• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

SA 9500II Issue: New Thread Old Problem: I Need Help Please!!!!

Mark, thanks for understanding that for me slower is better. I'm in no hurry. This is a verification that everything on the AWH-052 Right has been checked and is found to be correct. A reminder that Q1 through Q4 are original and that Q 5 and Q6 are new OnSemi mje15032/33s. This is true of both cards. Everything is symmetrical. I have metered and verified that the right board is the right channel. I have metered from the Amp-in to the board and traced through. I also traced from the board to the right speaker output, through the A+B switch to the speaker output on the back of the amp. It traces through. The issue is in the right board for sure. I also double checked by feeding source and pulling the input plug, verifying that right was right.

The Chinese threw me a curve at first because the cheap mini to rca adapter cord I was using as the input interconnect was unlabeled. To me, red has always been right and white has always been left. Apparently that is not true in China. This threw me for a loop last week until I finally figured out the problem.

As far as the resistors that were replaced, the originals were tested with with a meter, other than my Fluke. The meter was not autoranging and I forgot about that, having gotten so used to my Fluke. The resistors were replaced with carbon resistors and tested before installation. They were very close to spec.

I will pull R8 and R9, out of circuit, in a day or two, when I recoup from today's marathon work call. C5 has already been replaced with a second new cap just to be sure it was okay. This was in response to your previous post. There was no difference. I do not have a ESR or capacitance meter; so I have no way to test caps. I feel fairly certain that C5 is okay.

I did test the voltage of the input at both cards. I used a tone generator on the computer. Both sides tested the same at the input but the output voltage was a substantial difference. IIRC left speaker output came in at 495mv and the right speaker output came in at about 20mv. I have no method of calibrating the output other than testing both sides at the same output level, which I did.

I have to go back from lunch now. I hope this information is helpful in some way. Thanks.
 
Last edited:
Very very helpful.

The best plan of (eventual) attack is to measure at the amplifier inputs (base lead) what the actual voltage level of the signal that is coming in is at for each amplifier channel, and what the output voltage is for each channel.

That cheap chinese adapter could very well be the problem... we will be proving it isn't... it has to do with the input resistors etc.. the pins at the edge of the board could measure ok, yet the amp is deaf.

We could try just "turning it around" - flipping left for right at the Amp-in RCA plugs, that might reveal something...

THEN maybe we fool around with measuring input and feedback resistors...
 
Last edited:
Mark and Goat67, this is a verification that the updated reading from the LEFT board in post #9 is correct. That said, I will still continue to concentrate on the right side.

Goat67, to answer the questions you posted:

Question did you have to replace the outputs that are connected to the left board due to a bias setting mishap?
Click to expand...

All outputs have been replaced with FJP1943OTUs and FJP5200OTUs, which are Fairchild products rated at 80w. These are TO-220 style replacing the old Batwing transistors, which I can't remember the style designation for.

Did you replace any other transistors on this board?
Click to expand...

Q5 and Q6 on both boards are now MJE15033/33s.

I seem to recall you had issues soldering on the left board. Verify you do not have any solder bridges.
Click to expand...

All boards were inspected, with a magnifying lens, before any reassembly was done. The soldering was clean. Very good question!!

Check connections and joints around Q1
Click to expand...

Done and looks clean and clear.

Now to Mark. you said in post #22:

The best plan of (eventual) attack is to measure at the amplifier inputs (base lead) what the actual voltage level of the signal that is coming in is at for each amplifier channel, and what the output voltage is for each channel.
Click to expand...

Is this what you mean taken from my post on #21? I do not remember what the input voltage was but it was the same for both sides.

I did test the voltage of the input at both cards. I used a tone generator on the computer. Both sides tested the same at the input but the output voltage was a substantial difference. IIRC left speaker output came in at 495mv and the right speaker output came in at about 20mv. I have no method of calibrating the output other than testing both sides at the same output level, which I did.
Click to expand...

As far as this:

We could try just "turning it around" - flipping left for right at the Amp-in RCA plugs, that might reveal something...
Click to expand...

Already done, with nothing changing as far as output.

As far as this:

That cheap chinese adapter could very well be the problem... we will be proving it isn't... it has to do with the input resistors etc.. the pins at the edge of the board could measure ok, yet the amp is deaf.
Click to expand...

I tried that adapter on my G 8000, that is on the workbench too, it played just fine through the Sansui. After I get the chance to check R8 and R9 I would be glad to try some measuring past the pins and into the board. I'll post as soon as I get the R8+9 check done. Thanks to you both for your questions and concern. I'll update asap.
 
OK, the adapter is close** to exonerated, but you are missing one point that I have been asking:

The input levels I have been asking for measurements of are important because of:
1. the actual AC voltage of the input signal
2. It's the input signal INSIDE of the circuitry at the BASE of the differential transistor input.

You see the base on the one side of Q1 is the input point, and the base on the other side is the feedback signal.

The feedback signal is a DC isolated(c5) voltage divider (r8 / r9) that we know the "input voltage" of - it's the amplifier output.

The "amplification" is because Q1 is trying to match BOTH base signals. if the output of the amplifier were connected DIRECTLY to the base of Q1 (R9 = 0 ohms) the output voltage of the amp would be (almost) EXACTLY the input. Just a higher current drive capability.

Think of the r8/r9 voltage divider as kind of a lever.

** (close to exonerated means that there are input impedance differences between the two receivers that could make a difference - probably not - but stranger things have happened - I can tell you strange stories... and this could end up being another one of them - but I hope not - until we prove otherwise)
 
I am sure you could tell some strange stories, Mark. :D More on those later. :yes: I was not missing your point; I was just not there yet, so to speak. I do understand what you are saying, to a degree, and I get the point. My hope is that once I can get to the checking of the resistors that we can start the process of which you speak.

If I was using, say a 1K tone, what would be a reasonable input voltage, at the input pin, to try to achieve for the purposes of tracing? Since I will be using a computer, not a real live tone generator, I will have no way to calibrate the output the level of the source from the creator of the source, ie. the computer. I can, however, measure the voltage of the output from the computer output and use that as my 0 point as it were. Does this question regarding what voltage to start with make sense to you and, if so, what would be a good voltage range to shoot for? Also is a 1K tone a good frequency for testing?

Thanks again for your attention and, BTW, what are you doing up at this time? You usually seem to keep my kind of night owl hours. :scratch2: :D
 
Last edited:
Resistors 8 and 9 have been checked. R9, a 62K, comes in at 62.5K. R8, a 2.4K, comes in at 2.42K.

I now await any instructions for tracing and am ready to proceed at your leisure.
 
Last edited:
I started missing daylight, so i forced a change in my hours. it takes a lot of resolve not to slip back in to the old hours.

The problem with not posting in the wee hours is that I just can't go through one pass and let the answers stand as people wake up to them. There are follow up questions (even though I try to give them enough to do to keep em busy for a while...) that appear quickly and it's hard for me to stop answering and to move back to the bench - or whatever else I should be doing.

If I try to get everything done for the day, THEN hit the forums, it usually results in me staying up much later than I wanted. I'll figure out a balance somehow...

Not complaining, it's just the way things are - I'm getting in less bench time than before.

1/2 volt is reasonable. 1 volt is supposed to be full power out.

part of it is being able to calculate the actual gain numbers and try to get some clues from the results.


edit: see what I mean. do the voltage measurements. that amp should be PERFECT....
 
Would it be possible to give me the path of points to follow; or should I just start at pin 22, the input and record the voltages from point to point. I am a virgin at this so I need a bit of hand holding at first until I get the hang of it. Maybe you could slip me a couple of drinks and I'll loosen up a little, wink wink....:naughty:

Enough of this Tom foolery. I'm off to set this up. I shall return.

:stupid: I am now setup and have a few basic questions that may appear silly to you but are not clearly obvious to me. First, should the amp be turned on? Second, do I need to test voltages at the points where capacitors bridge from one trace to another or do I follow the bread crumb trail of the resistors where they lead? Third, should I be A/B ing between the right and left boards, as a point of comparison?

Due to the fact that the resolution of the mouse is limited, the closest I can get is 503.8mv for the input. The frequency is 1k. Just for the hell of it, I swapped the input RCAs to see if there was much difference. The left RCA is sending 499.6mv so the difference is only 4.2mv, in favor of the right side. I would say that is not enough to worry about the cable/adapter.
 
Last edited:
Maybe we'd better turn it down to about 100 millivolts if you can. 1/2 volt may be too much. We don't need to have it swinging that hard. I'd worry about clipping under certain circumstances - which without a scope to catch it, would really skew the readings.

measurement point one:

the common trace connecting R2, R3, C2, c3 & base of q1 (the input)

measurement point two:

the common trace connecting R8, R9 and the OTHER base of Q1 (the feedback)

Measurement point three:

the common trace connecting R9, R16, R17, R18, R19, R26, R27, R28 & L1 . (the output)

One set of readings for the left channel.
Another set of readings for the right channel.

take your time, I'll be back tomorrow. Too much has been ignored today here at the casa mtf.
 
This I understand and I will comply, which will be tomorrow sometime. Thanks for the hand holding.

Mark, an additional question, if you please. In bmac's thread about shorting a SX 850 you suggested replacing the zeners as a matter of course due to the negative nature of the shorting experience. Would this be advisable for me to do the same, even though they test out good? Thanks.
 
Last edited:
9500

KingBubba said:
This I understand and I will comply, which will be tomorrow sometime. Thanks for the hand holding.
Click to expand...

Trust you are good lookin enough to hold hands.:D
Just a thought/observation/question. Could there be some difference between the 4 versions of the 9500 that are the problem?
ie: KU = USA
Kc = Canada
HG = Eur. (2 different voltage sw)
S = general export ( 4 different voltage sw)
I have a Ku & by definition my second amp is an S ( multiple voltage switch).
Believe I have noticed some transistors have different numbers for different versions.(on the schematics). Any bearing on any thing? :scratch2:
 
A difference between models and especially between the service manuals and the actual units is very possible. However in my case, I have two boards with entirely symmetrical parts distribution that are acting differently. All parts test out good with a meter when they are cold so now it is time to go in hot and follow the voltages from one point to the next to determine the differences. Being that the gentlemen who are assisting me are electrical engineers or similarly inclined, they are able to determine the potential problems by doing the math on the components and voltages and determine if the parts are reacting the way they are supposed to.

I am afraid that it took me three times to get through algebra I in high school. I even took remedial math and that did not help. If someone had shown me ohm's law back then, I most likely would have "gotten it" better because I would have had something to apply the math to. I can tear down a desktop or laptop or a theatrical lighting system and repair it with great fluency, but that mostly comes down to parts swapping and problem identification due to logic. When it comes to math, I fail miserably.

I don't know which model I have but at this point it would not really make much difference. Glad to hear your unit is up and on its feet.
 
Last edited:
Well, here goes. I, for one, do not know how anyone could make sense of these numbers.

Trace one is not complete in the sense that caps 1>3 do not have enough tail showing to read off of. This means that somehow I have to turn this thing over without shorting out the boards to read from the foil side of the PCB. I decided to read all I could face up and do the bottom all at once. In removing an unattached probe from R8 I touched the nearby plain jumper and a spark ensued. Everything I tested at this point seems unaffected by this mishap, but I imagine I will be going back and rechecking things. Echowars comment about this being like playing the game "Operation", when you were a kid, is appropriate, but on steroids. I used the foil diagram for the purposes of recording these numbers. For the sake of understanding I will describe the component position as either top, bottom, left or right as it is shown on the diagram.

Pin 22 106.9mv
pin 23 .3mv
R2 bottom 106.9
R2 top 3.7mv
R1 bottom 106.3mv
R1 top .3mv
R3 bottom 3.7mv
R3 top .3mv
Q1 bottom base 3.7mv
Caps will be added later after the flip

Trace #2

Q1 top base 3.7mv
R8 left 3.7mv
R8 right .3mv
R9 right 3.7mv
R9 left 98.9mv

Trace #3

R9 left 98.9mv
R16 top 99mv
R16 bottom 99mv
R17 top 99mv
R17 bottom 98.9mv
R18 top 99mv
R18 bottom 99mv
R19 top 98.9mv
R19 bottom 98.9mv
R26 top 98.9mv
R26 bottom 98.9mv
R27 top .3mv
R27 bottom .3mv ( I could find no commonality in R27 but recorded it anyway.)
R28 top 98.8mv
R28 bottom 98.8mv.
L1 top 98.9mv
L1 bottom 98.9mv

I will do the left side tomorrow, if possible. One comment about something that has me a bit confused. The meter showed higher millivoltage disconnected than it did connected. This make no sense to me. Another thing that confuses is that whenever connected to ground, pin 23, or any component traced to pin 23 the voltage started high, 3 -4 mv, and then slowly settled at a low of .3mv. After that point, I did not want to sit and wait any longer. I suspect that had I waited, the voltage would have gone lower. This is confusing to me.

I do not know if these readings are helpful to you all, but they were a puzzlement to me. :scratch2:
 
Last edited:
KB, the earth is chock full of spurious DC voltage, IIRC there is usually a stray 20 volts or so always around. On my Fluke 179 on the bench, with leads lying loose< i get 2 to 110 or so mv dc. After hooking the leads to a test point they no longer respond to the stray signals, only to the DUT, greater bandwidth I suppose even though it may be at a lower voltage.
I'm now ready to have my theory shot down in flames by, say, the meter was just excited?
 
markthefixer said:
Maybe we'd better turn it down to about 100 millivolts if you can. 1/2 volt may be too much. We don't need to have it swinging that hard. I'd worry about clipping under certain circumstances - which without a scope to catch it, would really skew the readings.

measurement point one:

the common trace connecting R2, R3, C2, c3 & base of q1 (the input)

measurement point two:

the common trace connecting R8, R9 and the OTHER base of Q1 (the feedback)

Measurement point three:

the common trace connecting R9, R16, R17, R18, R19, R26, R27, R28 & L1 . (the output)

One set of readings for the left channel.
Another set of readings for the right channel.

take your time, I'll be back tomorrow. Too much has been ignored today here at the casa mtf.
Click to expand...


KingBubba
This may clarify Mark would like 3 readings from each board at these Test Points (TP)
If you can from the component side if possible if thats easier

So from your post I would bet your readings equal this

R2 top 3.7mv TP 1
R9 right 3.7mv TP 2
R9 left 98.9mv TP 3

Hope this helps

attachment.php
 

Attachments

  • sa9500ii.jpg
    sa9500ii.jpg
    57.2 KB · Views: 81
Last edited:
Goat67 is right on the money with understanding how I did my readings. Top, bottom, left, right, just how it appears on the foil side diagram. I hope I was clear enough. I could scan my printouts that I used for collecting the readings if that would be of any assistance. Let me know.
 
What I was trying to point out is you do not have to get readings
from all the components top bottom left right,
just the three test points I pointed out. Make sense?
The less probing you do the better, you do not want to repeat the probe shorting out pins like happened on R8.

BTW I do not think that oops caused any damage lets hope.

To make it simple

TP 1 Top Side of R2
TP 2 Right side of R9
TP 3 Left Side of R9
 
I understand, to a degree, what you are trying to clarify. The problem I am having is trying to find the commonality of some of the points. An example would be test 3 - R27. What does R27 have to do with the trace that is common to all the other resistors.

Is this all just to find out if that particular trace is whole?

If that is the reason behind these procedures, then I am clear with the process. If there is more to it than seeing if the trace is whole, then I am still unsure of what I am looking for and why.
 
I think you are reading to much into this.
I would just measure the Test points I specified.

From what I got from Mark's request was he was just telling you the electronic point in
the schematic not to have you verify the connections or the trace.
 
You must log in or register to reply here.
Back
Top Bottom