Pioneer SX-3900 quit mid album

My SX-3900 died the other day mid play and I'm hoping that someone out there can help shed a little light on whats going on. All lights on face and selectors still work but I have no output from either channel or phones. Power output meters show no action either. Figuring it might be the normal cold solder joints that these are none for I started tearing it apart looking for the obvious; nothing of that nature. I did find the small transformer on the power supply board had leaked a bunch of it's wax down into the cabinet. I would imagine that is one of it's problems. Could anyone clue me in as to how to test this transformer, and what values it should have. I'm kind of new at trying this depth of repair on this vintage stuff. I have schematics, service manuals (thanks to the AK Pioneer Archive), a Fluke meter, and no fear, at this point anyway. Maybe I can get lucky. Thanks in advance, Glenn.

Do you hear a click from INSIDE several seconds after powerup?

Also does the front panel react when tuning on FM? (tuning, signal strength, stereo indicator, quartz lock indicator).

You can try to exonerate the tuner and pre-amp sections by removing the pre-out / amp-in jumpers and connecting the pre-outs to another amplifier.

You have a Fluke, so you are in good shape there to check the power supply and amplifier output voltages.

You need to locate the amplifier outputs before they are disconnected by the protection relay on the amplifier assembly, and read the DC voltages present on the outputs to determine the state of health of the individual amplifier channels. I may have to post a picture and point...
edit: let's give this a verbal shot, The amplifier board is the big board mounted vertically next to the box containing the heatsinks (the finned things), In the center of the amplifier board is the protection relay, and on each side of the relay is a coil of wire. the connection you need to measure is the coils lead that is closer to the relay on each side. It could be insulated, preventing the reading, there are alternatives, but to get more detail, I do need a picture...

According to the schematic, you are looking for the power supply voltages of: + / - 54 volts (22;S/20;U), + / - 32 volts(23;B/24;A), +12.8 volts (19?;E) and +5.5 volts (11;D) at these (terminals) on the power supply board AWR-210. The numbers are fuzzy so don't worry if they don't make sense at first. There are a few other odd voltages, if necessary we'll chase them later. One set may go to the phono amp.

Download the sx-3800's manual as well, there are enough similarities between the two (and the sx-3800's scans are better) to make it worthwhile if we are component troubleshooting.

Don't interpret any long gaps in my responses as loss of interest, anger or anything else, the fact IS that I am engaged in hand to pipe combat with a MAJOR bathroom repair.... I have an sx-3900 in, waiting for me to fix, and am getting a 3800 in soon that is ALSO inoperative, so we should be able to get you squared away... with a little patience...

Hi Mark,
Yes, all the FM functions seem to work except the Quartz Lock light, I suspect a burnt lamp. I do get the click from the relay on the AWR-214 rectifier board, but was also getting a click from the relay on the AWH-097 Amplifier board when unit worked, the protection circuit relay, since the unit quit the click coming from the amplifier board is no longer there. There is also some flickering of ALL the lights and displays on the front panel from time to time. That also happened right after the unit died.

I should also mention that this is the variable input voltage "SG" unit.

I just looked at the machine, and the schematic, all you've told me makes a little since now. The coils near the protection relay are not insulated so I will be able to check them.

There is 0v at the coils on the amplifier board, protection circuit relay open? The voltages on AWR-210 are as follows:
22 - 20 = 45.3v
23 - 24 = 63.4v (way off, checked 3 times)
19 = 12.83v
11 = 5.25v

Im still wondering if that small transformer on AWR-210 that leaked all over could somehow be mixed up in all this, what do you think?

Will try to hook up another amp to pre-out but am wondering if it will work with the protection circuit open.

Thanks for all your help. Looking forward to the next step. Glenn.

garagmahal@netz said:

Hi Mark,
Yes, all the FM functions seem to work except the Quartz Lock light, I suspect a burnt lamp. I do get the click from the relay on the AWR-214 rectifier board, but was also getting a click from the relay on the AWH-097 Amplifier board when unit worked, the protection circuit relay, since the unit quit the click coming from the amplifier board is no longer there. There is also some flickering of ALL the lights and displays on the front panel from time to time. That also happened right after the unit died.

I should also mention that this is the variable input voltage "SG" unit.

I just looked at the machine, and the schematic, all you've told me makes a little since now. The coils near the protection relay are not insulated so I will be able to check them.

There is 0v at the coils on the amplifier board, protection circuit relay open? The voltages on AWR-210 are as follows:
22 - 20 = 45.3v
23 - 24 = 63.4v (way off, checked 3 times)
19 = 12.83v
11 = 5.25v

Im still wondering if that small transformer on AWR-210 that leaked all over could somehow be mixed up in all this, what do you think?

Will try to hook up another amp to pre-out but am wondering if it will work with the protection circuit open.

Thanks for all your help. Looking forward to the next step. Glenn.

Click to expand...

The coils have varnish, the color is darker than bare copper. The foil pattern is inaccessible, right? There are resistor lead holes straight UP (about an inch?) above the coil's holes that are on the same pattern(s). You should get repeatable millivolt level measurements at least... (if the protection circuit failed while the amplifiers are ok, which CAN happen)

The small transformer could have overheated, or it could be glue to hold it on for production wavesoldering and vibration protection in the field. There usually ISN'T a lot of gunk in a transformer to melt out, so my estimate is glue. Pictures would tell the tale...

The pre- outs are NOT affected by protection, they WILL work.

The power supplies need attention (probably first). be back soon with info..

Ok, you are going to have to find D6 on the power supply and measure it to see if there is 14 volts across it. The negative 54 volt supply tracks the positive 54 volt supply, so if we fix one, the other should fall into line, also both the 32 volt supplies also track the positive supply as well.

Hi Mark,
You're right about the varnish on the coils; I was able to get down behind with my probe and got 52.3v on both coils. My board doesn't seem to have that matching hole pattern.

After looking at the transformer goo it does look, and feel like glue, transformer is soldered to board.

I hooked up my Denon PMA-750 to the pre-out; everything in the Pioneer tuner area seems to work properly, except that quartz lock light. When everything is working properly, does combining 2 machines like that add thier output power together? What is the point of seperating the power output section from the other functions?

I found D6 on the power supply board, it reads 13.99v across. I'm not sure what you mean by the + & - supplies tracking each other.

I hope all this is painting a picture of the problem for you, let me know what else you need. Getting this fixed would be a great victory in the stereo war. Thanks, Glenn.

The 52.3 is baaaad, that's why the protection circuit cut in. The fact that it is on BOTH is interesting... What is the number stamped on the amplifier board? It will be three letters dash three numbers and maybe a last letter.
BUT, until we get the power supply voltages correct, nothing else is valid.

The pre-out stuff is just to see WHAT is and isn't working....

D6 has good voltage across it. IT is the "master" reference for all the power supplies. Tracking means that the negative 54 volt supply "mirrors" the positive 54 volt supply, anything the positive does, the negative copies... BOTH of the 32 volt regulators are set to deliver a proportion of their input voltage, so at 54 volts they would be (+/-) 32 volts... sort of like dominos, or a house of cards, when one goes the rest follow.

Just checking here, you are making all these voltage readings (except d6 since I said differently) referenced to the chassis ground (black meter lead)?

On the awr-210 power supply board (IS it that number?), could you read the three voltages EACH on Q1 (SGD), Q3 (BCE, metal tab = C) and Q4 (ECB). Be careful, the heatsink for Q3 (2sd313) could be at 70 volts or so if it is not insulated from the transistor's metal mounting tab, which is also a collector connection.
Q1 = 2sk34 an n channel jfet (SGD layout) a little black plastic one
Q3 = 2sd313 an NPN power transistor (BCE layout) about 1/2 inch square with probably a heatsink, technically a TO-220 case
Q4 - 2sc1885 an NPN smaller signal transistor (ECB layout) a little black plastic one.

Hi Mark,

Power Supply Board is AWR-210B. Amp Board is AWH-097B. All values measured to chassis ground, except D6.

I've found and identified Q1, 3, & 4, now I just have to figure out how to get to them with the power on! Will get back to you in a little bit with the values. These are all checked against chassis ground right?

Had to take a break from all this this afternoon, went down to the local thrift store and found a beautiful pair of Akai SW-160 speakers. Grills and cabs are in great shape, inards look good too. Listening to them right now, they make my Infinitys sound tinny. Running them with a Sansui 9090DB. May have to declare them a "score" when I get enough time to clean them up and check them out thoroughly. Thanks, Glenn.

Hey Mark, I can't get into the area with the short probes I have. Will go get some long ones tomorrow. Will send values then. Thanks, Glenn.

Make sure you insulate all but the very tips of the probes...

There may be a way to raise up the entire heatsink/amplifier assembly to work on it.

I thought the power supply was horizontal, I now see it is vertical. There should be a way to raise it too.

My computer's giving me fits, I need that like another hole in my head.

By the way, at least one of those components has to be bad, and will need to be removed and replaced, so the effort to expose the power supply regulator board will not be wasted.

Hey Mark,
Sorry it took so long to get back, hope you didn't waste to much time checking for the results. Spent most of the day looking for extended probes. After multiple strikeouts, I finally just made one out of .040 Stainless TIG welding rod covered with heatshrink, and viola, I got the values;

Q1 = both print callout locations read 55.1v, the other is 79.8v
Q3 = print callouts locations read, left 54.9, right 54.3, other 80.2
Q4 = top 14.16v, middle 55.0v, bottom 14.16v

They all look pretty close to what the print says. Did get a pleasent surprise though, between Q3 and Q1 I had to reposition the unit, I shut off the power and when I turned it back on to take the reading the protection relay on the Amp board closed with an audible click. I checked the output meter on the front, it was functioning so I plugged in my phones and unit worked. Cold solder joint maybe? Must have moved something with my probe tip.

Tried to lift both the Amp board, with heatsink attached, and the Power supply board when I had it torn apart the first time. Due to the amount of stuff around each one they wouldn't move but a 1/2" or so, not really much good. I even had a bunch of the single wires detached, and all the ribbon cables too. It's a jungle in there! Let me know what you think. Thanks, Glenn.


Just tried it again, doesn't work!

It's GOT to be a cold solder joint, PROBABLY at Q3, the combination of heat and mechanical stress would be just right for that location physically, and electronically it couldn't be at a much worse point.

It will be a pain to find, try watching the joints with a bright light as you wiggle the transistor. You may only FEEL the movement...

Q3 is a power transistor, and since I can't see it, IS it Heatsinked? is the heatsink used to mount other transistors as well? is it insulated from Q3?

The voltage readings are pretty good on the transistors, with Q3 and Q4 being bipolar transistors, there will (on a GOOD, operating transistor) be two voltages about 0.6 volt difference, with the third voltage generally being greater than the other two. If you're not watching for that 0.6 volt difference, it can slip by you. But I DON'T mention it at first, as to NOT prejudice anyone on the first readings...

No, I didn't waste time looking for you, I wasted it with a computer crash.... I have been S#*&&%$# bricks since 10 pm last night...

Hey Mark,

Q3 is heatsinked with an aluminum channel shaped piece that appeared to be glued to the board at one time (mounting tab of the transistor is screwed to inside of channel). When I had it apart I noticed that the glue is totally dried out and flaking off. Q6 right next to Q3 is the same way, dried out but still adhered to the board.

Can you recommend the proper glue to reattach Q3 heatsink to board?

Will tear it apart to check solder connections and let you know how it turns out. Thanks again, Glenn.

garagmahal@netz said:

Hey Mark,

Q3 is heatsinked with an aluminum channel shaped piece that appeared to be glued to the board at one time (mounting tab of the transistor is screwed to inside of channel). When I had it apart I noticed that the glue is totally dried out and flaking off. Q6 right next to Q3 is the same way, dried out but still adhered to the board.

Can you recommend the proper glue to reattach Q3 heatsink to board?

Will tear it apart to check solder connections and let you know how it turns out. Thanks again, Glenn.

Click to expand...

Heatsinks like that usually are allowed to 'float", gluing was a production step to hold things together for automated wave soldering.

When you access the pads for soldering, this whole situation can be eliminated permanently by desoldering the pad, carefully bending the leads over about 90 degrees, to make them flat to the copper pattern, and aligned on the longest axis of the pad, then soldered back together with a good, but not TOO generously filled, solder bead..

Is the screw attachment to the channel insulated? I am concerned about an electrically energized heatsink. with power OFF, you can do an ohmmeter test to the center lead, or the metal tab, and to the heatsink channel and see if there is continuity.

Go carefully, take small reversible steps, take your time... be sure you can back out gracefully with whatever you do... that's how I do it...

I checked it withpower on, heatsink reads 83v. Apparently they are not insulated from each other. I checked Q6, the one thats right next to Q3 and it's the same way, 83v. Should these be insulated with a fiber washer or something along those lines?

garagmahal@netz said:

I checked it withpower on, heatsink reads 83v. Apparently they are not insulated from each other. I checked Q6, the one thats right next to Q3 and it's the same way, 83v. Should these be insulated with a fiber washer or something along those lines?

Click to expand...

No it just means that you should excercise extreme caution around them, retrofitting insulation can be done, but it isn't a slam dunk. Between the tab and the heatsink generally there would be a thin mica insulator, with a thin white goop called thermal transfer compound, and a nylon shoulder washer in the screw hole, it looks a bit like a tophat, insulates the sides of the hole and also the screw head surface where it would contact the tab.

Sometimes, because not a HUGE amount of force is needed, a plastic screw is used instead, but that may be harder to find than that shoulder washer.

The transistor type is technically called a TO-220, and you would need what is called a TO-220 mounting kit.

(BTW the Q6 case would have been negative 83 volts...)