Vintage Audio
this section covers some restoration work that I did on my vintage HI-FI gear
Main Internet site | FAQ | Personal
Article #1 | 10:47 AM Thursday 08/02/2018

AKAI GXC-760D: Introduction

An AKAI cassette deck! From 1976 I believe. This is one of the good AKAI cassette tape decks. It is equipped with three GX heads, closed-loop dual capstan, a massive AC servo motor for capstan drive. In addition it has direct drive reels powered by DC motors.

I finally managed to raise money to buy one of these. It was barely working but in great cosmetic condition. Some of the incandescent bulbs were dead on arrival and the capstan belt was replaced but with mismatched size. The pinch rollers were defective as well. The left one has this weird wobble which also causes poor sound reproduction. At least one transistor has been replaced in the past with a genuine Japanese part. But the work looks like amateur. Thankfully nothing else was damaged. The PCB tracks are all OK. Cables as well. All seems original.

Even with these small inconveniences, the price was a premium. But it worths every cent. I like the sound. It resembles to that reproduced by their old reel to reel tape decks. The more recent GX cassette decks have a more analytical sound. But this one sounds vintage.

The construction of this unit is way overengineered. And the looks... Have you seen one of these in reality?

Article #2 | 11:00 AM Thursday 08/02/2018

AKAI GXC-760D: Technical Data

This cassette deck has the following technical characteristics.

Track System4 track 2 channel stereo system
Tape Speed4.76 cm/sec
Wow and Flutter<0.06% WRMS
<0.17% (DIN 45500)
Frequency Response30 Hz - 15,000 Hz (±3dB) - Fe2O3 tape
30 Hz - 16,000 Hz (±3dB) - CrO2 tape
30 Hz - 19,000 Hz (±3dB) - FeCr tape
Distortion<1% (1,000 Hz / 0 VU) - Fe2O3 tape
Signal to Noise RatioBetter than 51 dB (tape peak of +5 dB)
Erase RatioBetter than 70 dB
Bias Frequency100 kHz
Heads1 x GX record head
1 x GX playback head
1 x ferrite erase head
Motors1 x AC servo outer-rotor motor for capstan drive
2 x DC motor for reel drive
SemiconductorsTransistors: 83 / FET: 6 / Diodes: 129
Power Voltage220 V at 50 Hz
Dimensions / Weight142 mm (H), 440 mm (W), 306 mm (D) / 11.1 kg

Article #3 | 20:00 PM Thursday 08/02/2018

AKAI GXC-760D: Parts List

I have assembled a parts list for this cassette deck.

Descriptions and Parts Listings

The Power Supply and Oscillator P.C. Board is coded CB-5002 and is located on the lower right part of the unit as you look at it from above. This board implements the power supply -- obvious -- and the record pre-magnetization oscillator used to erase the tape when in record mode.

Power Supply and Oscillator P.C. Board (CB-5002)
IdentifierSchematicRecommendedMouser Number
C1100 uF / 25 V647-UBT1E101MPD8
C82200 uF / 50 V647-UPW1H222MHD
C9220 uF / 35 V647-UBT1V221MPD8
C10220 uF / 35 V647-UBT1V221MPD8
C111000 uF / 25 V1000 uF / 35 V647-UBT1V102MHD
C121000 uF / 25 V1000 uF / 35 V647-UBT1V102MHD
C13470 uF / 6.3 V470 uF / 16 V647-UBT1C471MPD1TD
C1410 uF / 25 V10 uF / 50 V647-UBT1H100MPD
C1533 uF / 25 V33 uF / 50 V647-UBT1H330MPD

The Switch P.C. Board is coded CB-2001 and is located in the lower middle position of the steel chassis as you look at it from above. It is mounted with the tracks side upwards.

Switch P.C. Board (CB-2001)
IdentifierSchematicRecommendedMouser Number
C147 uF / 25 V647-UPW1E470MDD1TD
C5 (?)47 uF / 16 V47 uF / 25 V647-UPW1E470MDD1TD
C10100 uF / 25 V100 uF / 35 V647-UPW1V101MPD
C15100 uF / 6.3 V100 uF / 16 V647-UPW1C101MED
C161 uF / 50 V647-UPV1H010MFD

The Noise Filter P.C. Board is coded CB-2027 and is located above the two flywheels. It is a small board by all means and it houses four filter coils and four capacitors. The reel motor wires are soldered to this board.

Noise Filter P.C. Board (CB-2027)
IdentifierSchematicRecommendedMouser Number
C10.47 uF / 50 V470 nF / 50 V MLCC80-C315C474K5R-TR
C20.47 uF / 50 V470 nF / 50 V MLCC80-C315C474K5R-TR
C30.47 uF / 50 V470 nF / 50 V MLCC80-C315C474K5R-TR
C40.47 uF / 50 V470 nF / 50 V MLCC80-C315C474K5R-TR

The Relay P.C. Board is coded CB-5001 and is located underneath the steel chassis. You can recognize it quickly due to the big relay and the single electrolytic capacitor.

Relay P.C. Board (CB-5001)
IdentifierSchematicRecommendedMouser Number
C147 uF / 25 V647-UPW1E470MDD1TD

The Servo P.C. Board is coded CB-2025 and is located in the upper left corner of the unit as you look at it from above. It is located near the big AC capstan motor.

Servo P.C. Board (CB-2025)
IdentifierSchematicRecommendedMouser Number
C10.47 uF / 50 V470 nF / 50 V MLCC80-C315C474K5R-TR
C31 uF / 50 V1 uF / 50 V FILMN/A
C410 uF / 16 V10 uF / 50 V647-UPW1H100MDD
C1347 uF / 10 V47 uF / 25 V647-UPW1E470MDD1TD
C141 uF / 100 V647-UPJ2A010MDD
C15220 uF / 25 V220 uF / 35 V647-UPW1V221MPD1TD

The Peak Meter P.C. Board is coded CB-5031 and is located behind the two VU-Meter indicators.

Parts need to be ordered twice for this board.

Peak Meter P.C. Board (CB-5031)
IdentifierSchematicRecommendedMouser Number
C12.2 uF / 50 V647-UPW1H2R2MDD
C20.47 uF / 50 V470 nF / 50 V MLCC80-C315C474K5R-TR
C310 uF / 16 V10 uF / 50 V647-UPW1H100MDD
C410 uF / 25 V10 uF / 50 V647-UPW1H100MDD
C510 uF / 25 V10 uF / 50 V647-UPW1H100MDD
C610 uF / 16 V10 uF / 50 V647-UPW1H100MDD
C70.47 uF / 25 V TC470 nF / 50 V MLCC80-C315C474K5R-TR
C80.1 uF / 25 V TC100 nF / 50 V MLCC80-C320C104K5R
C94.7 uF / 25 V4.7 uF / 50 V647-UPW1H4R7MDD

The SYSCON P.C. Board is coded CB-5004 and it occupies the upper middle part of the steel chassis. You can't miss it as it is the first big printed circuit board that you see when you take the wooden lid off.

SYSCON P.C. Board (CB-5004)
IdentifierSchematicRecommendedMouser Number
C133 uF / 25 V33 uF / 50 V647-UBT1H330MPD
C233 uF / 25 V33 uF / 50 V647-UBT1H330MPD
C34.7 uF / 6.3 V4.7 uF / 50 V647-UBT1H4R7MPD
C422 uF / 25 V22 uF / 50 V647-UBT1H220MPD
C6100 uF / 25 V647-UBT1E101MPD8
C7100 uF / 25 V647-UBT1E101MPD8
C81 uF / 50 V647-UBT1H010MPD
C933 uF / 25 V33 uF / 50 V647-UBT1H330MPD
C1033 uF / 25 V33 uF / 50 V647-UBT1H330MPD
C1122 uF / 25 V22 uF / 50 V647-UBT1H220MPD
C124.7 uF / 25 V4.7 uF / 50 V647-UBT1H4R7MPD
C134.7 uF / 25 V4.7 uF / 50 V647-UBT1H4R7MPD
C141 uF / 50 V647-UBT1H010MPD
C151 uF / 50 V647-UBT1H010MPD
C161 uF / 50 V647-UBT1H010MPD
C171 uF / 50 V647-UBT1H010MPD
C181 uF / 50 V647-UBT1H010MPD
C191 uF / 50 V647-UBT1H010MPD
C201 uF / 50 V647-UBT1H010MPD
C211 uF / 50 V647-UBT1H010MPD
C224.7 uF / 160 V647-UPW2C4R7MPD

The Preamplifier P.C. Board is coded CA-5205 and it occupies the upper middle part of the underneath of the steel chassis. You can't miss it as it is the biggest printed circuit board that you see when you look at the steel chassis turned upside-down. It is also the most complicated board of them all.

Parts need to be ordered twice for this board.

Preamplifier P.C. Board (CB-5004)
IdentifierSchematicRecommendedMouser Number
C2220 uF / 25 V647-UKZ1E221MHM
C310 uF / 16 V647-UKL1C100KDDANA
C547 uF / 10 V47 uF / 25 V647-UFG1E470MEM
C747 uF / 16 V47 uF / 25 V647-UPW1E470MDD1TD
C94.7 uF / 25 V4.7 uF / 50 V647-UFG1H4R7MDM
C104.7 uF / 25 V4.7 uF / 50 V647-UFG1H4R7MDM
C1510 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C1610 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C1810 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C1910 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C2010 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C24220 uF / 16 V647-UPW1C221MPD
C2510 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C2847 uF / 10 V47 uF / 25 V647-UFG1E470MEM
C2910 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C32100 uF / 10 V100 uF / 16 V647-UPW1C101MED1TD
C3547 uF / 16 V47 uF / 25 V647-UPW1E470MDD1TD
C3947 uF / 16 V47 uF / 25 V647-UPW1E470MDD1TD
C4010 uF / 16 V647-UKL1C100KDDANA
C4110 uF / 16 V10 uF / 50 V647-UFG1H100MDM
C42100 uF / 10 V100 uF / 16 V647-UPW1C101MED1TD
C4310 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C4447 uF / 25 V647-UPW1E470MDD1TD
C4610 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C471 uF / 25 V1 uF / 50 V FILMN/A
C4810 uF / 16 V10 uF / 50 V647-UFG1H100MDM
C50100 uF / 10 V100 uF / 16 V647-UPW1C101MED1TD
C5210 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C531 uF / 25 V1 uF / 50 V FILMN/A
C5410 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C5510 uF / 16 V10 uF / 50 V647-UFG1H100MDM
C5747 uF / 25 V647-UPW1E470MDD1TD
C5810 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C6510 uF / 16 V10 uF / 50 V647-UFG1H100MDM
C6847 uF / 10 V47 uF / 25 V647-UFG1E470MEM
C6910 uF / 25 V10 uF / 50 V647-UFG1H100MDM
C72100 uF / 10 V100 uF / 16 V647-UPW1C101MED1TD
C7547 uF / 10 V47 uF / 25 V647-UPW1E470MDD1TD

From an electronics point of view, this deck has a complicated construction. It rises big problems for a full maintenance job. There is an insanely large number of electrolytic capacitors. Where it was possibly I have chosen a film or a tantalum capacitor. The density of electronic parts on these printed circuit boards is very high for a 1970s design. Oh, I almost forgot: there is no silkscreen. This means you need to double and triple check the position and polarity of each capacitor.

As always, let me throw a piece of advice. Do not hurry, take your time and do the job once. And do it well.

Article #4 | 10:30 AM Saturday 10/02/2018

AKAI GXC-760D: Replacing the Lamps

This cassette deck suffers from 40+ years old burnt lamps syndrome. Time to change all of them. There are five miniature bulbs rated 24 V / 40 mA in the command buttons. Another two axial lamps are in the cassette compartment. The remaining two axial lamps are above the VU meters, behind a prism. All axial lamps are rated 8 V / 200 mA. I sourced all these bulbs then proceeded to disassembling the unit.

First off I have addressed the control buttons bulbs. The rewind and pause bulbs were still functioning. A a best practice, I replaced all bulbs in this machine.

This is the controls block.

This block is disassembled by removing the two screws holding the black mask. Then you carefully remove each button by gently pulling it out. A spring is beneath every button. Remove these as well. Then again three screws are securing the lights block to the chassis. Remove the top two screws and the bottom screw. The entire plastic holder gets out. Now another two screws are holding the printed circuit board to the plastic holder. Remove these as well. And...

Clearly someone has been here before me. Check out the 3 mm LED in place of the record bulb. Dirty job. The guy even managed to break two printed circuit tracks. Which I fixed as well with miniature copper tape. I also cleaned the lights board with isopropanol. The result is as follows.

Here are the new miniature light bulbs.

I have then cleaned all the old AKAI corrosive glue with D509 thinner. Lengthy job. And smelly too. Well ventilated room is a must. Then I have inserted all the five new miniature bulbs and soldered them in place.

Next follows the trimming of the bulb leads. This is how the bulb array looks on the other side. All nice and clean.

Detailed miniature bulb view.

All good on these control buttons bulbs. I have reassembled everything in reverse order.

Next up is the replacement of the cassette compartment bulbs. You can see here that only the left original bulb is still lighting. The other one is burnt.

Removing this steel part is easy enough. Two screws are securing it in place. Here is the steel holder with the old lamps. New lamps are visible also.

I have then reassembled the steel part in reverse order and moved on to the VU meters illumination.

Just above them there is a small screw that is holding up the light dispersing prism. I have removed this and put the prism in a safe place. Each VU meter has its own lamp mounted on independent printed circuit boards. Each board is secured in place with two screws. I removed the screws and desoldered the lamps. Unfortunately the same corrosive glue attacked two of the screws. I gave them a good bath in a mild acid that eats rust. The screws came out clean.

Before reassembling the prism I fired the unit up. The lights came on as expected.

I have secured the prism back in its place then I have proceeded to the reassembly of the unit. This is the lamp aftermath. All junk.

Here are some final pictures.

I have chosen not to replace the lamps with modern LEDs because I really like the warmth of incandescent light. I hope the new bulbs will last at least as long as the old ones.

Article #5 | 07:30 PM Friday 16/02/2018

AKAI GXC-760D: Replacing the Pinch Rollers

Because the pinch rollers on my unit were worn out, I have acquired new parts custom built for this model. Here are the old rollers versus the new ones.

You can clearly spot the worn out surface of the old pinch rollers in the pictures below.

I think that the old ones were not original. Why is that? Because most certainly the original rollers had a bronze inner bush. Exactly like the new custom built ones.

Pinch roller replacement is a fairly easy operation on this machine. First I have removed the springy metal C-clips and then I have pulled out the old rollers. I used isopropanol to clean the shafts of old grease and residue. I have applied a slight amount of quality fine grease to the metal shafts and then installed the new rollers. I have used some nylon washers between on each side of the rollers. Next up I have secured the rollers in place with the C-clips.

All good. Deck plays perfectly now. I can play again 100 min. BASF tapes without fears of tape skewing.

However there is a design drawback. When the deck is idling, like in 90% of its life, the left pinch roller is pressed by the tape guidance mobile armature. There are two springs behind this armature. One is regulating the tension on the pinch roller and the other smaller one is regulating the tension on the mobile armature.

So this is how the roller appears is pressed when idling. This is not good on the longterm. The rubber on the tape contact path will deform and distortions due to wow & flutter will increase.

I have found a silly simple solution to this problem. Thus I have installed a piece of standard printing paper folded in six directly between the tape guidance arm's resting piece and the metal shaft where it rests. It is hard to explain but simple to figure out while you are looking at the tape transport with a pen flashlight.

Now the pinch roller is no more pressed by the tape guide when the deck is idling.

Here is the paper.

And the assembly in play position.

Up until I will come back with a better solution, I am happy with this.

Article #6 | 10:00 AM Saturday 17/02/2018

AKAI GXC-760D: Replacing the Belts

Next follow the belts. In this unit there are only two rubber belts. One is for the closed-loop dual capstan drive system. And the other one is for the counter and auto-stop mechanism.

The capstan belt is wider. Here is how it looks like.

The counter belt has a normal cassette tape deck belt profile and it looks like this.

The replacement is a relatively simple operation. Once you get out the mecha-block steel backplane that holds the capstan drive AC servo motor. As for the counter belt, I have just persuaded the old belt to exit via space between the reel hub and the metal cassette assembly frame. The new one was mounted in reverse.

The new belts have no notable improvement over playback quality. The machine was operating fine with the old ones as well. But knowing that new rubber parts have been installed, should keep me on the safe side for years to come.

Article #7 | 14:50 PM Saturday 24/02/2018

AKAI GXC-760D: Adjusting Playback Speed

This cassette deck has a FG servo controlled AC capstan motor. Thus it has a small printed circuit board that performs the task of the speed controller.

Adjusting the playback speed is done by connecting the output of the deck to the input of the frequency counter while playing a reference tape containing a 3,150 Hz sine-wave recorded at 0 dB.

I had to use an insulated screwdriver to rotate the trimmer. With a metallic screwdriver there were some unusual anomalies in motor speed control. I found out that the FG servo circuit is fairly stable. Less stable than my aging Hewlett Packard sine-wave signal generator. But this is as expected.

Playback speed is now accordingly set thus I can enjoy good music reproduction on this unit.

Article #8 | 09:15 AM Monday 14/01/2019

AKAI GXC-760D: Servicing the Reel Motors

While using this tape deck I have noticed that the tape transport stops completely from time to time. As I have changed the only two belts in this machine there is nothing else suspect but the reel motors. In my case the take-up reel motor. So it is time to dismantle it and see whether I can fix it.

In this deck, the reel motors are of type GSM-300 and have three functions: motor, generator, and magnetic brake. These motors are now no longer available (NLA). So fixing them is imperative in order to use the GXC-760D machine. Otherwise another solution would be to find a similar deck for spare parts and recover the motors hoping they won't exhibit the same issues.

Let's go with the problematic reel motor first. I won't detail the removal procedure since it is straightforward if you study the mechanical block.

This is the motor outer shell. There are two screws that secure the bottom cap in place.

This is the motor inner shell. There are three screws that secure the bottom cap in place.

This is the stator with the beautiful copper wire. You can clearly spot the carbon and oxidation build-up on the collector. I am going to remove this altogether and clean everything.

The bottom cap contains the carbon brushes. In this motor the brushes are quite worn. But there is nothing I can do to straighten this up. What I can do however is to replace the dampening sponge material between each brush arm and the tensioning spring. Well it is not quite a spring but it works similarly. This material is worn already and it easily disintegrates when touching it. I have cut two pieces from a similar material that I had around and then glued them in place of the old dampening material.

This is the restored motor stator.

The motor is reassembled.

I have reapplied a drop of quality grease to the bronze bushings. The motor was installed back from where it came from.

Since I was at this point, I decided to work on the supply reel motor as well. The steps are identical so I will just put some pictures.

The motor servicing operation is lengthy and requires patience. It took me about one hour per motor from disassembly to reassembly. You need an additional hour to disassemble the mechanical block and about the same time to reassemble it as it was.

While working on these motors I also preventively changed two electrolytic capacitors on the motor power supply distribution printed circuit board that is bolted to the mechanical block. These are rated 0.47 uF / 50 V. I have replaced them with Kemet modern parts rated 470 nF / 50 V. Normally I would have changed these on a full capacitor replacement job. But given this PCB is hard to reach and I already had the mechanical block disassembled, the operation seemed straightforward.

This is the completed board.

Next I was eager to test the serviced motors. I have reassembled the mechanical block and loaded a cassette. The two motors accelerated very quickly and snapped my tape in a split-second. I observed that both the supply and the take-up reels were spinning in reverse. I cannot explain why this was happening since I noted carefully the way the wires were originally soldered. Upon re-soldering I have carefully respected these notes. Total disaster. After inverting the connections to both motors it appears that everything works as it should. Play mode delivers constant torque. Fast forward and fast rewind is smooth again.

Now I will throw-in my 2 cents opinion. One should never ever run these motors on full speed without a tape loaded. The motors will accelerate very fast and the brushes will quickly deteriorate. The collector will become clogged in carbon dust and spark debris. In the end the motors will either work intermittently or loose torque. Or both at the same time. I would go even further to say that it wouldn't be advised to fast forward or fast rewind tapes on this old deck. Once the brushes are completely deteriorated it will be very hard to replace them. Assuming that this operation can be done.

That's about it. Off I go to listen some tapes now.

Article #9 | 08:58 PM Tuesday 22/01/2019

AKAI GXC-760D: Wow & Flutter Curiosity

This evening I loaded a Sony Metal cassette tape in the GXC-760D machine. After the first couple of seconds I was like jaw drop: an insane amount of wow & flutter was omnipresent. I mean what the heck! New pinch rollers, serviced motors, new capstan belt, clean and lubricated mechanism. And junk mechanical performance.

No matter what, the Sony Metal cassette had this massive wow & flutter over the recorded music. Especially on the beginning of the tape. I loaded then a 100 minute BASF CrO2 tape and there was absolutely no audible wow & flutter. I really cannot explain what is wrong.

The weird thing is that this Sony Metal cassette works flawlessly in my Sony TC-K850ES recorder. It is obviously a mechanical problem but I cannot identify it. Pinch roller pressure is good, otherwise it would chew up longer tapes such as 100 and 120 minute cassette tapes.

I guess this machine doesn't like metal tapes. Unfortunately I cannot postulate this theory since I only have two metal cassettes: both being cheap Sony Metal XR types.

At this very moment I am like 20 minutes deep into side A of the 100 minute BASF cassette and everything is perfect sound-wise. You can never be bored with vintage stuff. Either you repair one module then another one fails elsewhere either there are incompatibilities. Oh the joy!

Article #10 | 07:24 PM Friday 25/01/2019

AKAI GXC-760D: Wow & Flutter Follow-up

In a madness access I have disassembled the mecha-block and cleaned the capstan sintered bearings and shafts. Then I applied 0W20 fully synthetic oil with the aid of a needle connected to a 5 ml syringe. The oil was applied through the oil hole under each bearing housing. Each capstan assembly ate about 1.5 ml of oil. I don't know where this oil is now -- obviously inside the bearing housings. It doesn't drip outside via the oil hole nor does it escape though the microscopic space between the capstan and the sintered bronze housing.

I put a tiny drop of the same oil inside each sintered bronze pinch roller bearing for better lubrication. The fine grease that I used last year when I mounted the new pinch rollers was already dried up and evaporated for some reason. The oil is here to stay. Hopefully the capillary action will draw and hold the oil molecules inside the bronze bearings for a long time.

The result is that now even the junk Sony Metal XR tape doesn't have noticeable wow & flutter while listening on the headphones.

What have I learned? Never trust the vintage HI-FI sellers. The guy from whom I bought this deck back in 2018 said the mecha-block was fully serviced and oiled up. I think he was dreaming as the capstan bearings were running dry. I was mislead by what he said and I thought the mecha-block was OK. Apparently it really needed maintenance.

12.25 AM Later Edit: Here I am deep in the night listening to a BASF Chrome Super II old tape with selections of The 3rd and the Mortal. I almost forgot how atmospheric this band was. I don't know if it still exists at the moment. I played the 90 minutes tape twice in a row.

If you will

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