Pioneer SX-9 Where to start, if at all

[markthefixer]

IF you wish to compare voltages first, that's fine with me.

The 8 + 2 transistors per side that we have talked about.
two channels. at least 3 readings per.

But it's a lot harder for me than the simple test of swapping 4 last transistors.
ESPECIALLY if the fault lies in the driver transistors. I have to specify all these measurements, figure out what's significant.

A compromise: measure the base to base and emitter to emitter voltages on each channel for the drivers.

q133 base to q135 base voltage
q133 emitter to q135 emitter voltage
q134 base to q136 base voltage
q134 emitter to q136 emitter voltage

we'll see if the channels are being driven asymmetrically.

not quite as good as swapping them, but for now...

 

[weeber]

I will swap the driver transistors and lift all the components that I cold tested and test them with the DMM.

You know the best way to trouble shoot and if that's what you'd do then that's what I'll do.

It'll probably be a week before I can get back to it.

 

[weeber]

IF you wish to compare voltages first, that's fine with me.

The 8 + 2 transistors per side that we have talked about.
two channels. at least 3 readings per.

But it's a lot harder for me than the simple test of swapping 4 last transistors.
ESPECIALLY if the fault lies in the driver transistors. I have to specify all these measurements, figure out what's significant.

A compromise: measure the base to base and emitter to emitter voltages on each channel for the drivers.

q133 base to q135 base voltage
q133 emitter to q135 emitter voltage
q134 base to q136 base voltage
q134 emitter to q136 emitter voltage

we'll see if the channels are being driven asymmetrically.

not quite as good as swapping them, but for now...

Mark, I have a question.

Currently I have all the idle pots set to zero. The left side of course has too high a voltage even set at zero. Should the right side be set to the correct idle voltages before I take the above voltage readings? As a guess I would say yes?

 

[markthefixer]

Both sides should be set to zero for a valid comparison. sorry...

 

[weeber]

Q133 base 1.07mV emitter 0.50mV
Q135 base -0.99mV emitter -0.42mV

Q134 base 1.25mV emitter 0.67mV
Q136 base -1.20mV emitter -0.63mV

These were with vr3, vr5, vr4 and vr6 all turned to their zero positions.

 

[markthefixer]

A compromise: measure the base to base and emitter to emitter voltages on each channel for the drivers.

q133 base to q135 base voltage
q133 emitter to q135 emitter voltage
q134 base to q136 base voltage
q134 emitter to q136 emitter voltage

example:
q133 base to q135 base voltage

red probe on q133 base, black probe on q135 base THEN take measurement.

 

[weeber]

Got it.

Q133 to q135 base to base 2.0v
Q133 to q135 emitter to emitter 0.92v

Q134 to q136 base to base 2.4v
Q134 to q136 emitter to emitter 1.2v

 

[weeber]

I found some time today the swap the driver transistors q133<>q134 q135<>q136 side to side, but it made no difference.

I also double checked the following components by unsoldering one leg and measuring:
d112,d114,d116 and d136 all checked OK
I dont know what th102 is supposed to measure at, but it is 11.5Kohms
r128 218ohms(220)
r132 219ohms (220)
r136 751ohms(750)
r138 755ohms(750)
r140 219ohms(220)
r142 220ohms(220)
r144 6.18Kohms(6.2K)
r146 2.2Kohms(2.2K)
r200 100ohms(100)

 

[markthefixer]

the diodes, you NEEDED to list the forward conduction voltages (like 0.612) displayed by the meter during the diode test.

and you needed to compare the readings of the diodes AND the resistors for BOTH sides.

example:
r128 218ohms(220) r129 221ohms
d112 0.612 d113 0.597


for d114, d116, d113, d115 *** these are IMPORTANT!!! ***

 

[weeber]

I didnt write down the values for the diodes, but they were all 0.6 something. I will get the values for the right side and post them, and redo the left side diodes.

Looking at the schematic, are there any other components, in your opinion, i should also check aside from the ones Im already doing?

 

[markthefixer]

edit - in retrospect, r129, r130, r131, r132 as well.

No more surprises, after the readings are in, if they don't illuminate something, we will go live with a dump truck load of voltage measurements.

I could see how the drivers were off in one channel and on in the bad channel - but as you proved by swapping them, they are only following directions.

I have high hopes for this:

for d114, d116, d113, d115 *** these are IMPORTANT!!! ***

edit 2 - when cold, then when warmed up I am reminding myself to ask for the stv-2 diode voltages from both channels.

 

[weeber]

r130 218ohms(220) r129 221ohms
r132 219ohms (220) r131 220ohms
r136 751ohms(750) r135 752ohms
r138 755ohms(750) r137 753ohms
r140 219ohms(220) r139 219ohms
r142 220ohms(220) r141 220ohms
r144 6.18Kohms(6.2K) r143 6.2Kohms
r146 2.2Kohms(2.2K) r145 2.18Kohms
r200 100ohms(100) r201 99.8ohms

d112 0.548 d111 0.541
d114 0.605 d113 0.582
d116 0.609 d115 0.583
d136 0.631 d135 0.101

Here's the measurements for both sides.

They look pretty even EXCEPT d136 & d135.

Thoughts?

d115 & d135 were the diodes I replaced on the RIGHT side when I blew the amp up.

 

[weeber]

Bump

 

[markthefixer]

D135 is really 0.101 volts??? this was taken while out of circuit?
and the other diodes too?

0.1v is a damaged diode, BADLY damaged.

D135, D136 are 4.6v zener diodes.

They are supposed to be a "safety catch", but 0.1v is damaged. did you get a reverse reading of OL on D135?

otherwise the distribution of the diode voltages is swinging a bit hard the wrong way.

D111 / D112 is ok, but that's a different area of the circuit.

114 + 116 = 0.605 + 0.609 = 1.214
113 + 115 = 0.582 + 0.583 = 1.165
-----------------------------------------------
1,214 - 1.165 = 0.049 that's 50 millivolts where they really count.

since D135, D136 are supposed to be a "safety catch", and weren't there in the SX-3900, you can try just leaving D135 out - BUT if there is something else it is "catching" we could have a problem.

 

[weeber]

Yep, d135 is toast. Conducts in both directions. This one I replaced and probably replaced it with something inappropriate. This is what I bought, the original part number is in brackets
AWH-114: D135: 954-E-183 (WZ-046)

All the diodes and resistors where tested with one leg lifted.

Would swapping d113 & d114 or d115 & d116 even things out?

I can see what you mean about it working without d135 and d136 as d135 is no good, yet that channel still adjusts fine. But what made it go bad, I wonder?

 

[markthefixer]

AWH-114: D135: 954-E-183 (WZ-046)

WHERE did you GET that number??? for a 4.6v 500mW axial leaded zener diode???

OK, I see, I missed the check request post at the end of November.. Post 71...
That's also where the trimpots came from.... :tears:

the 954-E-183 is a CURRENT REGULATOR DIODE, 50v 18mA minimum limiting voltage 4.6 volts.

the 512-1N5230B at 4.7 volts 500mw would have been more appropriate.

1s1555 hmmm, that's 35v at 100ma
we generally use 1n4148's at 100v and 200mA.

perhaps you should get a bunch of 1n4148's, you can get 100 of them for $1.80

Then we get some that are close together in forward conduction voltage.

I really don't like those voltages. And a "lightning bolt" did go through some of them.

We have to nail something down - which channel has the high idle current again... and which diodes of the ones you/we tested are associated with it?

d115 & d135 were the diodes I replaced on the RIGHT side when I blew the amp up.

The RIGHT channel, which is BEHAVING itself.

Is your "side" terminology the same as your "channel" terminology?

VR4 and VR6 (d114 and D116) is the channel that gets too high, the LEFT channel.

D114 + D116 = 0.605 + 0.609 = 1.214 right channel, ends up with idle current too high
D113 + D115 = 0.582 + 0.583 = 1.165 left channel, stays stable (more or less)
-----------------------------------------------
1.214 - 1.165 = 0.049 that's 50 millivolts where they really count. on the channel that has too much idle current. Suspicious huh?

YOU could swap 113/114 AND 115/116 and see if the problem moves.

OR, power it up, and take initial voltage readings across all 4 diodes (live) watch it increase it's minimum idle current and re-take those readings.

I would just glom in four new 1n4148's (after I checked that the new diodes had reasonably close diode forward voltage readings out of circuit) and be done with it.

Two of the four double resistors were ok? The ones on the LEFT channel?

This is why, with the NSA amps I did, that they got all new diodes and transistors, and the resistors all tested with one leg lifted. It's faster than troubleshooting weird symptoms.

 

[weeber]

We have to nail something down - which channel has the high idle current again... and which diodes of the ones you/we tested are associated with it?

The LEFT Channel has the high idle current. The diodes tested for the LEFT channel were the ones with even numbers D114, D116, D136

Is your "side" terminology the same as your "channel" terminology?

When i say LEFT, I mean the LEFT channel (components with even numbers)

VR4 and VR6 (d114 and D116) is the channel that gets too high, the LEFT channel.

D114 + D116 = 0.605 + 0.609 = 1.214 right channel, ends up with idle current too high
D113 + D115 = 0.582 + 0.583 = 1.165 left channel, stays stable (more or less)
-----------------------------------------------
1.214 - 1.165 = 0.049 that's 50 millivolts where they really count. on the channel that has too much idle current. Suspicious huh?

YOU could swap 113/114 AND 115/116 and see if the problem moves.

OR, power it up, and take initial voltage readings across all 4 diodes (live) watch it increase it's minimum idle current and re-take those readings.

I would just glom in four new 1n4148's (after I checked that the new diodes had reasonably close diode forward voltage readings out of circuit) and be done with it.

I've ordered new diodes of both types, and a few so I can find ones with similar voltages. I will just replace all 6 diodes on both channels, unless you feel taking the voltages as above is really important.

Two of the four double resistors were ok? The ones on the LEFT channel?

The double resistors on the LEFT channel both tested OK.

The double resistors on the RIGHT channel did not (one tested OK, one did not)

All 4 were replaced with new.

This is why, with the NSA amps I did, that they got all new diodes and transistors, and the resistors all tested with one leg lifted. It's faster than troubleshooting weird symptoms.

Learning experience. Next time I will do just that. I did test all resistors out of circuit, but I only tested the diodes for functionality, not voltage. Most of these parts are so inexpensive, and you have to take them out to test them anyway, it makes sense to just replace. :scratch2:

 

[markthefixer]

Most of these parts are so inexpensive, and you have to take them out to test them anyway, it makes sense to just replace. :scratch2:

Ding ding ding, we have a winner. :thmbsp:

The common parts, anyway, when you start fooling with replacing resistors, there are other details besides resistance to be considered, that can bite you on a willy nilly replacement. And once you cross the "resistance barrier" where the worst case voltages available aren't sufficient to cause wattage induced failure (heat) there is little to harm them electrically, so if they haven't drifted or been physically damaged, ARE left alone. We tolerate a small risk of damage when lifting the lead, to get an accurate resistance reading .


Now, for the rest:

Rereading the voltages isn't necessary. Might be entertaining, but the small risk isn't necessary.

The THEORY is that in that part of the circuit, the 100k variable resistor varies the current source's (there are two, they see-saw their output voltages to compensate for the amplifier's output voltage) current output, which flow through those diodes. The variation of current through those diodes affects the forward voltage developed across them. Which shifts the operating thresholds for the NSA circuit that prevents inappropriately low voltages from getting to the individual driver halves and completely shutting down the output transistors. This is what gets the amp from about 0.05% distortion to 0.005% distortion, by eliminating that complete shutdown of the output transistors during parts of the waveform.

Think of it as placing a ruler across the waveform and saying "you will go no lower", and sending it to the NPN half (positive going half) of the outputs,
and then doing the same for the negative half of the waveform saying "you will go no higher", and sending THAT to the PNP half (negative going half) of the outputs.
Kind of splitting the signal into two parts, for the two halves of the output stages.
THAT is the essence of the NSA portion of the amp.

edit - added the image below:

 

[weeber]

Cool. Thanks for the info Mark. I like getting the background as whats actually happening in the circuits.

Im a little frustrated though.
I got some 4148 diodes and found 4 that all measured the same at 0.588 for d113 to d116
I then got 2 5230b that both measured 0.697 for d135 and d136.

Now, the right channel starts at 53mV and slowly climbs to 66mV with all pots set to zero.
The left channel starts at 53mV and climbs to 61mV.

The DC offsets were L 20mV R 11mV.

What now?

 

[weeber]

I had such high hopes the diode replacements would do the trick.

Now both sides are misbehaving. At least they're misbehaving equally.