Attempting SX3800 power amp repair

Herrenvolk

New Member
Hoping this is the right board for this.
This began about a month ago, I accidentally ordered a multimeter from Hong Kong and it took this long to arrive. I've no exact idea what may have caused my problems, so I'll post everything significant that's happened with this unit. When I bought it, the sx3800 had a problem where the left channel had more bass than the right channel. About a month ago I took it to a gathering of friends and it has two unfortunate things happen to it, a pizza box was left over it's vent for ten to fifteen minutes and it suffered a ~2ft tumble from a stationary car. After finding the pizza box on it, I turned it off and let it sit to cool off for twenty or so minutes, throughout the next few hours it got some light use and didn't exhibit any operational abnormalities.
The incident with the car happened on arrival home. I checked inside it immediately after it fell from the car, it hadn't and hasn't suffered any apparent damage from the shock, and it continued to function until early the next day.
I'd been listening to it for a short while when the right channel gives a pop and degrades in quality, and the unit strong burning smell(not the smell from over volted, burning electronics. I've smelt this before, from functional units under normal operating conditions, just not as strong as here. I don't know what components give it off, so I can't describe it further than that.). I quickly turned it off, and after leaving it for a minute, turn it on again. All the lights still come on, but the relay no longer clicks. I used some low wattage speakers to test the preamp and tuner, and both seem to function without problems. A few days ago I tried to take the voltages from the power amp to compare to some listings that I found on this forum, using the case as a ground. On testing pin 1, the voltage hovered around 23V before jumping into somewhere in the range of 40V, and the board emitted a burning electronics smell, which returns now whenever I power up the unit. Using the case as ground was guesswork, I hadn't seen a ground pin specified where the voltages were listed, but I wouldn't imagine that testing anywhere with a volt meter could cause a short, everywhere I've read about them states that a volt meter does not complete a circuit, which leads me to fear that my hands were shaky enough to cross pins 1 and 2.
The preamp still works,
Here is the manual I've been using, http://www.kallhovde.com/pioneer/sx-3800sm.pdf
I've seen people post about voltages, so if anyone has documents or knowledge to add to that, it would be appreciated, particularly voltage lists, part lists, and circuit diagrams.
Was I right in my method of measuring voltages, or terribly wrong, and if so, what is the correct way?
I still would like to check the power supply to be completely sure that the problem is the power amp/protection circuit.
I've misplaced the link from which I got my target voltages, so I can't post those.
 
The jarring could have displaced cracked solder joints, which are now "burning" with the heat of poor connections. The heat also could have aggravated that.

Look at the TO-220 power transistors on the power supply and amplifier boards, as well as on the amp board there are 4 smaller transistors with thicker center leads - they heat up and unsolder themselves.
 
I've unscrewed the heatsink from the case and unscrewed both boards from the heatsink. I don't see any broken solder joints on the TO-220's, or anywhere else on either board.
I was able to get some shots of the power supply board, but I can't get a good angle on the power amp board at the moment, so I'll post back later today with shots of that when I can get some help.
Should I try to discharge the capacitors on the power supply board, or are they safe? I've been careful around some of the connections on this, but if the capacitors are safe, I could try to remove the heatsink.
 

Attachments

  • DSC00269.jpg
    DSC00269.jpg
    84.4 KB · Views: 73
  • DSC00270.jpg
    DSC00270.jpg
    102.6 KB · Views: 52
  • DSC00271.jpg
    DSC00271.jpg
    87 KB · Views: 57
  • DSC00272.jpg
    DSC00272.jpg
    78.6 KB · Views: 64
IIRC, this model uses plastic connecting plugs to the large heat-sink-mounted output transistors, and over time, these connectors become very brittle and crack open, usually leading to shorted outputs. How do the connectors look on yours?
 
One of the plugs is cracked and not making a good connection. Would intermittent connections there cause damage elsewhere?
And here are those pictures of the power amp.
 

Attachments

  • DSC00279.jpg
    DSC00279.jpg
    72.1 KB · Views: 45
  • DSC00278.jpg
    DSC00278.jpg
    57.6 KB · Views: 38
  • DSC00277.jpg
    DSC00277.jpg
    71.7 KB · Views: 42
  • DSC00275.jpg
    DSC00275.jpg
    63.3 KB · Views: 50
  • DSC00274.jpg
    DSC00274.jpg
    87 KB · Views: 66
  • DSC00280.jpg
    DSC00280.jpg
    56.7 KB · Views: 41
Thanks, I'll order a few of those to keep on hand. Whenever I can get one in, I'll post back here whether or not that fixes it.
 
There is more, and I need to analyze the first post you made, and think on it...

The questions you asked: the answers can be dug out of the service manual.

Once I have the situation firmly in mind, i will come up with measurements to make or parts to replace. Usually the power supply board needs rebuilding.
Then there are some weaknesses in the amp board.
 
You NEED to describe WHAT is happening, is ANYTHING AT ALL lighting up. It is IMPORTANT!!

Now we need measurements. Start with the GWR-122 power supply board, connect the DMM black lead to a bare metal chassis ground, probe with red probe and INSULATE all but the tiniest portion of the probe tip.

pin 20 +47v dc regulated
pin 22 _32v dc regulated
pin 25 -??v dc regulated
pin 24 -??v dc regulated
pin 23 -32v dc regulated
pin 18 -47v dc regulated
pin 15 +12.8v dc regulated
pin 17 +5.5v dc regulated
pin 13 dc unregulated volts

then on the heat sink, the three voltages each on the pins of the four power transistors.
 
Everything lights up except the speaker A indicator, that was broken when I bought the unit. No lights have gone out since I bought it.
Voltages are:
20
+46.2V

22
+31.6V

25
-14.3V

24
-20.6V

23
-32.0V

18
-46.9V

15
+12.9V

17
+5.1V

13
+29.6V

The transistors, beginning at the front and going back, the noticeably broken socket had been on the front most transistor.

green
+7.1V
yellow
+7.0V
red
+46.0V

blue
+6.8V
brown
+7.1V
orange
Blue and brown would climb by 0.1v/4s, when left on it reached 8.4v and was still rising. When turned off briefly and turned on again, the increased voltage remains, but if the unit is left of for several minutes the voltage begins at ~7V again. The readings seem to decrease over time, rather than resetting immediately at a certain time.

blue
0.0V
brown
-0.7V
orange
-46.7V

green
+.06V
yellow
+0.7V
red
+45.0V
 
Ok, power supply is behaving, and we have all appropriate lights.

First two transistor readings are high, second two readings are reasonable.

On the side (it's "split" right down the middle) of the amp card where the two transistors are high, try to adjust the DC balance/offset to 0.000v, using the manual's procedure. It might not adjust far enough, so record the highest and lowest it can go using the DC balance/offset adjustment using that side's vr1 OR vr2.

The fly in the ointment is that pins 22 or 23 as referenced in the procedure will NOT be live to measure since there is no sound - the relay on the board ("protection") is OPEN, disconnecting the speakers (and 22, 23) from the amp.

SO, the easiest point to describe where to find it is the center wire of the dual resistors R165, R167, NEAR those pins 22 and 23 , the resistors are white rectangles about an inch wide, 3/4 in tall and 1/4 inch thick with three wires.

The diagram on page 22 of the downloadable manual shows the trace from the "amplifier", as opposed to "speaker" side of the relay contacts and the few components it goes through (L1, L2, R173, R174) on it's way to the relay contacts.

on the "good" side, that resistor has blue 0.0V on one side and green +.06V on the other side. The center would probably read: 0.030 or 30mV.

from your posting (these are the reasonable voltage readings) :
blue
0.0V
brown

-0.7V
orange
-46.7V

green
+.06V

yellow
+0.7V
red
+45.0V

while the first two are too high, and we are going to try to adjust them down, IF it can be done:
green
+7.1V
yellow
+7.0V
red
+46.0V

blue
+6.8V
brown
+7.1V
orange
 
Last edited:
Measuring from the center pin of the resistor on the bad side of the amp(vr1), I got 7.4V with the constantly rising behavior I described on the transistor pins, the voltage did not change with vr1 turned fully either way.
 
I haven't been posting as much, other things interfered.

This channel of the sx-3800 has a problem.

This amp will "balance" if both power transistors are disconnected (easy with sockets) as long as there is NO load, NOT EVEN headphones.

So disconnect the power transistors and repeat the procedure of the 8/1 post.
 
The reading started at 7.4V, no change from turning VR1 either way. It's smelling strongly of electrical burning while it's on, no load's on it and the volume's all the way down.
 
OK, we have to start testing transistors, diodes and resistors to trace the damage through the circuit on that channel. It pops up in surprising places.

Look for burned looking resistors first.

BTW this is NOT "busy work" - I do it TOO. And I have done a LOT of them. Successfully.
 
There aren't visibly burned components on the board, I'll check through with a multimeter tomorrow.
No need to worry about me, I understand how repairs like this can be tedious.
Also, if Google serves me, transistors can be tested with the diode test function on a basic multimeter, right?
 
There aren't visibly burned components on the board, I'll check through with a multimeter tomorrow.
No need to worry about me, I understand how repairs like this can be tedious.
Also, if Google serves me, transistors can be tested with the diode test function on a basic multimeter, right?

Bipolar Junction Transistor Testing Basics

Post 11 has a specific procedure and typical good npn/pnp results.

Resistors and diodes need one leg lifted for accurate readings, transistors generally need to be out of circuit.

AND do NOT power up after repairs until we discuss how to power it up without killing it, and it's NOT just on the DBT, but also with the wrong initial pot settings sending it up in smoke. Not that powering it up will be soon, just getting the warning in early enough to do some good.
 
Back
Top Bottom