AU888 restore discussion

gslikker

Super Member
I bought an AU888 amp.
Having the restorations from others and the modifications by @kevzep in mind, I decided to dig into the mods and probably recap the unit.
So I started removing some caps and resistors and before I knew it I had wiped all the parts from the circuit cards accessible and visible from the top......

Some tinkering paranoia here. My working time on it is limited so it may take all winter. I have all kinds of NOS and salvaged components at my disposal.

A general remark beforehand is that part of the boards are the AU999 type and some component differences are already to be found in the schematics.

The status at the moment is repopulating the two small power supply boards and having 6 empty other circuit cards.

My plan is to exchange some info and put the efforts in this topic.
I am looking forward to discuss this with the gurus and just go each board one by one when repopulating them with parts again. Please be aware I am an "educated guess guy" instead of a design EE.

So: below here is the first discussion subject, which is the small power supply feeding the main amps.
 
Last edited:
au888-F1293pwr.jpg

So, here is the power supply, which feeds the 4 differential pairs of the main amps only, quickly browsing the AU888 schematic.
Note, my plan is to use bigger caps (as I have them already) for the power amp, and as well for the B5 / B6 power attachments.
By quick glance, it looks the supply puts out some 30 to 40 milliamps total feeding the main amps differential pairs.

The idea is I am trying to figure out some risks and situations. I would like to stiffen the power supply without risks (remember, it is tinkering paranoia)
So, mind jogging the very purpose of this board/power supply setup looks only to have a little more driving voltage for the power amp, in a less easy way.

Things which get my attention are:
- What happens if one would put seriously more capacitance instead of the 4.7 microfarad caps at the B5 /B6 outputs.
What I have in mind, is there would be big voltage sag of the main smoothing caps during high current output of the main amps. It could result in B5/B6 voltages still high and get a wrong-way voltage over the power supply transistors momentarily.

- The collector-base resistors which differ by 50%, they could be for proper power-up/down behavior of the voltages.
Those are to be left original I think, but if anybody has other thoughts of the "why" of these resistor differences I would like to know!

- The usefulness of the circuit as a whole (3V AC only, single phase rectification to about 4 Volts only, losses over the diode and output resistor, losses over the transistors.)

Nice to know is the original transistors 8002 is HFE 60, the 9002 is 50, on the DCA75.
Alltransistor site states minimum of 8002 to be 40, and minimum for 9002 to be 80. (This indicates the 9002 could be deteriorated)
 
Last edited:
F1230 board, B7 and B8 power.
The F1230 is supposed to be on AU999 according schematics but also in this AU888 it is to be found, while the schematics call for a F1230A.
Apart from the C007 on the F1230 is 220 uF and 100 on F1230A schematic I see no difference.

Apart from doubling all capacitances except for the C011 ceramic (which stays a ceramic), C007 went from 100 to 330 uF, just because it did fit and no other of same leg width was available. Hopefully and probably it will not give any problems. As the old big caps had signs of heating by the R003 68 ohm resistor, I took a little bigger one and mounted it a few millimeters above the card te enable a little more airflow.
 
Last edited:
F1293 board, B5 and B6 main amp diff amps supply.
To gain 0.4 Volts on each supply, 1N5820 Schottky diodes were mounted, and I guessed the output resistors could be lowered to 39 ohms.
I decided upping the output caps to 22 uF would not give a problem (although this has to be proven later...)
The half wave rectifier caps went from 1000 to 3300 uF., NPN transistor became JAN2N2219A and PNP J2N3868.
 
Last edited:
Pictures of the two power boards.
The restuffed boards I cleaned off the solder first and then put on a little "SK10 Lotlack" spray.
This is a spray for circuits cards have it protected but still solderable afterwards.

I hardly do see any rosin after soldering so I just leave it, as the picture shows it was worse, before

2018_1201_201704.JPG
2018_1210_205413.JPG 2018_1210_205459.JPG

To be continued later.
 

Attachments

  • 2018_1201_201744.JPG
    2018_1201_201744.JPG
    53.3 KB · Views: 9
Last edited:
One last thing I did was measuring the coil which was removed from the tone board.
It is supposed to be 0.8H (according the schematic, at whatever frequency).
Now, one measured 0.792 which is very close (at 1 kHz) and the other 0.742 which is over 10% off.

Nice to know it has a SET SCREW on the side. It is LOCKED by laquer on the inside, so turning the set screw is RISKY because it is a ferrite and probably very vulnerable.
I managed to scratch off a little bit of the tiny amount from the end of the set screw and it was loose enough to adjust.
This can only be done if the coil is off the circuit card, from the bottom of the coil housing.
I did leave the coil (which looks like a little transformer) mounted in its round housing so the metal pieces coming out of the housing are not touched.

I will take a picture of the coil, another day.
 
Nice to know it has a SET SCREW on the side. It is LOCKED by laquer on the inside, so turning the set screw is RISKY because it is a ferrite and probably very vulnerable.
Interesting - I have my tone boards out of my 999 now and have often wondered about that inductor. I wonder if it is worth removing & measuring? No one I think has done that before on AK until you did it in your 888?
 
Interesting - I have my tone boards out of my 999 now and have often wondered about that inductor. I wonder if it is worth removing & measuring? No one I think has done that before on AK until you did it in your 888?
I am not capable of understanding the tone control. While the tone control is not symmetrical, I could imagine it has some measurable effect by default. The tweaking of the coil was rather sensitive and maybe it changes over the years. BTW nothing is connected when the circuit board is out of its connector, so it can be measured in-circuit. There is a series resistor value difference between AU888 and AU999 so I will get back to it when refurbishing those pcb's.
 
Last edited:
Today off day, so time to do the next easy board. The board is supposed to be F-1208A, but actually a F1208 is used.

Apart from renewing the two electrolytics, there was no real reason to repopulate the board, but while cleaning the others, this was done, too.

Most caps were renewed as well as the resistors. Part of the NOS resistors I are supposed to be metal films brand Corning Glassworks, I like them as they have thick tinned copper leads.

Due to the low current I presumed a high HFE transistor was needed so the 2SA640 found having HFE of 250 was replaced by a BC560C having HFE almost 600.
So, what is the noteworthy thing? A lot of parts are missing! This is because of only the F1208A parts are mounted on the F1208 board.

The missing ones are easily identified in the picture. All components were found as drawn in the F1208A schematic.
2018_1211_214546 (Medium).JPG
 
Discussion item.

There is no ground reference at the tone control circuitry, at least I do not see it.

Therefore, it looks to me, the DC voltage level in this circuitry mostly depends on leakage current.
I assume most DC leakage will come from the electrolytic caps C717 on the F1208A board and C719 on the F1209 board according this schematic.
Both capacitors of same size have their "+" side towards the tone control switches circuitry.

So my assumption is, the DC level is somewhere between their "-" connections at TR703 emitter and TR705 collector (mounted and marked this way in the real) and in fact this way one of those caps will be wrong way around.
I found this situation in another Sansui also.

Anybodies comments ????

12081209.jpg
 
Last edited:
next was F1228 TT input.

A modification was found on the bottom, which is a 100 ohms resistor in series with the output capacitor C613. I do not see it on any of the schematics of AU888 nor AU999. I looks a factory mod.

Also, the C615 , 6 nanofarad caps were a little off since one was about 5.8 and one 6.2.

Having nothing alike available, 2.7 and 3.3 nanofarad was put in and they just fit together in the holes. The 100 ohms series resistor is put on the component side now.
 

Attachments

  • f1228bottom.jpg
    f1228bottom.jpg
    90 KB · Views: 55
  • f1228top.jpg
    f1228top.jpg
    73.1 KB · Views: 63
One last thing I did was measuring the coil which was removed from the tone board.
It is supposed to be 0.8H (according the schematic, at whatever frequency).
Now, one measured 0.792 which is very close (at 1 kHz) and the other 0.742 which is over 10% off.

Nice to know it has a SET SCREW on the side. It is LOCKED by laquer on the inside, so turning the set screw is RISKY because it is a ferrite and probably very vulnerable.
I managed to scratch off a little bit of the tiny amount from the end of the set screw and it was loose enough to adjust.
This can only be done if the coil is off the circuit card, from the bottom of the coil housing.
I did leave the coil (which looks like a little transformer) mounted in its round housing so the metal pieces coming out of the housing are not touched.

I will take a picture of the coil, another day.
Sorry, but how is that over 10% off?
0.8 - 0.742 = .058
.058/0.8 = 7.25%

Am I missing something?
 
Nice - so the 888 has all carbon-comp resistors? I recently did that board on my 999 and it had all normal carbon types in it
 
Sorry, but how is that over 10% off?
0.8 - 0.742 = .058
.058/0.8 = 7.25%

Am I missing something?
Of course, you are right, my thinking was garbled at the time, I guess :dunno:

Nice - so the 888 has all carbon-comp resistors? I recently did that board on my 999 and it had all normal carbon types in it

I swiped everything off the boards so what you see is not original.
The new ones may look like carbon composites from a distance, however I was told a long time ago by the original owner they were CGW (Corning) metal films.
I think they are commercial style CGW RL20 ones, which look having about same body as CGW RN60D.
Although, maybe other people know which brand/type they actually are so a picture again:

You can see they have 4 holes in the lead attachment, and the left one is what remains when you burn the resistor using gas and clean off the debris.
For comparison, a grey CGW resistor having text on the right, that one does not have the 4 holes but apart from that it looks the same.

2018_1220_095448.JPG
 
Last edited:
In the meantime, the 1208 board has added 160 k resistors from the outputs (pin 8 and 11) to ground, to get rid of the lack of ground reference problem.

The 1209 Tone block board R753 is 18k. Note: it is 12k in the AU999 schematic.
The modification suggested by @kevzep was put in, I guessed a high HFE transistor would do in TR705 position so I put in a BC560. The other two are KSA1015.

This apart from higher value capacitors / slight resistor differences just as I desired/did and/or I did have available.

Another small difference from A999 schematicis R739 being 6.8k which is marked 8.2k on the AU999 1209 board schematic, and R747 is 10k versus something looking like 12k on the schematic of AU999

Biggest difference is R715, which gives less amplification in the output stage for the AU888.

See the AU888 and AU999 schematics for the differences.

The position changed R755 because of the modification is easy, just mount it 90 degrees off (of course, drill a hole), as can be seen on the picture.f1209.jpg
 
Last edited:
Fuses are relocated to make room for a possible extra circuit card for a loudspeaker relay.
Another diode bridge rectifier was mounted.

Below pictures are of the protection resistors soldered to old blown fuses during initial test of a main amp circuit card, and to show it is quite useful to clean up debris from the card edge connector sockets.
Just fold a piece of paper and spray some contact cleaner.

More to follow later.
fuseresistors.jpg cleanpaper2.jpg cleanpaper1.jpg
 
In the meantime, the power amp boards are repopulated and together with power boards mounted back in the amp, bias is 20 mA for the moment. Bias can be set precisely due to a 20 turn potentiometer.
The small signal transistors all were replaced.

The input circuit is modified to have more equal base current for the input pair and as well to have a precise zero offset adjustment, which easily lets the amps adjusted for lower than a millivolt DC.
R805 was 47k, it was modified to 46k4, and a 820 ohms resistor was put from the 5k (also 20 turn now) potentiometer to ground. So, the input transistor pair sees are more alike 47k resistor in total, and at the same time the offset pot is much less sensitive.


2019_0122_221208.JPG


I have a Burosch NF200 audio test piece. It consists of a sine/square wave generator, an analog distortion analyzer which can do 1kHz only, and internal 8 ohm loads.
It has analog watt meters and an analog distortion meter lowest setting 0.1% full scale.

The (expected) disappointing thing is that due to the low voltages of the power supply, at a first glance the max power output at 1 kHz BEFORE rising of distortion is only some 35 Watts.
At this level, distortion according the meter is a comfortable 0.03% (testing the main amps only, other cards are not in yet).
It does NOT give 2 x 45 Watt without signs of clipping on the oscilloscope, the DC voltage on the smoothing caps goes down to about 30 Volts (from 35), so the transformer is not really up to the task and the specifications are probably a bit "optimistic". At 45 Watts both channels driven distortion is over 3% which is max which can be displayed on the analog scale.
1kHz square wave looks very good.
Things will be tested more later, to see if any errors measuring were made.

(note: main amp transformer windings is just 1 Volts higher than on the AU666, 2x27 Volts versus 26 Volts...from which one already can conclude from 35 Watts to 45 is a big step........)

One more thing is the fact I still feel to have too much voltage loss over the F1293 board transistors and resistors, so I modify it again (a little lower transistor base resistors and again lower output resistors), to get the maximum out of this weird power arrangement).
2019_0122_220627.JPG
 

Attachments

  • 2019_0122_220938.JPG
    2019_0122_220938.JPG
    203.6 KB · Views: 9
Last edited:
View attachment 1352800

So, here is the power supply, which feeds the 4 differential pairs of the main amps only, quickly browsing the AU888 schematic.
Note, my plan is to use bigger caps (as I have them already) for the power amp, and as well for the B5 / B6 power attachments.
By quick glance, it looks the supply puts out some 30 to 40 milliamps total feeding the main amps differential pairs.

The idea is I am trying to figure out some risks and situations. I would like to stiffen the power supply without risks (remember, it is tinkering paranoia)
So, mind jogging the very purpose of this board/power supply setup looks only to have a little more driving voltage for the power amp, in a less easy way.

Things which get my attention are:
- What happens if one would put seriously more capacitance instead of the 4.7 microfarad caps at the B5 /B6 outputs.
What I have in mind, is there would be big voltage sag of the main smoothing caps during high current output of the main amps. It could result in B5/B6 voltages still high and get a wrong-way voltage over the power supply transistors momentarily.

- The collector-base resistors which differ by 50%, they could be for proper power-up/down behavior of the voltages.
Those are to be left original I think, but if anybody has other thoughts of the "why" of these resistor differences I would like to know!

- The usefulness of the circuit as a whole (3V AC only, single phase rectification to about 4 Volts only, losses over the diode and output resistor, losses over the transistors.)

Nice to know is the original transistors 8002 is HFE 60, the 9002 is 50, on the DCA75.
Alltransistor site states minimum of 8002 to be 40, and minimum for 9002 to be 80. (This indicates the 9002 could be deteriorated)
the previous post, I stated not that satisfied with the F1293 board, and I decided to modify it again.
In the end, it has over 2 Volts higher output than the main power transistors power supplies.
Still, I am not really satisfied, since I discovered during "rethinking" this circuit.

The reason for this rethinking is the base AND emitter of the transistors are GROUND-referenced because of the capacitors, but the Collectors are referenced to some 4 Volts higher above the main DC voltage values, which may vary a lot during high peak output. READ: the input to those stabilizing circuits will be compromised when it is needed most......

It looks IF the MAIN power supplies have a big sag due to some large power swing, supply to those regulator transistors will be gone and maybe even less than the output at the moment. Basically, the transistors CAN get reverse voltage and momentarily the 100 µF caps at transistor bases CAN be higher than collector voltage. So I decided to mount safety diodes over the collector-emitter junction, in case the capactior wants to deplete itself over the b to c transistor junction, I hope the diode forward voltage is less than that of the transistor. Maybe I am paranoid...

Anyway, after the mod I found the transistor for negative side more saturated than for the positive side, and also giving more ripple (40mV versus 10 mV) when the amp was full power/not clipping at 1 kHz.
I think I will calculate resistors to parallel to the caps at the base of the transistors to make sure the regulating transistors can not totally saturate.



ANY comments about this way of thinking by people into design welcome, since I am not an EE and I may be wrong...

Below the board pictures and modified parts list:

diodes on back
2019_0126_171613.JPG
2019_0126_171457.JPG
F-1293 Power block
original transistors CDC8002 HFE 61 (alltransistors states 80 minimum for this one, old one may be compromised)
CDC9002 HFE 50

Parts used.
C021 Panasonic FM 3300 / 10 was 1000 / 10
C022 Panasonic FM 3300 / 10 was 1000 / 10
C023 Panasonic EB 220 / 50 was 100 / 50
C024 Panasonic EB 220 / 50 was 100 / 50
C025 Panasonic FR 47 / 63 was 4.7 / 50
C026 Panasonic FR 47 / 63 was 4.7 / 50
C027 CK05BX103K 10 nF/ 100 ceramic alike it was
C028 CK05BX103K 10 nF/ 100 ceramic alike it was
R011 resistor 0.5W 6k8 was 10k
R012 resistor 0.5W 4k02 was 4k7
R013 resistor 0.5?W 27 ohms was 47
R014 resistor 0.5?W 27 ohms was 47
D003 Schottky diode 1N5820 was normal diode
D004 Schottky diode 1N5820 was normal diode
TR003 NPN transistor JAN2N2219 HFE 179
TR004 PNP transistor J2N3868 HFE 98
D? Diode 1N4004 parallel to R011
D? Diode 1N4004 parallel to R012
 
Last edited:
Back
Top Bottom