Repair Pioneer A9 NSA

[markthefixer]

Instead of measuring the resistance of the VR to get several different points on the graph, can I divide the voltage from min to max that drops across the VR into 4 equal intervals and then take the measurements.

YES

."or all this"

4 equal intervals will really be 5 points, inside of one of the intervals it will cross this point: Q25 emitter set to 1v and thus 1 mA
And that I expect will be found to be an internal result during an operating adjustment.

It is crucial in understanding some of the circuitry. R101, D11, D13 and R103, D17, D15 . The voltage drop across a resistor is set by ohm's law, but the voltage difference from the anode to cathode of a diode depends upon the current flowing through the diode. One can TRY to estimate it from the datasheets - but that rapidly piles assumptions on assumptions. One bad or inaccurate assumption brings the whole house of cards tumbling down.
Better to KNOW.

We / I can only guess until the outputs are in place and idle current flows.
I'm hoping that 1v point will show up when we are actually reading idle current.

 

[Carguy1394]

YES

."or all this"

4 equal intervals will really be 5 points, inside of one of the intervals it will cross this point: Q25 emitter set to 1v and thus 1 mA
And that I expect will be found to be an internal result during an operating adjustment.

It is crucial in understanding some of the circuitry. R101, D11, D13 and R103, D17, D15 . The voltage drop across a resistor is set by ohm's law, but the voltage difference from the anode to cathode of a diode depends upon the current flowing through the diode. One can TRY to estimate it from the datasheets - but that rapidly piles assumptions on assumptions. One bad or inaccurate assumption brings the whole house of cards tumbling down.
Better to KNOW.

We / I can only guess until the outputs are in place and idle current flows.
I'm hoping that 1v point will show up when we are actually reading idle current.

Should I then also include voltage measurements across the specified elementsR101, D11, D13 and R103, D17, D15 ,for each interval change?

 

[markthefixer]

yes please. the discovery of the 1mA readings is important, but the other data points need to be read to see what is in that range.

It would be helpful, if at the end of the tests, when you intend shutdown, that before shutdown the 1 mA point is set, so when power is scheduled to be off, the resistance needed to get that 1mA can be discovered.

 

[Carguy1394]

OK, I got it.
Let's see what happens.

 

[Carguy1394]

So. I made all the measurements from #155 except for the last phase of varying VR1 to maximum at 1 volt across R47.

We have the values VR1=0/ 0.125/0.300/0.450 V from 0.600Vmax measured at a 1 volt across R47 done

In this situation, VR3 is open at 3/4 of max (30k Ohms of 40k ohms max, measured on the board), and VR1 is also at 3/4 of max.
But the voltage is not in a linear relationship because when changing VR1 from 2/4 to 3/4 the voltage measured at both ends of R1 on the GWX-593 jumped from 3 volts to 6.62 volts and the drivers heated up to about 80C
I expect if I increase VR1 to the max voltage(VR1 to 0.600V) ,then across R1 (150 ohms =75+75 ohms) if it jumps by the same order to go around : 6.62 + (6.62-3)=10.24 volts.

Bearing in mind #157 :
"3v max stays, the 150 ohm emitter resistor will get hot, but nothing burns. More voltage WILL burn the 75 ohm + 75 ohm resistors, ..."

I ask this question:
Should I try to increase VR1 to max in this situation and make measurements if it lasts the board?

If during normal operation of the amplifier, at full amplification, the elements are loaded in this mode - I prefer now to test the circuit board will withstand, and not when I installed the PT.
If we don't risk it, I am ready to post all th measurements.
If we have to try VR max too let me know. I will wait instruction on this case

 

[markthefixer]

I was looking for any of these "gotcha's", which you have found.

They are something to avoid, at least until understood.

I have to carefully analyze this and everything else in your last post.

Deep thought. not tonight though, probably tomorrow.

 

[Carguy1394]

Here is a visualization of the question.
I have measured both channels and have indexed the rows with L and R for the left and right channels respectively. Odd numbers of transistors and resistors and diodes are for the left channel and even numbers are for the right channel.

I got over 6 volts on R1 and R2 and so I stopped and asked if I should turn VR 1/2 to max to fill the HL and HR rows. If something is unclear or wrong, ask for clarification. Maybe I should have written it during the day , but I think that is ok.
Just I don't know how to upload with a better resolution. If is unreadable let me know to think for something better

I copy-pasted from excel with a some correction of typo. to be more readable.. Let me know if you find some mistakes to correct them.
*************
First letter is the number of the procedure

Second letter is L or R and is Left or Right channel on the board GWH-146.

EXAMPLE:

AL-procedure A for Left channel

AR- procedure A for Right channel

--------------------------------------------------
AL-VR3 30kOhm (R47=1V) ,VR1=0
AR-VR4 30kOhm (R48=1V) , VR2=0
BL - VR 1,2,3,4 = 0 , R47=2,5V ,Q13C=1,781V
BR - VR1,2,3,4=0 , R48=2,54V , Q14C=1,753V
CL - VR1=0, VR3=max=79,4V
CR - VR2=0 , VR4=max=81,5V
DL=AL ,DR=AR
EL - VR3 (R47=1V) ,VR1=0,150V (23,5Ohm)
ER - VR3(R48=1V) VR2=0,150V (23.5 Ohm)
FL - VR3(r47=1V) ,VR1=0,300V
FR - VR4 (R48=1V) , VR2=0,300V
GL - VR3(R47=1V), R1=0,450V ……….6.62V on R1 150 ohm
GR - VR4(R48=1V) , R2=0,450V……….6,06V on R2 150 ohm


Indications of the schematic of the measurements:

1- Q13C/Q14 C red to GND black
2 - Q15C /Q16C red to GND black
3 - Q17 E /Q18 E red to Q21 E /Q22 E black
4 - Q17C /Q18C red to Q17E/Q18E black
5 - Q21E/Q22E red to Q21C/Q22C black
6 - Q25E/Q26E red to GND black
7 - Q25C/Q26C red to R7/R72 black
8 - Q27C/Q28C red to R71 /R72 black
9 - voltage on R47/R48
10 - voltage on R101/R102
11 - voltage on D11/D12
12 - voltage on D15/D16
13 - voltage on D13/D14
14 - voltage on R103/R104
15 - voltage on D17/D18

 

[Carguy1394]

I see. There is a mistakes. Too late for correction. New post corrected.


First letter is the number of the procedure
Second letter is L or R and is Left or Right channel on the board GWH-146.

EXAMPLE:

AL-procedure A for Left channel
AR- procedure A for Right channel

--------------------------------------------------
AL-VR3 30kOhm (R47=1V) ,VR1=0
AR-VR4 30kOhm (R48=1V) , VR2=0
BL - VR 1,2,3,4 = 0 , R47=2,5V ,Q13C=1,781V
BR - VR1,2,3,4=0 , R48=2,54V , Q14C=1,753V
CL - VR1=0, VR3=max=79,4V
CR - VR2=0 , VR4=max=81,5V
DL -VR3 30KOms( R47=1V) , VR1=0 -this is done already in AL
DR - VR4 30KOhms (R48=1V) VR2=0 -this is done already in AR
EL - VR3 (R47=1V) ,VR1=0,150V (23,5Ohm)
ER - VR4(R48=1V) VR2=0,150V (23.5 Ohm)
FL - VR3(r47=1V) ,VR1=0,300V
FR - VR4 (R48=1V) , VR2=0,300V
GL - VR3(R47=1V), R1=0,450V ……….6.62V on R1 150 ohm
GR - VR4(R48=1V) , R2=0,450V……….6,06V on R2 150 ohm


Indications of the schematic of the measurements:

1- Q13C/Q14 C red to GND black
2 - Q15C /Q16C red to GND black
3 - Q17 E /Q18 E red to Q21 E /Q22 E black
4 - Q17C /Q18C red to Q17E/Q18E black
5 - Q21E/Q22E red to Q21C/Q22C black
6 - Q25E/Q26E red to GND black
7 - Q25C/Q26C red to R71 /R72 black
8 - Q27C/Q28C red to R71 /R72 black
9 - voltage on R47/R48
10 - voltage on R101/R102
11 - voltage on D11/D12
12 - voltage on D15/D16
13 - voltage on D13/D14
14 - voltage on R103/R104
15 - voltage on D17/D18


This is the best I can do for the requirements in #155 and #168 combined. We have 15 measurement methods that vary in 12 different ways. What I have added in addition, but it is necessary for more accurate statistics, is the right channel indexed with R. There I hope I have not confused the transistors, because I have simply used the mirror symmetry and the system of their designation as everything in this amplifier which is for the left channel is marked with an odd number and everything for the right channel with an even number.

I haven't made this chart since the first time. I made more than 6 different variants, including 3D graphics, but of all the variants, this is the most compact and the easiest to understand if you read the instructions carefully. In total I have made over 900 measurements of which I have published here in this table only 180. The reason for this is that the measurements have to be made all at once for a very short time without interruptions because the system is constantly self-regulating around some equilibrium and the data is affected by lots of factors likes the mains voltage, which varies, the thermal drift of the transistors.
Also my hand over the board which has capacitance and the amp picks it up and interferes which immediately raises the voltage on Q13C/Q14C. For this I had to ground my hand with a wire to the heater. Several times my probes slipped on the solder joints and shorted between the legs of the transistors, but there was some protection left to work and the amplifier did not burn out. That was the main reason why I liked that scheme better.
Finally, after a lot of practice, I managed to measure all 180 measurements at once, and I am posting them here in the hope that they will be of some use.( Still no answer from markthefixer should I do the last measurements with VR1/VR2=max and R47/R48=1v)
The measurements are accurate and the chart works.
Type designation likes this:
7 - Q25C/Q26C red to R71/R72 black means the following:
Q25C red to R71 black and Q26C red to R72 black.
This analogy is followed in all such notations with a slash "/"

 

[Carguy1394]

I don't know what to do with this board and if it is suitable. The measurements show that the right channel is much better than the left. I don't like the left one at all in this way. I will change the transistors of the left channel again in unison with the right one on the hFE.. The other thing that is not clear to me is whether I should assemble the amplifier with this board and with these drivers from eBay, which may be fake, and what quality replacements I can currently find . I am still waiting for a response from MTF. He must be very busy, but I hope when he is free he will respond.

 

[markthefixer]

You may not see me, but I am working my way through the readings.
I have perhaps 6 different amplifier repair threads active, including a sx-3900 nsa amp. It's main difference is that yours has pre-drivers, necessitating an increase from +/- 1.2v to +/- 1.8v on yours. (as well as auto DC nulling and other stuff)
That may have resulted in that "gotcha" I talked about a few days back. ("I got over 6 volts on R1 and R2")
I am trying to wrap my mind around the NSA circuits as modified by the addition of (perhaps for "safety") D19, D21 2.2v zener diodes.
as well as understanding what or how the 6 volts turns up.

The "fake" drivers of which you speak, are they the ones from Israel?

perhaps it would be good now to bring the other three boards up to the same level of operation, finding and eliminating their faults. Even if it means individual transistor testing.

perhaps also I tend to devolve to 100% component (well bipolar t's, diodes and resistors) testing because I have a good desoldering setup where I am relatively confident of not causing damage.

 

[Carguy1394]

You may not see me, but I am working my way through the readings.
I have perhaps 6 different amplifier repair threads active, including a sx-3900 nsa amp. It's main difference is that yours has pre-drivers, necessitating an increase from +/- 1.2v to +/- 1.8v on yours. (as well as auto DC nulling and other stuff)
That may have resulted in that "gotcha" I talked about a few days back. ("I got over 6 volts on R1 and R2")
I am trying to wrap my mind around the NSA circuits as modified by the addition of (perhaps for "safety") D19, D21 2.2v zener diodes.
as well as understanding what or how the 6 volts turns up.

The "fake" drivers of which you speak, are they the ones from Israel?

perhaps it would be good now to bring the other three boards up to the same level of operation, finding and eliminating their faults. Even if it means individual transistor testing.

perhaps also I tend to devolve to 100% component (well bipolar t's, diodes and resistors) testing because I have a good desoldering setup where I am relatively confident of not causing damage.

As I thought.
It seems like the A9 is more complicated than a lot of other amps in this class, or no one has posted this in-depth about the repair.
Those transistors on the red plastic are the Israeli ones A1930/C5171 . I have not installed them yet because I kept them as very valuable. Are they fake? let me know if you have any info. They are all plastic-how they can disperse the heat??
These others are currently installed. I bought them from an eBay Chinese distributor very cheap. I wonder how they stand the tests. I got burned a lot before with short circuits and these are all I have left.
I think I'll start with the A1930/C5171, if there is no other choice.
The board is in functional condition but not very shiny , with lots of repaired stripes. That's how I bought these amplifiers, the other 2 are a little better.

It would really help me a lot to test some components like this to know which ones are good to use. And to bring the other boards in such functional condition You just have to tell how to make it happen and what components.

 

[Carguy1394]

A small improvement perhaps. I changed the transistors of the left channel Q9-Q15 and Q29,Q31, selecting them with the same hFE as on the right. Almost the same voltages as on the right channel were obtained . The difference is a maximum of 0.08 volts. I measured them at VR1,2,3,4=0.
I will also measure them with VR3/VR4 set so that there is 1 volt across R47/R48 and R1/R2 are 3/4 open which is 0.450 volts measured across them. That is, I will repeat the GL and GR measurements from #173

 

[markthefixer]

a good plan.

the Israeli ones A1930/C5171, Do you have the Peak Atlas DCA55 or DCA72? When yours tests Hfe, does it list an Ic?
I ask this in hopes of freeing up the A1930/C5171 from testing embargo. ( i.e. I test the ones I have)

 

[rcs16]

FYI I found what I hope to be as a reputable source for 2sc5171, to bad they do not stock the 2sa1930 as well
https://tvpartscanada.com/en/search...rway=desc&search_query=2sc5171&submit_search=
Good luck folks, I follow along this thread once and a while, admiring your dedication to sorting it out. I do not have much to offer as NSA designs are Marks expertise.

 

[rcs16]

another related find, listed as NOS
https://www.ebay.com/itm/3253596776...2OrAJ1FQ5grH0eL7Dhb4LHkMHV|tkp:Bk9SR4Dlx8TwYA

 

[Carguy1394]

another related find, listed as NOS
https://www.ebay.com/itm/325359677613?hash=item4bc0f2b0ad:g:-AEAAOSwCWdiuIe2&amdata=enc:AQAHAAAAsJPslLsTOISHmscnrsss8We1hOZ8b+sP74pOtZURoZ2mAy0JdBFdBY8AGUSg383k43uNL6D2iNRROS+UQtMpu+M+4ThVJp1k2BF+nuiawFkCMAY2R530ifiLN9bRkHpIv+tGvI/cRIZQehUPFN3Mr1gMV+qoASdQLTkPbMR8WXmA6LBRKBJly8Pp4KzNsAsE/uQDPektnySoxmuExN2OrAJ1FQ5grH0eL7Dhb4LHkMHV|tkp:Bk9SR4Dlx8TwYA

Thank you.
So where is the romance and joy of a successful repair if I buy this board? I'd rather buy a whole working amplifier instead of a circuit board. But that's not interesting to me because I have 5 more high-end amps than this one working and I want to see how a tuned amp with my input will work. And especially the same model with different transistors

 

[Carguy1394]

a good plan.

the Israeli ones A1930/C5171, Do you have the Peak Atlas DCA55 or DCA72? When yours tests Hfe, does it list an Ic?
I ask this in hopes of freeing up the A1930/C5171 from testing embargo. ( i.e. I test the ones I have)

So.
I took measurements on the left and right channels after replacing the transistors on the left channel and selected them according to the hFE.
I am posting them together with the previous measurements so that the results can be easily compared. I have marked them in red and they are only for the GL and GR modes marked in red.


The chart is an addition and part of the previous post #173 and the same explanations and abbreviations apply to it. Only the lines in red are different. ( it is so simple that eater I start understand it)

Good voltages are also obtained on the pre-drivers and drivers, which I have not included here, but I may post them later.

"Do you have the Peak Atlas DCA55 or DCA72? When yours tests Hfe, does it list an Ic?"

- I don't have a Peak Atlas nor an oscilloscope nor a tone generator nor a klirrfactor meter ... I wish I had them a lot. But even with them I don't know what I would do in a sea of
fakes transistors.Multi tester TC1 some time list Ic or Ie, currently is burnt ,and I am locking for another one.
Reputable vendors are already working with Chinese transistors from long time ago (my opinion ). Just from the better quality. Actually I am impressed from the work of the 100% certified fake transistors , bought from eBay , that I am using it and they survived all this tests and also additional lots of short circuits.
What is the difference between a Chinese fake and the original - the Chinese fake can give good amplification curves when testing with Peak Atlas, but when it sees 60 volts it can smoke, or it can withstand, but when it sees the amperage for which it is prescribed - then burns.

What I need to know is what is the maximum voltage that goes through R1/R2 150 ohm to see if the drivers and predrivers will withstand by full power of the amp and to tested the boards with this voltage.. The other transistors before that, even if they are fake, will last probably .
The hot temperature of about 80C on the drivers scare me and make me to stop rising the voltage by the measurements. So I don't know did I reach the voltage of maximum power of the amp. ,and did this transistors will survive in by maximum power.
So did some one already tested A1930/C5171 how it works likes a drivers on Pioneer amp, or I will be the first one to use it?
I will do the next and will build the other 2 boards.

 

[Carguy1394]

I need sub for 2sa995A and 2SC2291A.
From AK:

"2SA995 if you need a replacement, you should find something. (KSA1381, complement KSC3503 from ON Fairchild) "
Is this rely work?

 

[markthefixer]

AS a GENERAL statement, these are dual transistors in one case.

In these cases we have found "matched pairs", and heat shrinked them in close physical proximity.
matching is quite a divisive concept, arguing between Hfe matching and Vbe matching. Since Hfe varies with collector current,it's almost hopeless, and the best would be to find the current they are operated at and use that for matching Hfe.

Transistors purchased as tape mounted have been observed to have parameters similar to each other.

Since this amp has servo dc offset I wouldn't expend much effort in choosing matches, the system will null it out.

Usually when the "base stealer" manual offset adjust is used, THOSE are the transistors we want more closely matched.

have you found defective 2sa995 or 2sc2291a transistors?

 

[Carguy1394]

AS a GENERAL statement, these are dual transistors in one case.

In these cases we have found "matched pairs", and heat shrinked them in close physical proximity.
matching is quite a divisive concept, arguing between Hfe matching and Vbe matching. Since Hfe varies with collector current,it's almost hopeless, and the best would be to find the current they are operated at and use that for matching Hfe.

Transistors purchased as tape mounted have been observed to have parameters similar to each other.

Since this amp has servo dc offset I wouldn't expend much effort in choosing matches, the system will null it out.

Usually when the "base stealer" manual offset adjust is used, THOSE are the transistors we want more closely matched.

have you found defective 2sa995 or 2sc2291a transistors?

Yes, I found 2SC2291 with a difference in hFE between the left and right sides - hFE 140/422. .This is on one of the next two boards I'm getting ready to assemble. On other boards these transistors have hFE matched to unity.