's are still noisy, I may try the 2SA1873 option, but then again, the noise may indicate just another issue somewhere else in this receiver.

*Update*

I swung over to Audiokarma for a little while regarding the SX-580. The SX-580 is back up and running with the help of Mr. Merlynski 

https://audiokarma.org/forums/index.php?threads/a-pioneer-sx-580-with-low-volume.912613/

The problem was with diode D403. It was intermittent. It controlled the voltage of the collector of Q502. I replaced several other 1S2437's too with 1N4148's and the voltages were put back on track! All I can say is, check all of the diodes within these old receivers, especially in the power supply area. But, really, you should replace every glass diode that you see.

The final waveform of the receiver's output, just under the point of distortion, read 13VDC, ~23 Watts per channel.

I may fiddle with other 2SA798's to get the DC Offset down in the low 20's or single digits.

The 8 Ohm dummy load is done too!

Thanks again for all of the help everyone