SX3800 Filter Cap Smoke

[aldena007]

I have been working on two SX3800's that needed the usual going over to bring them back to life. I followed the excellent prior post to bring these two back to life. The procedure I followed was to pull both the power supply board [GWR-122] and the Power Amp Ass'y [GWH-139] out for testing and rebuild. This included testing each resistor, replacing all diodes, transistors, and caps. The main filter caps were not replaced in each unit. The first SX3800 powered up and I was able to set the DC Balance and Idle current with no problem. This unit is now in production and sounding fine.
The second unit which would not power up at all was given the same treatment as the first. I replaced the power supply board in the heat sink and decided to test the voltages before I put in the Power Amp Ass'y. I powered up the unit on Dim Bulb and the unit passed that step. I then powered up the unit with line voltage and was going to let it warm up and test voltages. The front panel lights came up and the flouro meters came on. Also the FM tuner was lit up and the stereo indicator was illuminated. Good sign. Suddenly I heard a "pop" and the top of one of the filter caps came off and clouds of white smoke vented out of both capacitors. The main power fuse blew. I have examined the board and there are no external signs any other components were damaged.
I plan to pull the board tomorrow and check for other damage, but now I think powering up the unit without the power amplifier installed may have been a mistake. Where should I start on this.

 

[markthefixer]

Ok, what did you replace the main rectifiers with?

D209 - D212 SR3AM- 4

With BOTH 12,000uf 56v caps C225 & C226 venting, I strongly suspect that the main rectifier diodes were replaced backwards.

It wasn't because of not having the power amp installed.

 

[aldena007]

Those Diodes were replaced with four Radio Shack Rectifier Diodes (Cat# 276-1143) that I bought about 12 years ago. Package says they are 3 amp/ 200piv. According to the information on the package they are supposed to be a replacement for IN5402. I checked the orientation of the diodes according to the silkscreen on the PCB and that looks good. Could the diode body be incorrectly marked for anode/cathode? I have new 1N5404 diodes I bought from Digikey for the other SX3800.

 

[aldena007]

I am sending the before and after pix to show the orientation of the new and old. You can see the vented Caps to the right.

 

[aldena007]

I pulled the Power Supply Board this morning and checked the diodes D209 - 212. Here is the results of diode testing:
D209 1N5406 0.455V Black Body
D210 FR306 0.388V Blue Body
D211 1N5406 0.455V Black Body
D212 FR306 0.376V Blue Body
All four diodes had the same catalog number but the diodes do not seem to be the same. I am now of the opinion the rectifier diodes should be identical or near so. I don't know if this was a factor in my main power supply caps venting. The event was fairly tame with a faint pop and clouds of what seemed to be steam or white vapor. No odor however. I have attached some more pictures for reference.

 

[markthefixer]

Check them against the 1n5404's. I would just have used the 1n5404's. You don't explicitly mention OL reading on reverse of diode test.

The orientations in the pictures show no obvious faults, thus the diodes are the suspects.

Unless....
The ORIGINAL caps smoked? Those are very hard to install reverse polarity - not so much so with replacement caps.

What would I do? Four new 1n5404's 2 new caps and when they worked, crush the RS crap in a vise.

 

[aldena007]

The reverse diode test was "OL" for each RS diode. The vented caps that were in the unit were the originals. My plan will be to do as you suggest and use four 1n5404 new diodes and replace the ruined main caps. I have two SX-D7000 units that need overhaul and was "warming up" on the SX-3800's. I will repost my outcome on this unit in a few days. Thanks for the information.

 

[aldena007]

Well good news and bad news. Replaced the D209 - 212 diode with new 1N5404 diodes. Replaced the vented caps with a couple of spare original caps from a donor SX3800. Prior to reinstalling the GWR122 Power supply board I was reviewing my notes and pictures and found a BIG discrepancy between my notes and what I did when wiring up the pcb the other day. The power cables from the transformer were in reversed. See picture.



The wires connect into pin position 26, 27,28 on GWR-122. They are a pair of red wires carrying 104 VAC and the black is the center tap at 52 VAC all off the main transformer. Pin 26 has the black center tap connected to it but it should be connected to pin 28 near filter cap whose base is in the lower right in the picture and the red wire in position 26. When I desoldered these three wires they were mounted in a plastic header which broke apart. Thus I had three loose wires and just mixed them up when I reconnected them singly leading to the loss of two filter caps on power up. The next picture is the wiring done correctly.



Reinstalled the pcb and brought it up on the DBT which passed and then brought it up on full power. I then tested the voltages on all the pins 1 - 28. The PCB was tested without connecting the amplifier board.
The results follows:

Pin 1 102 VAC Pin16 12.63 VDC with tuner active
Pin 2 102 VAC Pin 17 5.12 VDC with tuner active
Pin 3 51.7 VAC Pin 18 0.00 VDC <=Big Problem Should be -47 VDC!
Pin 4 12.21 VAC Pin 19 - Ground
Pin 5 12.21 VAC Pin 20 46.6 VDC
Pin 6 2.1 VAC Pin 21 0.00 V
Pin 7 9.17 VAC Pin 22 +31.91 VDC
Pin 8 1.43 VAC Pin 23 -31.73 VDC
Pin 9 1.43 VAC Pin 24 -21.15 VDC
Pin 10 0.00 V Pin 25 -15.14 VDC
Pin 11 -13.27 Pin 26 69.5 VAC
Pin 12 -13.28 Pin 27 69.5 VAC
Pin 13 -4.90 VDC Pin 28 34.43 VAC
Pin 14 23.22 VDC
Pin 15 12.63 VDC

On pin 7 there is a small problem. There is a 10 watt dropping resistor that should be 4.7 ohm but is actually applying 5.6 ohm to the circuit. I have a new one on order.

On pin 13 I think it should be 21 VDC but I don't know where I got that number in my notes.

The obvious problem is pin 18 which should have a -47 VDC. I will pull the board tomorrow and need to get an opinion on the best approach here. I am looking at Q205 but I am also considering a full component check and shotgunning everything. Any ideas or guidance is welcome.

 

[goat67]

Fixed the previous post formating

Reinstalled the pcb and brought it up on the DBT which passed and then brought it up on full power. I then tested the voltages on all the pins 1 - 28. The PCB was tested without connecting the amplifier board.
The results follows:

Pin 1 102 VAC
Pin 2 102 VAC
Pin 3 51.7 VAC
Pin 4 12.21 VAC
Pin 5 12.21 VAC
Pin 6 2.1 VAC
Pin 7 9.17 VAC
Pin 8 1.43 VAC
Pin 9 1.43 VAC
Pin 10 0.00 V
Pin 11 -13.27
Pin 12 -13.28
Pin 13 -4.90 VDC
Pin 14 23.22 VDC
Pin 15 12.63 VDC
Pin 16 12.63 VDC with tuner active
Pin 17 5.12 VDC with tuner active
Pin 18 0.00 VDC <=Big Problem Should be -47 VDC!
Pin 19 - Ground
Pin 20 46.6 VDC
Pin 21 0.00 V
Pin 22 +31.91 VDC
Pin 23 -31.73 VDC
Pin 24 -21.15 VDC
Pin 25 -15.14 VDC
Pin 26 69.5 VAC
Pin 27 69.5 VAC
Pin 28 34.43 VAC

On pin 7 there is a small problem. There is a 10 watt dropping resistor that should be 4.7 ohm but is actually applying 5.6 ohm to the circuit. I have a new one on order.

On pin 13 I think it should be 21 VDC but I don't know where I got that number in my notes.

The obvious problem is pin 18 which should have a -47 VDC. I will pull the board tomorrow and need to get an opinion on the best approach here. I am looking at Q205 but I am also considering a full component check and shotgunning everything. Any ideas or guidance is welcome.

 

[zebulon1]

If your getting voltage on pins 23, 24, and 25 you should be getting voltage on pin 18. Those voltages are derived from the -47v supply. You may of missed a solder (trace) connection.

I also wanted to mention that removing the boards to work by un-soldering the harness wires or unwrapping the wire wraps is strictly forbidden in this forum. It causes more problems than it's worth. Takes time away from more needed issues. I just finished a complete restore of a 3800 and only de-soldered one pin terminal.

The work is well done so far but in the end you can achieve the same results with alot less effort.
Nice thread.

 

[aldena007]

Thanks for the information. I will double check the the work on the pcb's. I definitely don't unwrap the wires but some broke or were broken at the pins when I first started working on this one. They are very brittle and any flexing seems to cause wires to break off at the pin. I will try to handle with more care. Thanks for the heads up.

 

[aldena007]

Good news! She is up and running. Zebulon1 was correct about the voltage on Pin18. It was there all along, I guess I just missed the pin or did not get good ground when I checked voltages. The next issue I had was in setting the idle. While letting the unit warm up the heat sink got really hot in just a few minutes. I shut it down and waited for it to cool and then set the DC Balance and then shut it down to cool. I did the same with the Idle current for each channel. I then powered up with no load and the main heat sink began to warm up at a leisurely rate and stabilized around 112 degrees measured at the top of the heatsink. I plugged in headphones and the sound was clear and strong. I think I let this one intimidate me a little bit. Attached are a couple of pictures:


Tuned to a local FM station [No antenna attached] Got stereo and quartz lock [the bulb is a little dim].



My main project for the last three weeks. The power supply and power amp boards. This unit would not power up when I started [Bad Q211 on GWR-122 pcb.



Laser thermometer reading on the top of the main heatsink. There are several little things I need to do to get this one back into production mainly the indicator lamps for the different operation modes and speaker indicator lamps. I will be wrapping this thread up for now. My next project is the two SX-D7000's. Both have power supply issues causing them to suddenly quit or not come on at all. Thanks for the help on these two.

 

[markthefixer]

SX-D7000 power supply has restricted ventilation around it due to cabling. Hot transistor joints break causing your symptoms.

After a FULL rebuild (as long as the JFETS are working and retained) and judicious bundling of cables to allow for airflow, it should be more reliable.

 

[zebulon1]

Mark, could you comment on the idle adjustment procedure and your observations on the NSA outputs. Mostly concerning the heat generated by them?
I doubled checked the idle settings several times and wondering why it runs so hot. I mean hot, like to the touch.
The OP "aldena007" 3800 is running much cooler than mine.
This is the first foray into these NSA receiver's for me. Not sure what to expect?

Thread hijack? Maybe a little bit.

 

[aldena007]

The temps I was getting on the unit discussed above I was measuring the top of the main heat exchanger without the bottom and bonnet was on. I think with the unit buttoned up and running the temps will be much hotter. I was mainly concerned with the main heat exchanger heating up rapidly (too hot to touch) before I made the idle adjustments. The other SX3800 I worked on and repaired runs about 135 degrees F at the top of the unit. I will be reporting the final temps after the unit is fully encased with bottom and bonnet.

 

[aldena007]

After thinking about zebulon1's heat question I decided to try and establish some sort of baseline for evaluating temps from these Pioneer units. I have been zapping individual components with my laser thermometer and found components that were really hot and some areas that were not. I decided to measure the heat being passed off the unit by measuring the air temp coming through the top vent on the case with the bottom and top cover in place. So I set up this experiment as the following pictures demonstrate.





I needed some sort of target for my laser thermometer and selected some cpu heatsinks above the power supply board [to the left], main power transistor heatsink

, and power amp pcb [to the right]. The top of the picture represents the rear bulkhead area and the foreground toward the front. I powered up the unit and let it run at moderate volume into some Sansui SPX7000 speakers for an hour. Ambient air temp was 85.6 degrees F. The following data was captured:
Power Supply Side Back to Front
115.5 degrees - 85.6 = 29.9 net change
129.5 degrees - 85.6 = 43.9 <- This is above Q201 and Q205 which measured individually run between 145 ~ 152 on this unit [too hot to touch]
121.5 degrees - 85.6 = 35.4
Power Transistors Heatsink Back to Front
130.5 degrees - 85.6 = 44.9 net change
131.5 degrees - 85.6 = 45.9
131.6 degrees - 85.6 = 46.0
Power Amp PCB Back to Front
123.5 degrees - 85.6 = 37.9
124.0 degrees - 85.6 = 38.4
125.5 degrees - 85.6 = 39.9

What does all this mean? I'm not sure but I feel that the temps coming out of the unit are more significant than measuring individual components. I have checked two transistors on the power supply which are well over 140 degrees. Why the Pioneer units run so hot is just the design in my mind. I am used to the temps off of my Sansui 9090, 9090DB and QRX9001 way less than these Pioneers. These temps do demonstrate that unobstructed air flow is essential but other than that not much else can be done except maybe a fan.​

 

[zebulon1]

That seems normal. Well at least it's the same as mine running at 138 to 140F at normal listening levels.These NSA's run at crazy frequency's hence the heat.
I have read several of Marks posts on this subject but they are in bits a pieces due to the nature of problems the members were encountering with their NSA's.

 

[markthefixer]

Heat sink orientation when setting the idle current is CRITICAL - these skivied sinks are very sensitive to mis-orientation.

When troubleshooting on the bench, I have a rig of slow fans that force ventilate the sinks since they are perpendicular to their normal operating orientation.

 

[aldena007]

How much "warm up" time do these units need before the idle current is set? I have been working on two different SX-3800's and one is serial BCnnnnnnnn [March 81] and this most recent one is AAnnnnnnnn [January 1980]. The power supply board and the power amp board are numbered different. On the BC unit the power amp board is GWR-140 and it has "cut outs" where you can connect to jumper wires to set the idle with the unit setting level with the skivied heat sinks vertical. This one is easy to adjust but the earlier model with the GWR-139 power amp board is much more difficult to situate [I set the receiver on it's side] to make adjustments but this puts the heat sink horizontal and I let the heat sink get to about 120 degrees and quickly set the idle. Should I rig a fan and reset the idle?.

 

[aldena007]

Just to bring this thread to a close and inform those who helped get this SX3800 back into working condition I have provided a picture of the subject unit on the bench in full operation. All indicator and illumination lights function and it sounds great. See picture.