Vintage Audio
this section covers some restoration work that I did on my vintage HI-FI gear
Article #1 | 10:00 PM Sunday 03/01/2016

Sansui AU-9500: Introduction

This is our living room amplifier. A true ruling authority amongst our other amplifiers.

When I purchased this amplifier, it was in full working order. This is amazing for its age. But it had and still has an annoying hissing noise that amplifies with the volume. The sound is sublime, even with 40+ years old electrolytic capacitors and old failing transistors. But I can not afford frying scarce transistors and adjacent parts by means of an electrolytic capacitor imminent failure. So I have decided to restore this amplifier to its former glory.

This will be a long process but I will document it accordingly. Maybe you find my work inspiring or helpful. Or my descendants, during a distant future maintenance, should this amplifier survive the years to come...


The following articles are not to be treated as do-it-yourself tutorials on how to fix, restore, rebuild, or improve the unit in cause. This was not my initial intention. But you can consider this whole content as a general guideline, should you decide to launch into such an adventure.

The entire documentation is just a reflection of my work and I cannot be held responsible if you damage your unit, or even harm yourself in the process.

Article #2 | 11:00 PM Sunday 03/01/2016

Sansui AU-9500: Technical Data

This amplifier has the following technical characteristics.

Continuous RMS power (both channels)80 W + 80 W (8 Ω at 1,000 Hz)
Total Harmonic Distortion<0.1% (at rated output)
Intermodulation Distortion (70 Hz : 7,000 Hz = 4:1 SMPTE method)<0.1% (at rated output)
IHF power bandwidth (each channel driven at 8 Ω)15 Hz - 40,000 Hz
Frequency Response (at normal listening level)15 Hz - 40,000 Hz +0 dB / -1 dB
Load Impedance4 Ω - 16 Ω
Damping Factorapproximately 50 at 8 Ω load
SemiconductorsTransistors: a lot (!) / Diodes: sufficient (!)
Power Voltage100 V, 110 V, 117 V, 127 V, 220 V, 230 V, 240 V, 250 V at 50/60 Hz
Power Consumption550 W (max) / 205 W (rated)
Dimensions / Weight140 mm (H), 500 mm (W), 347 mm (D) / 23.3 kg

Article #3 | 08:10 PM Monday 04/01/2016

Sansui AU-9500: Parts List

I have assembled a parts list for this amplifier. My restoration touches some almost all of the original transistors as well, even though I consider them irreplaceable for an authentic Sansui sound. I don't believe that transistors actually contribute to any specific sound signature. But 2SA726 are prone to failure. So are 2SC1313. I am very interested in exchanging the electrolytic capacitors because this amplifier is experiencing electrolytic capacitor issues. So this is what you will mostly find listed below.

The schematic value corresponds to what normally can be found in the electrical schematics. The recommended value is what I used as replacement for the former part. The BP inscription signifies a bipolar capacitor. Where I found appropriate, I have chosen a film capacitor replacement instead of an electrolytic. My reasons are reliability in time. Less electrolytic capacitors, less time-ticking bombs.

Descriptions and Parts Listings

The Power Supply capacitors mounted on the Chassis are for decoupling purposes. They are located near the blue sockets beneath the steel chassis. These capacitors are directly soldered on various pins of the blue sockets.

Power Supply / Chassis
IdentifierSchematicRecommendedMouser Number
C0104.7 uF / 100 V10 uF / 100 V647-UKZ2A100MPM
C0114.7 uF / 100 V10 uF / 100 V647-UKZ2A100MPM
C0124.7 uF / 100 V10 uF / 100 V647-UKZ2A100MPM
C0134.7 uF / 100 V10 uF / 100 V647-UKZ2A100MPM

The Power Supply and Protector board is coded F-2030-2 and is located in the front-center of the amplifier, under the black metallic cage. You can quickly recognize it due to the big protector relay and stabilizer transistor cooled by a small U-shaped aluminum radiator.

Power Supply / Protector Block (F-2030-2)
IdentifierSchematicRecommendedMouser Number
C052100 uF / 75 V100 uF / 100 V647-UKZ2A101MHM
C054220 uF / 75 V220 uF / 100 V647-UKZ2A221MHM
C055100 uF / 75 V100 uF / 100 V647-UKZ2A101MHM
C057470 uF / 16 V470 uF / 25 V647-UKZ1E471MHM
C058220 uF / 75 V220 uF / 100 V647-UKZ2A221MHM
C060220 uF / 50 V647-UKZ1H221MHM
C062220 uF / 50 V647-UKZ1H221MHM
C90147 uF / 10 V BP47 uF / 16 V BP647-UES1C470MPM
C90247 uF / 10 V BP47 uF / 16 V BP647-UES1C470MPM
C9061000 uF / 6.3 V1000 uF / 16 V75-517D108M016CG6AE3
C907100 uF / 50 V647-UKZ1H101MHM
C9081 uF / 50 V BP1 uF / 50 V FILMN/A
RL901DC24V MY-2 RelayOMRON MY2-02-DC24653-MY2-02DC24

There are two audio Driver boards in this amplifier. Each one of them is coded F-2029. They can be found symmetrically mounted near the big black power transistor aluminum radiators. These two boards are shielded by their respective black metallic cages. Keep in mind that from the table below, you need to double order each part.

Driver Block (F-2029)
IdentifierSchematicRecommendedMouser Number
C8012.2 uF / 50 V2.2 uF / 50 V BP647-UES1H2R2MDM
C807220 uF / 6.3 V220 uF / 16 V BP647-UES1C221MPM
C81547 uF / 80 V47 uF / 100 V647-UKZ2A470MHM
C81747 uF / 80 V47 uF / 100 V647-UKZ2A470MHM
VR8015 kΩ652-3386H-1-502LF
VR803200 Ω652-3386H-1-201LF

The Equalizer Block is coded F-2006A and implements the Phono preamplifier and the RIAA curve corrector. This board is left positioned, under the black metallic shield that also covers the Power Supply and Protector board. You can not miss it once you remove that heavy Faraday steel shield.

Equalizer Block (F-2006A)
IdentifierSchematicRecommendedMouser Number
C6011 uF / 50 V1 uF / 50 V FILMN/A
C6021 uF / 50 V1 uF / 50 V FILMN/A
C60310 uF / 10 V10 uF / 16 V BP647-UES1C100MDM
C60410 uF / 10 V10 uF / 16 V BP647-UES1C100MDM
C60747 uF / 6.3 V47 uF / 16 V BP647-UES1C470MPM
C60847 uF / 6.3 V47 uF / 16 V BP647-UES1C470MPM
C61347 uF / 6.3 V47 uF / 16 V647-UFG1C470MEM
C61447 uF / 6.3 V47 uF / 16 V647-UFG1C470MEM
C6154.7 uF / 50 V4.7 uF / 50 V BP647-UES1H4R7MEM
C6164.7 uF / 50 V4.7 uF / 50 V BP647-UES1H4R7MEM
C6173.3 uF / 50 V3.3 uF / 50 V BP647-UES1H3R3MDM
C6183.3 uF / 50 V3.3 uF / 50 V BP647-UES1H3R3MDM

The Tone Control Block is coded F-2019B and is located beneath the steel chassis, under the Power Supply and Protector board. This board is the most complicated of them all as it contains a lot of electrolytic capacitors. I list them all below. Note that TC signifies a tantalum capacitor.

Tone Control Block (F-2019B)
IdentifierSchematicRecommendedMouser Number
C701470 nF / 25 V TC470 nF / 50 V FILM647-QYX1H474KTP
C702470 nF / 25 V TC470 nF / 50 V FILM647-QYX1H474KTP
C70347 uF / 6.3 V47 uF / 25 V647-UKZ1E470MPM
C70447 uF / 6.3 V47 uF / 25 V647-UKZ1E470MPM
C70547 uF / 6.3 V47 uF / 25 V647-UKZ1E470MPM
C70647 uF / 6.3 V47 uF / 25 V647-UKZ1E470MPM
C7094.7 uF / 50 V647-UFG1H4R7MDM
C7104.7 uF / 50 V647-UFG1H4R7MDM
C7172.2 uF / 50 V647-UFG1H2R2MDM
C7182.2 uF / 50 V647-UFG1H2R2MDM
C71947 uF / 6.3 V47 uF / 25 V647-UKZ1E470MPM
C72047 uF / 6.3 V47 uF / 25 V647-UKZ1E470MPM
C7234.7 uF / 50 V647-UFG1H4R7MDM
C7244.7 uF / 50 V647-UFG1H4R7MDM
C7314.7 uF / 50 V647-UFG1H4R7MDM
C7324.7 uF / 50 V647-UFG1H4R7MDM
C733220 uF / 35 V220 uF / 50 V647-UKZ1H221MHM
C734220 uF / 35 V220 uF / 50 V647-UKZ1H221MHM
C799220 uF / 25 V647-UKZ1E221MHM

As you can see, there are a lot of electrolytic capacitors to be replaced. While this amplifier desires a lot of respect, I cannot but advice you to take your time and observe the electrical schematics and my capacitor lists. You can stick to the original values and types or you can follow my list. It is up to you. But the results will be similar. For best results, don't forget about the small signal transistors designated as 2SA726 and 2SC1313. Your amplifier will shine again and will produce that great Sansui sound. After all, this is a top performer and will surely reward your ears.

Do not hurry, take your time and do the job once. And do it well.

Article #4 | 20:00 PM Tuesday 10/10/2017

Sansui AU-9500: Restoration

Normally when starting the restoration process I organize all the parts so that I can quickly replace the faulty ones in the shortest time possible. And this is what I did in this case.

Here is a batch of Panasonic stacked film non polarized capacitors. Very good for stage coupling. They are rated 1 uF / 50 V. As you can see, two of them are already missing because they were just mounted in the equalizer block.

And here are about a hundred KSA992 transistors. I am using them in hFE matched pairs that I discover with my transistor tester. I ordered a lot of them because they are very cheap and having more than necessary is perfect for sorting purposes. Careful with these transistors as their terminals are designated as ECB. The old 2SA726 are BCE. So the replacements must be mounted in reverse.

Next, I scored about a hundred 2SC1845 transistors. I am using them in sorted by hFE pairs as well. Also, care should be taken with these transistors as their terminals are designated as ECB. The old 2SC1313 are BCE. So the replacements must be mounted in reverse. A great but weird particularity of these transistors is that I have measured about 20 of them from the entire lot and all have the same hFE of 428. So no need to sort. I will however test them before soldering.

Sorting transistors, however, is a boring task that requires a lot of patience and time.

This thing is dwarfing my workplace. I had a hard time positioning the unit on my bench.

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 with 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 a standard and a precision desoldering pumps and desoldering wick in various widths. To clean the flux, I use isopropyl alcohol and high purity acetone.

Empirically, I found that working with 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 in 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.

Power Supply / Chassis Restoration

On the chassis there are four really large filtering capacitors. I have tested these against the specifications and they are doing fine. So no replacement for these yet. But I am already searching at my favorite electronic parts store for future-proof replacements. While ordering spare parts I need to be really careful so that they will be identical in diameter and hopefully better than the originals on ripple current filtering. I am still searching and once I will find what I am looking for, I will revisit this page.

However there are some capacitors that I have decided to change. I am talking about C010, C011, C012, C013, rated 4.7 uF / 100 V in the schematic. In my unit they are 3.3 uF / 100 V. I have used Nichicon MUSE KZ series replacements rated 10 uF / 100 V. These capacitors provide local filtering and decoupling of the supply rails for the two driver boards. So bigger is better in this case. I was thinking about bypassing these with 100nF / 100 V non-polarized polypropylene capacitors to cut up on high-frequency oscillations, should they appear. But I have had a hard time observing such oscillations with my 100 MHz oscilloscope. Thus, I quickly abandoned this idea.

Left chanel before.

And after.

Right chanel before.

And after.

It is fairly easy to work on the chassis. But I have used a high power 80 W soldering iron because of the steel chassis. It works like a massive heatsink and the smaller irons will be quickly drained of caloric energy. A large tip helps as well.

F-2030-2 Power Supply / Protector Board Restoration

This board is really easy to remove but you need to note the order of the wires that are attached to it. Besides four soldered wires, all the other wires are using connectors. I have used a 1 mm permanent marker to mark the order of the cables. Before mounting the board back, I will use isopropyl alcohol to wipe clean the cables so that my markings will be erased. The four heavy duty power wires which are soldered are very easy to desolder using a vacuum pump and a powerful iron. I managed to easily desolder them without damaging anything, in about two to three seconds of heat application.

Inspecting the board reveals that there is a number of twelve electrolytic capacitors. I have decided to let alone most of the transistors on this board. Beware that this board has very large ground planes so soldering and desoldering capacitors is difficult to impossible for a low power soldering tool. What I use is a Weller 80 W electronically controlled temperature soldering station with appropriate tips. You do not want to damage tracks of these unobtainium printed circuit boards. So good tools and a lot of precaution is necessary for a clean job.

The board prior to the restoration.

Detailed view on the relay and the coils. In my case the OMRON relay is in very good working order up until one point. And it appears it is the original part. The issue is that from times to times, the relay armature is not assuring a perfect electrical contact and the right channel drops. I have decided to replace the relay with the OMRON MY2-02-DC24 modern counterpart. It is expensive but it worths every cent.

I used Nichicon MUSE KZ series capacitor replacements for the old ELNA ones. These capacitors are fairly large in diameter and also in height. I dare to say that in most cases they are even the same size as the original parts. So I changed the two 220 uF / 75 V capacitors with 220 uF / 100 V capacitors. Then I changed the two 100 uF / 75 V capacitors with 100 uF / 100 V parts. C057 was rated at 470 uF / 16 V. The replacement is 470 uF / 25 V. Regarding the replacements for C060 and C062, I have used the same 220 uF / 50 V rating for both of them. C907 is used as filter for the protector circuit. It was originally rated at 100 uF / 50 V. I replaced it with the same rating modern counterpart. In the protection circuit, C901 and C902 were bipolar capacitors, originally rated at 47 uF / 10 V. I replaced both of them with Nichicon ES series 47 uF / 16 V bipolar parts. C908 is a 1 uF / 50 V bipolar capacitor. I have replaced it with the same type modern part. C906 is part of the relay delay circuit and was originally rated at 1000 uF / 6.3 V. My replacement is a Sprague part rated at 1000 uF / 16 V and 105 ℃.

The board after the job.

The new relay.

And the relay mounted on the printed circuit board.

Then I have proceeded to renewing of all solder joints on the back side of the board. I am using quality eutectic solder.

F-2029 Driver Board Restoration

These two power amplifier sections are the easiest printed circuit boards to work on of them all. There are only four capacitors and two transistors to change per board. There is also a thermal fuse glued on top of a small signal transistor. If you want you can take care of removing that one too. I did as there is no apparent need for a thermal fuse in this part of the circuit. Furthermore some of these amplifiers come from the factory without thermal fuses. I think my unit is part of a newer revision.

Here is the board prior to refurbishing.

The filtering capacitors are designated as C815 and C817. Originally they were rated at 47 uF / 80 V. I have used Nichicon MUSE KZ series rated at 47 uF / 100 V. They fit like a glove. C801 is used as an input capacitor and is rated at 2.2 uF / 50 V. I used a Nichicon FG series capacitor of the same characteristics. C807 is originally rated at 220 uF / 6.3 V. I have replaced it with a Nichicon MUSE KZ series capacitor rated at 220 uF / 25 V.

Transistors TR801 and TR803 are the input differential pair and are of type 2SA726. They are known to develop annoying hiss and crackling sound in time. As a matter of fact, this Sansui amplifier cracks after a few minutes of use. Also, it has a disturbing hiss when idling. I don't want to take any chances and I have decided to replace these transistors as well. I have used hFE matched KSA992 transistors.

As said above, I also took care of the fuse. The glue is very corrosive so I thoroughly removed it with isopropyl alcohol. The fuse is mounted in series with a resistor. I have removed the thermal fuse and soldered the resistor directly in its place, as factory-marked on the board.

And here is the board after the restoration.

Here you can see a detailed view of one of the 47 uF / 100 V replacement capacitors.

On the soldering side I carefully looked for faulty joints. There were none. But I am preparing for the long future so I renewed the solder joints using the eutectic soldering alloy.

The other channel is following as well. This is before restoration.

Detailed view on the thermal fuse. I experimented a bit after dismantling this fuse. So I triggered it with the soldering iron. Once popped there is no means to reset. Or so I think.

And after the job.

Detailed view of the hFE matched KSA992 transistors.

The other channel enjoys the same soldering renewal treatment.

F-2006A Equalizer Board Restoration

Removal of the board is extremely simple. The procedure is to first remove the two side screws holding the steel Faraday cage. The the cage itself. After that you gently pull the board upright out of its socket.

Working on this board is easy as well. There are a twelve capacitors and at least two transistors to change. Why at least? Because there are two 2SA726 transistors which are mandatory to change. But there are another two 2SC1313 transistors. While these are not yet causing issues, as a prevention, you can change these with 2SC1845. I decided to change all of them.

This is the board before restoration.

C601 and C602 are the input stage coupling capacitors. Originally they were 1 uF / 50 V, electrolytic. I have changed them with 1 uF / 50 V Panasonic stacked film capacitors. C603 and C604 which were originally 10 uF / 10 V were replaced with 10 uF / 16 V Nichicon ES series bipolar capacitors. C607 and C608, originally 47 uF / 6.3 V were replaced with 47 uF / 16 V Nichicon ES series bipolar counterparts. C613 and C614 are used to stabilize the static functioning point of two transistors and were originally rated at 47 uF / 6.3 V. I have used 47 uF / 16 V Nichicon FG series replacements. C615 and C616 are part of the RIAA correction network and originally were 4.7 uF / 50 V. I have replaced them with 4.7 uF / 50 V Nichicon ES series bipolar parts. The output coupling stage capacitors, originally rated at 3.3 uF / 50 V were replaced with 3.3 uF / 50 V Nichicon ES series bipolar capacitors.

And after.

On the back side I have renewed the solder joints.

F-2019B Tone Control Board Restoration

Removal of the board is extremely difficult. The procedure involves removing of the power supply board. Then the metallic shielding cage. After you have access to the solder side of the board which is bolted to the chassis with six screws. There is another fixing point in the left side, near the local filtering capacitors. This one requires a powerful soldering iron to detach. But this is not all. There are also a whole lot of wires attached to the solder side of this board. So I decided to let this board alone and adopt a different approach. I have positioned the amplifier so that I can easily reach the components from beneath, while having enough space to work on the solder side. It is an awkward job but the wires were untouched.

This is the board before restoration.

C701 and C702 are the input stage coupling capacitors. They are rated 470 nF / 25 V and are of tantalum type. I really cannot imagine why they have chosen this type of capacitor when it is clear that had the oversized mylar film green drops in production. Better dielectric for audio purposes and less problems caused should these capacitors fail. I have measured these old tantalum ELNA capacitors with my tester. Capacity is ranging from 460 nF (left) to 570 nF (right) while ESR is ranging from 29 R (left) to 120 R (right). Clearly out of specifications. Because I forgot to order replacements for these along with all the other parts, I have sorted through my capacitors box and found two Philips MKT orange drops rated 470 nF / 63 V. They fit like a glove on the PCB. On the tester, ESR is 0.9 R and leakage is 0 for both. Measured capacity is 467 nF. Even though they are not new, I believe they do qualify. Regarding their sound, it is fine. I like it. I have some experience with these capacitors in various tube amplifiers that I have built in the past. No disappointments yet.

The local emitter decoupling capacitors C703, C704, C705, C706, C719, C720 were originally rated 47 uF / 6.3 V. I have used Nichicon MUSE KZ series replacements rated 47 uF / 25 V. Stage coupling capacitors C709, C710, C723, C724, C731, C732 were rated 4.7 uF / 50 V. Nichicon FG series 4.7 uF / 50 V were used as replacement. I have used the same counterparts for C717, C718 2.2 uF / 50 V stage coupling capacitors. The local filtering capacitors C733 and C734 were 220 uF / 35 V. I have used Nichicon MUSE KZ series capacitors rated 220 uF / 50 V. The same for C799, originally 220 uF / 25 V, now replaced with new MUSE KZ series part of the same rating.

There are a also six 2SA726 and six 2SC1313 transistors to change on this board. I have used hFE matched pairs as before. In my unit the first preamplification stage was very noisy. It is constructed around two 2SK30 FET transistors. I have used NOS Toshiba 2SK30Y replacements. There are also two 2SC1364 transistors on this board. I have replaced these with KSC1845 which apparently work very well in this circuit.

Overall this board is very boring to work on. It took me at least four hours to exchange all the parts, due to the difficult clearance and maintenance access. Progress is very slow and barely visible. But I finished it after all. Here it is.

On the back side I have renewed the solder joints.

Beautiful track layout.

In respect to Kosaku Kikuchi and his team that designed this amplifier, I didn't want to have European parts in this amplifier. So I have sourced two Nichicon mylar film capacitors to replace the Philips MKT parts in the input coupling stage.

A little bit bulky but they do fit.

This board concludes the replacement of parts in this audio amplifier.


While having this amplifier on my bench I also redid the soldering of various other joints on all the other boards scattered around. I am not going to step into details as there is nothing interesting. But here is a picture.

Also, these packs of RCA connectors were suffering from transportation damage.

The repair was done with spare parts from disassembled units. I got the parts from eBay. An array of RCA jacks came from dismantling of a Sansui AU-101 and another array came from a Sansui 8080DB unit. When dismantled, they look as such.

Wire forest nightmare.

Fabricating the ground planes from pure copper multi-wire cable which I tinned with leaded solder.

Soldering the wires to the new RCA array.

More wires...

Done soldering the wires. The array looks like this.

In the end it looks like this.

RIP RCA array #1.

The other one comes up next. This is before.

From the other side of the array we can see that the board has developed some big transversal cracks. This array is compromised.

Upon dismantling the wires, the array board disintegrated when I wanted to get it out.

RIP RCA array #2.

Now let me show how I slowly rebuilt the RCA bracket. I have salvaged all the components from the old array and resoldered everything in place.

Oh No! More Wires.

All soldered back where they should be.

This is how the RCA array looks now.


Old parts.

This is everything that I did on this amplifier in the electrical field.

Electrical Settings

A final touch of the restoration is 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 and clearly detailed in the manual. 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 of the left channel up until the voltmeter reads 0 mV ±10 mV. Repeat this step for the right channel. This time adjust VR801 of the right channel. I have managed to set the DC offset to 0.3 mV. on both channels

Adjusting the idling current of the power amplifier is done by removing the speaker fuses on the back of the unit. First, unscrew the holder for fuse F002. Turn Speakers selector to OFF position. Connect DC ammeter on the terminals of F002 fuse. Adjust VR803 of the left channel for a reading between 30 mA and 40 mA. Repeat the procedure for the right channel. This time remove fuse F003 and adjust trimmer VR803 of the right channel. I have set the bias to 35.0 mA on both channels.

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.

Electrical Measurements

I decided to test this amplifier against the rated specifications using the signal generator, frequency counter, and oscilloscope. My setup is as follows, given that I conduct the measurements one channel at a time.

  1. Power on test equipment and let them warm up for half an hour or so.
  2. Power on audio amplifier and let it warm up for at least 10 minutes.
  3. Set signal generator output voltage to .775 V.
  4. Connect signal generator output to amplifier AUX input.
  5. Connect frequency counter to Speaker output.
  6. Connect oscilloscope to Speaker output.
  7. Start the measurements.

Measuring 12 Hz sinewave. In the service manual the rated bandwidth starts at 15 Hz.

Measuring 1 kHz sinewave.

Measuring 40 kHz sinewave. Amplifier output drops considerably but it still amplifies a clean sinewave.

As a bonus, measuring 1 kHz squarewave. Looks funny but then again, my signal generator is not very good with the square wave function.

I tried measuring sinewaves up until 80 kHz. Surprise, the amplifier still does its job. However oscillations start to appear and output drops a lot.

Cleaning the Front Panel

If you ask me, besides the sonic qualities, this is one of the most important things. I cannot stand dirty units. I believe keeping them dirty is equivalent to disrespecting them and the work that was put on them. So I proceeded to disassemble everything on the front panel and clean up all the parts.

Beautiful detail.

The black felt between the up-down switches has been cleaned with isopropyl alcohol and was installed back in its place.

The cleaning of the knobs comes up next. Unfortunately one of them is scratched and I don't have the means to repair it.

Cleaning of the up-down switch caps follows next.

Now it is time to put back the cover of this amplifier and off it goes for another 40 years, in my living room this time.

Article #5 | 10:00 AM Sunday 10/12/2017

Sansui AU-9500: Impressions

I have used this amplifier for at least two years prior to the restoration so I really know what it is up to in terms of sound. Low frequencies had a very nice ringing quality that I really appreciated. Like a very low frequency reverberation that made bass guitar sounds seem very close to you in terms of space. Mid frequencies were rendered very soft and pleasant. The high frequencies were not rendered that well. The sound was mushy in this part of the spectrum. Overall the sound was very pleasant to my ears. I hear very well in the high frequency spectrum so there was really no need for me to top up the high frequency knob. I was using it at about +2 dB.

After the restoration I had this urge to test the unit as soon as possible. This means that I quickly installed it on a cloth and wired the speakers and other cables. Powered the unit on and after the short relay delay, the familiar power hiss and some 50 Hz noise was heard from the speakers. I was blown away. So much work for ... worse results? Then I quickly realized that in my hurry to test the unit, I have forgot about the Faraday cage effect of the bottom steel plate. Thus, I secured it in place with its bolts and the hiss reduced a lot. On the other side, the noise disappeared completely. So I was happy to start playing some music on it.

I have played some known tracks ranging from Bach classical orchestra music, going through the rock register supported by Pink Floyd, King Crimson, Black Sabbath, Anathema, Metallica, and Burzum, ending up with the hip-hop register dominated by Eminem, Ice Cube, Dr. Dre. Then I shifted the register again to slower music such as Leonard Cohen, Cesaria Evora, Engelbert Humperdinck, and Frank Sinatra. The amplifier behave as expected, performing like a true artist on the stage. I have used mostly vinyl records through my very old direct-drive Micro-Seiki derived turntable.

Now should I try an explanation of the sonic characteristics? OK, let me try. The low frequencies are rendered in a more controlled manner. The bass has lost a little of the ringing quality. I can say that I liked it better before. But honestly it sounds more precise now. The medium frequencies are rendered the same as before. My ear cannot spot differences in this portion of the audio spectrum. Now the real surprise comes in the high-end of the audio spectrum. High frequencies are rendered with great emphasis. I had to lower the High knob to 0 dB to compensate. The sound is non-fatiguing as always. Overall, less sound coloring and more neutrality. It is also true that I don't have a pair of quality speakers but in the mean time I listen to what I have. In the meantime I have acquired a pair of JBL 4312A speakers: Leonard Cohen is now in my living room.

The amplifier started to sound even better after a few hours -- maybe twenty? -- of playing music. I cannot explain this phenomenon but I guess it is the capacitor break-in period. My engineering background says that capacitors do require formatting. And I also know several things that happen inside an electrolytic capacitor during the format procedure -- self healing would be one phenomenon. But in the end I just trust my ears and my knowledge.

To sum up all this work, there is a faint high frequency hiss present in the speakers or headphones. It doesn't bother me as the amplitude is very low and you have to stick your ears to the audio drivers to hear it. I believe this is acceptable for such an old design. After all this amplifier was engineered in 1974. That is 43 years at this very moment in time. Make it 44 by the time this thread received the first revision.

Audio friendship. These units sound good together, especially on extremely low audio levels -- < 0.5 W -- with the preamplifier COMP switch turned on. If increasing the volume, then COMP should be turned off. The audio preamplifier is feeding the MAIN IN input on the AU-9500. Alternatively I am using the AU-9500 as an integrated amplifier. This means that everything passes not through the external preamplifier but through its internal preamplifier. A joy to listen to the Sansui AU-9500 whatever combination I choose to use.

Overall I declare the restoration a success. Until the next maintenance, this topic is closed.

Article #6 | 11:32 PM Wednesday 01/01/2020

Sansui AU-9500: Further Audio Faults

A couple of years have passed since the restoration. I enjoyed many hours of audio program through this amplifier. But now the unit began producing an annoying FM-like sound only on start-up. This annoyance ceases after a couple of seconds. It is independent of the position of the volume potentiometer and sometimes is very loud. I suspect some transistors that I was very confident that wouldn't fail on each of the driver PCBs.

Somehow it is my own fault since I was overconfident that some transistors in some parts of the schematic wouldn't fail. But apparently I was awfully wrong. As much as I hate coming back on projects that I completed at some point in time, I really need to open this unit up and service it again. The good thing is that I don't have to remove it from its place and I don't have to disconnect any of the RCA signal cables nor the speaker cables. I just need to remove the top cover then the two steel shielding cages of the audio drivers and then I can quickly and safely remove the cards from their sockets.

Let's proceed then.

F-2029 Driver Board

TR805 and TR807 originally 2SC983 was replaced with KSC2383. TR809 originally 2SC735 was replaced with BC337 while TR811 originally 2SA562 was replaced with BC327. TR813 was 2SC1124 and replaced with KSC2690. TR815 and TR831 were originally 2SA706 parts and were replaced with KSA1220.

Trimmer resistors VR801 and VR803 were replaced with Bourns cermet type modern parts of the same ratings. Below you can see the difference between old and new trimmer resistors. The old ones exhibit a big amount of corrosion.

While here I decided to do a small improvement. I have replaced C801 with a Nichicon ES series bipolar capacitor of the same 2.2 uF / 50 V rating. On the other hand I have replaced the formerly KZ series C807 with a Nichicon ES bipolar capacitor rated at 220 uF / 16 V. I believe this update will bring slight total harmonic distortion (THD) improvements.

Solder side.

The other channel undergoes a similar approach.

View from a different angle.

Solder side.

F-2030-2 Power Supply / Protector Board

While here I decided to preventively replace some of the transistors and diodes on the power supply and protector board. I don't get it why I haven't done it in the first place anyway.

So I have replaced 2SC1124 with KSC2690 (TR051), 2SA706 with KSA1220 (TR052), 2SC734 with KSC945 (TR054), and 2SC1364 with KSC945 (TR901). Originally TR902 was also a 2SC1364 but I opted for a more robust KSC2383 as replacement. Diodes D051, D052, D053, and D054 were originally 10D-4 and were replaced with 1N5395. All other 10D-1 diodes (D908 and D909) were replaced with 1N5392.

Then I changed the 1 uF bipolar electrolytic capacitor with a Panasonic stacked film part of the same rating.

Solder side.


Old parts from the driver boards.

Old parts from the power supply board.

Old parts versus new parts.

Next, I assembled everything back in place and then performed the DC offset (0.1 - 0.3 mV) and idling current (35.0 mA) adjustments again. The amplification section is stable, as expected. Power up noise disappeared and the idling noise floor greatly improved.

In the end I have also updated the parts lists in Article #3 above to accommodate these replacements.

One interesting thing that happened after this service was that the relay armature clicking sound changed drastically. Now I can hear it firmly engaging at power on. That's strange. Maybe the relay driver transistor was weak after all. It's only a supposition.

I believe that's it for now.

Article #7 | 09:35 PM Monday 17/02/2020

Sansui AU-9500: Further Impressions

I had time to enjoy the amplifier for the last couple of weeks and all I can say is that I like it even better than before. It produces the same authoritative sound while maintaining a great stereo image separation. All this with a minimal noise floor compared to what it was with the old transistors.

The sensation of brute -- but refined -- power is present even at very low audio levels and the bass reserve is there. It is true that I only listened to audio cassettes through it. I haven't played a vinyl disc, yet. But since this is a living room amplifier, I have watched some documentaries on Youtube with the sound amplified through it. A nice experience overall. Before I was always hearing the annoying hiss while watching TV-like content with spoken soundtrack and no background music. At times it was unbearable. Imagine a documentary about the Planets and Stars of the Universe and long speaking breaks. The hiss... horror. Annoying as crap.

But getting back to the music, a '90s Eurodance mixtape has just ended moments ago. As much as I'd like to hear another one, I have some other stuff to do as well.

There is something that freaks me out though. The tape deck monitor level potentiometer on the back of the unit is squeaky and interrupts randomly if tackled. But I found a position where both tape channels are reproduced correctly in terms of audio level intensity. Although I did it in the past and I know I can, I don't really like to service potentiometers and I hope I don't have to do it anytime soon.

your help matters

Please note that all the work presented herein this site is non-commercial. This is my hobby and I am doing this in my spare time. Through this page I freely share my knowledge with you. But if you like my work, please consider helping me buy a transistor or a capacitor for my projects.

Thank you!

Copyright © 1998- Alexandru Groza