Kenwood KA907 repair Manual
Pagina 1 di 8Kenwood KA-907 Audio Amplifier Repaired
Kenwood KA-907 repair
The Kenwood KA907 is fast high performance well twenty years old Hifi amplifier. The repair of a defective amplifier is shown and as with the repair step for
step was proceeded. To see is, how from the repair a restoration became. Subsequently, the amplifier was measured on its characteristics as audio amplifier.
Automatic website translation from German to English done by www.world.altavista.com
Error tracing and cleaning
The amplifier had accumulated a long and hard
life behind, for much dust and traces of rust.
The equipment serves as a spare part donor. It is
actually culpably equipment in such a way comes
to leave. After cleaning a few years can never
harm.
A second Kenwood KA907 amplifier is to be
repaired. It still is in a tidy condition. Broken, at
the output stages the exit hangs firmly on +60
V. Without the inserted protective circuit it
would have been the end for a DC coupled
woofer.
Output stage plate solder side. To their in the past clearly
already one soldered. To see down on the right of also the
semiconductor cemetery, all defective output stage
transistors. With many a low impedance connection
between basis and collector. One thereby without shortcircuit, large currents must have flowed.
The orginal output stage transistors are the PNP types 2SA1065 and NPN types 2SC2489. Both are very fast, them are fast bipolar power transistors
(150V/10A/120W). Unfortunately all are defective. A few new 2SC2489 and a set of 2SB681 PNP and 2SD551 NPN (150V/12A/100W) are provided. At first sight
these fit as replacement quite well, the amplifier however not for these types were compensated. If the original typ is still available hardly, the usual transistor
comparing lists help. These lists do not beru-hit a corner-sight however always all parameters. If another transitor type is used, always exactly superior at which
screw thereby one turns. In many cases is it possible, often is even improvements thereby attainable, in addition, the opposite is possible. The amplifier can
reciprocate and be able even still more semiconductor broken to go.
After the cleaning the printed circuit board is again clean. To see on the right of and on the left of the TO-3 versions for the output stage transistors, which are
put and over the two screws the collector (housing) contacted. The put variant is very low-maintenance. White resistances are 0.47 ohms of resistances, them
serve for the local inverse feedback and simplify also that join in parallel the transistors. The driver transistors apparent get along with small radiator box.
Fortunately is present service the manual, without connection diagrams would be the repair a toilsome work.
To fast searches of errors in the circuit a simple circuit analyzer already meets:
thus the pns of transitions with for instance 600mV at basis emitters and basis collector can be examined. Rough Transitordefekte is so recognizable.
Diodes in passage direction show for instance 300-600mV
Resistances may be never many high impedance as the nominal value, otherwise it is surely broken. Often the ohm measurer shows however a lower value, is
usually because of the parallel connection with other elements. Many defective of resistances show "become high impedance" up to the complete loss. The
21/11/2005http://www.amplifier.cd/Verstaerker/Kenwood/KA907_english.htm
Pagina 2 di 8Kenwood KA-907 Audio Amplifier Repaired
devil is however in the detail, sometimes gauckeln the resistance condensers joined in parallel the circuit analyzer wrong values forwards, it can several
seconds take to the announcement stands.
in the inserted condition DC voltage measure. However the oscilloscope, it is clearly safer shows the cam shape, the circuit analyzer is unsuitable in this
relationship.
At first sight - OK ONE - the output stage was attached now at will at the
left channel. It equipped with the provided 2SB681 and 2SD551
Transitoren. Those are not the correct, but the only sentence, which are
still functional. For tests is correct even if it hops go could. Volume
controls on zero, entrances leave and power ON open. Often an oscillation
begins not immediately with the smaller amplitude to reciprocate but only
with larger - thus a small signal put on and volume slowly untwist and at the
oscilloscope observe which happened, if the sine begins easily - danger! can have gone to volume back - if the scope then becomes nevertheless
suddenly completely "green" is often too late it - the transistors broken.
The risk of an immediate oscillation exists.
After first switching on on is the amplifier stably,
off approx.. it begins to reciprocate to 25 V easily.2
kHz without load at the output.
Now if the amplitude is still continued to increase
(10V/DIV), the oscillation begins, very clearly
particularly in the negative half wave.
Japanese Toshiba power bipolar transistors in the TO-3 housing. Installed with
thermal compound and mica disks for the electrical isolation of the collector to the
radiator box. When the assembling of the insulating washers gladly errors are made,
therefore with the ohm meter for high impedanceness to examine. The thermal
compound is widerliches things at the fingers that then later also everywhere at the
whole table lubricates. Lay on only as much that the surface scratches and
unevenness in the insulating material, housing and radiator box are filled out. The
paste is not to form an additional developing layer, which worsens the thermal
resistance, really only for gaps to fill. Also fully do not tighten the screws at the
beginning, so that few days later after the first enterprise still a something can be
pulled tight. It lasts always something to itself the paste completely sets.
the Transitoren became warmer, by the
permanently flowing idle current. The oscillation
grows, slowly it off time will switch.
The causes for an oscillation are variously, probable a defect in the first part of the output stage (long plate behind the large electrolytic caps). Whereby the
wrong output stage transistors this oscillation with to favour can.
Now an attempt was started, as the
last remaining 2SC2489 orginal
transistors affects the output.
Power Amp (X00-2010) part (B)
Now it applies to examine the long plate behind the large caps.
Meanwhile times the oscilloscope exchanged one with CRT
readout. The picture shows the amplifier in the no-load
operation with PNP spare transistor and the orginal NPN in
the positive half wave.
When putting on 4 ohms a load the amplifier becomes
stable. Only strangely enough it already begins with
approx.. to clippen -13 V. Still defects must be
present.
In the long printed circuit board (X00-2010) (B)
directly behind the large caps very probably is an
error. The two channels are symetrically developed,
quite short layout, hardly wire entanglement. The
oblong, narrow, small Transitoren is Doppelfet's in the
first stage (for instance in the picture center on the
right of and left), in a housing that is favourably e.g.
because of the drift characteristics over
temperature. Which with the years, the decreased
capacitance of the elektrolyt capacitors became
possibly critical. Also the small potentiometers (above)
are gladly a problem, only to turn if necessarily,
outside of the place where the middle contact all the
years is confessed, the surface are gladly dirty, there
must few times turn to become around the deposits
with the middle contact to remove.
the first error already found, a 180 ohm
resistance more largely than the
measuring range of 2 kOhm. That is
broken. The surrounding construction
units, in particular semiconductors applies
it now to naturally also examine. Down with
the printed circuit board, sometimes more
simply said than done. First a few pictures
make, serves later than assistance around
to know which wire again where be ranksoldered must, without studying
laboriously the connection diagram. If the
wires away are, still the plastic tie-clips
come.
The amplifier became completely dirty with the years.
Down one sees the four electric rectifiers capacitors. In
the present condensers with same capacity became
importantly smaller and also better. Nevertheless I see no
reason it exchanging with those for the time being, am
nearly still ok, am also not so cheap. Besides also here
something Orginalzustand is lost, although is rather all the
same to me otherwise actually. Measuring of the capacity
would be interesting, but only if they out-solder
themselves leave. Who does not have a capacity measurer
for high capacities, this can judge also in such a way:
Amplifiers e.g. with 4 ohms load (sufficiently strong
loudspeaker goes naturally also), to untwist times
correctly. Observe a sine signal and the appearing ripple
voltage at the condenser with the oscilloscope as signal
source. With a bad capacitor with falling capacity a high
ripple will anyhow have to be seen.
21/11/2005http://www.amplifier.cd/Verstaerker/Kenwood/KA907_english.htm
Pagina 3 di 8Kenwood KA-907 Audio Amplifier Repaired
Only no inhibitions, the dirt must simply down of the
printed circuit board. A mild seifenloesung makes
nothing at all for the printed circuit board and
electronics. A cleaner is also favourably, since it dries
arrears-free and well, like the table-ware also. Of
course also white spirits go, are however already
stronger in the effect, give gladly easily matte of
arrears on the printed circuit board and are the
recommendation for the second wash course. The spray
"LR" printed circuit board cleaner with plastic brush in
the spray button integrated, of the contact chemistry
supplies good cleaning results. For a strongly
verdreckte plate one needs however much of it. LR is in
the best way suitable also for soldering arrears of all
kinds. There are possibilities a printed circuit board to
be cleaned. It can already be that water is clean-run
somewhere into a construction unit, where it would
when new not clean-run it a certain risk. A usual
construction unit that because of washing now
accelerated broken would go, soll's nevertheless calmly
directly broken go, I remove it, rather with me than in
the next year with you in the living room.
After washing the printed circuit board is drying to
foehn, but not with the 2000 Watts, but like that
that one can hold the printed circuit board thereby
still comfortably with the hand. Do not adjust thus
too hot. A compressed air pistol dries also well. Only
at air to leave go also, the landing on water esters
dry, one do not see remaining however for a very
long time not receive, are not not so mad when
switching on on. A heating element does good
services. Landing on water esters form e.g. in the
housings of the small potentiometers, these must in
any case have drained before switching on on. The
cleaning already requires a certain expert
knowledge. A clean printed circuit board makes
possible also an optical inspection. Even if it does
not harm functioned, it these also under a
magnifying glass to regard correctly. Fine tears in
the conductive strips are to be never excluded, and
mechanically damaged construction units better
recognizably. Can the soldered connection one so
also better observe or after-solder.
The printed circuit board (X00-2010) (B) looks
now again clean. The remaining defective
construction units on this printed circuit board,
Transitor Q20, one NPN 2SC2633 and two
pertinent of resistances R48, 50 with ever 180
ohms. In the other channel Transitor Q19
2SC2633 was defective, but resistances the
ok. Why also always the power stage transitors
went broken, very probably a defect caused the
other one. By the one internally short circuit
transistor Q20 for example a too high tension
lies at the resistance, the high achievement
makes for it from and it goes with defective. It
is not always foreseeable whether a transistor
in the case of defect short circuit or becomes
high impedance, depends on the error situation,
in addition, on the internal structure, which
gives way e.g. first: the bond wires (becomes
high impedance) or the silicon (can melt, with
low impedance).
If already already, everything is made clean. Do you see like copper shine again beautifully?
The internal printed circuit boards are all down, there are already worth it the baseplate
to deseam itself. Put on the water, soap and the brush. Before however thoroughly superior
where only little water drauf is. E.G. do not ask the trafos and relays. Naturally as few as
possible into the potentiometers or the volume control, which leave themselves front part
(potentiometer and switch) and rear part (preamplifier) well to take off here.
spare parts for the defective Transitoren Q19 and Q20.
There see, one both to the output stage with the spare tansistors runs at both channels of the long PCB. The photo shows the output voltage at 4 ohms a load. I
have here times untwisted on 36 V amplitude with a frequency of 2600 cycles per second. Is small still little more amplitude, afterwards the amplifier begins
however with clipping. It brings these 36 V at 4 ohms at the scope with that orginal transitors. The radiator box will very warmly not to recommend for continuous
operation with such to a sinusoidal load.
First no further defective construction units more found, the time to the assembly tingles/swarms
in the fingers.
The safety devices shine even again little. On the large plate is only
rough wood at construction units drauf, for which the water does
not make anything. Complete dividing and the large plate, would be
so much work, which hardly counts itself, cleans the whole ten
thousand wires and cables to solder on off and again - thanks no goes here also times differently.
Maximum performance are:
36 V * 36 V/4 ohms = 324 Watts (during the point of the sine)
(36 Volt/1.414) * (36 Volt/1.414)/4 ohms = 162 Watts effective achievement (RMS)
RMS of Watt thereby is meant the achievement in warmth, which would convert the sine equivalent DC voltage in the resistance in warmth. In other words: the
sinus-wave voltage shown with the amplitude of 36 V would heat the load resistance on a certain temperature. DC voltage of 36 Volt/1.414 = 24.445 V would heat
the load resistance exactly on the same temperature as the sinus-wave voltage.
Loudspeaker connections
As carriers now a small pretty peace of
A desire the orginal loudspeaker cables connecting
terminals had to be replaced with banana sockets.
The black socket carrier must be expenditure-soldered,
the banana sockets can be mounted simply not on it.
wood is to serve that then from the rear
on the rear wall one screws. Thus is
written "chamfers" correctly.
About the measure with the wood one can argue. Purely electrically (Ohm's resistance) it does not bring to anything, perhaps that here: a gilded banana plug and a
gilded socket do not oxidize over years, the contact resistance remain constant, as long as the spring action remains (nickel plated goes naturally also). Some cclamps loosen with the time a little, do not control there occasionally did not operate. Favourably is the banana plugs solution also, since a short-circuit actually
only heavily in addition-get is. If one considers the fact that with both amplifiers the output stage transistors went broken then is the risk of short circuit not at
all on the light shoulder to take times in such a way. How quickly does it happen that a few kupferadern are not under-wedged correctly and pluses and minus to
touch oneself? Particularly then if one has it hasty, the light bad is etc.. Yes, that is an advantage of the banana sockets, conclusion with the Gefummel, different
21/11/2005http://www.amplifier.cd/Verstaerker/Kenwood/KA907_english.htm
Loading...