Article #4 | 10:00 AM Friday 16/10/2018
Restoration
As always before I start working on a vintage unit restoration, I organize all the parts that will be changed. Further more I am sorting transistors based on the hFE factor. This is a boring job but it pays for itself. Careful with Fairchild KSA992 transistors as their terminals are designated as ECB. The old 2SA493 or 2SA726 are BCE. Same for KSC1845. So the replacements must be mounted in reverse.
General Considerations
Working on this unit exposes you to electrical hazards. There are lethal voltages inside.
Severe accidents and possibly death by electrocution might occur. I am qualified and skilled at electronics and I have been doing audio gear repairs for over 20 years. If you lack experience, please take these articles as just a knowledge base. Do not attempt to repair something that you cannot handle as there is a high chance of doing further damage while also possibly suffering accidents.
Good tools are a must for a quality restoration. I use eutectic soldering alloy and a temperature-controlled soldering station equipped with various tip shapes. I have both standard and precision desoldering pumps and desoldering wicks of various widths. To clean the flux, I use isopropyl alcohol and high-purity acetone.
Empirically, I found that soldering at a temperature of exactly 300 °C is safe for these vintage printed circuit boards. I have never lifted any pads and I never wait more than a couple of seconds with the hot tip on any pad. While working on the chassis, I use between 360 and 440 °C. Flux fumes are extremely toxic and should be avoided at all costs.
Every replacement part is brand new, from a reputable manufacturer, ordered from the U.S.A., Japan, or Germany. In addition, I only use parts that are suitable for specific circuit sections, after inspecting and comprehending the original schematic diagrams. Last but not least, I have years of experience backing up my choices and actions.
Sometimes, I like to take unnecessary actions, such as replacing perfectly working components for futureproofing, or completely stripping down PCBs of parts for a better cleaning. While tedious, I have fun while doing so, and I also learn a thing or two in the process.
Chassis Restoration
On the chassis there are only the power transistors that I am aiming to clean and reapply thermal grease. Thankfully, there is not much to do here since the transistors are all original. However, they are all dirty, and the old mica is very brittle; it breaks easily when I touch it. I forgot to take pictures in a before state. But here is the power transistors array after the service.
Under each transistor and under the mica itself there is a drop of non-conductive computer-grade thermal paste. This helps with the heat transfer.
These two filtering capacitors and their bleeding resistors follow up next. Somebody replaced the original ELNA parts with these Kendeil ones. They might be of good quality but I really don't know anything about them. Or if they were new or used when they were installed, although the 0216 marking sounds like a very recent date.
The previous technician managed to chip some of the epoxy covering one of the bleeding resistors.
The Nichicon KG replacement capacitors are a bit bulkier in diameter. You will see what I mean.
When mounted on the chassis, the metal clips don't fit the pre-drilled holes. That must be one of those weird cases when new capacitors of the same rating, are bulkier than old ones. I also had to use longer screws to tighten the clips over the capacitors.
I installed new bleeding resistors and replaced the old negative power supply rail flat cable with a new one.
I don't like that the new capacitors don't fit correctly on the chassis. At the same time I am thinking that maybe the metal clips are the problem. Maybe I will investigate whether there are metal clips that fit. But that is a lower priority project on my list.
F-2013-1A Power Supply Board Restoration
The power supply printed circuit board is the most damaged board of them all. The former technician worked carelessly and lifted tracks. He also used inappropriate replacement parts. Here is the board before the restoration.
Mismatched Zenner diode.
Very messy on the tracks side.
The board has been stripped of parts. The brownish spot is due to regulator transistor overheating. Using a small radiator is not part of a good engineering design.
I cleaned the pads. Nice and shine.
The previous messy technician did a very ugly job here. Take a closer look at those lifted tracks.
For C009, C010, C012 I have used Nichicon KZ MUSE series rated 10 uF / 100 V. C014 was replaced with a Nichicon KZ MUSE modern counterpart rated the same as the original. The ceramic capacitors received glass beads as spacers from the printed circuit board.
Resistor C004 was formerly replaced with a mediocre quality resistor which I ditched and replaced with a Dale RN60 series resistor.
Transistor TR001, originally 2SB507 was replaced by the former technician with a BD244 part. I replaced it with KSA940. Transistors TR002 and TR003, originally 2SA678 were replaced by the former technician with 2N3904 parts. I have used KSA733 parts. Be sure to order the transistors with the C suffix. This means center collector.
Zener diode ZD was replaced with a 15 V / 0.5 W Zener by the former technician. I replaced it with a Nexperia 14.1 V / 1.3 W modern part. The original was a 14 V / 1.5 W part. Diodes D002 through D005 were formerly replaced with 1N4004 parts. I ditched them and replaced with 1N5392 parts. These appear to be more of an equivalent to the original 10D-1 parts. Diodes received glass tubes as separators. Why? Rectifier diodes produce heat. While it is true that in this case, the heat is not sufficient to damage the printed circuit board, good engineering practices say to use separators. Germanium diodes are susceptible to heat. A longer lead will act like a heatsink preventing damage to the junction when soldering. Then again I have never damaged a Germanium diode or transistor by soldering since my childhood when I built my first diode-detector radio. I guess this comes with practice.
Replanting the parts...
...while seating the small capacitors on glass beads.
More parts have been soldered in place.
Diodes spaced from the printed circuit board by glass tubes.
Finished PCB assembly.
Hopefully this new heatsink would help with heat radiation.
Solder side. This is the best I could do. As I previously said, this PCB was in a very bad shape as the previous technician damaged a lot of tracks.
F-2041 / F-1215-A Protector Board Restoration
F-2041 is shared with the AU-7500 model and is still a joy to restore. A few capacitors and the coupling relay are targeted here.
This is before.
The solder side looks very neglected.
C901 was a 47 uF / 10 V bipolar capacitor and it was replaced with a Nichicon ES bipolar series modern part rated 47 uF / 16 V. C902 acts as the relay turn-on delay and it was rated 220 uF / 6.3 V. I replaced it with a modern part rated 220 uF / 25 V. C903 is a supply rail filter capacitor for this circuit. I replaced it with the same 100 uF / 50 V rating capacitor from the Nichicon KZ MUSE series. For C906 rated 1 uF / 50 V I have used a Panasonic stacked film capacitor of the same rating. C907 is a 1 uF / 50 V bipolar capacitor that was replaced with a Nichicon ES series counterpart of the same rating. The two relay driver transistors were replaced with KSC1845 parts. Even though these are small signal audio semiconductors, they can handle the relay load as well.
Diodes D901 through D904 form a sensible rectifying bridge for the protector circuit. They were originally 1N34 Germanium diodes. The former technician replaced these with RF detection Schottky diodes of type 1N5711. While these would work OK in this circuit, I have decided to use NOS 1S188FM Germanium parts that I have around. I thought I had a stash of NOS and recovered 1N34 diodes but I cannot find them anymore.
Diodes D905 and D906 were originally 10D-1. I have changed these with 1N5392 parts.
All these diodes received glass tubes as separators from the printed circuit board.
The relay can be ordered by the OMRON MY4-02-DC24 part number. This fits perfectly in place of the old relay. In fact it is identical despite the 40+ years that have passed.
Everything went away in the first place. After cleaning, this board appears in almost new condition.
On this board the track pads cleaned very well.
And, thankfully, there are no lifted pads, or damaged traces.
Mismatched transistors and diodes.
Sub-miniature black glass tubes that will be used as diode spacers.
Replanting the parts on the board as follows.
Glass beads detail.
Finished.
Now, the solder side is in perfect shape.
F-1215-A is very easy to work on because there is exactly one capacitor to be replaced. Removal is easy as well. Only one screw to remove. Here are some pictures before the job.
Decades-old dirt, dust-contaminated flux, and a dodgy solder job.
I have replaced the old 220 uF / 35 V capacitor with a modern Nichicon KZ MUSE series counterpart rated 220 uF / 50 V. The old transistor with the corroded legs was replaced with a Fairchild KSC1845.
This kind of riveted construction reminds me of ancient PCBs from the '60s.
Five components in total have been soldered to the PCB.
The entire assembly looks so retro, as if it were a subcomponent of the Lunar landing module.
I also wanted to tin the etched text, but I thought there would've been a risk of damaging them.
Thermo tube on wire endpoints help not to damage the wire insulation while soldering.
While putting this board back, pay attention to the metal fixture. It has a tendency to run off and misalign with the screw hole. Nothing to worry about. With minimal care I have secured it back in place in no time.
F-2034-A Driver Board Restoration
The driver board is just near the power transistors array. Very easy to service if you turn the unit upside down to have access to the soldering side while holding parts with your left hand from beneath the chassis. Mine was very dirty so I decided to remove it completely from the chassis.
This is the board prior to restoration.
It's clear to me that the previous repairmen tried wasn't very attentive. While the job was functional, the appearance (e.g.: his business card) is lackluster.
The same old dirt is omnipresent.
On the other side, the previous technician did a very dirty job.
Uh... huh...
An entire track lifted. How did he manage to do this!?
Input coupling capacitors C801 and C802 were originally rated 1 uF / 50 V. I have used same rating Panasonic stacked film capacitor replacements. C805, C806, C811, and C812 were rated 47 uF / 50 V and were replaced with same rating Nichicon KZ MUSE series modern capacitors. The bipolar capacitors in positions C807 and C807 were rated 47 uF / 10 V. I have replaced them with Nichicon ES bipolar series capacitors rated 47 uF / 16 V. Power supply rail decoupling capacitors C815, C816, C821, and C822 were rated 100 uF / 50 V and were replaced with the same type modern parts from the Nichicon KZ MUSE series. C819 and C820 are used for emitter decoupling of transistors TR805 and TR806 for static functioning point stabilization. These were 220 uF / 6.3 V and were replaced with Nichicon KZ MUSE series parts rated 220 uF / 25 V.
I have also replaced all four 2SA493 transistors in positions TR801, TR802, TR803, and TR804. hFE matched Fairchild KSA992 transistors were used instead.
Transistors TR805 and TR806 were originally Sony 2SC1124 parts. The previous technician replaced them with BD241C. I think Kosaku Kikuchi would strongly disagree. But hopefully he'd forgive me for using KSC2690A in place of the original parts.
But this ugly situation is about to change as I slowly progress with the restoration.
I love the tan color of these old printed circuit boards. And the dark-green silkscreen complements the PCB color so well.
Lifted pads.
hFE matched KSA992 small signal transistors.
One channel is restored.
Variable resistors were installed later on.
Difference between the capacitors used in this restoration and those used in the previous restoration.
The other channel is worse.
I recovered and secured the pad for the upcoming repair.
Which looks as such. Can you spot it? I'll make it easier by drawing an ellipse over the repaired area.
This driver board is now restored. However a few parts are still pending placement.
But the PCB already looks promising.
Ugly and unfit BD241C bias transistor was replaced with KSC2690A.
Which when installed on the board appears as such.
Both channels received the same transistors -- hFE matched of course.
Compared to the right side of the PCB, on the left side, the KSC2690A transistor is a little closer to those electrolytic capacitors.
On the tracks side everything looks good now.
15 pF capacitors are soldered in place.
I should have probably used C0G MLCC parts in lieu of the ceramic capacitors. Well, maybe next time.
This driver board resembles the F-2034 board that can be found in Sansui AU-7500 units. However some parts are different and some are mounted in different other places.
F-2028-A Equalizer Board Restoration
The equalizer board is socketed and can be found in the rightmost side of this amplifier. It has a thick steel Faraday cage protecting the circuit of electromagnetic radiation. Removal of the shield is done by removing the two screws securing it in place. Gently pull the board out of its socket presents you with this. In my unit there was a F-2028 board, sans the "-A" designation. I have inspected the F-2028 board installed in the AU-7500 series amplifier and while they are identical, certain parts have slightly different values. Furthermore some parts are missing on the board installed in AU-6500 series.
Prior to restoration.
Ugly flux residues leftover from the factory, and tantalum capacitors.
Even though all electrolytic capacitors have already been replaced on this PCB, I will drop them them, and do the job right.
Have I mentioned dirt?
Input capacitors C601 and C602 were of tantalum type, rated 3.3 uF / 25 V. However, on my board, they were of 2.2 uF / 25 V type. I have replaced these with bipolar Nichicon ES series modern parts rated 3.3 uF / 50 V. C603 and C604 in the RIAA correction loop were polarized parts rated 10 uF / 10 V. They were replaced with Nichicon FG series parts rated 10 uF / 50 V. C609 and C610 are rated 47 uF / 16 V. I have replaced these with Nichicon KZ MUSE series capacitors rated 47 uF / 25 V. C611 and C612 were rated 33 uF / 10 V and were replaced with Nichicon KZ MUSE series modern parts rated 33 uF / 25 V. C623 and C624 were marked as 10 uF / 25 V and were replaced with Nichicon FG series rated 10 uF / 100 V. Output coupling is done through C625 and C626, rated 1 uF / 50 V. I replaced them with Panasonic stacked film capacitors rated 1 uF / 50 V. All fault-prone transistors were replaced with modern Fairchild parts, hFE matched. Transistor replacements, as follows.
- TR601, TR602, TR605, and TR606, originally 2SA726 (2SA493 on my board), with Fairchild KSA992.
- TR603 and TR604, originally 2SC1313 (2SC1000 on my board), with Fairchild KSC1845.
The dirt on this board pushed me to disassemble everything and wash all parts with isopropyl alcohol. Re-planting of all the parts on the clean printed circuit board should be rewarding in the end.
I have removed the components halfway through.
Then I have thoroughly cleaned the pads.
The difference is like night and day.
The other channel follows shortly.
Both channels are done.
By my standards, this is how this PCB assembly would be restored.
I say this looks like a good job.
Needless to say, there are no lifted pads.
Previous restoration work included tantalum capacitors in the signal path on the phono preamplifier. I am normally using these in digital circuits for supply rail decoupling purposes. Or in various other digital circuits. For audio I would pass.
F-2045 Tone Control Board Restoration
This board was the most complicated to work on of them all. This is because of wire forest that is soldered directly on the solder-side of the board. I had to tag all the wires first so that I will know where they connect. There is a handful of capacitors and a eight small signal transistors that are subject to replace. This printed circuit board was not that dirty. But since I took a no-reserve disassemble and clean-all operation, I have removed it as well. Naturally every electronic part had to go off the board, get cleaned up then re-planted.
Here is a picture of this board prior to servicing.
On the solder side, there's a thick layer of dirt and some careless soldering.
C703 and C704 were rated 47 uF / 16 V and were replaced with Nichicon KZ series parts rated 47 uF / 25 V. C707 and C708, rated 33 uF / 16 V were replaced with Nichicon KZ MUSE series capacitors rated the same as the originals. Coupling capacitors C709 and C710 were rated 4.7 uF / 50 V and were replaced with same rating Nichicon ES series bipolar parts. C717, C718, C725, and C726 were 1 uF / 50 V polarized capacitors. I have replaced these with Panasonic stacked film 1 uF / 50 V capacitors. C719, C720, C723, and C724 were 10 uF / 16 V and were replaced with 10 uF / 50 V Nichicon FG series parts. All transistors were replaced with modern Fairchild parts, hFE matched as follows.
- TR701, TR702, TR705, TR706, TR707, and TR708 originally 2SA493, with Fairchild KSA992.
- TR703 and TR704, originally 2SC1000, with Fairchild KSC1845.
As a side note, the former technician replaced 2SA493 transistors with 2SA970 and 2SC1000 transistors with 2SC2240 parts. Not a bad selection by any means but I preferred to stick to the trusted Fairchild parts. I have them sorted in pre-matched pairs anyway.
I have stripped all the parts from one channel then thoroughly cleaned the printed circuit board with isopropyl alcohol.
On the tracks side I cleaned everything.
Cleaning the dirt revealed some scratches, and some other minor defects.
But there is one surprise: one partially lifted pad and one partially lifted and missing pad. Nothing I cannot repair anyway.
Parts are mounted back on the PCB assembly.
Together with some of the components that will be part of the other channel restoration.
Glass beads.
Can you spot them?
I borrowed the idea from the old Kensonic / Accuphase gear. I liked the looks, but also the ease of soldering smaller components placed on glass spacers.
Solder side (upper section) looks good again.
Lifted pads are now repaired.
The other channel follows next. The soldermask paint is very brittle on this channel. I don't know why but it appears to be attacked by a strong solvent somewhere in the past. The other channel doesn't have this problem. Interestingly enough since Sansui normally used a heavy duty soldermask paint.
Tracks side has been cleaned as well.
This circuit section has less optical defects than its homologue.
All components are soldered in place.
The PCB assembly looks brand new.
On the solder side everything looks ready to go. Or is it ... to sing?
Detail.
F-2036 Accessory Board
Normally there is nothing to restore on this particular board. But given the fact this board has the bottom-most position of them all, it attracted a lot of dirt.
Here is a picture of this board prior to servicing.
Then everything went out for a clean.
Tracks side has been cleaned very well.
All parts are back on the printed circuit board.
Those styroflex capacitors are of very good quality.
And the solder side is good again. This is the board mounted on the chassis with all the wires soldered-in.
Aftermath
Old parts. Er, these capacitors are not so old, and could be reused for lesser projects. Thus, I donated them to a close friend.
Now these are old parts indeed. Thus, I binned all of them.
Electrical Settings
The restoration cannot be complete without performing the electrical settings as illustrated in the service manual. Thus, there are two main settings that I need to touch before putting the case back on.
- Output of Power Amplifier Section
- Current Alignment of Power Amplifier Section
The procedure for setting the output of the power amplifier DC voltage is very simple. First, commute the Speakers selector to SYSTEM-A position. Then connect a DC voltmeter to the SYSTEM-A left terminals. Set the scale of the DC voltmeter to 20 mV. Adjust VR801 up until the voltmeter reads 0 mV ±10 mV. Repeat this step for the right channel. This time adjust VR802.
Adjusting the idling current of the power amplifier is done by removing the speaker fuse cover on the back of the unit. The next step is to remove fuse F002. Turn Speakers selector to OFF position. Connect DC ammeter on the terminals of F002 fuse. Adjust VR803 for a reading of 30 mA ±2 mA. Repeat the procedure for the right channel. This time remove fuse F003 and adjust trimmer VR804.
Needless to say that you need to power the unit off each time you remove the fuses. It can be done with the unit powered but there could be a risk to short-circuit terminals and blow off the nice output power transistors.
